[00:00:17] *** estibi has joined #opensolaris
[00:00:54] <myrkraverk> hmm, if I want all my zpools to exist in /pools/ do I use the -m option to zpool create ?
[00:02:02] <myrkraverk> aparently so
[00:05:53] <richlowe> alanc: I'm not sure you can do either, re your case.
[00:05:57] <jamesd> zpool create pools  raidz|mirror  drives....
[00:06:04] <richlowe> alanc: I see your mail on dtrace-discuss, but I see nothing out of osol-arc/psarc-ext about it.
[00:06:17] <alanc> PSARC-EXT wasn'
[00:06:22] <alanc> t a choice when I filed it
[00:06:22] <richlowe> alanc: ... because you sent it to PSARC. :)
[00:06:32] <alanc> and plocher's office was empty
[00:06:38] <alanc> when I went to ask him
[00:06:40] <richlowe> I'm totally lost with that stuff.
[00:06:49] <richlowe> 2006/569 kinda dried up without anything actually happening.
[00:07:22] <alanc> yeah - I just submitted it as is to PSARC and figured there was a low enough chance of comments beyond "+1" that it wasn't worth more hassle to figure out how to open it as an open case
[00:07:47] <gisburn> alanc: ping!
[00:08:13] <alanc> sorry gisburn, I have disappeared in the 30 seconds since I last typed a message
[00:08:16] <alanc> 8-)
[00:08:30] * alanc runs away!
[00:09:10] * gisburn chases alanc
[00:09:21] * gisburn grabs a spear and throws it after alanc
[00:09:26] <gisburn> NO RUNNING!
[00:09:51] <alanc> okay, you caught me, what do you want?
[00:10:02] <gisburn> alanc: wanna try -xpagesize_stack=64K and check whether this is good for Xsun/Xorg ?
[00:10:34] <gisburn> alanc: I've tried this for ksh93 and it improved the performance quite good (see my comments about the "global var" patch for the ksh93 libcmd)
[00:11:31] <alanc> I can add it to the list, but won't have time for performance testing that soon
[00:11:38] *** GmanAKF is now known as Gman
[00:12:02] <alanc> oh, that's a Sun cc flag isn't it?   hmm, wonder if I can fake it with a mapfile for gcc
[00:12:32] <gisburn> alanc: erm, I doubt mapfiles are involved for this.
[00:12:58] <gisburn> alanc: just add it to the Sun Workshop/forte/Studio build flags now :-)
[00:13:11] <alanc> if I can't do it with gcc, then it's going to have 0 effect on Xorg
[00:13:24] <gisburn> alanc: why ?
[00:13:27] <gisburn> oh
[00:13:29] <gisburn> wait
[00:13:33] <boyd> I thought there was a way to do it with env vars for the loader
[00:13:34] <gisburn> ASM for Xrender ?
[00:13:36] <alanc> my Xorg builds are already using gcc
[00:13:44] * gisburn hits alanc
[00:13:53] <Gman> alanc, aww, you didn't send the dtrace xorg probes through psarc-ext :)
[00:14:27] <alanc> I tried to resist as long as possible, I begged the Studio team to give me the compatibility needed, but I've finally given in to the dark side
[00:14:38] <gisburn> alanc: traitor!
[00:15:04] <alanc> I was even the first consolidation owner to adopt Studio 10 for 32-bit builds
[00:15:09] * gisburn grumbles and throws stones after alanc
[00:15:19] <alanc> Gman: I couldn't figure out ho
[00:15:21] <alanc> w
[00:15:35] <gisburn> alanc: yes, I know... but the gcc thing for Xorg sucks.
[00:15:37] <alanc> and couldn't find plocher when I went to ask
[00:15:42] <Gman> alanc, you don't just send it to the alias and everything just magically works? :)
[00:15:43] <boyd> Won't that make the performance, like, you know... suck
[00:15:44] <gisburn> alanc: can you add it to the Xsun flags, please ?
[00:16:21] <alanc> boyd: performance is actually the biggest reason to switch to gcc for Xorg
[00:16:22] <delewis> boyd: for SPARC, but Xorg isn't very useful on SPARC, anyway :-)
[00:16:38] <delewis> gcc code generation on x86 and x86_64 is fairly decent
[00:16:48] <delewis> but I think Sun Studio generates superior code for x86_64, nowadays.
[00:17:03] <alanc> Xorg has gcc-format inline assembly including MMX optimizations that beat the code generated by either compiler
[00:17:03] <boyd> Actually, I was surprised at the recently posted benchmarks showing how much better s11 was than gcc
[00:17:06] <gisburn> delewis: the problem is that the MMX/SSE code used by the Xrender extension uses gcc asm
[00:17:16] <boyd> Ah
[00:17:42] <boyd> Also, I think a lot of the benefit (and it was a *lot*) from s11 was from auto-parallelisation
[00:18:14] <myrkraverk> hmm, is setxkbmap not part of EXCR?
[00:18:31] * alanc runs away!
[00:18:31] * richlowe cackles
[00:18:45] * gisburn prepares another spear
[00:18:57] <richlowe> alanc: it's great to know it's not just me though. :)
[00:18:58] *** RaD|Tz has joined #opensolaris
[00:19:05] <gisburn> alanc: please please add -xpagesize_stack=64K -xpagesize_heap=64K to the Xsun build flags, Ok ?
[00:19:19] <richlowe> myrkraverk: to give you an answer, no it isn't.
[00:19:28] <myrkraverk> richlowe: thank you ;)
[00:19:28] <richlowe> myrkraverk: I've been bugging alanc about it for at least 18 months now.
[00:19:33] <richlowe> myrkraverk: I think he's avoiding it just to piss me off :)
[00:19:38] <alanc> gisburn: the next time I have time to look at Xsun performance, I'll try it
[00:19:42] <myrkraverk> richlowe: ok, I'll start to bug him too ;)
[00:20:11] <myrkraverk> alanc: we want setxkbmap! we want setxkbmap! we want setxkbmap! we want setxkbmap! we want setxkbmap!
[00:20:12] <alanc> but given the #1 task on my Xsun todo list is "File ARC case to announce EOL of Xsun", that may be a while
[00:20:27] <gisburn> alanc: these options are for SPARC
[00:20:28] <richlowe> alanc: now *that* one should go through the -EXT stuff. :)
[00:20:34] <myrkraverk> you're killing Xsun?
[00:20:44] <gisburn> myrkraverk: yes, no X11 support on Solaris
[00:20:49] <alanc> just as soon as we can get everyone moved to Xorg
[00:20:52] <boyd> So lemme get this straight... s11 compliing Xorg on SPARC would be a nice thing, right?
[00:21:00] * alanc thwacks gisburn
[00:21:04] <gisburn> boyd: YES
[00:21:12] * gisburn thwacks alanc back.
[00:21:13] <richlowe> Yeah, you don't mean "X11 support"
[00:21:16] <richlowe> you mean "Windowing System support!"
[00:21:17] <gisburn> What does "twack" mean ?
[00:21:19] <delewis> alanc: are Xsun drivers going to be ported to Xorg (SPARC framebuffers are what I'm getting at)?
[00:21:22] <myrkraverk> gisburn: ah, good - what are we using instead? the haiku app server?
[00:21:23] <boyd> Esp since the inline asm is presumably for x86
[00:21:38] <delewis> or is Sun EOLing all framebuffers except the recent ones?
[00:21:53] <gisburn> delewis: Everything gets EOL'ed.
[00:21:54] <alanc> we haven't decided yet which compiler to use for SPARC
[00:22:04] <delewis> I think the answer is pretty obvious
[00:22:05] <boyd> All output will henceforth be on fanfold paper
[00:22:08] <richlowe> delewis: my guess would be a combination of the two.
[00:22:08] <delewis> Sun Studio is necessary.
[00:22:17] <richlowe> at least, one would hope "Common" rather than "Recent", in the above.
[00:22:19] * gisburn plays the "little mischief"-game with delewis
[00:22:24] <delewis> gcc code generation on SPARC *sucks*
[00:22:25] <delewis> to put it mildly.
[00:22:41] *** hspaans has quit IRC
[00:22:44] <gisburn> delewis: +1
[00:22:49] <gisburn> delewis: gcc sucks in general
[00:22:49] <delewis> maybe with the Sun Studio back-end things wouldn't be so bad
[00:22:53] <alanc> delewis: the SPARC graphics group proposed EOL'ing most old frame buffers (all UPA models and early PCI models) before we started pushing them to Xorg
[00:23:06] <alanc> we're still waiting to determine what will happen
[00:23:09] <delewis> alanc: will Xsun stay in the consolidation or will it be removed (similar to gcc)?
[00:23:15] <delewis> er
[00:23:18] <delewis> s/gcc/cde/
[00:23:27] <delewis> not sure where that came from :-)
[00:23:40] <alanc> which is why the ARC case is simply to publish the announcement that Xsun may not be around in the future, not to actually remove it
[00:23:50] <alanc> the plan is to eventually remove it
[00:23:55] <richlowe> right after CDE ;)
[00:23:56] <delewis> oh, this is terrible :-(
[00:23:57] <boyd> ... like CDE
[00:24:00] * richlowe whistles
[00:24:12] * gisburn hits alanc with -xpagesize_stack=64K -xpagesize_heap=64K
[00:24:23] <myrkraverk> richlowe: well, CDE is an option for everyone capable of using XIG
[00:24:31] <boyd> Anyone have any news on Trusted JDS integration timeframe plans?
[00:24:40] <myrkraverk> (that'll not be me though, since I have an nvidia card)
[00:24:44] <alanc> boyd: already done
[00:24:48] <boyd> !?
[00:24:53] <boyd> What build?
[00:25:08] <alanc> unless you mean re-integrating into the GNOME 2.14/2.16 builds
[00:25:18] <alanc> it was done for 2.6
[00:25:27] <boyd> multil-level sessions?
[00:25:39] <alanc> yep - it's in S10U3
[00:25:47] <boyd> Wow
[00:25:52] <boyd> .... but not in SX I guess
[00:26:01] <boyd> which is annoying
[00:26:06] <alanc> it was in SX until GNOME got upgraded
[00:26:11] <boyd> indeed
[00:26:16] <alanc> have to ping Gman on the re-integration plans
[00:26:19] <boyd> but TX wasn't in then :)
[00:26:20] <axisys> i am getting ready to upgrade my ultra 20 to solaris express.. i already have the dvd iso for b48.. should i rather get the b50 instead to install?
[00:26:32] * boyd pings Gman
[00:28:30] <boyd> alanc: So if I read this right you're saying that the trusted JDS stuff went into SX before the TX stuff did, but by the time TX went in Trusted JDS was out?
[00:28:38] * Gman reads back
[00:28:45] <richlowe> heh, another warlock heads-up, and still no real reply on getting it outside.
[00:28:55] <alanc> no, the integration order was TX -> TJDS -> GNOME 2.14
[00:29:04] <Gman> oh, the trusted stuff was a bit of a fuckup imho
[00:29:12] <Gman> we integrated it into nevada for 2 builds
[00:29:13] <boyd> alanc: Oh, I didn't realise that TX had been there so long
[00:29:19] <Gman> the required soak time for it going into an update
[00:29:25] <alanc> they specifically delayed 2.14 integration until after TX & TJDS had a bit of soak time so it could go into S10u3
[00:29:26] <boyd> Ah.
[00:29:29] <richlowe> and then killed it.
[00:29:31] <Gman> then we installed a new version of gnome into nevada, which didn't have the trusted bits ported
[00:29:35] <richlowe> Gman: so, when can we expect it back? :)
[00:29:41] <Gman> heh, hopefully b53
[00:29:41] <boyd> indeed.
[00:29:46] <Gman> that's what they're targetting
[00:29:51] <Gman> though i don't think any of the patches are ported yet
[00:30:14] <boyd> Ok... I'm about to start experimenting with TX on 49. I won't wait for TJDS
[00:30:22] <alanc> the TJDS bits are still in xscreensaver 8-)
[00:30:23] <Gman> yeah, best not
[00:30:30] <Gman> heh
[00:30:31] * boyd sighs and fires up CDE
[00:30:35] <gisburn> alanc: will the switch from Xsun to Xorg on sparc affect Ultra5 machines with m64 framebuffers ?
[00:30:37] <myrkraverk> erm, is there an API to interface zfs - for grachical apps or some such stuff?
[00:30:40] <Gman> we have a few paches that actually apply
[00:30:44] <Gman> but i doubt it works
[00:30:55] <alanc> gisburn: TBD
[00:30:57] <boyd> myrkraverk: There is a lib but I don't think it's considered public
[00:31:09] <richlowe> it isn't.
[00:31:16] <alanc> well, I suppose, really, yes - either they'll be supported by Xorg or EOL'ed
[00:31:19] <myrkraverk> boyd: ok - it's just for me anyway
[00:31:24] <gisburn> alanc: TBD ?
[00:31:31] <boyd> myrkraverk: You are aware that there is a web interface for it already, right?
[00:31:32] <Gman> alanc, xorg has any selinux patches yet?
[00:31:33] <alanc> but m64 is the first frame buffer Jay has gotten Xorg running on
[00:31:34] <richlowe> I think the intent is to make it so, but there's stuff that'd have to be done.
[00:31:43] * richlowe is going on memory of zfs-discuss
[00:31:47] <alanc> TBD = "To Be Determined"
[00:31:58] <richlowe> Gman: alanc blogged about that stuff a while ago, I think.
[00:32:01] <richlowe> someone did, anyway.
[00:32:04] <alanc> Gman: the selinux stuff is still in a branch, planning to land for 7.3
[00:32:05] <myrkraverk> boyd: no, I'm just a n00b ;)
[00:32:33] <gisburn> delewis: please activate emergency prodedure "2"
[00:32:34] <alanc> I blogged about XACE, which is the common layer TX & SELinux will both use to plug in their policy checks in mostly the same places
[00:32:55] <boyd> I think it's not running by default, but "smcwebserver enable" then smcwebserver start" then connect to https://localhost:6789 I think will work. Not sure on the port number :(
[00:33:18] <alanc> it was kinda cool last week to see the GNOME desktop resize control panel working on an SB1500 with m64 card
[00:33:22] <myrkraverk> well, I'm going to sleep now; night all
[00:33:27] <boyd> myrkraverk: ^^ also, definitely https not http
[00:34:00] <myrkraverk> boyd: well, I'm interested in the API - not the gui tools ;P
[00:34:10] <boyd> myrkraverk: Fair enough
[00:34:39] <boyd> Ok, since we have a well-informed population here today, I have 2 random questions:
[00:34:52] <Tpent1> alanc: re your mail about 64 bit having to be twice as good as 32, .... do you have a volume control that goes to 11 to? ;)
[00:35:04] <alanc> heh
[00:35:05] <richlowe> myrkraverk: most of it's in $SRC/lib/libzfs, the bits you see prototypes for, but no implementation are in $SRC/common/zfs
[00:35:07] <Gman> alanc, ahh
[00:35:09] <boyd> 1) has anyonw heard if there is any work at all on supporting graphical boot (i.e. fall-back to text mode)
[00:35:23] * Gman rather interested to know what the gnome community make of our trusted patches
[00:35:33] * Gman suspects 'arse' will come into it somewhere
[00:35:41] <myrkraverk> richlowe: k ;)
[00:35:42] <boyd> I'd say they will make a boat... or a puppy
[00:36:01] <Tpent1> or a snake or a giraffe or a butterfly
[00:36:04] <boyd> Sorry... Movie joke..
[00:36:07] <boyd> Hi Tpenta
[00:36:08] <richlowe> or a pterodactyl.
[00:36:09] <Gman> boyd, graphical boot? yeah, the install group, no?
[00:36:15] <Gman> cayman project
[00:36:26] <alanc> boyd: 1) graphical boot is on the longer term todo list of the CoolStart project - mostly Jan & Liane, but it's a ways off still
[00:36:33] <richlowe> I wouldn't think Caiman and graphical boot were related.
[00:36:36] <Gman> ahh, ok
[00:36:39] <Gman> i thought they were
[00:36:40] <richlowe> since the semi-existing stuff pre-dates caiman.
[00:36:50] *** tmarble has quit IRC
[00:36:51] <boyd> Cool... thanks for the info guys
[00:36:58] * Tpent1 always thought cayman was a tax haven
[00:36:58] <Gman> oh, tmarble was around
[00:37:00] <Gman> nice.
[00:37:08] <boyd> I wondered what liane was doing... she seems to have been a bit quiet
[00:37:21] <Gman> proposing stupid projects like smf-doc? :)
[00:37:30] <boyd> Gman: now now :)
[00:37:36] <Gman> sorry, uncalled for
[00:37:41] <Gman> but it did seem particularly pointless.
[00:37:42] *** axisys has quit IRC
[00:37:46] * boyd nods
[00:37:53] <richlowe> Gman: you should ask stevel about the webapp project.
[00:38:04] <Gman> stevel, what about the webapp project?
[00:38:09] <boyd> hehe
[00:38:15] <richlowe> Gman: ... wasn't it you wondering what was happening with it?
[00:38:24] <Gman> yes, i was
[00:38:44] <Gman> but i don't expect stevel to come up with the cards all the time
[00:38:55] <boyd> Ok, question 2... does anyone off the top know what nics the U20M2 and U45 have, and is either of them currently supported under crossbow?
[00:39:00] <alanc> and when will we finally get the make-bugs.opensolaris.org-not-suck project?
[00:39:15] <Gman> doubtful that's going to happen anytime soon :/
[00:39:19] <jmcp_> boyd: nge
[00:39:26] <boyd> Hi jmcp_
[00:39:26] <jmcp_> dunno about crossbow though
[00:39:29] <jmcp_> boyd: hi
[00:39:30] * Gman notices that sch has gone quiet too
[00:39:37] <jmcp_> boyd: the u20m2 has two nge, the u20 only has one
[00:39:41] <Gman> hurray, beat up the tonic team again ;)
[00:39:43] <richlowe> Yeah, sch dried right up after the last argument.
[00:39:46] <boyd> Hmm.. I think not for xbow I think that's bde, ce and e1000g only ATM
[00:39:47] <richlowe> I guess I just have a way with people...
[00:39:58] *** nwf has joined #opensolaris
[00:40:03] <boyd> err s/bde/bge
[00:40:09] <Gman> richlowe, wouldn't worry about it too much - you're right most of the time
[00:40:12] <Gman> and they know that too
[00:40:20] <boyd> Anyone know about the U45?
[00:40:27] <richlowe> boyd: I thought it was nemo in general?
[00:40:36] <richlowe> (and I didn't think ce was nemoized, either)
[00:40:46] <boyd> richlowe: Yeah... you're right
[00:41:09] <boyd> untill universal nemoizationinficationality
[00:42:10] <Tpent1> ce is definitely absolutely positively NOT nemoized
[00:42:13] *** piwi has joined #opensolaris
[00:42:32] <boyd> .... which if course precedes network-stack-ksh93ification
[00:43:02] * boyd ducks
[00:43:03] *** piwi has quit IRC
[00:43:27] *** estibi has left #opensolaris
[00:43:32] *** slowhog has joined #opensolaris
[00:44:44] *** Mazon is now known as mazon
[00:45:22] <Gman> stevel, is there a repo for the starterkit? or do i have to download a monster tarball?
[00:47:37] <stevel> gman: do you mean the installation materials/html/docs/etc. for the starter kit?  or the final big isos?
[00:47:50] <stevel> well, in either case - it's just a big monster tarball unfortunately
[00:47:53] <stevel> the docs people put it together
[00:48:05] <Gman> yep, mostly enough to create a screenie to replace the current one on the project page
[00:48:24] <Gman> i can probably unpack a tarball remotely, and grab the docs though
[00:48:47] <stevel> yup
[00:49:18] <Gman> happen to know the internal location?
[00:49:48] <stevel> yeah, it lives on my machine at the moment
[00:49:55] <stevel> /net/donuthole.sfbay/export/opensolaris/starterkit.zip
[00:50:03] <Gman> sweet, thanks
[00:51:07] * Gman starts copying over
[00:53:49] *** mnowak has quit IRC
[00:54:41] *** _william_ has quit IRC
[00:59:20] *** mnowak has joined #opensolaris
[01:00:37] *** Cybernd has quit IRC
[01:02:42] <Gman> sahafeez, did you log 6487968 ?
[01:03:32] *** nwf has quit IRC
[01:03:43] <sahafeez> gman: i logged one today. did not look at the number. checking
[01:03:56] <Gman> probably yours
[01:04:03] <Gman> sounds like it's sparc specific alright
[01:04:12] <Gman> which is prompting me to actually boot up my sb100 :)
[01:04:29] <delewis> drag-and-drop?
[01:04:36] *** nwf has joined #opensolaris
[01:05:24] <sahafeez> hum. does not come up in the search btw. stevel was saying it was fixed in 2.16 gnome. but i did a fresh install of snv50 and it did not work. so i then did an install of vermillion over that and it did not work. also i tried with and without xinerama enabled
[01:05:39] <sahafeez> vermillion_52s
[01:05:46] <Gman> delewis, yep
[01:05:53] *** kupfer has joined #opensolaris
[01:05:53] *** ChanServ sets mode: +o kupfer
[01:06:00] <delewis> I'm able to reproduce it in b50 (and it was present in b48, as well)
[01:06:03] <Gman> sahafeez, you got all the hal bits?
[01:06:28] <sahafeez> hal bits? more info please
[01:06:29] <boyd> stevel: Am I correct in assuming that now that there is a repo there will not be any more complete bundles for hg? I ask since I have a location where http downloads of the bundle would be a much faster way to get close to up-to-date than pulling the repo via ssh
[01:06:43] * gisburn wishes he had a NIgara1 box at home for development.
[01:07:04] <boyd> Don't we all... except for the noise :)
[01:07:05] <Gman> sahafeez, hal = hardware abstraction layer, contains a daemon, part of the tamarack project
[01:07:18] <gdamore> are the T1's noisy?
[01:07:34] <stevel> boyd: http://www.opensolaris.org/jive/thread.jspa?threadID=16063&tstart=0
[01:07:37] <Gman> sahafeez, has an smf service system/hal:default
[01:07:40] <gisburn> gdamore: I don't care.
[01:07:47] <stevel> boyd: and yeah, that's the idea.
[01:07:54] <sahafeez> i installed what i said, b50 and v52s. is it in there?
[01:07:55] <dwc-> gdamore: yes, kinda
[01:08:00] <boyd> stevel: Thanks, I missed that post
[01:08:27] <gdamore> stevel++
[01:08:36] <stevel> boyd: i'll still publish the full Hg bundles for the build deliveries if that helps you any
[01:08:39] <gdamore> (wrt to trimming down deliveries)
[01:09:06] <stevel> (though, because of respins/backouts - if you try to unbundle that and then pull the repo on top of it - you may grow another head)
[01:09:24] <gdamore> gisburn: you should just take that 60 day free trial offer for Niagra systems. :-)
[01:09:37] <delewis> gdamore: it's not exactly "free"
[01:09:44] <gisburn> gdamore: <rant> I HATE TO WAIT TWO DAMN GODFORSAKEN DAYS TO COMPILE OS/NET ON MY DAMN GODFORSAKEN CRAPPY ultra5</rant> (sorry for the strong language but there is no other way to vent my wrath about 1) ctf*junk sucking-up CPU time and 2) stupid gcc shadow sucking up CPU time, too. DAMN IT)
[01:09:46] <boyd> stevel: I'm not sure that I'm enough to justify that... in fact, I not sure what the volumes are like... would publishing every 10th bundle or something make a difference?
[01:09:48] <delewis> you still have to have the necessary credit limit to receive one
[01:10:09] <gdamore> gisburn: WTF are you using shadow compiles?
[01:10:27] <stevel> boyd: well, sch has said he'd like to keep the build-Hg-bundles too - so it's not just you
[01:10:38] <sahafeez> Gman: so did i miss this hal?
[01:10:54] <dwc-> you could probably get a blade 100 pretty cheap
[01:11:06] <delewis> and a Blade 100 would suck the same amount as an Ultra 5.
[01:11:11] <Gman> sahafeez, svcs -a | grep hal
[01:11:13] <delewis> even a Blade 150.
[01:11:16] <Gman> does it say online?
[01:11:19] <gdamore> i do a lot of compiles on blade 100 and on 500.  it isn't that bad, as long as you aren't doing full builds
[01:11:20] <Gman> [just sanity checking, that's all]
[01:11:21] <dwc-> or if you're using gcc, a cross-compiler on a faster system
[01:11:28] <gisburn> gdamore: my university is "not interested" in sponsoring "... any more Sun-related 'adventures'" and my credit card doesn't have that much money left. And you need a valid credit card with enougth credit to get a niagara test box. Otherwise I would have begged to get one from kupfer  already
[01:11:31] <gdamore> cross compiler won't help
[01:11:47] <dwc-> delewis: experience has shown me to build on the blade 100s, not the ultra 5s around here
[01:11:54] <gdamore> cross compilation isn't practical with ON.  :-(
[01:12:04] <delewis> dwc-: on an NFS export it probably wouldn't be that bad
[01:12:14] <gdamore> the blade 100s have better IO (marginally), and typically faster CPUs.
[01:12:14] <gisburn> gdamore: I need shadow builds to catch gcc erors and full builds to catch errors in the l10n locale catalog generation.
[01:12:17] <delewis> but the SB100 still uses the same cmdk chipset that does not support DMA as the Ultra 5.
[01:12:28] <delewis> gdamore: I attribute the better IO to the faster CPU and faster disk.
[01:12:39] <gdamore> the PCI bus is different, IIRC.
[01:12:53] <delewis> it's still unified, I think.
[01:13:01] <sahafeez> Gman: that box is nuked right now. i was going to put Sol10ga on it.  i will reinstall 50+v52s and let you know tonight sometime
[01:13:05] <gdamore> stick the data on a PC-based NFS server. :-)
[01:13:12] <Gman> sahafeez, hrm, i've just drag and dropped on vermillion
[01:13:14] <Gman> and it's working fine
[01:13:25] <Gman> i'm on snv48
[01:13:34] <boyd> stevel: Well, If you change it, I'll use it. In the meantime I'll do what I have been doing... clone and make my own bundles on a box that's got faster net access
[01:13:38] *** swa has joined #opensolaris
[01:14:11] <gdamore> gisburn: if you want to pay for the shipping and power, I'm pretty sure I can get you a 4-way E450 clone on loan. :-)
[01:14:32] <gisburn> gdamore: you can't imagine how SICK I feed about waiting days for sparc build to complete.
[01:14:39] <sahafeez> well it did not work on 46, do not think i did 48. did not work on 49 or 50. i can make it happen everytime. there are a few tickets open for it and others hear have said that they have the problem
[01:14:41] <delewis> shipping to Germany for an E450 clone would be pretty high, I think :-)
[01:14:52] <delewis> gisburn: don't feel bad
[01:14:54] <gdamore> yeah, not cheap. :-)
[01:14:55] <sahafeez> what was the logic of using a ide chipset with no dma.
[01:14:57] <gisburn> gdamore: I guess it's more tha $100, right ?
[01:15:02] <delewis> I thought 12 hours on my SB1000 for an ON build was bad.
[01:15:05] <gdamore> NFI
[01:15:09] <gisburn> gdamore: and I guess I need an insureance, right ?
[01:15:14] <delewis> sahafeez: shitty IDE support in Solaris.
[01:15:25] <delewis> and no one has since bothered to add DMA support to cmdk
[01:15:33] <sahafeez> how parallel is the ON build? i have 4 cpus
[01:15:36] <gdamore> insurance wouldn't be too expensive.  and it is smalelr than a "real" E450.
[01:15:45] <gdamore> its a 2U rack mount server
[01:15:46] <sahafeez> the u5 and u10 were crap.
[01:15:56] <gisburn> delewis: note that the linux IDE driver for Ultra5/10 is faster than the solaris ones.
[01:16:03] <delewis> gisburn: of course it is.
[01:16:16] <gisburn> delewis: but... why ?
[01:16:21] <delewis> prior to Solaris 10, Solaris IDE support was horrific, as I said a few lines above.
[01:16:36] <delewis> gisburn: lack of interest in quality IDE drivers, I suppose.
[01:17:01] <delewis> afterall, the majority of systems Sun does sell are fibre-channel, SCSI, (and now) SATA.
[01:17:13] <gisburn> delewis: this is Solaris 11/B48 and IDE performance is still 60%-70% of the linux one.
[01:17:26] <delewis> gisburn: I haven't tested in awhile :-)
[01:17:27] <gdamore> heh.  i know about horror stories with FCAL drives if you go back a ways.
[01:17:33] * delewis owns mostly fibre-channel and SCSI systems
[01:17:38] <gdamore> it takes sun a while to get their storage drivers worked out, always.
[01:17:45] <delewis> with the exception of my Packard Bell router (which, ironically, uses cmdk without DMA, as well)
[01:17:48] <Gman> sahafeez, are you drag and dropping any particular thing?
[01:18:13] <sahafeez> Gman, anything
[01:18:28] <delewis> Gman: try draggin something in nautilus' file browser
[01:18:33] <delewis> dragging, rather.
[01:18:44] <Gman> works fine here
[01:18:53] <Gman> this is on sunblade100, with dual heads
[01:19:33] <Gman> tried draging and dropping to and from a file open dialog too, and that works fine
[01:19:44] <gdamore> delewis: how do you identify the cmdk chipset on sparc?
[01:20:11] <delewis> hmm, not sure. on x86 it shows up on boot and is therefore in the dmesg output
[01:20:42] <gdamore> cmdk is a different disk driver than the sparc driver.  i'm not entirely convinced you're right about cmdk on sparc
[01:22:19] <delewis> I'm certain the Ultra 5s have a CMD-646 in them
[01:22:23] <boyd> I thought it was always dad on sparc
[01:22:37] <boyd> Certainly is on a U5
[01:23:02] *** nwf has quit IRC
[01:23:25] <delewis> cmdk, instance #1
[01:23:35] <delewis> that's on my x86 system
[01:23:36] <sahafeez> Gman, i am going to go to the trouble of reinstalling this box and breaking it. what do you need from me to see what the issue is?
[01:23:41] <delewis> maybe they are different drivers
[01:23:47] <delewis> still, it's a limitation of the chipset
[01:24:09] <boyd> Yeah, there is a historical diference in the IDE drivers from sparc/x86
[01:24:10] <Gman> sahafeez, before you do that, others might be able to reproduce it here
[01:24:13] <Gman> anyone'?
[01:24:24] <delewis> reproduce what?
[01:24:28] <delewis> the drag-and-drop bug?
[01:24:44] <boyd> Yeah... I'll try... are there, like, steps?
[01:24:48] *** nwf has joined #opensolaris
[01:24:49] <Gman> delewis, yeah
[01:24:52] <boyd> (I have b49 here)
[01:24:53] <delewis> boyd: drag and drop a file in Nautilus
[01:25:02] <boyd> 'kay
[01:25:03] <Gman> boyd, do you have the vermillion bits?
[01:25:06] <delewis> Gman: I reproduced it in snv48 and snv50
[01:25:07] <sahafeez> on sparc?
[01:25:18] <sahafeez> delewis, thanks!
[01:25:23] <boyd> I have stock b49 on x86 :(
[01:25:27] <delewis> (I said that *way* earlier)
[01:25:42] <Gman> delewis, do you have the vermillion bits?
[01:25:56] <delewis> Gman: whatever was shipping in snv48 and snv50 is what I have
[01:26:00] <Gman> svcs -a | grep hal - does it show 'online'
[01:26:05] <Gman> delewis, ok, that's a no
[01:26:15] <delewis> online         Oct_26   svc:/system/hal:default
[01:26:16] <Gman> that's based on gnome2.14, rather than 2.16 that i have here
[01:26:19] <gdamore> i don't know about the ultra5s, but i'm checking my other USIIi and IIe based systems now.
[01:26:30] <boyd> So, we can't help?
[01:26:31] <Gman> thanks all
[01:26:53] <Gman> given we're so close to pushing back, i don't think we'll backport a fix
[01:27:04] <boyd> gdamore: You're looking for cmdk on sparc?
[01:27:36] * Gman checks the ChangeLog - i know padraig put back a fix on something recently
[01:27:39] <gdamore> no, i'm trying to determine the chipset that is used, to validate whether or not DMA is used
[01:28:02] <boyd> Ah
[01:28:12] <dwc-> I have an sb100 with dad
[01:28:19] <gdamore> SPARCLE500 (SB100-ish clone) uses an ALI M5229 UDMA IDE controller
[01:28:26] <dwc-> I think they turned off the ultra 5
[01:28:26] <boyd> gdamore: I have a U10 handy, if there's something you want me to do
[01:28:31] <dwc-> and the ultra 1s
[01:28:32] <gdamore> just a sec.
[01:29:11] <gdamore> the SB100 also has an ALI M5229 UDMA controller.
[01:29:42] <gdamore> boyd: what I need is the PCI vendor and device ids for the 'ide' node in the device tree.
[01:30:05] <gdamore> you can get this by doing prtconf -vp, and search for the one that says "name:  'ide'"
[01:30:17] <gdamore> then look at device-id and vendor-id fields
[01:30:33] <gdamore> hmm... i suppose I could just power on my own U10. :-)
[01:30:58] <boyd> pci1095,646
[01:32:23] <gdamore> that's a cmdtech 646.  so delewis is right for u10 but not sb100
[01:32:56] <delewis> :-(
[01:33:01] *** Fish- has quit IRC
[01:33:08] <gdamore> interestingly enough, netbsd on my Ultra 10 seems to enable DMA on that chip
[01:33:52] <delewis> I believe some of the later Ultra 5s and Ultra 10s shipped a CMD-646 that supported DMA
[01:34:00] <delewis> I've heard the earlier models did not
[01:34:19] <delewis> regardless, I'm fairly sure the Solaris CMD-646 driver never enabled DMA, regardless.
[01:34:32] <gdamore> i'd have guessed my U10 was an early model, but i could be mistaken
[01:34:43] * boyd looks at some other boxes to be sure
[01:34:46] <dwc-> now that I think about it, I've gotten DMA errors from the sb100s
[01:34:55] <dwc-> so those surely have it
[01:35:31] <gdamore> absolutely.  its "udma", although i'm not sure what exactly is Ultra about it. :-)  I assume that Ultra DMA is some IDE/EIDE thingy
[01:36:37] *** uebayasi has joined #opensolaris
[01:36:49] <twincest> i think DMA IDE is a very old technology, and UDMA is the useful one
[01:37:26] <boyd> All the U5's I have available have the same IDE controller
[01:38:06] <gisburn> oh
[01:38:09] <gisburn> kupfer: Hi!
[01:38:10] *** stevel has quit IRC
[01:38:17] <kupfer> hi Roland!
[01:38:17] <dwc-> udma came in around 33mhz, but there was dma before that
[01:38:22] <dwc-> I think
[01:38:30] <dwc-> err not 33mhz, 33mb/s
[01:38:39] <gisburn> kupfer: Do you have an UltraSPARC-3 non-Cu machine around (e..g. 600MHz US-3) ??
[01:38:49] <gisburn> s/\?\?/\?/
[01:38:54] <gdamore> it would be pretty easy to validate this.  try doing some disk IO.  If it saturates your CPU, then you are using PIO and yech!  If you have cycles left, then you must be using DMA. :-)
[01:38:55] <kupfer> beats me.
[01:39:03] <gisburn> kupfer: ?!
[01:39:11] <gdamore> US3 started at 750MHz, IIRC.
[01:39:14] <delewis> gisburn: I do
[01:39:17] <delewis> gdamore: no
[01:39:18] <delewis> 600MHz
[01:39:30] <delewis> CU variants started at 900MHz
[01:39:31] <gdamore> really?  I have never seen/heard of a 600MHz US3.
[01:39:32] <kupfer> gisburn, sorry, "beats me" means "I have no idea"
[01:39:40] <delewis> gdamore: they are very rare
[01:39:47] <delewis> Auralis has a 600 and 750 in her SB1000
[01:40:07] <gdamore> did sun ever actually sell 600MHz cpus?
[01:40:22] <gdamore> i seem to recall the 900s came out right on the heels of the 750s.
[01:40:41] <gdamore> then we had to "wait" for US3i to get over 1GHz.
[01:40:54] <gisburn> delewis: I guess the last commit in http://polaris.blastwave.org/log/on/branches/ksh93/gisburn/prototype004/?verbose=on is... uhm... not good for you... :-)
[01:41:16] <delewis> gdamore: http://sunsolve.sun.com/handbook_pub/Systems/SunBlade1000/spec.html
[01:41:47] <gisburn> delewis: but this will only be a problem if you have more than 16 64k pages for a specific process and use 4M pages in the same process.
[01:41:53] <delewis> gdamore: there are two 900 variants -- cu and non-cu
[01:41:58] <delewis> the CU variant came quite a bit later, IIRC.
[01:41:59] <gdamore> ah yes.
[01:42:00] <Stric>   The sparcv9 processor operates at 600 MHz,
[01:42:06] <Stric> in a blade1000 ;)
[01:42:20] <delewis> gisburn: :-(
[01:42:27] <gdamore> you gotta be kidding me.  we are doing special page mapping for ksh93!?!
[01:43:04] <gisburn> delewis: actually I do not belive this is harmfull US-3 CPus prior Cu
[01:43:12] <gisburn> gdamore: erm, yes.
[01:43:20] <gisburn> gdamore: but this is IMO harmless.
[01:43:28] <gisburn> gdamore: read the mailinglist :-)
[01:43:54] * delewis doubts gdamore uses ksh93
[01:43:55] <sahafeez> Gman so do you need me to reinstall this box?
[01:43:56] <gdamore> i'm just horrified about the amount of engineering work going to optimize long running shell scripts.
[01:43:57] *** kupfer has quit IRC
[01:44:19] *** jsubl2 has joined #opensolaris
[01:44:22] * gdamore gdamore believes long running programs should not be shell scripts.
[01:44:22] <Gman> sahafeez, not for the moment, i'll see if i can get padraig to take another look
[01:44:34] <gisburn> gdamore: the engineering work consists of benchmarking, a compiler option and some other minor tweaks.
[01:44:35] * delewis will be able to say someday he remembered when `the great ksh93 migration' began.
[01:44:46] <gdamore> hahah.
[01:45:06] <sahafeez> Gman, thanks this is driving me nuts. it makes me want to try linux on the box (shutters)
[01:45:12] <alanc> but when will it end?
[01:45:16] <Error_404> feh, ksh
[01:45:19] <Error_404> pointless crap
[01:45:20] <delewis> sahafeez: no, you don't want to try Linux on SPARC.
[01:45:25] <delewis> especially anything that's even relatively modern.
[01:45:26] <gdamore> alanc++
[01:45:33] <Error_404> sahafeez: no, you really don't... linux/sparc sucks
[01:45:34] <Gman> sahafeez, no worries - i'm pretty certain it's fixed in our latest stuff, so that's likely to be at least b53
[01:45:40] <gisburn> gdamore: read http://mail.opensolaris.org/pipermail/ksh93-integration-discuss/2006-October/001631.html
[01:45:52] <Error_404> it kpanicked all over the place for me
[01:45:56] * delewis watches Error_404 be burnt alive
[01:46:01] <gdamore> gisburn: work on _integrating_ ksh93 into opensolaris first, I think.  the performance tweaks should come later, IMO.
[01:46:02] <dwc-> linux/sparc is good for a gateway... nobody tries sparc exploits on a linux machine
[01:46:26] <Error_404> you could also just use a more secure OS
[01:46:31] <delewis> gdamore: I'm sure PSARC won't like it if ksh93 doesn't meet ksh88's performance.
[01:46:37] <delewis> and thus, won't allow it to be putback.
[01:46:59] <delewis> talk about a catch-22.
[01:47:00] <gdamore> delewis: yes, but if you have to do MMU tricks to get there, then something is wrong, IMO.
[01:47:05] <dwc-> Error_404: sure, but what's more likely to have a 0-day... solaris on sparc/x86? linux on x86? linux on sparc? gnu/hurd on anything ;)
[01:47:06] <gisburn> gdamore: sorry but I am STARVING from doing anything. I really like to work in the integration but the truth is: I have to wait for others like Glenn&&Sun.
[01:47:07] <delewis> gdamore: :-)
[01:47:38] <delewis> dwc-: you forgot Plan 9
[01:47:41] <gisburn> gdamore: ksh93 already outperforms the solaris ksh
[01:47:42] <gdamore> gisburn: the more you change the code, the more QA has to keep regression testing, the longer it will take to integrate.
[01:48:19] <gisburn> gdamore: which regression testing do you mean ? ksh93 has no prior history in OS/Net.
[01:48:30] <dwc-> yea, you could be user #6 of plan 9 ;)
[01:48:31] <delewis> gisburn: it will once it is integrated
[01:48:37] <gdamore> well, initial acceptance testing then.
[01:48:40] <delewis> dwc-: it isn't *that* bad, you know. :-)
[01:48:44] <gisburn> delewis: yes, yes
[01:48:49] <gisburn> gdamore: ?!
[01:49:05] <gisburn> gdamore: I guess we did that thing to death, that's why we have /etc/ksh.kshrc&co.
[01:49:14] <gdamore> but you keep tweaking!
[01:49:16] <gisburn> gdamore: to get happy users
[01:49:35] <gisburn> gdamore: what else should I do ? I am waiting for the next tarball from Glenn...
[01:49:43] * delewis would like to find a happy ksh93 user
[01:49:49] * delewis ducks
[01:49:58] <gdamore> i suspect your users will be happier if you get it _done_
[01:49:59] <Tpent1> delewis: why do you think PSARC would get upset about a perf issiu? PSARCs brief is interfaces
[01:50:05] <gisburn> I am starving from doing any work. I'd like to do something usefull but the options are limited.
[01:50:23] <gdamore> work on other /usr/xpg4/bin stuff? :-)
[01:50:24] <delewis> Tpent1: good point. Who would get upset about a performance issue/
[01:50:26] <gisburn> Tpent1: s/PSARC/RTI approval gremium overloards/
[01:50:30] <twincest> Tpenta: who would usually be the one to obje... yes, what he said
[01:50:33] *** simfordWFH has joined #OpenSolaris
[01:50:45] <delewis> s/\//\?/
[01:50:57] <Tpent1> I'm not actually sure; but you could be certain of getting a bug logged against it right quick
[01:51:03] <delewis> hehe
[01:51:08] <gisburn> Tpent1: ?!
[01:51:09] <gdamore> the goal of ksh93 is to replace ksh88, so ksh93 had better be better. :-)
[01:51:35] <delewis> "ksh93: the answer to life"
[01:51:36] <gisburn> gdamore: we already kick ksh88's b*tt when it comes to i18n support.
[01:51:37] *** Tpenta has quit IRC
[01:51:44] *** Tpent1 is now known as Tpenta
[01:51:51] *** ChanServ sets mode: +o Tpenta
[01:51:52] <gisburn> delewis: 48
[01:52:02] *** RaD|Tz has quit IRC
[01:52:22] <Tpenta> how about "trusting to the integrity of the project team" ???
[01:52:27] <gisburn> delewis: and don't tell me you don't know what "48" means...
[01:52:31] <gisburn> Tpenta: what ?!
[01:52:45] <gisburn> Tpenta: what do you mean ?
[01:53:04] <delewis> gisburn: is that some variant of 69?
[01:53:04] * gisburn looks at Tpenta ...
[01:53:09] <Tpenta> answering delewis question about who should be responsible for ensuring the performance is what it should be
[01:53:24] <twincest> i don't thihnk he was criticising
[01:53:47] <gisburn> erm...
[01:53:50] <Tpenta> believe it or not it was actually a form of compliment. I believe that there are many things for which we need to trust the project teams
[01:53:53] <gisburn> which question was that again...
[01:53:59] <Gman> i'm sure ksh93 sucks though :)
[01:54:10] *** Tpent1 has joined #opensolaris
[01:54:16] * gisburn growls
[01:54:23] * gisburn looks at Gman
[01:54:44] <Tpenta> (11:50:12) delewis: Tpent1: good point. Who would get upset about a performance issue
[01:55:02] <gisburn> uhm
[01:55:15] <gisburn> I would be upset about a performance issue in ksh93.
[01:55:21] <gdamore> ksh93 "long running script performance" is a non-issue.
[01:55:34] <gdamore> far, far more important is the performance of short lived scripts, i think.
[01:55:42] <twincest> agree
[01:55:44] <gdamore> at least if ksh93 is ever going to become a default shell.
[01:55:56] <gisburn> gdamore: even that is addressed via setting the default stack page size to 64k
[01:56:07] <gisburn> gdamore: read my posting about builtin commands in libcmd
[01:56:11] <gdamore> that's not what your bug report says.
[01:56:30] <gdamore> and before changing stuff for performance optimizations, you really should have some benchmarking to justify the optimization.
[01:56:41] <gisburn> gdamore: no, but I tweaked libcmd to allocate more memory from stack instead of global vars, therefore using 64k pages
[01:56:43] <twincest> uhm, 64K default pages for a shell?
[01:57:13] <gdamore> it _shouldn't_ matter, because it is "just virtual memory", but yeah, it surprises me too
[01:57:51] <gdamore> in any case, i think this kind of changed should be backed by real performance tests, not just swags and guesses.
[01:57:59] <gisburn> gdamore: it matters because you don't take the memory access pattern into account.
[01:58:10] <gisburn> gdamore: I did real performance testing.
[01:58:22] <gisburn> gdamore: but only on a Ultra5 and a Blade1500
[01:58:50] <gisburn> gdamore: do you really think I am doing such changes ina blind flight ?
[01:58:51] <gdamore> how much did you save?
[01:58:53] <twincest> gis: did you try using a block allocator or caching free?
[01:58:56] * gisburn kicks gdamore
[01:59:02] <gdamore> haha.
[01:59:07] <gdamore> can i have another, sir?
[01:59:09] <twincest> that seems like a better solution than changing the page size
[01:59:35] <gisburn> twincest: ksh93 uses the libast allocator which does caching and pooling.
[01:59:40] <gdamore> the page size change smells to me like microoptimization, and i'm not convinced it is worthwhile right now.
[01:59:59] <gdamore> otherwise, why not change _all_ programs to use it?
[02:00:10] *** mnowak has quit IRC
[02:00:14] <gisburn> gdamore: it's no microoptimisation when your script juggles around 20MB of data in associatve arrays.
[02:00:27] <gdamore> wtf is your scripting doing juggling around 20MB of data? :-)
[02:00:28] <twincest> oh, i thought we were talking about scripts that people actually use
[02:01:05] <gdamore> i guess this is just another case of trying to make ksh93 into a programming language instead of script parser. :-)
[02:01:11] <gisburn> gdamore: it depends on the usage. in theory it may be a good idea to make 64k the default stack size in solaris/sparc.
[02:01:12] <dwc-> I thought everyone does mandelbrot in ksh
[02:01:24] <gdamore> (so when is the emacs port to ksh93 scheduled?)
[02:01:30] <Stric> I have met mandelbrot :)
[02:02:18] <gisburn> gdamore: at least ksh93 uses setjmp/longjmp and that causes MANY page hits for 8k pages but only 3-4 for 64k pages in the stack.
[02:02:27] <Tpenta> way cool alanc: PSARC/2006/609 Xserver provider for DTrace
[02:02:30] <Godsey> is there any sort of waiting list for hardware I can put my email address on for notification?
[02:02:38] <Godsey> hardware support that is :)
[02:02:57] <gdamore> hmm.... is that an attribute of how ksh93 is designed?
[02:03:13] <gisburn> gdamore: what ? setjump/longjmp ?
[02:03:19] <gdamore> Godsey: I want to be on a waiting list for hardware.  Not support, just hardware. :-)
[02:03:34] <gdamore> yes, the heavy usage of setjmp/longjmp
[02:03:38] <Godsey> I fubared
[02:03:39] * Tpenta wants to be on the waiting list for a u20m2
[02:03:40] <twincest> gda: does the free hardware foundation still exist? :)
[02:03:46] <gdamore> :-)
[02:03:47] <gisburn> gdamore: this is needed for supporting aborts from $( ... ) - remeber ksh93 does not create subprocesses for this kind unless & is used, too.
[02:04:00] <Godsey> my dell has the MPT1068 not MPT1068E which is supported
[02:04:07] <twincest> or if the $() construct forks
[02:04:20] <Godsey> er SAS1068
[02:04:30] <gisburn> twincest: forking for $( ... ) is hirrible slow.
[02:04:34] <gdamore> so the ksh93 $() improved implementation is a major optimization, agreed.  But is it now appropriate to do microoptimization?
[02:05:00] <gisburn> gdamore: depending on the type of optimisation, yes
[02:05:03] <gdamore> btw, can you cite any actual performance gains as a %?
[02:05:06] <Tpenta> many sins are committed in the name of optimisation in the absence of full analysis
[02:05:11] <gdamore> Tpenta++
[02:05:14] <twincest> gis: but necessarywhen you want to use it to run a program
[02:05:26] <gisburn> gdamore: the global vars are in the way to make libshell threadsafe to I move them to the stack.
[02:05:40] *** mnowak has joined #opensolaris
[02:05:41] <Godsey> I'm installing CentOS then vmware server
[02:05:51] <Godsey> I hope I can install solaris 6/06 under that :)
[02:05:56] <gisburn> gdamore: and that way we get a cheap way to map these data with 64k pages via making the stack mapped with 64k pages.
[02:06:12] <gdamore> libshell being threadsafe is a nicety, but if it forces you to change MMU settings, then maybe its not the right solution.  does it really _need_ to be threadsafe (at first blush?)
[02:06:22] <gisburn> gdamore: depending on the script it'sin the range of 0%-8% with the current patches in my tree.
[02:06:30] <twincest> maybe he's going to implement forking builtins as threads
[02:06:39] <gdamore> hehe.
[02:07:12] <twincest> (i'd rather see that fixed in the kernel, if it's so slow..)
[02:07:24] <gdamore> which, forking?
[02:07:25] <gisburn> gdamore: the "threadsafe libshell" project is independant from this part except that they share some common obstaches (e.g. global vars)
[02:07:34] <twincest> gda: this in paricular: < gisburn> twincest: forking for $( ... ) is hirrible slow.
[02:07:51] <gdamore> well, fork() is slow.  compared to not forking, anyway.
[02:08:12] <twincest> compared to thr_create, i'm sure
[02:08:18] <twincest> but does it have to be? :)
[02:08:22] <gdamore> esp. if you don't need (or can't benefit from) parallel execution (e.g. you're just going to wait())
[02:08:38] <twincest> vfork ;-)
[02:08:49] <gdamore> even thr_create is probably slow compared to just setjmp/longjmp
[02:09:04] <twincest> yes but setjmp/longjmp can't implment $()&
[02:09:05] <gdamore> (given you don't need actual parallel execution)
[02:09:18] <gdamore> true.
[02:09:27] <twincest> well "command & wait" is a pointless construct, isn't it?
[02:09:38] * Godsey should have got the damn 2200 M2
[02:10:04] <gdamore> no, but if $( some test )  is not pointless
[02:10:13] <twincest> but that doesn't fork, according to gisburn
[02:10:47] <gdamore> right.  but what he was saying, i think, was that the fact that it doesn't fork is one of the new ksh93 improvements.
[02:10:50] <gisburn> twincest: using ( mycmd )  # does not fork() either unless you really start a subproicess via &
[02:11:12] <twincest> i think i lost my point somewhere
[02:11:19] <gdamore> haha.
[02:11:30] <twincest> i was saying thatfor when it does fork, i'd rather see the fork speed increased than add threading to ksh
[02:11:39] * gisburn sends twincest some "...........................................................................................................................................................................................................................................................................................................................................................................................
[02:11:41] * gisburn ................................................................................................................................................................................................................................................................................................................................................................................................................
[02:11:42] * gisburn ................................................................................................................................................................................................................................................................................................................"
[02:12:09] * gdamore smacks gisburn with polka-dot covered komodo dragon
[02:12:21] <gisburn> twincest: I was saying to make libshell threadsafe. Threading support is somethung different.
[02:12:36] <twincest> i know
[02:12:53] <twincest> i was just saying, humourously, that you were going to fix $() forks by turning them into threads
[02:13:01] <gisburn> nah
[02:13:02] <gdamore> gisburn: i suggest, that at this point you avoid any gratuituous changes to ksh93, and only deal with bugs, at least until after it integrates
[02:13:34] <twincest> harder to change things after it integrates though..
[02:13:41] <gdamore> not necessarily.
[02:13:41] * gisburn prays for another alpha source tarball from upstream
[02:13:48] * gisburn is starving.
[02:13:53] <gdamore> i think perfect is the enemy of the good in this case.
[02:14:00] <twincest> how can ksh_93_ be alpha?
[02:14:10] <gdamore> gisburn, go work on some other bugs.  :-)
[02:14:31] <gdamore> see, for example, my most recent stuff to fix xargs and id.  trivial stupid stuff, but someone should still do them.
[02:14:48] <gdamore> twincest: 2093.
[02:14:49] <twincest> i really need to install solaris somewhere and do the SCA thing
[02:14:56] <gisburn> twincest: ksh93a, b, c, d, ... q, r, s, are the versions. ksh93s- is alpha, ksh93s (no +/-) is stable and ksh93s+ is from the stable branch
[02:15:01] <gdamore> or is it 2193?
[02:15:16] <twincest> gis: when does the date change?
[02:15:17] <gisburn> gdamore: ksh93 has "id" as a builtin command.
[02:15:26] <gisburn> twincest: never
[02:15:35] <gdamore> that doesn't mean it can replace /usr/bin/id.
[02:15:36] <gisburn> twincest: ksh93 is ksh, spec=1993
[02:15:37] <twincest> so ksh93 doesn't exist, and we're still on ksh88?
[02:15:47] <twincest> okay, so the date changes when the spec changes
[02:15:56] <gisburn> twincest: yes
[02:16:03] <gisburn> gdamore: ?!
[02:16:06] <gdamore>  id has non-posix behavior that can't easily change
[02:16:18] <gisburn> gdamore: the "id" builtin in ksh93 implements POSIX behaviour.
[02:16:30] <gisburn> gdamore: in theory we could use the builtin command
[02:16:38] <gdamore> yes, but /usr/bin/id has behavior that is _different_ from the posix behavior
[02:16:51] <gdamore> and you can't change that behavior without pissing off existing sun customers.
[02:16:56] <gisburn> gdamore: didn't your asc fasttrack fix that ?
[02:16:57] *** uebayasi has quit IRC
[02:17:04] <gdamore> you can add posix semantics, but you can't change existing semantics
[02:17:15] <gisburn> grumpf
[02:17:24] <gdamore> no, my fast track _added_ posix semantics.  but unfortuantely /usr/bin/id has to remain separate from /usr/xpg4/bin/id
[02:17:26] <boyd> Isn't that what /usr/xpg? is for?
[02:17:35] <boyd> Ah, quite
[02:18:08] <gdamore> you are going to find a bunch things that can't just replace /usr/bin, because the ksh93 builtin does not provide _identical_ (or at least 100% compatible SunOS behavior)
[02:18:53] <gdamore> i'm also not entirely sure I like the idea of forking off a large shell just to do something that a 14k ish app can do.  i think the shell may cost more.
[02:19:06] <gisburn> gdamore: erm.
[02:19:07] <gdamore> but this is the thing that performance analysis is supposed to answer.
[02:19:14] <twincest> not if you have a ksh93d running all the time that can run builtins via a door
[02:19:15] <twincest> ;-D
[02:19:22] * gisburn thinks gdamore doesn't know how builtins work.
[02:19:35] <gisburn> gdamore: you can call the builtin without starting ksh93
[02:19:46] * gisburn kicks gdamore
[02:19:49] <gdamore> but you still have to exec ksh93, right?
[02:19:52] <gisburn> no
[02:19:54] <gisburn> why ?
[02:20:11] <gdamore> how do you call a builtin from another program?
[02:20:22] <gdamore> remember, ksh93 will not necesarily be the parent process
[02:20:23] <Tpenta> I imagine that would be libcmd?
[02:20:26] <gisburn> gdamore: you just need a binary wrapper which calls the |b_*(int ac, char **av, Shell_t*context);| function
[02:20:43] <gisburn> where * is the name of the bulitin
[02:21:07] <gdamore> okay, so you have main() { b_basename(...); } or somesuch.
[02:21:11] <gisburn> yes
[02:21:38] <gdamore> hmmm... this means mmaping libshell.  not sure about whether that is a net win compared to a small application or not.
[02:21:56] <gisburn> gdamore: no, you only need libshell for builtins which are in libshell
[02:22:05] <gisburn> gdamore: you need libcmd for buildins in libcmd
[02:22:11] <Tpenta> gdamore: you map it, but you would not necessarily page it all in
[02:22:26] <richlowe> gdamore: depends on the application, too.
[02:22:35] <richlowe> true(1)? Well, no...
[02:22:39] <gdamore> this feels like "swagging" again.
[02:22:46] <gisburn> swagging=?
[02:22:52] <Tpenta> then again, as shared text, the odds would be that it would already be paged in
[02:22:53] <gdamore> guesswork, educated or otherwise
[02:23:00] <richlowe> Tpenta: I'm not sure I agree there.
[02:23:09] <twincest> gda: in many cases the parent shell will provide an equivalent builtin, at least if /bin/sh becomes ksh
[02:23:15] <gdamore> i think this has to be benchmarked to be sure.
[02:23:16] <richlowe> Tpenta: bits maybe because of the wordexp() bug, otherwise you're making guesses at the frequency of use of ksh.
[02:23:27] <Tpenta> let me rephrase
[02:23:29] <gdamore> yes, but there are many programs that fork()/exec() that are not the shell.
[02:23:38] <twincest> gda: to call trivial builtins?
[02:23:39] <richlowe> gdamore: and libc`wordexp ;)
[02:24:02] <twincest> i don't think i've ever forked to call basename
[02:24:05] <gdamore> sometimes, yes.  :-)
[02:24:12] <gdamore> have a Makefile that does it?
[02:24:22] <richlowe> especially back when basename was a script.
[02:24:24] <twincest> my makefiles invoke C++ compilers :)
[02:24:25] <Tpenta> the probability of the portions of the library being actually already in memory for the function that the wrapper calls would be largely increased over having the portions of the discrete application already being in memory
[02:24:32] *** foldingstock has quit IRC
[02:24:36] <Tpenta> better?
[02:24:40] <gisburn> yes
[02:24:45] <gdamore> Makefiles do exec a lot of other programs, some of them "trivial"
[02:24:55] <richlowe> Tpenta: I was *solely* advocating giving this more thought than, "It's possible, and it maybe good"
[02:25:05] <Tpenta> fair enough
[02:25:22] <gdamore> richlowe: I think we may be in violent agreement.
[02:25:25] *** unixconsole has joined #opensolaris
[02:25:27] <Tpenta> i'd also want to see some hard data, but the possibilities are, .... "interesting"
[02:25:28] <gisburn> gdamore: no, make executes the shell which then exec()'s the requested command.
[02:25:39] <richlowe> I spent the majority of my day getting to the point pcred no longer thinks my shell is suid.
[02:25:47] <richlowe> I'm not exactly in the best of moods :)
[02:25:52] <gisburn> Guys, you get your toy soon enougth. You can the benchmark it to death.
[02:25:57] <gdamore> gisburn: i don't think that is necessarily true.
[02:26:06] <gdamore> (the bit about make, I mean)
[02:26:54] <gdamore> I'd rather have it delivered without microoptimizations, gisburn, was my only statement.
[02:27:15] <gisburn> gdamore: 8% are IMHO not a microoptimisation.
[02:27:30] <gdamore> MMU tweaks are by default microoptimization. :-)
[02:27:37] <gisburn> gdamore: and we finally get an application to exploit the 64k pages by default... :-)
[02:27:56] <gdamore> if you need MMU tweaks to get 8%, then something is wrong with either the defualt MMU setup, or with how you are using your data.
[02:28:02] <gisburn> gdamore: it's the same category of microoptimisation as the isaexec() thing you complained about.
[02:28:15] <Tpenta> for the most part I would argue that shell optimisations really won't give you a lot (obviously there are exceptions). Most things that the shell does are short lived. Do you *really* care if your script takes 20 or 40ms to run?
[02:28:24] <gdamore> isaexec run over lots and lots of shell scripts in loops is not microoptimization
[02:28:51] <gdamore> if that script runs in a tight loop i might.  but that's not the 8% case, I bet.
[02:28:56] <gisburn> gdamore: that was a commen about 32bit vs. 64bit in our last discussion.
[02:29:08] <Tpenta> in the general case that is; obviously if your script is executed 10 million times, then yea you probably care (and perhaps you should not have written a shellscript)
[02:29:32] <gisburn> gdamore: the 8% were peak improvement fro a testcase which runs quicksort on an associative array with 20MB data inside.
[02:29:48] *** sparc-kly_ has joined #opensolaris
[02:29:50] <gdamore> that, is exactly what i mean by "non default case"
[02:30:02] <gisburn> gdamore: I know.
[02:30:45] <gdamore> i'm far more concerned about the cumulative impact of 8 gazillion scripts that already exist and are designed around short lifetimes and memory sizes.
[02:30:46] <gisburn> Tpenta: I know. We could debate over the 64k default pag size for heap. however the 64k default stack size still makes sense in any case.
[02:31:26] <gisburn> gdamore: BTW: the quicksort() is a builtin command.
[02:31:30] <richlowe> You'd have to give a better reason than "Because", preferably with measurements.
[02:31:59] * gisburn prays to satan...
[02:32:17] <gdamore> should have known it was builtin.
[02:34:17] * gisburn consideres to spam opensolaris-discuss@ with some pointless benchmarks.
[02:34:28] <gdamore> to compare.  forks of the shell for short lived purposes (e.g. in make) are critical, and widely used.  adding 20 msec to the run time of /bin/true will probably add hours to the time it takes to build Solaris.  by comparision, saving 8% on quicksort in ksh93 is unlikely to be noticed by anyone except very dedicated ksh hackers (and maybe pacman/GNAW players)
[02:34:29] <Gman> richlowe, you see my mail on website-discuss? :)
[02:35:00] *** sommerfeld has quit IRC
[02:35:10] <richlowe> Gman: not yet.
[02:35:17] <gdamore> gisburn: as I said already, if you are bored, go work on something else until ksh93 integrates.  it needs the dust to settle, IMO, without constant tweaking, before it integrates, anyway.
[02:35:21] <richlowe> Gman: when/where?
[02:35:33] <Gman> richlowe, probably hasn't hit the archives yet
[02:35:35] <gisburn> gdamore: how often should I explain that there are scripts which do more than these things and have large datasets ?
[02:35:48] <richlowe> Gman: I'm actually on website-discuss. :)
[02:36:07] <richlowe> Gman: ah, I see it now.
[02:37:10] <gdamore> i agree that these scripts exist.  i don't think they are "typical".  isaexec will hurt far more users than the 8% spent on 8k pages will, I think.
[02:38:32] <gdamore> i also _rather strongly_ suspect that folks doing "real programming(!?)" in ksh93 are already accustomed to downloading and building their own version of it, and are even less likely to be impacted than users of the "default ksh" :-)
[02:39:03] <sahafeez> programing in a shell is just wrong.
[02:39:08] <gdamore> anyway, time to go now.  I need to spend some time with the family and stop worrying about a shell that i might never use. :-)
[02:39:15] <gdamore> sahafeez++
[02:39:27] <richlowe> Quite true, but people do it.
[02:39:32] <richlowe> ... take a look at wx, sometime. :)
[02:39:35] <richlowe> and that's "simple" :)
[02:39:38] <gisburn> gdamore: which point is it to have a 64bit OS which can't use it's full abilities just because the binaries are limited to 32bit ? THat is just wrong.
[02:39:51] <richlowe> limited to what now?
[02:40:01] <gisburn> richlowe: 2GB data.
[02:40:09] <gdamore> gisburn is complaining that he wants to create > 2GB arrays in ksh93 scripts.
[02:40:12] <richlowe> ... are we still talking about ksh?
[02:40:15] <gdamore> which is just effing insane.
[02:40:28] <richlowe> Yeah, I have no experience in that area, so I'll keep out.
[02:40:30] <gdamore> to do this, he will force an extra exec on _all_ users of ksh.
[02:40:31] <gisburn> gdamore: people already requested a 64bit perl
[02:40:36] <delewis> I could think of more vivid adjectives to describe it ;-)
[02:40:38] <richlowe> I agree with gdamore for what it's worth, but that means I've never actually done it...
[02:40:52] <dwc-> yea, but most people munge way more data in perl than they do in ksh
[02:40:53] <delewis> gisburn: for a completely different reason
[02:40:55] <gdamore> (i.e. to isaexec a 64-bit ksh93 or 32-bit appropriately)
[02:41:19] <delewis> a 64-bit Perl is required access the 64-bit libraries of a database, like Oracle, to connect a 64-bit database isntance
[02:41:22] <gisburn> dwc-: I have seen biosh processes with more than 8GB heap... :-)
[02:41:25] <richlowe> people have requested a 64bit perl, and got told "No".
[02:41:28] <gdamore> perl is not the "default short-lived shell" though.  who cares if perl startup takes 10ms longer?
[02:41:30] <richlowe> for reasons I'm still not sure I follow.
[02:41:43] <delewis> richlowe: yes
[02:41:44] <richlowe> the comments the CRs are closed with, and what delewis just said, disagree entirely.
[02:41:45] <delewis> I've requested it before
[02:41:46] <dwc-> can't the 64-bit users just set a path to include $PATH/$bits/
[02:41:49] <delewis> as have 20 people before me
[02:41:50] <dwc-> at the beginning of their path or something?
[02:42:04] <gdamore> dwc:- that would be fine with me.
[02:42:11] <richlowe> delewis: did $VENDOR stop shipping 32bit client libs?
[02:42:24] <gdamore> i don't object to shipping a 64-bit ksh.  i object to wrapping all execs of ksh with isaexec.
[02:42:45] <gisburn> gdamore: does it matter on 2 1GHz machine ?
[02:42:50] <gdamore> ABSOLUTELY!
[02:42:53] <gisburn> s/2//
[02:43:06] <delewis> richlowe: you can access a 64-bit database instance with 32-bit libs, yes, but it is nice to have 64-bit libs on some occasions, especially when you're working with a lot of data.
[02:43:07] <Stric> Why not ship a 64 bit only?
[02:43:10] <gdamore> you already complain about make taking a long time.  adding 10 msec will _kill_ solaris compilation time.
[02:43:12] <delewis> even IBM bothers to ship a 64-bit Perl
[02:43:17] <delewis> exactly for database reasons
[02:43:31] <dwc-> sun already has /usr/lib/64
[02:43:34] <dwc-> just add a /usr/bin/64
[02:43:39] <richlowe> gdamore: for what it's worth, I'm currently unable to measure the difference isaexec makes.
[02:43:45] <richlowe> I suspect you'd have to loop over a long long time, then average.
[02:43:56] <gdamore> just like ... er... make does.....
[02:43:57] <dwc-> richlowe: what? no dtrace script to tell you how many ns it takes?
[02:44:30] <richlowe> dwc-: /usr/bin/$64bitarchname
[02:44:34] <richlowe> that's how isaexec works. :)
[02:44:41] <richlowe> there's no reason you can't run stuff straight out of there.
[02:44:42] <gdamore> i can tell you that eliminating a simple call to "pwd" in Makefiles saved a _lot_  of time in the build.
[02:45:00] <gdamore> right, /usr/bin/sparcv9/ksh93
[02:45:03] <richlowe> gdamore: yeah, I was trying to get a sane measure without sucking up a bunch of my time waiting for it. :0
[02:45:06] <dwc-> exactly. and just stuff that into your path
[02:45:25] <richlowe> a very very quick glance makes me think around 9 thousandths of a second.
[02:45:36] <gdamore> richlowe: wrap /bin/sh and compare build time of Solaris. :-)
[02:45:38] *** sparc-kly has quit IRC
[02:45:56] *** lacaAFK is now known as laca
[02:45:59] <richlowe> which is actually pretty bad.
[02:46:01] <dwc-> richlowe: I know what isaexec does ... I've borrowed the concept to write wrappers for sun4/i86.pc/etc.
[02:46:01] <richlowe> Hm.
[02:46:02] <gdamore> 9 msec adds up very quickly in tight loops
[02:46:05] <dwc-> but I had to write mine in shell script
[02:47:02] <gisburn> richlowe: note that isaexec runs faster, much faster when it finds a "hit" early without stat()'ing the whole list of isalist down.
[02:47:13] <gdamore> compared to most programs, the isaexec overhead is negligible.  but there are a few programs that are executed a _lot_, and will suffer badly.  i contend the default shell is one of them.
[02:47:23] <gisburn> richlowe: and 9ms sounds like a blade100
[02:47:56] <gdamore> hey, i still use a blade 100!
[02:48:22] <gdamore> the reality is that it is all percentages that matter, not absolute times, anyway.
[02:48:27] <gisburn> richlowe: or did you put the test application into /usr/bin/sparcv7/ and let isaexec stat()' all the way down from sparcv9&co. ?
[02:48:33] <dwc-> firestorm [5] uname -ri
[02:48:34] <dwc-> 5.8 SUNW,Sun-Blade-100
[02:48:37] *** Godsey has left #opensolaris
[02:48:52] <gisburn> umpf
[02:49:13] <gisburn> dwc-: solaris 8 predates my perf, tweaks
[02:49:24] <dwc-> fine
[02:49:25] <dwc-> war# uname -ri
[02:49:25] <dwc-> 5.10 SUNW,Sun-Blade-100
[02:49:33] <delewis> $ uname -ri
[02:49:35] <gisburn> heh
[02:49:35] <delewis> 5.11 SUNW,Sun-Blade-1000
[02:49:38] <gisburn> better
[02:49:39] <delewis> an extra 0 makes all the difference :-)
[02:49:40] <gdamore> anyway, isaexec for ksh93 to accomodate a few bizarro users who want to do unusual things, is a bad idea, if it it will make everyone else suffer.
[02:50:02] <gisburn> my god. stop it.
[02:50:28] <gdamore> okay, time to go carve pumpkins anyway. :-)
[02:50:31] <gisburn> gdamore: if you dislike it make /usr/bin/ksh93 a hardlink to /usr/bin/sparcv9/ksh93
[02:50:57] <dwc-> make SHELL=/usr/bin/sparcv9/ksh93
[02:51:10] <gisburn> I think the "default" should work for everyone, even those "bizaro" people who have zillions of dollars
[02:51:33] <dwc-> eh, if they want it changed, they can file an RFE
[02:51:37] * gdamore suddenly knows why solaris has turned into a bloated pig.
[02:51:50] * dwc- pokes solaris and hears it squeal
[02:52:10] <sahafeez> everyone should just use bash is what i say (ducks!)
[02:52:23] 
[02:52:27] <dwc-> or ship SUNWksh93-64
[02:52:33] <gisburn> umpf
[02:52:50] <dwc-> and let the sysadmin replace SUNWksh93 it if he so chooses
[02:52:55] <Stric> or make 64bit default on 64bit platforms..
[02:53:05] <gisburn> Stric: that IS the default
[02:53:10] <gdamore> i'm leaving for real now, but i'll post my objections to wrapping ksh93 with isaexec in shell-discuss
[02:53:14] <Stric> so what are we arguing about? :)
[02:54:02] <delewis> gisburn: perhaps we should bring up Xsun removal again :-)
[02:56:23] *** EdLin has joined #opensolaris
[02:59:43] *** yongsun has joined #opensolaris
[03:00:18] <unixconsole> xsun can be removed once the framebuffers on sparc's is supported on xorg..
[03:00:26] <richlowe> gdamore: ksh93-integration-discuss maybe better.
[03:00:33] <delewis> unixconsole: that's the whole point.
[03:00:41] <delewis> not *all* of the framebuffers are going to be ported.
[03:00:42] <gisburn> richlowe: please no.
[03:00:45] <richlowe> delewis: Given that alanc said they had no real idea *what* would be supported.
[03:00:52] <richlowe> delewis: and that EOL != EOF
[03:00:57] <richlowe> I'm not so sure now is the time to argue about it.
[03:01:22] <gisburn> gdamore: and I disagree with you. executing ONE single external, non-builtin command outwheights the price of isaexec() by magnitudes
[03:01:23] <richlowe> gisburn: if people have concerns, they have concerns.
[03:01:25] <unixconsole> well either sun ports the drivers for the cards over the past like 5 years.. or they get the nvidia cards to work on sparc;)
[03:01:26] <delewis> richlowe: it's at least should be known that SPARC framebuffer support is looking more and more finite.
[03:01:30] <richlowe> gisburn: coercing them into silence is not the way to deal with those concerns.
[03:01:40] <delewis> unixconsole: no
[03:01:47] <gisburn> gdamore: since many common commands are builtins in ksh93 it is for /dev/null to complain about isaexec
[03:01:52] <delewis> I want my TGX and Elite3D framebuffers to work as they work now 5 years down the road.
[03:01:58] <delewis> and there's a lot of people who want that, too.
[03:02:09] <gisburn> richlowe: shell-discuss@ is IMO suited better for this.
[03:02:10] <sahafeez> I just care about the Elite3d
[03:02:21] <richlowe> delewis: you have to admit, cg6 has had quite the lifetime.
[03:02:40] <unixconsole> I just care about my expert3d lite card..
[03:02:41] <richlowe> and it *may* still work "5 years down the road"
[03:02:46] <delewis> richlowe: absolutely, but killing off the SBus framebuffers would make a lot of UltraSPARC-II useless on the desktop.
[03:02:47] <richlowe> consider how long CDE has been EOL'd for at this point.
[03:02:48] <sahafeez> time have changed tho. hardware has started to be come useless faster
[03:02:53] <richlowe> ... it's not gone yet, has it?
[03:03:05] <dwc-> I thought CDE was required for "unix" compliance
[03:03:11] <sahafeez> like what? only the Ultra 1 was SBUS right?
[03:03:12] <delewis> dwc-: only UNIX98
[03:03:19] <delewis> and the UNIX98 Workstation spec. more specifically
[03:03:20] <richlowe> sahafeez: ultra2, as well.
[03:03:23] <delewis> UNIX2003 no longer requires it.
[03:03:29] <richlowe> of the machines still supported, only the ultra2 is sbus.
[03:03:32] <delewis> as well as the E-class servers.
[03:03:33] <dwc-> did they mandate a replacement?
[03:03:37] <richlowe> and certain servers.
[03:03:37] <delewis> dwc-: no
[03:03:38] <sahafeez> UPA and up would be fine..
[03:03:41] <richlowe> ... which don't exactly need the FB support.
[03:03:42] <delewis> they completely dropped the Workstation spec.
[03:03:48] <delewis> richlowe: but still
[03:03:59] <delewis> if I want to put a framebuffer in my E4500 and use it, I should have that ability.
[03:04:11] <richlowe> dropping cg6, and the earlier c3d models wouldn't make me sad.
[03:04:21] <richlowe> dropping anything post xvr-100 timeframe, I think could be a mistake.
[03:04:28] <Stric> dropping c3d would make me sad :/
[03:04:28] <richlowe> but as I said, it all depends on when this would actually *happen*
[03:04:29] <hile_> when did you get a 4500, delewis?
[03:04:31] <sahafeez> what do you need to use an U2 for w/a FB? a desktop? come on now..
[03:04:34] <richlowe> and since none of us know either of these things.
[03:04:37] <richlowe> what is there to argue about?
[03:05:04] <Error_404> sahafeez: what's wrong with a u2 as a desktop?
[03:05:09] <delewis> sahafeez: I can use it as a desktop if I wanted.
[03:05:12] <Error_404> they're nice little machines
[03:05:13] <delewis> that's the point.
[03:05:16] <richlowe> actually, scratch that, I do have an early c3d here somewhere, so yeah, I care a little ;)
[03:05:22] *** esaxe has left #opensolaris
[03:05:23] <richlowe> only a very little though.
[03:05:30] <sahafeez> it is slow and old and there are faster machines out there for cheap
[03:05:41] <richlowe> you can guess how much I'd care based on how little idea I have of exactly where it is :)
[03:05:43] <delewis> sahafeez: I'm not interested
[03:06:13] <richlowe> this is just the US-I discussion all over again :(
[03:06:14] <delewis> my SB1000 is more than capable of doing everything I do.
[03:06:25] <unixconsole> For the pci based cards, support is definitely required. I wouldn't mind the nvidia cards on sparc though. some parity between the platforms would be nice:)
[03:06:38] <richlowe> delewis: your SB1000 probably doesn't have anything less than an xvr100 or pgx64 in it.
[03:06:47] <richlowe> (the xvr100 isn't technically supported in a b1k, but it works...)
[03:06:54] <delewis> richlowe: Elite3D
[03:06:57] <spawrq> ungh
[03:07:02] <spawrq> the xvr100's are crap
[03:07:06] <unixconsole> how is the sb1000? I was thinking up either getting on of those or an ultra20m2 to replace my ultra60.
[03:07:09] <Stric> we have a bunch of dual sb1000's with dual c3d (and dual mice+keyboard)
[03:07:14] <richlowe> spawrq: it's a radeon7k
[03:07:19] <richlowe> and exactly what's in my x86 machine too ;)
[03:07:22] <spawrq> its crap in any language.
[03:07:34] <spawrq> we've had lots of annoying problems with them.
[03:07:41] <delewis> I really don't do a lot of graphics-intensive work.
[03:07:41] <Stric> richlowe: and it's in sunray1 too I think
[03:07:59] <delewis> programming, communication, document composition, and watch video.
[03:08:06] <delewis> the Elite3D handles all of that fairly well.
[03:08:11] <Stric> for 2d, radeon7k is just fine
[03:08:23] <delewis> as would most framebuffers (at least ones capable of doing 24-bit colour, which excludes the cg6)
[03:08:24] *** halton has joined #opensolaris
[03:08:25] *** halton has left #opensolaris
[03:08:30] *** halton has joined #opensolaris
[03:08:36] <Stric> it kicks the crap out of c3d :P
[03:08:43] <richlowe> the 3d wouldn't be *that* bad.
[03:08:49] <richlowe> but of course, the xvr-100 is a '2d framebuffer'
[03:08:54] <delewis> Creator3D is better than the Elite3D in all but a few cases.
[03:08:57] <richlowe> which is also spelled "they never bothered with GL drivers"
[03:09:18] <sahafeez> i thought it was the other way around. i have 2 of the elite3d-m6
[03:09:38] <delewis> sahafeez: Creator3Ds are capable of running at a much higher resolution than the Elite3Ds at the sacrifice of 3D.
[03:09:39] <unixconsole> the elite3d-m6 is better than the c3d's
[03:09:58] <delewis> maybe it's the Expert3D I'm thinking about that has superior texture memory.
[03:10:16] <unixconsole> the expert3d is the way to go if you can find one on ebay.
[03:10:23] <delewis> regardless, the Creator 3Ds will still do a higher resolution than the Elite3Ds
[03:10:28] <delewis> which are limited 1280x1024
[03:10:32] <unixconsole> a lot faster than the elite or creator cards.
[03:10:37] <sahafeez> expert3d is pci..
[03:10:40] <delewis> unixconsole: XVR-1000 is the way to go :-)
[03:10:59] <sahafeez> i just sold one on ebay. it only went for like $30
[03:11:08] <unixconsole> delewis: of course.. if you can find a cheap xvr-1000. I got my expert3d for $120.
[03:11:13] <sahafeez> i had the xvr-1000. sold it for $290
[03:11:16] * gisburn is back
[03:11:19] <delewis> unixconsole: sounds overpriced.
[03:11:34] <delewis> most XVR-1000s tend to go for about $300
[03:11:37] <unixconsole> delewis: that was almost 3 years ago..
[03:11:48] <delewis> unixconsole: you failed to specify that :-)
[03:12:07] <unixconsole> delewis: well.. that's when i upgraded my ultra60;)
[03:12:28] *** terinjokes has joined #opensolaris
[03:12:43] <delewis> well, I had quite an upgrade from an SS5 (which wasn't really my primary workstation to begin with) to an SB1000.
[03:12:45] <sahafeez> i did not see a reason for the xvr-1000. it made jds faster then the creator but i sold the xvr-1000 and got an u80 for the money
[03:12:53] <terinjokes> hey, i'm interested in learning more about solaris
[03:13:12] <unixconsole> delewis: lets see.. expert3d's going for 29.99 and xvr-1000's going for 300 today..
[03:13:15] <delewis> sahafeez: the only reason I'd like an XVR-1000 is DVI and the fact it can use the Dell widescreen displays (the 24" model)
[03:13:23] <sahafeez> ah.
[03:13:34] <delewis> it can also do full-screen OpenGL output that's very texture intensive.
[03:13:36] <delewis> like video
[03:13:39] <sahafeez> i have 2 of the elite3d with 2 23 inch crts
[03:14:06] <sahafeez> never tried to play video beyond gootube on my elites
[03:14:13] <unixconsole> delewis: do you like the sb1000? is it loud?
[03:14:18] <delewis> unixconsole: very quiet
[03:14:18] *** terinjokes has left #opensolaris
[03:14:31] <delewis> checkout the SunSolve specs on the noise
[03:14:42] <sahafeez> bought a lotto ticket. if i win i will by everyone that has helped me a new sun ;)
[03:14:42] <delewis> probably one of Sun's most quiet workstations.
[03:14:43] <Stric> way lower than u10s for instance
[03:14:52] <delewis> then again, I'm also sitting behind a pSeries p640
[03:14:57] <sahafeez> ok. going home.
[03:14:58] <delewis> so I'm probably deaf in most regards.
[03:14:59] <unixconsole> delewis: hmm.. that's good to know. I had to install the noise reduction kit on the u60 to make it quiet.
[03:15:16] <sahafeez> my u80 is pretty quiet
[03:15:31] <Stric> u80 and b1k has the ~same chassis, right?
[03:15:35] <delewis> the SB1000 is somewhat more quiet than the U80 according to SunSolve
[03:15:43] <delewis> Stric: roughly, very roughly I think
[03:15:52] <sahafeez> looks the same. different colour
[03:15:53] <Stric> at least they look very similar
[03:15:56] <delewis> it's better to say the SB1000 and SB2000 have identical chassis.
[03:15:58] <delewis> :-)
[03:15:59] <sahafeez> heavy system.
[03:16:08] <sahafeez> ok. off to home.
[03:16:09] <delewis> yeah, mine weighs about 80LBs
[03:16:14] <delewis> the U80 should weigh slightly more
[03:16:23] *** sahafeez has quit IRC
[03:16:34] <unixconsole> a vendor made an offer to sell a sb1000 for just under $600 to me. I'm thinking about taking them up on the offer. which cpu's do you have?
[03:16:43] <delewis> unixconsole: that's a bit overpriced.
[03:16:50] <delewis> you could get a decent SB2000 configuration for that much
[03:17:07] <delewis> unixconsole: I have 2x750MHz USIII (non-CU)
[03:17:17] <unixconsole> 2x750Mhz, 2GB ram, 2x36GB hdd's.
[03:17:19] <delewis> I'd like to upgrade to the 900MHz CUs :_)
[03:17:32] <delewis> unixconsole: not bad, but I'd urge you to look at the current SB2000 offerings on Ebay
[03:17:39] <delewis> they range from about $400-$800
[03:17:40] <unixconsole> let me see..
[03:17:44] <delewis> and most of them include 900MHz CUs
[03:18:03] <delewis> which are *much* faster than even the 900MHz non-CUs, let alone the 750MHz non-CUs.
[03:18:04] <unixconsole> hmm.. I'd rather have the 900's over the 750's..
[03:19:12] *** slowhog has quit IRC
[03:19:51] *** slowhog has joined #opensolaris
[03:19:53] <unixconsole> hmm.. 2x900Mhz's, 2gb ram, 1x36gb hdd.. for $450 plus $95 shipping.. not bad.
[03:20:20] <delewis> unixconsole: yeah, I was thinking about getting one myself, but I decided to go with something a bit more server-like
[03:20:32] <delewis> I'll just upgrade my SB1000 to an SB2000
[03:20:50] <delewis> (which are the same system except for the higher spec'd procs)
[03:20:57] <unixconsole> I was thinking about getting an ultra20m2, but I still like sparc:(
[03:21:09] <delewis> unixconsole: I'm in the same boat, myself.
[03:21:18] <delewis> and I'm going to stay with SPARC, I think.
[03:21:21] <delewis> call me silly. :-)
[03:21:28] <richlowe> Ok. :)
[03:21:30] <unixconsole> I know how you feel.
[03:23:28] <unixconsole> I'm trying to see if I can get a discount from a friend at sun on the ultra20m2. if it comes out cheaper, I'll probably go that route. otherwise, I think the sb2k would work.
[03:23:54] <unixconsole> originally, I was thinking of get a mac mini, but the video card sucks.
[03:24:29] <delewis> I've just become accustomed to having a real, UNIX workstation and it seems wrong to go any other way.
[03:25:02] <richlowe> I was only being sarcastic, if you were replying to me :)
[03:25:16] <delewis> AMD64 is nice, yes, but I just can't get past some of the x86_32 and x86_16 cruft that's still left.
[03:25:19] <delewis> richlowe: :-)
[03:25:32] <unixconsole> I know what you mean.. not having obp really bugs me.
[03:26:06] <delewis> and I really don't expect a lot from my hardware to begin with -- I just want it do whatever it does well.
[03:26:17] <unixconsole> too bad sun doesn't sell a cheap niagara workstation.
[03:26:20] <delewis> programming, document composition, communication, and watching the occasional video.
[03:26:37] <delewis> pretty simple stuff, really.
[03:27:01] <delewis> unixconsole: some of the portable SPARC vendors may be looking in that area.
[03:27:12] <unixconsole> reall?
[03:27:14] <unixconsole> really?
[03:27:20] <delewis> this is speculation, but I've heard it mentioned once or twice.
[03:27:37] <delewis> and regardless, I doubt I'll be able to afford it in any case :_)
[03:27:46] * delewis just bought an SB1000 a year ago
[03:27:47] <unixconsole> true..
[03:28:51] <unixconsole> sun should take a lowend 4 core niagara and put it into an ultra20 like case. sell it for $800.. that would get ppl all over the niagara platform really quick.
[03:29:10] <delewis> unixconsole: we may see something like that with the Niagra2
[03:29:22] <delewis> the Niagra1 isn't really suited for a workstation-like environment (FP performance sucks)
[03:29:26] <unixconsole> one can only hope.
[03:29:35] <delewis> and the lack of VIS doesn't make things much better.
[03:29:35] <unixconsole> true.
[03:29:37] <alanc> delewis: cg6 drivers are open source, so if people care, they can continue to make drivers after Sun gives up on them - but it's getting really hard to justify supporting sbus framebuffers that were last sold a decade ago - it's difficult to find ones we can test on
[03:29:40] <delewis> hardware VIS, anyway.
[03:29:57] <delewis> alanc: I'll send you my cg6 :-)
[03:30:00] <unixconsole> yeah.. vis is nice:)
[03:30:02] <richlowe> alanc: I have a bunch. :)
[03:30:17] <alanc> I've got one Ultra 2 in the lab with cg6
[03:30:32] <alanc> that's the only machine in there that has working sbus that can run Nevada
[03:30:32] <richlowe> alanc: I'm assuming that the porting work would be on the X side.
[03:30:33] <Stric> I can spare some turbogx'es ;)
[03:30:43] <richlowe> alanc: Xorg drivers speaking to the same kernel fb interface
[03:30:48] <delewis> alanc: don't get me wrong -- I can completely understand that Sun has limited staff to apply to every area of development, like porting SPARC graphics drivers, but the thought of my SPARC hardware becoming useless is quite.. troubling.
[03:30:51] *** dunc has quit IRC
[03:31:02] <richlowe> alanc: ... which gives you another reason to yell at sparc graphics (or hardware in general) :)
[03:31:28] <alanc> if it was up to me, we'd make as much as possible work - but most SPARC graphics drivers are locked up in the SPARC organization, far from my control
[03:31:50] <alanc> the real solution is to hope they take up gdamore on his offer to try to get those unlocked
[03:31:52] <gdamore> alanc: i'm trying to fix that for you.
[03:32:02] <elektronkind> it sucks that everything is so compartmentalized there
[03:32:06] <unixconsole> seems that very little has changed at sun when it comes to things like that:(
[03:32:07] <delewis> well, I just wanted to get the issue out there, and show that there are still some people that do use SPARC as a personal workstation.
[03:32:07] <richlowe> alanc: frankly, I'd like gdamore to ask the vendors *anyway*
[03:32:16] <richlowe> alanc: that way he can wave it around.
[03:32:29] * richlowe is still wondering who one should pester regarding cassini and co.
[03:32:31] <gdamore> haha.
[03:32:44] <gdamore> you do _not_ want to see cassini source.
[03:32:47] <unixconsole> delewis: indeed.. I think sun forgets about ppl like us sometimes.
[03:32:56] <gdamore> it will make your head explode
[03:32:59] <Stric> gdamore: can't be worse than procmail :)
[03:32:59] <richlowe> gdamore: I don't like code being closed for seemingly no reason.
[03:33:05] <gdamore> oh yes it can.
[03:33:06] <richlowe> and since iirc cassini is a fully Sun thing...
[03:33:07] <alanc> our group has been using the Xorg open source ATI drivers to get m64 & pfb boards up and running without having to deal with that, while the SPARC group concentrates on XVR-2500
[03:33:18] <richlowe> Stric: the real question is, can it be worse than fetchmail? :)
[03:33:26] <Stric> which has nice code like
[03:33:28] <gdamore> the nics group still think they have some "competitive advantage" to protect.
[03:33:29] <Stric> if (
[03:33:30] <Stric> #else
[03:33:46] *** ndroux has quit IRC
[03:33:47] <boyd> No wanting to pay for the work required to open code is, of course *some* reason
[03:33:48] <richlowe> Is gem their's too?
[03:33:53] <richlowe> I used to have this noted down somewhere...
[03:33:57] <delewis> I just love an abundant use of pre-processor directives.
[03:33:58] <gdamore> those are just implementation flaws.  you haven't seen brain damage until you look at the code. :-)
[03:34:00] <gdamore> gem is Sun's.
[03:34:22] <gdamore> the gem driver is actually not too bad, although it suffers from a certain amount of hme heritage.
[03:34:27] <elektronkind> gem as in the original gig-e nic?
[03:34:32] <gdamore> right.
[03:34:33] <elektronkind> ah yes
[03:34:36] <gdamore> well, there was a vge before tht.
[03:34:39] <EdLin> when is approximately the next release of Solaris 10 planned?
[03:34:42] <gdamore> s/tht/that/
[03:34:45] <delewis> don't forget Happy Meal :-)
[03:34:57] <gdamore> hme was 10/100 only
[03:35:04] <gdamore> gem is basically hme++
[03:35:13] <elektronkind> hme was a set of differnt phy's tho, if I'm not mistaken
[03:35:16] <delewis> where does eri fit in (which is RIO-based or something, IIRC)
[03:35:17] <delewis> ?
[03:35:19] <gdamore> right.
[03:35:21] <boyd> EdLin: Appoximately between now and the end of the year. It's denoted 11/06
[03:35:22] <unixconsole> update 3 is probably going to happen in december at his point..
[03:35:24] <unixconsole> this..
[03:35:25] <richlowe> delewis: eri is open, and in ON
[03:35:31] <delewis> richlowe: nice
[03:35:33] <EdLin> boyd: thanks
[03:35:40] <gdamore> eri is the hme modified and put into a common core.  it shares a gem core, IIRC.
[03:35:46] <elektronkind> thre's also dmfe which was on the X1/V100
[03:35:48] <EdLin> boyd: any features planned?
[03:36:03] <gdamore> dmfe on the X-1 was a Sun brain fart.
[03:36:06] <boyd> EdLin: Several... anyone know of a list somewhere?
[03:36:26] <richlowe> dmfe is open, too.
[03:36:30] *** Dr_Jekyl1 has joined #opensolaris
[03:36:31] <unixconsole> yeah, the dmfe's were a little different:-P
[03:36:40] <EdLin> I tried googling about a release schedule and didn't find anything...
[03:36:58] <gdamore> the original dmfe driver was hacked from dnet, a really bad idea.  it has since been greatly improved.
[03:37:23] <EdLin> maybe I'll find it if I google for Solaris 11/06....
[03:37:27] <elektronkind> oh yeah. I remember the copious miseed interupt 4 messages
[03:37:32] <elektronkind> s/miseed/missed
[03:37:33] <richlowe> now there's screwy.
[03:37:35] <gdamore> i think the deal was that the platform group was able to save like $.25 or somesuch by not using an eri but using dmfe instead. :-)
[03:37:52] <boyd> EdLin: Off the top of my head: Trusted Extensions, new zfs bits and pieces...
[03:38:11] <boyd> umm...
[03:38:12] * gdamore is still amazed dnet has EOF'd yet.
[03:38:19] <richlowe> this doesn't break after the first run, but after the first *successful* run.
[03:38:25] <unixconsole> I'm in the beta for update3.. so you'll see things like secure by default, ldom support(hopefully), tx, zone features, zfs features..
[03:38:26] *** nwf has quit IRC
[03:38:29] <elektronkind> then there's cassini (ce) wich seems to have turned into a bastard child
[03:38:36] <EdLin> boyd: I just found info on trusted extentions.
[03:38:47] *** Dr_Jekyll has quit IRC
[03:38:56] <gdamore> ce was a complete rewrite, but was done, from what i can tell, by looking at the gem sources.
[03:39:05] <boyd> zfs changes include raidz2, hot spares..
[03:39:17] <EdLin> unixconsole: secure by default is a real plus.
[03:39:19] <boyd> Yeah, some talk about ldom support... dunno
[03:39:22] <gdamore> it has some really, really screwball stuff in it.  that doesn't belong in a NIC driver at all.
[03:39:43] <richlowe> boyd: there's stiff an LDOM-putback-a-day, ish, just to ON fixing various things.
[03:39:45] <gdamore> huge portions of the driver are #if 0'd. :-)
[03:39:46] <richlowe> 'still'
[03:40:03] <unixconsole> yeah, secure by default is good. there are a lot of features backported for zones, zfs, least privs, etc.
[03:40:12] <boyd> richlowe: I still couldn't parse that after the correction
[03:40:33] <unixconsole> well the ldom stuff is on a seperate beta that starts tomorrow.
[03:40:36] <richlowe> boyd: there's somewhere close to a putback-per-day fixing LDOM issues, just in ON.
[03:40:36] <elektronkind> perhaps one day someone will have the intestinal fortitude to rewrite these forlorn interfaces (ce, dmfe, and so on) in gldv3 with all the crap refactored and cleaned up
[03:40:42] <gdamore> they're doing secure by default in an S10 update?!?  how did that get past ARC?  don't get me wrong, i love it, i just can't believe the change is appropriate for a patch release
[03:40:48] <boyd> richlowe: Ah... thanks :)
[03:40:51] <richlowe> gdamore: two stage.
[03:40:58] <Tpenta> i can tell you how it got thru arc
[03:41:00] <richlowe> gdamore: not so much 'by default' in s10, iirc, it asks you.
[03:41:08] <Tpenta> for an upgrade the default behaviour is not to change
[03:41:09] <gdamore> elektronkind: maybe.  there is some chance I might the opporuntity. :-)
[03:41:16] <Tpenta> for an install you are asked a question
[03:41:30] <gdamore> Tpenta: okay, that makes sense.  tks.
[03:41:30] <boyd> I assume there's a new sysid keyword for it too
[03:41:39] <richlowe> boyd: I'm not sure if that's a sign they're beating it clean to ship it, or a sign it's not good to ship ;)
[03:41:43] * Tpenta was at that meeting :)
[03:41:54] <richlowe> Tpenta: you get all the fun.
[03:41:57] <unixconsole> hopefully, I'll get a t2k at work to test out the ldom stuff. that will be a lot more interesting than update 3.
[03:41:58] <boyd> richlowe: Hmm... I saw something on a mailing list a couple of days ago
[03:42:01] <gdamore> richlowe++
[03:42:13] <Tpenta> I also get to write opinions and chase up project teams
[03:42:27] <richlowe> Tpenta: yeah, but the 2006/550 concall was a really useful thing.
[03:42:30] <richlowe> Tpenta: ... and then such things ceased.
[03:42:35] <gdamore> Tpenta:  I think richlowe has volunteered implicitly to sit on ARC... :-)
[03:42:36] <boyd> Bah! Unexplained failure playing with TX
[03:42:47] <richlowe> Tpenta: if the 2006/569 call is closed, I'm going to raise hell.
[03:42:50] <richlowe> Tpenta: just as prior warning.
[03:42:51] <richlowe> :)
[03:42:54] <elektronkind> boyd: don't mess with texas
[03:42:56] <boyd> richlowe: ++
[03:42:59] <boyd> elektronkind: :)
[03:43:07] <gdamore> btw, I've volunteered _explicitly_ to do the same. :-)
[03:43:10] *** EdLin has left #opensolaris
[03:44:01] <unixconsole> brb..
[03:44:08] <Tpenta> richlowe, it's the general case that a fastrack lives in email only. If concsensus is reached, then it is pretty much automatically approved at the next meeting. the ksh93 fast track had a call because it looked like we could not reach concensus and there were outstanding issues
[03:44:31] <Tpenta> if you have somethingto say, then it needs to be in the email thread
[03:44:38] <Tpenta> (which i know you have already contributed to)
[03:44:42] <gdamore> i'm surprised ksh93 was able to be a fasttrack.
[03:45:12] <gdamore> its a major change that i would have thought would have required pretty much a regular case.
[03:45:51] <Tpenta> it should not have been as contentious as it ended up. It was a fair intial submission as a fast track.
[03:46:09] <gdamore> okay.
[03:46:30] * gdamore has a couple of items in the fasttrack queue himself. :-)
[03:46:38] <Tpenta> the only issues I really had was that it came before the arc after a lot of work had already been done, so it was very easy for egos to get out of joint by people asking questions
[03:47:01] <gdamore> i think i understand that. :-)
[03:47:06] <Tpenta> we have a saying, ... "arc early arc often". it makes things easier on everyone
[03:47:14] <alanc> adding new commands that follow existing precedent from other systems is usually fasttrackable unless they're really complex or strange
[03:47:23] <gisburn> Tpenta: well, arc early didn't work.
[03:47:25] <gdamore> ksh93 is pretty complex. :-)
[03:47:55] <Tpenta> roland, i think we must agree to disagree
[03:47:58] <gisburn> Tpenta: we couldn't arc it early. if we did that this way we would have endedf like prototype001 which was totallly for the trashcan.
[03:47:58] <alanc> (say, Xorg, which was a full case, but more because of the implications of the change to our long-term X server strategy - most of the X commands we add are fasttracks)
[03:48:03] <gdamore> and, there have been some details just discussed here that i think were contentious enough that fasttrack might not have worked.
[03:48:17] <alanc> (like say, if I was to ever file a fasttrack for setxkbmap)
[03:48:18] <gisburn> Tpenta: the prototype001 was a disaster.
[03:48:31] <gdamore> setxkbmap would be fasttrackable. :-)
[03:48:34] <gdamore> i think.
[03:48:43] <Tpenta> my thoughts are that if it had been done as a full case and an inception been done, we would probably had a lot more folks being knowledgable about what was going on and fewer noses out of joint
[03:48:49] <gisburn> Tpenta: but it delivered the data to make a new prototype002, collect the new data and build the arc case from this.
[03:48:54] <gdamore> Tpenta++
[03:49:12] <richlowe> Tpenta: 569 hasn't reached anything though.
[03:49:18] <richlowe> Tpenta: it's gone silent waiting for spec.
[03:49:19] <gdamore> gisburn: this is why you need an inception review and a full review both.
[03:49:40] <gisburn> gdamore: which would have been useless in our case.
[03:49:43] <richlowe> Tpenta: I agree entirely.
[03:49:48] <richlowe> Tpenta: I'm somewhat annoyed that wasn't suggested.
[03:49:55] <gdamore> why?  inception lets you do an initial presentation and give the heads up.
[03:49:57] <richlowe> by the time those of us outside knew that was a good idea, it was way too late.
[03:49:58] <Tpenta> in a full case you have inception (very early on, discussing design etc before a lot of work is done) and committment (after most of the work is done)
[03:50:09] <gisburn> gdamore: the integration work was neccesary to find out the problems which needed to be arc'ed.
[03:50:12] <gdamore> and get feedback to know what likely concerns are and collect a first round of questions
[03:50:34] <richlowe> gisburn: what inception would have done, is meant that a lot of the things you were asked about, would have been known and talked about in advance.
[03:50:39] <gdamore> you could still have done an inception.  the final review could have covered the issues.
[03:50:49] <Tpenta> all up, I think the things that were contentious were trivial when looked at the size of the projkect and the amount of work involved. The big arguments were over little things. That tells me that most of the work *was* actually done right.
[03:50:55] <richlowe> Though that reminds me.
[03:51:05] <gdamore> we call that "bikeshedding" :-)
[03:51:23] <gdamore> IIUC, ARC is just as subject to bikeshedding as any other technical committee. :-)
[03:51:48] <alanc> doubly so when commands everyone uses like shells or text editors come up
[03:52:06] <alanc> and when you discuss text editing modes in shells....watch out...
[03:52:15] *** fik has quit IRC
[03:52:41] <gdamore> hehe.
[03:52:45] <Tpenta> alanc: good to see the dtrace xorg case come up
[03:53:15] <alanc> I asked the JDS team when they'd be shipping Brendan's javascript probes in Firefox
[03:53:22] <gisburn> alanc: I am still waiting that the complainers show up and teach the 1st semesters "using the unix shell" in vi mode.
[03:53:26] <alanc> and their answer was "After you get your Xorg probes in!"
[03:53:40] <alanc> so I kinda had to finally get off my duff and do it
[03:54:15] <gdamore> gisburn:  arguments like that are a matter of opinion.  you have to be prepared for folks to have strong opinions.
[03:54:46] <gdamore> i wasn't present, but if you tried changing the default editor mode as part of your ksh93 putback... well... i'm not surprised it was painful
[03:55:23] <Tpenta> i woudl really like brendans stuff to be in there by default; that was a very smart 2 weeks of work on his part
[03:55:32] <alanc> gisburn: no one said it shouldn't be an option for you to configure at your site, but you can't expect every Solaris user in the world to use the defaults that make sense for your 1st semester newbies
[03:55:35] <gisburn> gdamore: POSIX demands that the plain shell without any startup files defaults to "none" editor.
[03:56:06] <gdamore> changing the default from existing solaris usage should have been a separate case.
[03:56:07] <Tpenta> Gentlemen, we do *not* need to open this argument again
[03:56:14] <gdamore> anyway, back to pumpkins for me.
[03:56:37] <alanc> and I need to head home
[03:56:40] <Tpenta> which reminds me, I have to get the halloween photos of how we did our house up for tonight online
[03:56:45] <Auralis> delewis: you have a dell 24" on your xvr-1000?
[03:56:50] *** alanc is now known as alanc-away
[03:57:02] <delewis> Auralis: no, but if I ever do get a Dell 24" I'll need an XVR-1000.
[03:57:09] <delewis> it's in the upgrade path.. somewhere. :-)
[03:57:18] <delewis> probably after I upgrade to CU procs, first.
[03:58:18] <richlowe> This is behaving so utterly nonsensically. :\
[03:58:32] <richlowe> and debugging the broken debugger with the debugger I broke is no fun *at all*
[03:58:55] <Auralis> ah, ok
[04:01:43] *** solarisjon has left #opensolaris
[04:06:44] <unixconsole> back..
[04:08:07] <twincest> rich: you don't have another debugger?
[04:09:18] *** mnowak has quit IRC
[04:09:49] *** mnowak has joined #opensolaris
[04:16:06] *** paxc has joined #opensolaris
[04:20:09] *** jmcp_ has quit IRC
[04:21:04] *** ndroux has joined #opensolaris
[04:26:57] *** unixconsole has left #opensolaris
[04:28:31] *** mrdeviant has joined #opensolaris
[04:29:12] *** sparc-kly has joined #opensolaris
[04:29:12] *** ChanServ sets mode: +o sparc-kly
[04:32:43] <Gman> hrm, looks like there's definitely some issues with the opensolaris mail server
[04:32:50] <Gman> mails are really slow coming through today
[04:33:00] <richlowe> A lot of stuff has seemed slower than usual today.
[04:33:38] <richlowe> wish I knew what caused firefox to suck CPU when I wasn't doing anything.
[04:33:55] <Gman> heh - it's probably caching things again :)
[04:34:09] <gisburn> richlowe: link prefetch
[04:34:16] <gisburn> richlowe: or flash plugin
[04:34:32] <gisburn> richlowe: or a javascript virus
[04:34:35] <gisburn> (no joke)
[04:35:07] * richlowe is waiting patiently for a little D-fu to get done.
[04:35:10] <richlowe> ... and now, again, I guess.
[04:35:16] * richlowe wasn't expecting stacks quite that long.
[04:35:59] <gisburn> richlowe: check whether you have any "extensions" listed which you didn't install. some sites try to install their own warez via firefox extensions
[04:36:05] <richlowe> gisburn wins at least a very small prize.
[04:36:12] <gisburn> richlowe: ?!
[04:36:15] <richlowe> Yeah, flash player, to a large degree.
[04:36:33] <richlowe> I'd bet the actual top offender was being caused by flash too, just not directly.
[04:37:20] <richlowe> Gman: since you guys ship this junk, I don't suppose you have a path to yell at adobe? :)
[04:37:34] * Gman reads up
[04:37:46] <Gman> ahh, right
[04:37:53] <gisburn> richlowe: wrong company
[04:37:55] <Gman> i actually think we have access to the source on nda
[04:38:02] <gisburn> richlowe: complain to mozilla.org
[04:38:09] *** axisys has joined #opensolaris
[04:38:14] <richlowe> gisburn: Not sure about that.
[04:38:14] <gisburn> Gman: it's a gecko problem, not a flash problem
[04:38:19] <gisburn> umpf
[04:38:20] <richlowe> No, it certainly isn't that.
[04:38:23] <richlowe> not exactly, anyway.
[04:38:32] <richlowe> libflashplayer is waking itself up a *whole* lot, and doing seemingly random stuff.
[04:38:42] <gisburn> richlowe: I've been woring in that code since many moons.
[04:38:44] <richlowe> especially since the firefox window(s) are unmapped, and there's no flash stuff currently visible.
[04:38:53] <gisburn> richlowe: see ?
[04:39:14] <gisburn> richlowe: gecko sometimes does not distribute this information to the plugin
[04:39:21] <richlowe> Well, it's libflashplayer that's choosing to do stuff for no good reason.
[04:39:21] *** eboutilier has joined #opensolaris
[04:39:34] <gisburn> old bug
[04:39:35] <richlowe> firefox hasn't told it the window is unmapped, sure.
[04:39:37] <gisburn> stupid bug
[04:39:40] <richlowe> but I don't *have any flash stuff loaded*
[04:39:49] <richlowe> so, flash shouldn't be doing anything anyway.
[04:39:52] <gisburn> richlowe: you did load it once
[04:40:22] <gisburn> richlowe: and the underlying DOM object containing the flash plugin was not destroyed yet
[04:40:43] <gisburn> richlowe: either leak or nsCOMPtr self-reference
[04:40:44] <boyd> It's adobe.... reader does stuff when your'e' not using it either. I think it want's the attention.
[04:40:56] *** paxc has quit IRC
[04:43:16] <richlowe> I'm happy to blame both of them :)
[04:45:27] <Gman> hey eboutilier
[04:46:28] *** Tpent1 has quit IRC
[04:50:44] <Error_404> yum, cold pluma moos
[04:51:02] *** chain-lightning has joined #opensolaris
[04:52:46] <axisys> what am I missing here http://www.rafb.net/paste/results/pOmmkK71.html
[04:53:05] <chain-lightning> I have a quick question; I'm trying to do a loopback mount of a file. I run "mount -F lofs <file> <mountpoint>" but <mountpoint> turns into a data file (according to file). How can I mount the original file so I can copy files from it? (eg, like linux's mount -o loopback option)
[04:53:05] <axisys> user.oracle is assigned 4G .. but it is taking 21G !!
[04:53:53] <axisys> chain-lightning: u have to use lofiadm
[04:54:06] <chain-lightning> axisys thank you!
[04:54:08] <axisys> lofiadm /path/to/image
[04:54:19] <eboutilier> Hi Glynn.
[04:54:22] <axisys> it will automatically create a dev link
[04:54:37] <axisys> and then just mount the /dev.. link to a /mnt point
[04:55:04] <richlowe> mount -F<fs of image> $(lofiadm -a /full/path/to/iso/file/or/similar)
[04:55:20] <richlowe> /mount/point
[04:55:57] <richlowe> (lofiadm -a will use the next available /dev/lofi dev, and print it to stdout, so you don't need to actually think about such things, beyond remembering lofiadm -d)  :)
[04:56:00] <axisys> so why the bxoracle zone here taking more memory than it is assigned w/ deny http://www.rafb.net/paste/results/pOmmkK71.html
[04:58:10] *** karrotx has joined #opensolaris
[04:59:39] *** sparc-kly_ has quit IRC
[05:00:27] *** Darwin_ has joined #opensolaris
[05:01:14] <axisys> i think i got it.. cuz i was not using rcap for user.oracle project
[05:03:19] *** gm152 has quit IRC
[05:03:58] *** gm152 has joined #opensolaris
[05:05:43] *** gm152 has quit IRC
[05:06:22] *** gm152 has joined #opensolaris
[05:08:36] *** axisys has quit IRC
[05:08:49] *** mrdeviant has quit IRC
[05:09:56] *** uebayasi has joined #opensolaris
[05:12:06] *** Darwin has quit IRC
[05:12:35] *** axisys has joined #opensolaris
[05:20:50] *** gisburn has quit IRC
[05:30:07] *** laca has quit IRC
[05:40:37] *** uebayasi has quit IRC
[05:43:03] *** sahafeez has joined #opensolaris
[05:44:10] <sahafeez> hey, real quick. i setup a system but forgot export/home. i just repartition the disk. can someone cat their /etc/vfstab and /etc/auto_mount and tell me what the default is for the nomral justed installed /export/home
[05:46:58] <richlowe> auto_home defaults to '+auto_home'
[05:47:18] <richlowe> I tend to add '*   <nodename>:/export/home/&', because I'm lazy.
[05:47:52] <sahafeez> ok, got that. how about the mnt line in vfstab. is it normal?
[05:48:18] <richlowe> dunno, all my machines have /export/home on ZFS.
[05:48:25] <richlowe> I believe so though, yeah.
[05:48:57] <sahafeez> i was under the impression it was differnet because of the auto_mount
[05:49:32] <twincest> i can't see why it would be
[05:49:39] <sahafeez> ok, thanks.
[05:49:40] <twincest> local automount just mounts one directory elsewhere
[05:49:44] <twincest> like /export/home/foo to /home/foo
[05:49:48] <richlowe> Hm, where's brendang when I need dtrace magic?
[05:50:00] <sahafeez> ok, cool. makes sense now. thank you
[05:50:43] *** uebayasi has joined #opensolaris
[05:50:50] *** sahafeez has quit IRC
[05:57:21] *** B|nTaRa has joined #opensolaris
[05:59:05] *** jsubl2 has quit IRC
[06:08:32] *** razrX has quit IRC
[06:08:53] *** razrX has joined #opensolaris
[06:10:19] *** EdLin has joined #opensolaris
[06:10:30] <Gman> richlowe, or for that matter bmc, ahl and mws
[06:11:01] <richlowe> Yeah.
[06:11:30] <Gman> SLACKARSES, ALL OF THEM!
[06:21:32] <soylent> is it safe to enable 'logging' for /
[06:21:56] <jamesd> its automatic on solaris 10 and later
[06:22:02] <richlowe> logging is (now) the default for /
[06:22:11] <soylent> well, I see - in mount options in /etc/vfstab
[06:22:13] <richlowe> I can't recall why that hasn't always been the case, though.
[06:22:20] <soylent> ah yeah
[06:22:23] <soylent> you're right
[06:22:29] <soylent> on /dev/dsk/c2d0s0 read/write/setuid/devices/intr/largefiles/logging/xattr/onerror=panic/dev=1980040
[06:22:32] <soylent> nm then
[06:23:09] <soylent> also
[06:23:12] <soylent> what's best
[06:23:17] <soylent> TCP_STRONG_ISS=1
[06:23:17] <soylent> or
[06:23:19] <soylent> TCP_STRONG_ISS=2
[06:23:20] <soylent> ?
[06:23:21] <Auralis> 2
[06:23:26] <soylent> that's what I thought
[06:23:38] <soylent> why isn't it default then?
[06:23:45] <Auralis> dunno
[06:32:04] <soylent> is
[06:32:07] <soylent> set noexec_user_stack=1
[06:32:08] <soylent> set noexec_user_stack_log=1
[06:32:11] <soylent> still required?
[06:32:16] <soylent> (in /etc/system)
[06:32:53] *** sparc-kly_ has joined #opensolaris
[06:37:06] *** spike723_ has joined #opensolaris
[06:39:25] <soylent> I guess it is.
[06:45:15] *** spike723 has quit IRC
[06:51:59] *** karrotx has quit IRC
[06:58:10] *** sparc-kly has quit IRC
[06:59:11] *** dclarke has joined #opensolaris
[07:04:11] *** gnu2it2 has quit IRC
[07:07:52] *** eboutilier has quit IRC
[07:08:35] *** uebayasi has quit IRC
[07:15:56] *** gregory has joined #opensolaris
[07:17:39] *** uebayasi has joined #opensolaris
[07:22:00] *** gregory is now known as MaGre
[07:25:51] *** tsoome has quit IRC
[07:29:08] <myrkraverk> hmm, I reinstalled sol (excr b50) and now my usb hdds don't show up - any ideas?
[07:29:19] <myrkraverk> (not even with format -e)
[07:29:34] <Auralis> whats messages says?
[07:32:30] <myrkraverk> what message?
[07:34:39] <myrkraverk> oh, strange dmesg claims there is no driver for my usb disk - it seems
[07:34:45] *** ndroux has quit IRC
[07:34:52] <myrkraverk> how do I find a driver?
[07:35:46] <myrkraverk> and why don't I have one *this* time?
[07:36:15] *** alobbs has quit IRC
[07:48:22] *** tsoome has joined #opensolaris
[07:50:32] *** Capricorn^80 has joined #opensolaris
[07:51:49] <Capricorn^80> how to specify /software/nss/ns-allinone-2.30/otcl-1.12,/software/nss/ns-allinone-2.30/lib,into your LD_LIBRARY_PATH environment variable.
[07:53:54] *** tw3rp has joined #opensolaris
[07:54:23] *** tw3rp has quit IRC
[07:58:08] *** sparc-kly__ has joined #opensolaris
[07:59:06] <dclarke> hello .. anyone around ?
[07:59:30] <dclarke> can someone do a traceroute for me ?  then email me the data to dclarke at blastwave dot org ?
[07:59:31] <myrkraverk> dclarke: no, not really ;/
[07:59:36] <dclarke> traceroute 66.225.151.252
[07:59:59] <dclarke> my bloody ISP is dealing with it but my connection is "flapping" off and on rapidly
[08:00:10] <dclarke> the web site works but a session .. no way
[08:00:32] <myrkraverk> dclarke: oh - btw, would you know where to find usb hdd drivers, when they seem to have escaped installation?
[08:00:35] <dclarke> any data I get will be helpful ..
[08:00:52] <myrkraverk> (it worked last time...)
[08:00:57] <dclarke> you mean you attach a USB device and its not detected ?
[08:01:04] <jamesd> dclarke, doing it now
[08:01:06] *** jmcp has left #opensolaris
[08:01:15] <dclarke> thanks jamesd
[08:01:18] *** jmcp has joined #opensolaris
[08:01:23] <myrkraverk> it's detected (of sorts) but no driver loads - no driver for device
[08:01:26] <dclarke> the ISP is telling me I need to buy a network managed device
[08:01:33] <myrkraverk> dclarke: am running it too, btw ;P
[08:01:37] <dclarke> baloney
[08:01:41] <dclarke> thanks !!
[08:01:51] <Capricorn^80> how to specify /software/nss/ns-allinone-2.30/otcl-1.12,/software/nss/ns-allinone-2.30/lib,into your LD_LIBRARY_PATH environment variable.
[08:01:52] <razrX> dclarke: do you want a traceroute from anywhere or from within the US only?
[08:01:58] *** PosixC has joined #opensolaris
[08:02:01] *** PosixCompliant has joined #opensolaris
[08:02:03] <dclarke> from anywhere !
[08:02:04] <myrkraverk> why would I be missing usb hdd driver *now* ?
[08:02:11] <dclarke> myrkraverk: thats weird
[08:02:16] *** PosixC has left #opensolaris
[08:02:17] <myrkraverk> dclarke: yes
[08:02:20] <dclarke> myrkraverk: do you have vold running ?
[08:02:35] <dclarke> myrkraverk: like .. what does svcs -av | grep vol   say ?
[08:03:17] <myrkraverk> I just did svcadm enable volfs
[08:03:36] <dclarke> myrkraverk: try a volcheck
[08:03:37] <razrX> dclarke: you got new email ;)
[08:03:40] <jamesd> dclarke, sent
[08:03:43] <twincest> dc: sent traces from 4 locations
[08:03:47] <twincest> (3 EU, one US)
[08:03:53] <dclarke> perfect !!
[08:03:58] <myrkraverk> dclarke: no luck ;(
[08:04:21] <dclarke> I get really miffed when I fork over money for a 10mbps connection and the ISP tells me on the phone that it "works for him"
[08:04:38] <dclarke> like gee .. the it works for the ISP .. thats nice
[08:04:42] <twincest> oh, have korea too
[08:05:09] <dclarke> cool !!
[08:05:12] <myrkraverk> dclarke: sent from iceland
[08:05:15] <jamesd> dclarke, i sent from two locations...  us, and centerbury england ...
[08:05:29] <dclarke> you see .. this is really great data
[08:05:42] <dclarke> how the heck can a ISP tech guy tell me its a problem on "my end"
[08:05:50] <jmcp> dclarke: my traceroute gets to static-66-225-157-141.ptr.terago.ca and then I get nuttin'
[08:05:53] <twincest> well, your router could be broken ;-)
[08:06:02] <jengelh> dclarke : he probably implied PEBKAC
[08:06:04] <jamesd> i get one from london england if you haven't gotten enough data yet
[08:06:05] <dclarke> but it works for the ISP
[08:06:10] <jengelh> (though he can't know)
[08:06:12] <razrX> the ISP tech guy should get brain in gear first before opening mouth
[08:06:31] <dclarke> well .. he just escalated it off of his desk
[08:06:42] <dclarke> Thanks You everyone !
[08:06:47] <razrX> yvw dclarke
[08:06:52] <dclarke> damn .. its 2AM .. I need sleep
[08:07:13] <razrX> dclarke: located at east coast somewhere?
[08:07:16] <myrkraverk> dclarke: volcheck doesn't help (and probably wouldn't, if the the usb hdd driver is missing)
[08:07:17] <dclarke> I think I will compile a single large email to the NOC at the ISP then go crash
[08:07:24] <twincest> curiously, this one stops at a different place
[08:07:43] <dclarke> myrkraverk: is there a way to do a ... devfsadm -C -v  ?
[08:07:46] <myrkraverk>   62.159.61.230 (62.159.61.230)  179.870 ms  175.730 ms  175.636 ms <--- this is the last line I see
[08:08:00] <dclarke> twincest: yes .. it is odd that it dies at different points from different sources
[08:08:03] <twincest> multihomed i guess
[08:08:08] <jengelh> note that traceroute usually depends on UDP
[08:08:10] <dclarke> twincest: I tried to point that out to no avial
[08:08:27] <dclarke> I opened up ICMP for the traceroutes to work
[08:08:39] <twincest> traceroute usually sends UDP by default, not ICMP
[08:08:41] <jengelh> well it uses both udp+icmp
[08:08:46] <dclarke> I think that traceroute will use ICMP by default
[08:08:47] <jengelh> sends udp, receives icmp
[08:08:49] <twincest> (it sends UDP and listens for ICMP)
[08:08:53] <jengelh> heh
[08:08:54] <twincest> dc: no, it won't
[08:08:54] <dclarke> ah
[08:09:09] <jengelh> there IS a "traceroute ICMP packet type", however
[08:09:14] <jengelh> type 30
[08:09:18] <twincest> traceroute -I (on solaris) does icmp
[08:09:19] <dclarke> perhaps I need to select a few options to traceroute
[08:09:19] <myrkraverk> dclarke: that gives me removing file: /dev/usb/device1 /dev/usb/device2
[08:09:51] <dclarke> myrkraverk: urk ..
[08:10:02] <myrkraverk> (I did mine from linux - my sol box does not have a working nic driver)
[08:10:09] <dclarke> myrkraverk: so now the device is really really gone because no driver existed to detect it
[08:10:41] <myrkraverk> dclarke: aparently, yes - but it worked on *this* box in my previous installation ;(
[08:10:52] <razrX> well, a traceroute -I to the addy gives the same result from a Sol 10 box. it stops dead at the .230 address
[08:11:11] <myrkraverk> the only thing I know I did different was not using the graphical installer - this time ;/
[08:11:31] <razrX> hey, what do you know ... at hop 13 it gets there dclarke
[08:11:42] <razrX> from my location
[08:11:55] <razrX> now let me do the trace again without the -I switch
[08:12:24] <razrX> and yes, my site does not block udp/icmp packets
[08:12:50] <asyd> \_o<
[08:12:59] <dclarke> traceroute -I -n -P 2 -q 5 -w 2 66.225.151.252
[08:13:08] <dclarke> thats what I am runnign now from Germany
[08:13:26] <razrX> without the -I I'm not reaching the destination
[08:13:38] <myrkraverk> I get there, with -I in hop 20
[08:13:47] <dclarke> damn
[08:13:52] <dclarke> I may be screwed then
[08:14:03] <dclarke> the problem is on "my end" perhaps
[08:14:13] <dclarke> my ssh sessions are dropping like flies
[08:15:35] <dclarke> razrX: the issue is UDP perhaps
[08:15:37] <myrkraverk> dclarke: you want the traceroute with -I ?
[08:15:44] <dclarke> myrkraverk: does it work ?
[08:15:59] <myrkraverk> dclarke: aparently, yes (am running again without)
[08:16:18] <dclarke> perhaps the issue is that I now need to buy a "managed network service" from the ISP just like they said
[08:16:25] <dclarke> great .. one more bill to pay
[08:17:21] <dclarke> okay ... well with a command line of  traceroute -I -n -P 2 -q 5 -w 2 66.225.151.252
[08:17:26] <jamesd> dclarke, the last command isn't working from my box..
[08:17:33] <dclarke> it works with only one hop full of asterisks
[08:17:47] <dclarke> jamesd : it fails ?
[08:17:54] <jamesd> i'm on my second line of asterisks now
[08:18:05] <jamesd> now 3
[08:18:24] *** slowhog has left #opensolaris
[08:18:34] <jamesd> okay after 3 lines of asterisks it works... but not exactly   expiring of trust
[08:19:01] <dclarke> okay .. that would be consistent with a flapping connection
[08:19:11] <dclarke> up and down every few seconds
[08:19:17] <jamesd> 10  154.11.63.18  198.925 ms  98.575 ms  39.902 ms  21.819 ms  22.486 ms
[08:19:17] <jamesd> 11  * * * * *
[08:19:17] <jamesd> 12  * * * * *
[08:19:17] <jamesd> 13  * * * * *
[08:19:17] <jamesd> 14  66.225.151.252  44.296 ms  43.192 ms  31.331 ms  32.580 ms  56.105 ms
[08:19:20] <Capricorn^80> how to specify /software/nss/ns-allinone-2.30/otcl-1.12,/software/nss/ns-allinone-2.30/lib,into your LD_LIBRARY_PATH environment variable.
[08:19:21] <razrX> same result here
[08:19:26] <dclarke> yeah .. I have nearly the same thing
[08:19:33] <razrX>  9  62.159.61.230  119.749 ms  104.670 ms  121.314 ms  118.493 ms  104.231 ms
[08:19:37] <razrX> 10  * * * * *
[08:19:40] <razrX> 11  * * * * *
[08:19:42] <razrX> 12  * * * * *
[08:19:45] <razrX> 13  66.225.151.252  128.471 ms  119.450 ms  127.372 ms  121.228 ms  118.269 ms
[08:20:07] <dclarke> I have seven lines of asterisks
[08:20:15] <dclarke> Capricorn^80: what do you need ?
[08:20:30] <dclarke> Capricorn^80: just set LD_LIBRARY_PATH=path1:path2:path3
[08:21:15] <Capricorn^80> should i use setenv or or just do PATH setting ?
[08:21:42] <razrX> Capricorn^80: use crle(1)
[08:21:59] *** sparc-kly_ has quit IRC
[08:22:53] <Capricorn^80> razrX: tell me how how to setup this . dont tell to read man crle
[08:23:22] *** Teltariat has quit IRC
[08:23:34] <dclarke> Capricorn^80: do this
[08:23:35] <razrX> crle -c /var/ld/ld.config -l /lib:/usr/lib:/software/nss/ns-allinone-2.30/otcl-1.12,/software/nss/ns-allinone-2.30/lib
[08:23:43] <chain-lightning> LD_LIBRARY_PATH=<SOMPATH>;export LD_LIBRARY_PATH
[08:23:58] <dclarke> Capricorn^80: LD_LIBRARY_PATH=path1:path2:path3;export LD_LIBRARY_PATH
[08:24:00] <chain-lightning> (if you're in a bourne shell)
[08:24:11] <dclarke> Capricorn^80: but ..  LD_LIBRARY_PATH is evil
[08:24:34] <dclarke> okay folks .. thanks for helping with the traceroute thing
[08:24:47] <dclarke> to me its clearly "not working" regardless what the ISP says
[08:25:19] <dclarke> I gotta get some sleep .. I have to drive in four hours
[08:25:30] <razrX> nn then dc
[08:25:31] <chain-lightning> Hope it works out for you
[08:25:37] *** dclarke is now known as dclarke_ZZZzzz
[08:26:32] *** sahafeez has joined #opensolaris
[08:27:10] <Capricorn^80> wat should i say as i installed NS2 n after that i got some configuration notes
[08:27:35] <chain-lightning> later, all
[08:27:38] *** chain-lightning has left #opensolaris
[08:27:46] <dclarke_ZZZzzz> oh .. last thought for the night
[08:27:55] <dclarke_ZZZzzz> root@enterprise # uptime
[08:27:55] <dclarke_ZZZzzz>   8:27am  up 570 day(s), 19:07,  5 users,  load average: 0.43, 0.24, 0.21
[08:28:04] <dclarke_ZZZzzz> ya gotta love Solaris stability
[08:28:09] <razrX> :)
[08:29:01] <dclarke_ZZZzzz> now I am off to sleep
[08:29:35] <sahafeez> running snv_50 on a p3 w/512mb ram, ide, an old ati. jds is usable. i am suprised.
[08:30:30] <Capricorn^80> dclarke: i m using bash
[08:35:08] <Capricorn^80> dclarke: is that command ok for bash ?
[08:47:56] *** sahafeez has quit IRC
[08:50:10] *** bengtf has joined #opensolaris
[08:52:38] *** miffe_ has joined #opensolaris
[08:56:04] <Gr|ffous> has anyone else noticed ZFS performance getting worse over time?
[08:57:08] <Gr|ffous> my desktop is definitely getting worse. I turned off compression, in case that was it. As an example, I did a background copy of 2 large files to my nfs server, and then tried running su on this box. It took 20 seconds for the password prompt to come up!
[08:58:28] *** mazon is now known as Mazon
[08:58:53] *** sahafeez has joined #opensolaris
[08:59:54] <Gr|ffous> this is interesting: # zfs snapshot angel@31/10/06
[08:59:55] <Gr|ffous> internal error: Invalid argument
[08:59:55] <Gr|ffous> Abort (core dumped)
[09:00:26] <Gr|ffous> I take it that I can't use a forwardslash - fine. But should it dump core? I'm newish to solaris - is that normal?
[09:00:55] <oxygene> no, not normal
[09:01:03] *** sylvain has joined #opensolaris
[09:01:40] <oxygene> Gr|ffous: how about using 2006-10-31 notation? sorts properly, too
[09:03:01] *** miffe_ has quit IRC
[09:03:03] <Gr|ffous> oh sure, that works. I only mention it because of the core dump thing.
[09:03:34] * Gr|ffous wonders if he should be filing a bug. I'm running snv45, so it's kinda old
[09:15:27] *** sylvain_4 has joined #opensolaris
[09:15:43] *** sahafeez has left #opensolaris
[09:19:17] *** Capricorn^80 has quit IRC
[09:19:21] *** dunc has joined #opensolaris
[09:23:35] *** trede has joined #opensolaris
[09:25:31] *** simford__ has joined #OpenSolaris
[09:28:25] *** sylvain has quit IRC
[09:30:19] *** simfordWFH has quit IRC
[09:33:35] *** Symmetria has quit IRC
[09:35:23] *** lplatypus has joined #opensolaris
[09:35:59] *** axxl has joined #opensolaris
[09:37:41] *** Dar_HOME is now known as Dar
[09:38:57] *** simford__ is now known as simfordWFH
[09:41:30] *** lplatypus has quit IRC
[09:41:48] *** |tsoome| has joined #opensolaris
[09:42:08] *** tsoome has quit IRC
[09:42:39] *** alobbs has joined #opensolaris
[09:43:33] *** [1]Cyl has joined #opensolaris
[09:44:57] *** vmhobbes has joined #OpenSolaris
[09:46:13] *** Symmetria has joined #opensolaris
[09:49:54] *** sylvain_4 is now known as sylvain
[09:52:12] *** uebayasi has quit IRC
[10:00:35] <asyd> s8
[10:00:37] <asyd> oups
[10:00:38] <Gman> richlowe, http://www.markshuttleworth.com/archives/80
[10:00:46] *** Cyl has quit IRC
[10:00:47] *** [1]Cyl is now known as Cyl
[10:00:59] <Gman> richlowe, the smart server stuff is pretty cool
[10:07:05] *** damienc has joined #opensolaris
[10:09:25] *** Capricorn^80 has joined #opensolaris
[10:09:34] <Capricorn^80> how to specify /software/nss/ns-allinone-2.30/tcl8.4.13/library into   TCL_LIBRARY environmental variable ?
[10:10:06] *** alanc_away has quit IRC
[10:10:07] *** Gman has quit IRC
[10:10:54] *** alanc_away has joined #opensolaris
[10:11:01] *** Gman has joined #opensolaris
[10:16:37] <Capricorn^80> anyone there ?
[10:17:18] *** uebayasi has joined #opensolaris
[10:21:13] <Capricorn^80> ?
[10:28:38] *** jamesd has quit IRC
[10:28:43] *** jamesd_ has joined #opensolaris
[10:31:47] *** halton has quit IRC
[10:32:40] *** |ReIkO| has joined #opensolaris
[10:32:43] <|ReIkO|> hi all
[10:33:04] <|ReIkO|> someone knows how can setup my laptop keyboard?
[10:33:11] <|ReIkO|> all works
[10:33:22] <|ReIkO|> but END and BEGIN
[10:33:27] <|ReIkO|> doesn't works
[10:33:31] *** Fish has joined #opensolaris
[10:33:44] <|ReIkO|> kdmconfig only support Generic Italian
[10:33:53] <Fish> hello
[10:34:03] <|ReIkO|> hello
[10:40:32] <phips> Morning all
[10:42:01] <raph_ael> hello
[10:44:14] *** vmhobbes has quit IRC
[10:47:38] *** |tsoome| has quit IRC
[10:51:21] *** MattMan has joined #opensolaris
[10:54:20] *** |ReIkO| has left #opensolaris
[10:56:21] *** calumb has joined #opensolaris
[10:56:37] <Gr|ffous> is "dumping core" a bit like a seg fault?
[10:58:45] *** |tsoome| has joined #opensolaris
[10:59:52] *** calumb is now known as calAFK
[11:17:36] <boyd> jamesd_: You half asleep tonight?
[11:25:08] *** vmhobbes has joined #OpenSolaris
[11:25:58] *** bengtf has quit IRC
[11:26:32] *** calAFK has quit IRC
[11:26:51] *** calumb has joined #opensolaris
[11:28:38] *** simfordWFH has quit IRC
[11:29:39] *** kimc has joined #opensolaris
[11:30:02] <kimc> good morning
[11:31:02] *** coolvibe has joined #opensolaris
[11:33:23] <Peanut> Hoi coolvibe
[11:33:50] <coolvibe> hey peanut
[11:33:55] <coolvibe> wassup
[11:34:09] <coolvibe> got a new job :)
[11:34:24] <Peanut> <AOL> Me too </AOL>
[11:34:29] <coolvibe> hahaha
[11:34:52] <coolvibe> I'm at TNT post, working as a Solaris admin
[11:35:14] <Symmetria> heh
[11:35:15] <Peanut> Nice, congratulations.
[11:35:22] <Symmetria> I wish I could find decent people to give new jobs to :(
[11:35:24] <coolvibe> thanks
[11:35:25] <razrX> gratz coolvibe
[11:35:42] <coolvibe> thanks razr
[11:35:49] <Peanut> I'm starting with JIVE (Joint Institute for VLBI Europe) next month, playing with radio-telescopes and huge bandwiths.
[11:35:51] <Symmetria> spent thousands of dollars advertising for good networking people and get absolutely bugger all in the way of decent responses in this country :(
[11:36:00] <coolvibe> Peanut: sounds like fun
[11:36:12] <Peanut> What country would that be, Symmetria ?
[11:36:24] <Symmetria> Peanut you into VLBI = very long baseline interferometory or something?
[11:36:27] <Symmetria> south africa
[11:36:43] <Gman> mmm, wonder when the companion will realize that jds is actually shipping the ogg/vorbis stuff
[11:36:50] <Peanut> Symmetria: yes.
[11:36:57] <Symmetria> heh, peanut, when you start working with JIVE, you can gimme access to astronomy data to add to my academic data mirror
[11:37:07] <Symmetria> I've got craploads of earth science data on there
[11:37:13] <Symmetria> got craploads of bioinformatics data
[11:37:18] <Symmetria> but very little astronomy stuff
[11:37:25] <coolvibe> damn, I'm just loving nomachine right now :)
[11:37:28] <Symmetria> and very little space based stuff
[11:38:06] <Symmetria> (I'm CTO of the south african research and education network, hence the academic data sets)
[11:38:16] <Peanut> Symmetria: Sorry, 'you can gimme access'? I think my new employer is not expecting me to treat his data quite like that.
[11:38:28] <Symmetria> heh Peanut no, but I'd assume that there are public data sets
[11:38:29] <Symmetria> they release
[11:38:35] <Symmetria> which I need mirror access to
[11:38:41] <Symmetria> because thats what most of those groups do :)
[11:39:02] <Symmetria> heh, and I dont mirror public data sets without permission :)
[11:39:20] <quasi> Peanut: sounds like a fun job
[11:39:31] <coolvibe> grr, damn IRIX compiler
[11:39:31] <kimc> Peanut: do you let the ham radio guys in on that ?
[11:39:48] <Peanut> kimc: Peanut == PE1NUT, 'nuff said ;-)
[11:39:57] <kimc> cool..
[11:40:13] <kimc> kimc == w8hd
[11:40:21] * quasi assumes that a place like that would have a good amount of data to churn and store
[11:40:40] <Peanut> They're actually restoring the old 25m dish for use by radio amateur and amateur radio astronomers because it's no good for current science anymore, but would be a shame to let rot away.
[11:41:09] <coolvibe> hm, IRIX has no err.h, damn.
[11:42:05] <boyd> Hmm... Just had a Dell multi-function 944 printer arrive for free... shame that linuxprinting.org describes it as a "paperweight" :(
[11:42:28] <Gman> heh
[11:42:41] <kimc> Peanut: 25m should work great for moonbounce :)
[11:43:44] <Peanut> http://nl.wikipedia.org/wiki/Afbeelding:DRT_2006_06.JPG
[11:44:43] *** swa has quit IRC
[11:46:07] <kimc> wow.. thats 25m..
[11:46:43] <Peanut> But currently also a 'paperweight', one of the azimuth motors is dead.
[11:50:54] <boyd> I bet your was more expensive than mine
[11:51:48] <kimc> hopefully someone has a line on a motor
[11:57:32] *** yongsun has quit IRC
[11:58:47] <kimc> a map shows a vlbi site in central america ..where is that ?
[12:01:46] *** uebayasi has quit IRC
[12:02:11] *** alienjr|szkola has joined #opensolaris
[12:02:19] <alienjr|szkola> czesc
[12:02:58] *** alienjr|szkola has quit IRC
[12:04:06] *** swa has joined #opensolaris
[12:04:34] *** alienjr|szkola has joined #opensolaris
[12:12:33] *** vmhobbes has quit IRC
[12:18:21] *** Gman is now known as GmanZZZ
[12:18:40] *** swa has quit IRC
[12:22:28] <kimc> time to get rolling.. cu all
[12:22:30] *** kimc has left #opensolaris
[12:23:01] *** IRCMonkey has joined #opensolaris
[12:23:26] *** FBdev has joined #opensolaris
[12:23:43] *** IRCMonkey has quit IRC
[12:23:45] *** Capricorn^80 has quit IRC
[12:24:00] *** jimgris has joined #opensolaris
[12:24:43] *** vmhobbes has joined #OpenSolaris
[12:29:57] *** jteo has joined #opensolaris
[12:32:07] *** GmanZZZ has quit IRC
[12:38:16] *** alienjr|szkola has quit IRC
[12:49:42] *** vmhobbes has quit IRC
[12:56:50] *** vmhobbes has joined #OpenSolaris
[12:58:00] *** Byron has joined #opensolaris
[13:05:29] *** polk__ has quit IRC
[13:05:54] *** polk_ has joined #opensolaris
[13:06:51] *** gm152 has quit IRC
[13:06:59] *** calumb has quit IRC
[13:07:00] *** mnowak has quit IRC
[13:07:09] *** mnowak has joined #opensolaris
[13:07:15] *** calumb has joined #opensolaris
[13:10:32] *** bernroth has joined #opensolaris
[13:10:39] <bernroth> hi!
[13:10:45] *** deather__ has joined #opensolaris
[13:10:48] <coolvibe> hey!
[13:11:48] <bernroth> i was told to find more solaris 9 experts here :)
[13:11:55] <bernroth> is that true?
[13:12:02] <coolvibe> bernroth: yeah
[13:12:22] <coolvibe> but we're kinda focusing on solaris 11 here ;)
[13:12:36] *** postwait has quit IRC
[13:12:36] *** pjd_ has quit IRC
[13:12:36] *** Symm has quit IRC
[13:12:36] *** deather_ has quit IRC
[13:12:36] *** bondolo has quit IRC
[13:12:36] *** |joni| has quit IRC
[13:12:36] *** lisppaste3 has quit IRC
[13:12:37] *** Drone has quit IRC
[13:12:37] *** svoboda has quit IRC
[13:12:37] *** switch has quit IRC
[13:12:37] *** AbeFroman has quit IRC
[13:12:48] <coolvibe> netsplit?
[13:12:52] <bernroth> hmm... our load balancers are running s9 since years now and it works so why change?
[13:13:05] <coolvibe> bernroth: nobody says you should
[13:13:34] <bernroth> perhaps to solve my arp problem... I don't know where it comes from but it seems static arp entries (flags SP) are getting overwritten with wrong mac addresses
[13:13:39] <coolvibe> this is the opensolaris channel, most people work on the open source solaris here
[13:14:05] *** postwait has joined #opensolaris
[13:14:05] *** pjd_ has joined #opensolaris
[13:14:06] *** Symm has joined #opensolaris
[13:14:06] *** deather_ has joined #opensolaris
[13:14:06] *** bondolo has joined #opensolaris
[13:14:06] *** |joni| has joined #opensolaris
[13:14:06] *** lisppaste3 has joined #opensolaris
[13:14:06] *** Drone has joined #opensolaris
[13:14:06] *** svoboda has joined #opensolaris
[13:14:06] *** switch has joined #opensolaris
[13:14:06] *** AbeFroman has joined #opensolaris
[13:14:55] *** deather_ has quit IRC
[13:15:00] *** AbeFroma1 has joined #opensolaris
[13:15:03] *** AbeFroman has quit IRC
[13:15:05] *** switch has quit IRC
[13:15:13] *** stefanjo has left #opensolaris
[13:15:34] <coolvibe> bernroth: ever tried snooping and seeing where those faulty mac adresses come from?
[13:15:41] *** switch has joined #opensolaris
[13:16:13] <coolvibe> bernroth: ethereal, snoop and tcpdump are your friends here
[13:16:43] <bernroth> the mac address is from the 2nd load balancer but snooping is impossible because when an ip swap happens ther are thousands of arp request/reply/collissions etc.. in the network.. and: it's in production use
[13:17:12] <bernroth> quite disadvantageous situation here :-/
[13:17:27] <lasseoe> indeed
[13:17:28] <coolvibe> why does it swap IP? is that intentional?
[13:17:43] <bernroth> hi lasseoe :)
[13:17:50] <lasseoe> moin bernroth
[13:17:58] <coolvibe> hey lasseoe
[13:18:04] <lasseoe> hey coolvibe
[13:19:21] <bernroth> IP swap may happen if the uplink of both servers is down and when the connectivity comes back. so for some seconds both servers have the same ip address and corrupting their arp table
[13:19:57] <bernroth> when the conflict is solved the arp table keeps corrupt and some service ip's are not responding anymore
[13:20:26] *** Symm has quit IRC
[13:20:38] *** Symm has joined #opensolaris
[13:20:41] <bernroth> my question is now if there is a way to avoid the update of static arp entries or/and disable arp announcements upon an IP conflict
[13:21:59] *** Trident has quit IRC
[13:24:44] <twincest> use a method of load balancing that gives the same MAC to both devices
[13:24:49] <twincest> (HSRP works this way, for example)
[13:25:18] <coolvibe> grrrr, why is that IRIX compiler so pedantic...
[13:25:31] <bernroth> we are using ZXTM from zeus as LB software
[13:26:18] *** Trident has joined #opensolaris
[13:30:52] <boyd> Ok, I need someone to bounce off... I have TX installed here, and a first labelled zone running. I can't log in at the zone console and I've  worked out that this is because the zone can't resolve it's own hostname. (i.e. getent hosts public returns nothing - public is the zone hostname). I've looked at /etc/inet/hosts and /etc/nsswitch.conf but they look fine... I think my brain is fading..
[13:31:21] <boyd> Anyone got an idea they want to throw at me>
[13:31:22] *** |tsoome| has quit IRC
[13:31:22] <boyd> ?
[13:31:47] <lasseoe> boyd, resolv.conf is ok as well ?
[13:32:04] <boyd> I have no dns set up at all
[13:32:19] * coolvibe bitches at FreeBSD/sparc64 for destroying the labels of my disks. Now Solaris install can't see 'em. Fiddlesticks
[13:32:27] <jamesd_> zlogin -S  should get you in.
[13:32:32] <coolvibe> now I have to format(8) lots of disks :(
[13:32:49] <boyd> Yeah, I can log in like that but that doesn't fix the problem
[13:33:12] <boyd> coolvibe: fmthard -s /dev/null /dev/rdsk/c0t3d0s2
[13:33:40] <boyd> actually, plain zlogin (without -C) works... but that's not the point
[13:33:47] <coolvibe> boyd: just plunking a new label on them is enough, but I have to do that with 12 disks
[13:34:01] <boyd> coolvibe: That's what fmthard does
[13:34:08] <coolvibe> not the end of the world, but just irritating
[13:34:22] <boyd> for disk in /dev/rdsk/c1*s2; do fmthard....; done
[13:34:27] <coolvibe> boyd: ah cool, that could save some time
[13:34:32] <LeftWing> No DNS? =P
[13:34:49] <boyd> It's a VM with host-only networking for testing...
[13:34:52] <LeftWing> Ahh.
[13:35:16] <boyd> So, I feel like I'm going mad. nsswitch.conf has hosts: files
[13:35:26] <twincest> what about ipnodes?
[13:35:27] <boyd> hosts has 172.16.0.4 public
[13:35:28] <coolvibe> why /does/ solaris lose it
[13:35:39] <boyd> ipnodes -> hosts nowadays
[13:35:42] <damienc> boyd: I have pinged some TJDS engineers on #jds
[13:35:43] <coolvibe> why /does/ solaris lose it's cool when it sees disks without labels?
[13:35:51] <damienc> no reply yet - prob at lunch
[13:35:56] * boyd nods
[13:35:59] <boyd> Thanks damienc
[13:36:15] * boyd didn't know there *was* a #jds
[13:36:30] <twincest> i'm guessing sun-internal #jds :)
[13:36:34] <damienc> yes
[13:36:37] *** |tsoome| has joined #opensolaris
[13:36:38] <boyd> Ah :)
[13:36:49] <Auralis> so we can't go there and bitch all the time about how crap gnome2 is :)
[13:36:57] <boyd> hehe....
[13:37:08] <damienc> kick kick kick ;)
[13:37:14] <twincest> JDS4 will be a KDE clone written as a ksh93 builtin
[13:37:23] <lasseoe> haha
[13:37:26] <boyd> lol
[13:37:27] <twincest> libshell will be 150MB
[13:37:29] * LeftWing sprays twincest with deroland.
[13:37:30] <damienc> Auralis: or we might tell you to fix some bugs ;)
[13:37:45] <boyd> Ok, so nobody has an obvious idea... I'm not *completely* stupid yet...
[13:38:03] <Auralis> hehe, why should i bother ti fix bug in gnome2? you don't want a rewrite
[13:38:08] <LeftWing> When I get Trusted Extensions for 10U2 I'll have a look at them. =P
[13:38:11] <LeftWing> ... 10U3
[13:38:19] <twincest> boyd: /etc/inet/ipnodes?  or is that gone now?
[13:38:36] <boyd> It's a symlink to /etc/inet/hosts these days (hooray!)
[13:38:42] <twincest> finally :)
[13:40:08] *** ericr has quit IRC
[13:40:33] <boyd> Hmm... basically I'm getting the behaviour from the second last post of this thread... but the solutions suggested don't help... I think I'll post..
[13:40:36] <boyd> http://www.opensolaris.org/jive/thread.jspa?messageID=64747
[13:43:13] <boyd> Hmm.. maybe this...
[13:46:10] *** sickness has quit IRC
[13:47:10] *** sickness has joined #opensolaris
[13:48:33] *** sparc-kly_ has joined #opensolaris
[13:52:52] *** trede has quit IRC
[14:01:00] *** Trident has quit IRC
[14:01:09] *** Trident has joined #opensolaris
[14:01:28] <PosixCompliant> should b52 be out today?
[14:02:09] <damienc> I don't think so. snv51 was only released last Thursday
[14:02:26] *** damienc is now known as dcFood
[14:03:08] <PosixCompliant> the scheule says 10/30/2006
[14:03:25] <PosixCompliant> and it is usually released on tuesday
[14:03:36] <quasi> PosixCompliant: isn't that the internal date?
[14:04:11] <PosixCompliant> quasi, not sure
[14:05:57] <PosixCompliant> in fact I don't know
[14:08:38] *** laca has joined #opensolaris
[14:08:50] *** MaGre_ has joined #opensolaris
[14:08:51] *** MaGre has quit IRC
[14:08:55] *** MaGre_ is now known as MaGre
[14:09:37] *** laca has quit IRC
[14:10:21] *** laca has joined #opensolaris
[14:17:48] <coolvibe> what? snv51 is out? I just downloaded snv50
[14:17:52] <coolvibe> damnit
[14:18:05] <timeless> heh
[14:18:09] <coolvibe> great timing :)
[14:19:16] *** sparc-kly__ has quit IRC
[14:21:32] <timeless> what's a good way to find out if my process is doing anything useful? prstat says it's vaguely alive...
[14:21:38] <timeless>   6030 timeless 3951M 3100M sleep   60    0   3:02:04 0.1% glimpseindex/1
[14:22:38] <trygvis> truss it
[14:22:41] <trygvis> truss -p >pid>
[14:22:51] <trygvis> s,>,<,
[14:24:08] <timeless> swift% truss -p 6030
[14:24:08] <timeless> truss: no such process: 6030
[14:24:18] <timeless>   6030 timeless 3951M 3105M sleep   60    0   3:02:05 0.1% glimpseindex/1
[14:24:23] <timeless> prstat still shows it?
[14:25:02] <trygvis> ps too?
[14:25:31] <timeless> swift% ps -a|grep 6030
[14:25:31] <timeless>  6030 pts/9     182:06 glimpsei
[14:26:44] <timeless> swift% ls /proc/6030
[14:26:44] <timeless> /proc/6030: Permission denied
[14:26:44] <timeless> swift% whoami
[14:26:45] <timeless> timeless
[14:26:45] <timeless> swift% ls -l /proc|grep 6030
[14:26:46] <timeless> dr-x--x--x   5 timeless webservd     864 Oct 31 10:25 6030
[14:26:51] *** Slepek has joined #opensolaris
[14:27:04] <timeless> cute, i have to be the right group
[14:27:22] <trygvis> clever
[14:27:26] <timeless> swift% truss -p 6030
[14:27:27] <timeless> truss: no such process: 6030
[14:27:28] <timeless> nope
[14:27:35] *** Hunger- has quit IRC
[14:27:38] <trygvis> as root?
[14:27:39] <timeless> but at least i can read it if i'm the right group
[14:27:48] *** _Hunger- has joined #opensolaris
[14:28:29] <timeless> um, this may take a bit, my system is busy :)
[14:28:51] <timeless> # truss -p 6030
[14:28:52] <timeless> truss: no such process: 6030
[14:31:49] *** loke has joined #opensolaris
[14:32:28] * timeless contemplates mdb -p 6030
[14:32:49] <timeless> swift% mdb -p 6030
[14:32:50] <timeless> mdb: failed to open /proc/6030/object/a.out: No such file or directory
[14:32:56] <timeless> finally a useful error message :)
[14:33:23] <trygvis> dude, you should have truss'ed the truss!
[14:33:43] <timeless> yeah yeah
[14:34:00] <timeless> swift% cd /proc/6030
[14:34:00] <timeless> swift% ls -l |grep object
[14:34:00] <timeless> dr-x------   2 timeless webservd     544 Oct 31 10:25 object
[14:34:51] <timeless> http://pastebin.mozilla.org/1129
[14:36:22] <timeless> ok, that's funny
[14:36:41] <timeless> swift% ls -la /proc/6030/as
[14:37:04] * timeless wonders if solaris is really saying "i don't have another 4gb of memory for you to read that file)
[14:37:29] *** mrdeviant has joined #opensolaris
[14:37:29] <timeless> why does truss try to open it RDRW, isn't RD good enough?
[14:45:01] <|tsoome|> damit, whats happening, mysun is down, veritas.com does not work and so on....
[14:45:07] <|tsoome|> sick...
[14:46:10] <wilbury> hi |tsoome|
[14:47:24] *** solarisjon has joined #opensolaris
[14:48:22] <sickness> ?
[14:50:23] *** laca is now known as lacaAFK
[14:52:28] <timeless> http://pastebin.mozilla.org/1130
[14:52:37] <timeless> is it bad to not have a.out in proc? :)
[14:53:06] *** uebayasi has joined #opensolaris
[14:53:18] <timeless> ok, i don't get it
[14:53:34] <timeless> there is an object directory and it's dr-x, shouldn't root be able to cd to it? :)
[14:54:13] <|tsoome|> depends
[14:54:20] *** |tsoome| is now known as tsoome
[14:54:38] *** solarisjon has left #opensolaris
[14:54:58] <timeless> care to elaborate? :)
[14:55:37] <tsoome> it its mounted from nfs server, then default is to remap root to nobody in server, and therefore root can't access it
[14:55:50] <timeless> no nfs :)
[14:55:59] <tsoome> or in trusted solaris environment, the access may be limited as well...
[14:56:01] <timeless> just ufs, zfs, tcp/ip-v4
[14:56:06] <timeless> global zone
[14:56:09] *** sickness has quit IRC
[14:56:16] <tsoome> and root can't access it?
[14:56:22] <tsoome> what error?
[14:56:27] *** sickness has joined #opensolaris
[14:57:04] <tsoome> and what are perimssion bits for parent dir?
[14:57:09] <timeless> object: No such file or directory
[14:57:33] <tsoome> truss ls object ?
[14:59:36] <timeless> http://pastebin.mozilla.org/1131
[15:00:58] <timeless> doesn't seem like it was very useful
[15:02:46] <tsoome> what about the parent dir?
[15:02:55] <timeless> truss ls . #?
[15:03:18] <tsoome> and is it mount point?
[15:03:29] <timeless> it's /proc/6030
[15:04:07] *** axisys has quit IRC
[15:04:44] <coolvibe> sick filesystem?
[15:05:15] <timeless> paste will probably be (waiting for pastebin) http://pastebin.mozilla.org/1132
[15:05:21] <timeless> more like sick computer
[15:05:32] <timeless> /proc is the standard snv /proc
[15:05:51] <timeless> the process isn't quite dead but is using about 4gb of vm (3gb ram when i last checked)
[15:05:58] <timeless> kinda hurts my computer a bit
[15:06:23] <coolvibe> because it's trashing the swap I think :P
[15:06:52] <tsoome> well you cant open the file if dir does not have x bit set....
[15:07:11] <coolvibe> tsoome: you can open it, but you can't scan for it
[15:07:25] <tsoome> but I'm not exactly sure about procfs semantics
[15:08:54] <timeless> back in a bit
[15:09:13] <tsoome> coolvibe well, since there is no way to get inode number, the result is still the same;)
[15:09:40] <asyd> s7
[15:12:50] *** calumb is now known as calAFK
[15:17:04] *** axisys has joined #opensolaris
[15:18:20] <timeless> ok, paste did arrive
[15:18:30] <timeless> would iostat or something show me swapping?
[15:18:42] <jamesd_> vmstat
[15:19:09] <timeless>  0 0 2 6856824 1652128 2 124 11 15 17 316 691 17 -0 0 0 8509 17259 926 3  8 89
[15:19:29] <timeless> not that i can read that very well
[15:19:46] <jteo> vmstat -p 2
[15:19:48] <timeless> i mean, i know i'm *using* swap, but that's different than actively paging to it
[15:19:50] <jteo> look for scanrate.
[15:20:21] *** eboutilier has joined #opensolaris
[15:20:26] <jamesd_> timeless, http://astore.amazon.com/jamesdsworld-20  buy a copy of   performance  and tools from that link
[15:21:09] <tsoome> in current case, I would suggest to restart this webservd process;)
[15:21:19] <timeless> it took 3+ hrs
[15:21:21] <timeless> i'd rather not
[15:22:29] <timeless> so you mean sr? it was 691 for the first sample and dropped to 0 for the subsequent ones
[15:22:45] * timeless waits for google to explain vmstat output
[15:22:46] *** Slepek has quit IRC
[15:23:11] <elektronkind> the man page explains vmstat output
[15:23:20] <tsoome> sr is how active is system to seek for free pages....
[15:23:47] <hile_> jamesd: from a box @ my gig http://pastebin.ca/230897
[15:24:29] <jamesd_> hile_, damm you have enough swap on that box?
[15:25:30] *** elflord has joined #opensolaris
[15:26:00] <hile_> btw, those are 22GB LUNs on the power-pathed discs
[15:26:07] <hile_> i think.
[15:26:12] <hile_> and on that one, yes.
[15:26:16] <hile_> one some of the others, we'd need more.
[15:26:34] <hile_> and yes, we NEED that much swap
[15:27:50] <timeless> what are you doing?
[15:28:06] <timeless> i'm just trying to cross reference a half gigabyte source repository
[15:28:39] <jamesd_> hile_,  http://astore.amazon.com/jamesdsworld-20  feel free to buy some books from that link bill them to your new gig.... you must stay current  :-)
[15:32:17] <hile_> running a couple hundred instances of large binaries
[15:32:55] *** Symmetria has quit IRC
[15:33:52] <timeless> the same large binaries or lots of different ones?
[15:34:17] <hile_> a bunch of different ones, but some of said binaries have 40 or so instances running at any given time.
[15:34:34] *** r3boot has quit IRC
[15:35:27] *** calAFK is now known as calumb
[15:35:41] <timeless> is a large binary >1g? :)
[15:35:58] <hile_> one of them's about that (and there are about 40 instances of that running, i think.
[15:36:33] <hile_> suffice it to say that we have a specific fix from IBM for our AIX machines (which is about ahlf our infrastructure) to allow text segments > 256MB
[15:36:48] <timeless> heh
[15:37:03] *** r3boot has joined #opensolaris
[15:37:30] *** sparc-kly__ has joined #opensolaris
[15:39:10] *** slax[mtk] has joined #opensolaris
[15:42:38] <slax[mtk]> hi all! is there a way to install solaris-packages from dvd and dissolve the dependencies?
[15:53:11] *** r3boot has quit IRC
[15:53:33] *** r3boot has joined #opensolaris
[15:55:30] *** jwtodd has quit IRC
[15:59:06] *** r3boot has quit IRC
[15:59:17] *** r3boot has joined #opensolaris
[16:01:23] * coolvibe drools over thumper
[16:01:43] <coolvibe> sun has some cool hardware nowadays
[16:02:08] <jamesd_> coolvibe, drool over this http://milek.blogspot.com/2006/10/thumpers-first-impression.html
[16:02:17] <coolvibe> I wonder how much noise a thumper makes though
[16:02:25] <jamesd_> I recreated using the same disk for a one large stripe and got with single dd 1.35GB/s :)
[16:02:50] <AbeFroma1> i'm trying desperately to get mgmt to buy one for DR
[16:02:55] *** AbeFroma1 is now known as AbeFroman
[16:02:57] *** lacaAFK is now known as lacaMtg
[16:04:37] <wilbury> nick wilburyWC
[16:04:38] <wilbury> ops
[16:05:03] *** RaD|Tz has joined #opensolaris
[16:05:43] *** sparc-kly_ has quit IRC
[16:05:56] <mrdeviant> it amazes me the number of people who think dd is a benchmark
[16:06:15] <gdamore> hi *
[16:07:54] *** lacaMtg is now known as laca
[16:08:07] <slax[mtk]> and again! is there a way to install pkg-files from a mounted solaris10-dvd and dissolve the dependencies?
[16:08:45] <Auralis> yes and by hand
[16:09:56] *** Mazon is now known as mazon
[16:11:20] <slax[mtk]> Auralis: is there a better way. for example to run pkg-get against the dvd?
[16:11:45] <AbeFroman> when is the new x86 kernel patch coming out.  i gotta have the x4x00 long startup fix
[16:12:06] *** uebayasi has quit IRC
[16:12:07] *** OnkelSchorsch has joined #opensolaris
[16:12:19] <jteo> mrdeviant, it's a benchmark. for some people.
[16:12:20] <jteo> ;)
[16:12:23] <jteo> gdamore, hello! :)
[16:13:08] <hile_> g'morning mrdeviant
[16:13:12] <hile_> hihi Auralis
[16:13:30] <mrdeviant> hi hile_
[16:13:38] <mrdeviant> jteo, yea, for idiots
[16:13:41] <mrdeviant> :)
[16:13:52] <hile_> mrdeviant: did Keiser's package show up yesterday?
[16:14:06] <mrdeviant> the postoffice left us one of those "come fetch if yourself" slips
[16:14:17] <hile_> bastards.
[16:14:22] <hile_> it's ONLY 65 lbs :)
[16:16:58] <Auralis> slax[mtk]: there is no better way, pkg-get has ZERO do with the original sun pkgs
[16:17:52] *** glagasse has joined #opensolaris
[16:19:07] <mrdeviant> did you find a new apartment, hile_
[16:19:27] <slax[mtk]> Auralis: i know, but pkg-get can dissolve dependencies. Maybe you know a script (bigadmin,...) who solved this?
[16:19:31] <hile_> a new one? we haven't been looking.
[16:19:47] <mrdeviant> oh. i thought you were. i thought you had too much stuff in the place you're in now.
[16:19:55] <Symm> hrm
[16:20:03] *** MattMan has quit IRC
[16:20:13] <Symm> bash-3.00# wc -l mozilla_access_log
[16:20:14] <Symm>  1028639 mozilla_access_log
[16:20:14] <Symm> !
[16:20:17] <Symm> heh
[16:20:26] <Symm> my mozilla mirror crossed a million hits
[16:20:29] <Symm> in under 72 hours
[16:20:30] <Symm> damn
[16:20:55] <hile_> i'm running low on amperage
[16:21:06] <hile_> but i can't really fix that until we get a house
[16:21:36] <hile_> which we're not doing until we get sushma's immigration more permanent than her current visa
[16:21:51] <mrdeviant> ah i see
[16:27:02] <Stric> http://cvs.opensolaris.org/source/xref/on/usr/src/lib/libresolv/res_gethost.c#306 <- interesting.. linux and AIX doesn't do this at least.. and it kinda messes up stuff behind your back..
[16:28:32] <Stric> undocumented too. (except for that comment in the source ;)
[16:30:33] <Symm> argh
[16:30:42] <Symm> blastware's rsync is compiled v4 only
[16:33:46] <mrdeviant> file a bug
[16:33:53] <Symm> heh doing so :)
[16:38:10] *** RaD|Tz has quit IRC
[16:39:46] <Stric> http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6471607 seems to be about this problem..
[16:42:09] <dwc-> stric: glibc-2.5 seems to have this behavior
[16:42:22] <dwc-> compiled in if SUNSECURITY is defined
[16:42:49] <Stric> interesting..
[16:42:52] <gdamore> I think sendmail shouldn't have this problem, because it should be using the libresolv directly, rather than going thru nsswitch
[16:43:27] <dwc-> Stric: now, whether the linux distributors compile it with this in or not is another thing ....
[16:43:53] *** PosixCompliant has quit IRC
[16:43:58] <gdamore> the only bit that is weird to me is, that if the name-to-addr lookup fails, name is returned instead of NULL.  I don't understand that bit
[16:44:34] <quasi> gah, zfs-crypto-discuss seems to get flooded by all sorts of commit messages for stuff that has absolutely nothing to do with zfs-crypto
[16:44:51] <dwc-> I would suppose it's probably off by default
[16:44:53] <Stric> I still think it's a bit "overprotective behind my back".. although that's probably because the API doesn't allow for any other result..
[16:45:48] <gdamore> every application having to double check DNS results on its own is a PITA -- and a lot of said applications don't bother with it.  i'm not sure that i agree that the Solaris bheavior is a bug.
[16:46:52] <Stric> it's a co-sysadmin over here messing with antispam checks in sendmail.. and that check fails on solaris because the resolver already did that check and sent back NULL instead (indicating that the address is forged)
[16:47:09] <eboutilier> slax[mtk]: For tools that resolve dependencies, you might want to check out the following:
[16:47:16] <eboutilier> http://www.opensolaris.org/jive/thread.jspa?threadID=13015&tstart=60
[16:47:20] <eboutilier> and...
[16:47:30] <eboutilier> http://www.opensolaris.org/jive/thread.jspa?threadID=13197&tstart=60
[16:47:55] <slax[mtk]> eboutilier: let me check, thx a lot
[16:48:02] <eboutilier> np.
[16:50:24] <gdamore> i still don't understand though -- sendmail has historically been one of very applications which used DNS directly, rather than nsswitch routines.  (I.e. it uses libresolv.)  It has to do this to get at MX records.  Is there some reason that the spam milter or whatever isn't also using libresolv?
[16:50:32] <Symm> ok, stupid question, under solaris when trying to build source, what do I need to do to make it see libraries that were installed via blastwave
[16:51:35] *** nwf has joined #opensolaris
[16:52:34] <mrdeviant> LD_FLAGS="-R/opt/csw/lib -L/opt/csw/lib"
[16:52:50] <Symm> thanks
[16:55:13] <Symm> Undefined                       first referenced
[16:55:14] <Symm>  symbol                             in file
[16:55:14] <Symm> libiconv_open                       rsync.o
[16:55:14] <Symm> libiconv                            log.o
[16:55:14] <Symm> ld: fatal: Symbol referencing errors. No output written to rsync
[16:55:14] <Symm> collect2: ld returned 1 exit status
[16:55:17] <Symm> HRM
[16:55:47] <Symm> any ideas?
[16:56:36] <dwc-> LDFLAGS without the _
[16:56:43] <Symm> aahh
[16:56:44] <Symm> sorry :p
[16:57:09] <mrdeviant> yea it varies. some programs want the _ and some don't
[16:57:19] <Symm> same thing
[16:57:37] *** mathie has joined #opensolaris
[16:58:02] *** funky has joined #opensolaris
[16:58:04] <funky> hi
[16:58:18] *** tsoome has quit IRC
[16:59:49] *** ndroux has joined #opensolaris
[17:00:08] <Symm> libiconv_open                       rsync.o  (symbol belongs to implicit dependency /opt/csw/lib/libiconv.so.2)
[17:00:08] <Symm> libiconv                            log.o  (symbol belongs to implicit dependency /opt/csw/lib/libiconv.so.2)
[17:00:09] <Symm> heh
[17:00:13] <Symm> now I get a different error
[17:01:04] *** laca is now known as lacaMtg
[17:01:06] *** stevel has joined #opensolaris
[17:01:07] *** ChanServ sets mode: +o stevel
[17:01:33] <gdamore> nicolas!
[17:05:44] <coolvibe> jamesd: sweeeeet
[17:06:04] <gdamore> !seen ndroux
[17:06:05] <Drone> ndroux is currently online in #opensolaris, but never said anything that I saw.
[17:06:17] <gdamore>  /brick ndroux
[17:13:20] <Symm> hrm
[17:13:34] <gdamore> is there a list of ! commands that drone understands somewhere?
[17:15:01] <quasi> stevel: do you perhaps know why zfs-crypto-discuss@opensolaris is getting flooded by all commits going to the gate? as seen on http://www.opensolaris.org/jive/forum.jspa?forumID=105 there's quite a flood of them at the moment
[17:15:35] <stevel> quasi: no, someone else just notified me of it
[17:15:40] <stevel> i don't know what's going on, i'm looking into it
[17:16:12] <quasi> they all say Repository: /hg/zfs-crypto/gate
[17:17:58] <stevel> ugh i think i know what happened
[17:18:22] <stevel> someone created a Mercurial repository (zfs-crypto/gate), and set zfs-crypto-discuss as the notification list - and then imported ON into it
[17:18:41] <Symm> heh what does that "symbol belongs to implicit dependancy" mean when compiling something anyway
[17:19:34] <quasi> stevel: sounds plausible
[17:20:21] <gdamore> anyone know off hand how to submit an RFE for thunderbird?
[17:20:56] <movement> gdamore: bugzilla.mozilla.org ?
[17:21:03] <movement> gdamore: it's probably already there :)
[17:21:20] <gdamore> maybe.  my idea is to add an audio feature to the mozilla filter feature.
[17:22:10] *** sommerfeld has joined #opensolaris
[17:22:10] *** ChanServ sets mode: +o sommerfeld
[17:22:19] <eboutilier> gdamore: I'm pretty sure thunderbird officially belongs to the JDS consolidation.
[17:22:33] *** bernroth has quit IRC
[17:22:37] <gdamore> yeah, but i want to post the RFE upstream at mozilla.  just asking...
[17:22:54] <eboutilier> Ah.
[17:23:17] <Symm> heh
[17:23:26] <Symm> thunderbird/firefox downloads
[17:23:31] <Symm> eat a significant portion of my bandwidth
[17:23:44] <Symm> suprisingly
[17:24:16] *** coffman has joined #opensolaris
[17:25:21] <eboutilier> Symm: Not sure I can help, but I'm curious... what are you trying to build? I take it you're building from source because blastwave, sunfreeware, et al don't have it...
[17:26:24] <Symm> eboutilier attempting to rebuild rsync with v6 support
[17:26:41] <Symm> blastwave's packages, almost all of them that I've tried are compiled without v6
[17:27:18] <Symm> and hrm, when I try and compile it, it finds the library and then tells me its part of an implicit dependancy and compile dies
[17:27:30] <stevel> well - this is a good chance to test the scalability of our mailman server and see how well it handles having to send out 140,000 emails
[17:28:02] <Symm> heh solaris definately handles the laods Im throwing at it well so far
[17:28:15] *** alanc_away is now known as alanc
[17:28:28] <Symm> up to 15 hits a second on our mozilla mirror
[17:28:37] <quasi> stevel: testing my mailserver as well - it now has 800+ messages from that list that procmail can't keep up with ;)
[17:28:53] *** slax[mtk] has quit IRC
[17:29:08] <stevel> quasi: i think i'm going to forcibly unsubscribe everyone from zfs-crypto-discuss - let mailman catchup with all the processing, and then resubscribe after all the notifications are done
[17:29:27] <quasi> stevel: sounds like a good plan
[17:30:04] <Peanut> How many people are on zfs-crypto-discuss anyway? Darren gave a presentation on it at the Dutch OSUG meeting, which was very interesting but there didn't seem to be that many people involved in it.
[17:30:06] *** tmcmahon2 has joined #opensolaris
[17:30:39] <quasi> Peanut: yeah, that presentation was very interesting
[17:30:41] <tmcmahon2> Are any of the opensolaris.org web masters on by chance?
[17:30:58] <tmcmahon2> Or someone that can throttle an email list gone beserk?
[17:31:25] <eboutilier> Symm: Since rsync is in pmpkg (oxygene's nifty turnkey build-system), another approach would be to download pmpkg, tweak its default build parameters for rsync to add v6 support, and that should do it I would think.
[17:31:52] <quasi> Peanut: but it is probably not the easiest bits of the code to poke about in with it mixing crypto and zfs
[17:32:07] <eboutilier> Here's pmpkg's freshmeat entry: http://freshmeat.net/projects/pmpkg/
[17:33:18] <tmcmahon2> Anyone? Bueller? Bueller?
[17:34:00] <eboutilier> tmcmahon2: Gone berserk?
[17:34:10] <stevel> tmcmahon2: zfs-crypto-discuss, i know. i'm working on it
[17:34:37] <stevel> the problem is i can't reach darren - and we're both colliding in mailman
[17:34:39] <tmcmahon2> stevel: That's the one. Thanks. This is why <insert name of diety> created email sorting.
[17:35:20] <tmcmahon2> stevel: Want me to try his extentsion?
[17:35:28] <stevel> tmcmahon2: garypen and i have tried. no answer
[17:35:53] <tmcmahon2> I might be able to track his cell down
[17:36:06] <stevel> i couldn't find it in NF; if you have it - that would be great
[17:36:29] <tmcmahon2> I'll ping some of my UK contacts and see what I can find
[17:36:43] <stevel> thx
[17:36:53] <tmcmahon2> His admin might have it as well. (That number is in nf.)
[17:36:54] <jteo> what happens if a gmail account was subscribed to zfs-crypto-discuss? hmm...
[17:37:15] <stevel> jteo: subscribed? nothing; they'd just get it like normal
[17:37:34] <jteo> i was thinking of of the going berserk part vs the inexhaustible storage..
[17:37:50] <stevel> the problem is there are a lot of people in ON who haven't subscribed to zfs-crypto-discuss, and they're all getting bouncebacks saying they aren't subscribed
[17:38:11] <jteo> ugh.
[17:38:17] <jteo> the same issue as last time?
[17:38:18] <stevel> tmcmahon2: who's his admin?
[17:38:24] <stevel> jteo: slightly different.  this time i didn't do it ;-)
[17:38:37] <stevel> tmcmahon2: n/m.
[17:38:42] <tmcmahon2> ...k
[17:38:42] <jteo> stevel, that's irrelevant.
[17:38:43] <jteo> ;)
[17:39:08] <stevel> $10 says people will blame me anyway though
[17:39:22] <tmcmahon2> I'll take that $10 and raise you two shares of SUNW
[17:39:25] <tmcmahon2> :-P
[17:39:53] *** ndroux has quit IRC
[17:44:21] * Peanut starts looking for a plunger for stevel
[17:46:25] *** ndroux has joined #opensolaris
[17:47:55] <Peanut> stevel: http://www.dilbert.com/comics/dilbert/archive/dilbert-20061007.html
[17:49:31] <quasi> ;)
[17:50:09] <tmcmahon2> stevel: Looks like its fixed.
[17:50:27] *** lacaMtg is now known as laca
[17:50:38] <stevel> tmcmahon2: darren put it into emergency moderation mode - but i'm worried that that is then sending 'your message is in moderation' emails to everyone who putback now
[17:50:39] *** LordKing has joined #opensolaris
[17:51:10] <tmcmahon2> stevel: Probably. Or worse...its starting email exponential email loops.
[17:51:29] <tmcmahon2> stevel: I suggest - Nuke the site from orbit. Its the only way to be sure. ;)
[17:51:44] <stevel> i would love nothing more than to nuke our mailman server ;-)
[17:52:40] *** laca has quit IRC
[17:53:22] *** laca has joined #opensolaris
[17:54:30] *** mazon is now known as Mazon
[17:54:50] *** loke has quit IRC
[17:55:45] *** nwf has quit IRC
[17:56:00] *** sylvain has quit IRC
[17:56:49] *** Godsey has joined #OpenSolaris
[17:57:17] *** nwf has joined #opensolaris
[17:57:36] *** nwf has quit IRC
[17:57:53] <Godsey> can I get a discount on hardware if my home page has a Sun saved my sanity logo? :)
[17:58:08] *** nwf has joined #opensolaris
[17:58:53] <Symm> hrm how do I get make to build a static binary
[18:00:32] <Peanut> Symm: you mean how to change your makefile so that the compiler/linker produce a static bin?
[18:00:41] <Symm> yeah
[18:00:58] <Peanut> What compiler are you using?
[18:01:13] <Symm> the /usr/ccs/bin stuff
[18:01:28] <Peanut> That's not a compiler.
[18:01:42] *** tmcmahon2 has left #opensolaris
[18:01:47] <Symm> bash-3.00# gcc --version
[18:01:48] <Symm> gcc (GCC) 3.4.3 (csl-sol210-3_4-branch+sol_rpath)
[18:01:48] <Symm> Copyright (C) 2004 Free Software Foundation, Inc.
[18:01:48] <Symm> This is free software; see the source for copying conditions.  There is NO
[18:01:48] <Symm> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
[18:01:56] *** slowhog has joined #opensolaris
[18:02:09] <Peanut> Ah ok, it's gcc.
[18:02:54] <Peanut> gcc when used as a linker takes a '-static' flag.
[18:03:06] <Peanut> So you could try setting LD_FLAGS='-static'.
[18:03:24] *** vmhobbes has quit IRC
[18:04:00] * Symm tries that
[18:04:26] *** Fish- has joined #opensolaris
[18:04:29] <Symm> ld: fatal: library -lpopt: not found
[18:04:30] <Symm> ld: fatal: library -liconv: not found
[18:04:30] <Symm> ld: fatal: library -lsocket: not found
[18:04:30] <Symm> ld: fatal: library -lnsl: not found
[18:04:30] <Symm> ld: fatal: library -lc: not found
[18:04:31] <Symm> heh
[18:04:33] <eboutilier> Symm: With librsync and a lot of sources, you can add that to the flags you use when  you run configure, like this: ./configure --enable-static
[18:04:35] <Symm> whats what I get when I add that
[18:04:35] *** solarisjon has joined #opensolaris
[18:04:50] <jengelh> gcc also takes -static during -c (except that it's ignored)
[18:05:09] <Peanut> Ah wait - you're building on Solaris 10? There's no such thing as static anymore on Solaris 10 and newer.
[18:05:18] *** |tsoome| has joined #opensolaris
[18:05:19] <Symm> aahh
[18:05:26] <Symm> hrm
[18:05:28] <jengelh> huh
[18:05:37] <Symm> bash-3.00# ./rsync
[18:05:38] <Symm> ld.so.1: rsync: fatal: libiconv.so.2: open failed: No such file or directory
[18:05:50] <Symm> ^^^ thats why I wanna build it static
[18:05:54] <Godsey> does the X2200 M2 offer raid1 for the sata drives?
[18:05:55] <Symm> :p it builds fine, but it doesnt run
[18:06:06] <Peanut> Which libiconv did you build against, where does it live?
[18:06:09] <jteo> ldd rsync
[18:06:16] <Symm> its in /usr/local/lib
[18:06:32] <Peanut> export LD_LIBRARY_PATH=/usr/local/lib rsync
[18:06:37] <Peanut> Try if that helps?
[18:06:40] <hile_> oh jesus..
[18:06:48] <Symm> aaahh
[18:06:51] <Symm> no
[18:06:54] <Symm> heh I fixed it :p
[18:06:58] <Symm> bash-3.00# ln -s /usr/local/lib/libiconv.so.2 /lib/libiconv.so.2
[18:07:03] <Symm> :p now it works
[18:07:10] <hile_> That's not a fix in the least
[18:07:11] <Peanut> That's baaad though.
[18:07:18] <hile_> that's a worse hack than LD_LIBRARY_PATH
[18:07:21] <hile_> build your binary correctly
[18:07:25] <|tsoome|> for s10 there is no need for gnu iconv;)
[18:07:34] <Symm> hile, all I did was type ./configure and then run make
[18:07:37] *** mathie has quit IRC
[18:07:43] <hile_> like i said, build it correctly :)
[18:07:49] <hile_> pass the appropriate linker args
[18:07:52] <Symm> tsoome rsync refused to build without it
[18:07:58] <Symm> hile, and what ARE those linker args
[18:08:04] <jteo> LDFLAGs
[18:08:05] <Symm> I'll do that if you can tell me what linker args to pass :)
[18:08:11] <|tsoome|> then fix configure
[18:08:12] <hile_> the appropriate -L  and -R paths in LD_OPTIONS
[18:08:22] <hile_> or LDFLAGS
[18:08:25] <|tsoome|> iconv is built into libc in solaris
[18:09:02] *** vmhobbes has joined #OpenSolaris
[18:09:05] <eboutilier> Symm: I'd nuke your /usr/local/lib stuff and build rsync (and dependencies) using pmpkg, if I were you.
[18:09:11] <Symm>         libiconv.so.2 =>         /usr/local/lib/libiconv.so.2
[18:09:12] <Symm> there
[18:09:13] <Symm> :)
[18:09:32] <Symm> heh built it properly
[18:09:57] *** musasabi has left #opensolaris
[18:09:59] <hile_> see the documentation for -L and -R in ld(1)
[18:10:17] <Symm> thanks hile
[18:10:38] <eboutilier> pmpkg makes things transparent if you want, or teaches you what it does if you want. Either way, it's a great way to build rsync and lots of other stuff on Solaris 10 or Nevada.
[18:11:04] <hile_> pmpkg?
[18:11:32] <eboutilier> All the same is true of the JDS pkgbuild spec-files-extra system too. But  it's probably overkill for what you need.
[18:11:55] <eboutilier> hile_: http://freshmeat.net/projects/pmpkg/
[18:13:25] *** Fish has quit IRC
[18:14:20] *** solarisjon has left #opensolaris
[18:15:11] <eboutilier> In case anyone wants to know more about pkgbuild/spec-files-extra (SFE):
[18:15:14] <eboutilier> http://pkgbuild.sourceforge.net/spec-files-extra/
[18:15:35] *** sparc-kly_ has joined #opensolaris
[18:15:37] <Symm> hrm, heh, didnt help when I rebuilt rsync anyway :( its still claiming no v6
[18:16:10] *** opSuse has joined #opensolaris
[18:16:29] *** eboutilier has left #opensolaris
[18:16:56] *** eboutilier has joined #opensolaris
[18:17:07] <dwc-> check your config.log
[18:17:36] <dwc-> see why it's not finding ipv6 support
[18:18:13] <Symm> what the hell
[18:18:14] *** calumb has quit IRC
[18:18:17] <Symm> got a bigger problem
[18:18:21] <Symm> virtual memory exhausted: Resource temporarily unavailable
[18:18:25] <Symm> Memory: 4095M real, 1202M free, 1670M swap in use, 14M swap free
[18:18:30] <Symm> why the hell is it using all my swap
[18:18:32] <sickness> evening all
[18:18:35] <Symm> and causing things to die
[18:18:43] <Symm> while I still have over a gig of free ram
[18:18:57] <hile_> what the hell are you running?
[18:19:10] *** calumb has joined #opensolaris
[18:19:27] <Symm> hile, just apache and an ftp daemon on there
[18:19:38] <Symm> an d that was a compile message, but rsync, compiling, various things are dying
[18:19:40] <Symm> with that
[18:19:58] <dwc-> prstat -s size ?
[18:20:38] <Symm> hrm that produces an entire screen of http processes
[18:20:38] <Symm> Total: 503 processes, 602 lwps, load averages: 0.18, 0.17, 0.18
[18:21:08] <Symm> thing at the top is java using 117meg
[18:21:23] <dwc-> so probably httpd and java
[18:21:34] <dwc-> the ajva's probably the smc thing anyways .... and can be killed
[18:21:50] <Symm> yeah I killed the java but its still not helping
[18:21:57] <Symm> I dont understand why my swap is maxed out
[18:22:00] <jengelh> I knew Java was evil!
[18:22:01] <Symm> while I still have tons of free ram
[18:22:12] <jengelh> poold also pulls ram
[18:22:14] <dwc-> Symm: maybe something in active memory used up the ram
[18:22:19] <dwc-> and is now gone
[18:22:33] <Symm> dwc, well, errr, why is it showing as free?
[18:22:43] <dwc-> it died?
[18:22:50] <Symm> Memory: 4095M real, 1239M free, 1681M swap in use, 1020K swap free
[18:22:55] <Symm> and changing constantly
[18:22:57] <Symm> on top
[18:23:06] <Symm> but never dropping below 1200meg of free ram
[18:23:11] <Symm> but things are still dying
[18:23:21] <Symm> dwc, its still causing things to die
[18:23:23] <hile_> top is notoriously inaccurate
[18:23:44] <dwc-> top's memory output is fine
[18:23:45] <Symm> hrm ok, is there a way I can drastically increase my swap using another disk?
[18:24:04] <OnkelSchorsch> hm. what does /tmp look like? did you put some big files there?
[18:24:07] <dwc-> it's cpu usage is somewhat less than reliable
[18:24:07] <rydis> If your swap is full, /tmp will be full, which will mess stuff up.
[18:24:32] <Symm> hrm
[18:24:35] <Symm> how do I increase swap
[18:24:39] <dwc-> swap -a
[18:24:51] <Symm> bash-3.00# df -h
[18:24:52] <Symm> bash: fork: Not enough space
[18:24:52] <timeless> ouch
[18:24:53] <Symm> oh gawd
[18:24:54] <Symm> :p
[18:24:58] <dwc-> pkill httpd
[18:25:18] <Symm> dwc, so if I just swap -a with a file name
[18:25:19] <dwc-> mkfile to make a file, swap -a to add it
[18:25:22] *** eboutilier has quit IRC
[18:25:23] <Symm> ok
[18:25:26] <jengelh> heh
[18:25:30] <jengelh> dd is for files :)
[18:25:51] <timeless> fsck
[18:26:00] <opSuse> qUESTIOB: OpenSolaris source browser: http://cvs.opensolaris.org/source :
[18:26:07] <rydis> The dd.sh site disappeared, didn't it? Kind of cute, with an editor implemented using just dd and sh. :)
[18:26:22] <opSuse> To which version does it correspond??
[18:26:28] * timeless cries
[18:26:31] <Symm> hrm, though does anyone know why swap is running out when there is so much free ram?
[18:26:39] <richlowe> opSuse: Ideally, it's entirely up-to-date.-
[18:26:45] <richlowe> I'm not sure if that's currently true, however.
[18:27:00] <dwc-> df | grep swap ?
[18:27:01] <timeless> opsuse: in  my experience it's more current than the snv i was using
[18:27:06] <dwc-> that tells you what filesystems are tmpfs
[18:27:19] *** vmhobbes- has joined #OpenSolaris
[18:27:21] <Symm> bash-3.00# df |grep swap
[18:27:22] <Symm> /etc/svc/volatile  (swap              ):   67416 blocks   590975 files
[18:27:22] <Symm> /tmp               (swap              ):   67400 blocks   590975 files
[18:27:22] <Symm> /var/run           (swap              ):   67400 blocks   590975 files
[18:27:22] <timeless> i had snv 49 and the code had a fix that hadn't made it to snv :(
[18:27:28] <Symm> Im busy building another 8gig of swap
[18:27:29] <richlowe> it should at worst correspond to the most recent weekly source drop.
[18:27:34] <dwc-> so look for big files in those directores
[18:27:35] <Symm> with that mkfile command
[18:27:36] <richlowe> ideally, it'd be perfectly in sync with the gate.
[18:27:38] *** vmhobbes has quit IRC
[18:27:50] <opSuse> timeless, are you usign snv from hg ? subverison ? tar.gz ?
[18:27:53] *** sparc-kly__ has quit IRC
[18:27:59] <opSuse> because sometimes they are a bit different
[18:28:00] <timeless> dvd images
[18:28:15] <opSuse> timeless, ok , thnks
[18:28:16] <Symm> dwc none of those have any big files in them
[18:28:18] <richlowe> it will *always* be more recent than the DVD images.
[18:28:27] <richlowe> (if it's not, someone screwed up, rather)
[18:28:41] *** vmhobbes- is now known as vmhobbes
[18:28:56] <Symm> bash-3.00# for i in `df |grep swap |awk '{print $1}'`; do du -hs $i; done
[18:28:57] <Symm>  612K   /etc/svc/volatile
[18:28:57] <Symm>   28K   /tmp
[18:28:57] <Symm>   88K   /var/run
[18:29:38] <richlowe> opSuse: it's currently up to date enough to be showing St. Paul support.
[18:29:41] <richlowe> opSuse: (which went back yesterday)
[18:29:43] <richlowe> so.
[18:29:56] <opSuse> richlowe, :)
[18:29:56] <Symm> Memory: 4095M real, 1202M free, 1548M swap in use, 8300M swap free
[18:29:57] <Symm> heh
[18:29:59] <Symm> there
[18:31:04] <timeless> richlowe: does mdb have any ::analyze_crash magical things?
[18:31:13] <timeless> (windbg has !analyze)
[18:31:27] <richlowe> Not really.
[18:31:35] <richlowe> SolarisCAT has some stuff like that, though.
[18:31:56] <timeless> http://pastebin.mozilla.org/1137
[18:32:01] <timeless> not a very useful stack :(
[18:32:19] <rydis> People seem to differ very much on the issue on how much swap is adequate. I prefer having 4+ times memory, but then I use Linux a lot, and /hate/ that OOM killer. (I also tend to have a fair amount of biggish processes mostly idling, like lisp worlds. :)
[18:32:53] <Symm> that swap -a to add that swapfile
[18:32:54] <timeless> rydis: well, i have a 3.9G core file and would have had 2 , but i ran out of disk space for the second copy
[18:32:57] <richlowe> timeless: oh, then SolCAT won't help, it's kernel-fu
[18:33:02] <Symm> where do I configure it to start that automatically on boot up?
[18:33:10] <Symm> or do I just add another rc3 script
[18:33:22] <richlowe> timeless: best you could do would probably be to try and reproduce it on a build dbx still works on.
[18:33:25] <|tsoome|> man smf
[18:33:32] <timeless> richlowe: funny
[18:33:33] <|tsoome|> for symm
[18:33:45] *** bondolo has quit IRC
[18:33:48] *** laca is now known as lacaAFK
[18:34:09] <timeless> Tue Oct 31 10:25:24 2006
[18:34:10] <richlowe> timeless: well, or enjoy digging at it with mdb. :)
[18:34:14] <timeless> Tue Oct 31 18:16:56 2006
[18:34:15] <jengelh> Tue Oct 31 18:34:18 MET 2006
[18:34:22] <timeless> it took 8 hours to die
[18:34:26] <jengelh> i'd say your clock is screwed
[18:34:29] <timeless> i think that perhaps, oh
[18:34:35] <timeless> malloc failed after memory_usage = e487f56c Bytes
[18:34:47] <richlowe> timeless: "quit recursing like that"? :)
[18:34:56] *** jteo has quit IRC
[18:35:05] <timeless> richlowe: yeah, it's this stupdi symlink in my file system
[18:35:10] <timeless> i need to find a way to avoid it
[18:35:31] <timeless> richlowe: but seriously, shouldn't something like recognize stupidity and complain instead of killing me? :)
[18:35:49] <timeless> jengelh: why? i'm in finland
[18:36:38] *** sparc-kly__ has joined #opensolaris
[18:36:50] <timeless> richlowe: wait
[18:36:57] <timeless> does this mean that the code did something stupid like:
[18:37:05] *** solarisjon has joined #opensolaris
[18:37:12] <timeless> char *a = (char*)malloc(FOUR_GIGS);
[18:37:20] <timeless> strcpy(a, "hi");
[18:37:22] <Godsey> the X2100 M2 is much cheaper than the X2200 M2, will I even notice a difference between them?
[18:37:23] <timeless> without null checking a :)
[18:37:26] *** nachox has joined #opensolaris
[18:37:40] <Godsey> I want 9 zones
[18:39:13] <timeless> yeah, ok, that's exactly what happened :)
[18:39:35] <timeless> ok, so, i think what i should do is consider a ulimit on glimpse and maybe fix glimpse not to be stupid :)
[18:39:46] <timeless> (or i could waste a year trying to replace glimpse)
[18:41:30] *** alanc is now known as alanc_away
[18:41:30] <jengelh> timeless : remove the cast, for sake.
[18:41:48] <jengelh> for the sake of good programming
[18:41:52] <timeless> jengelh: malloc returns void*, in c++ the cast is required
[18:42:16] *** calumb has quit IRC
[18:42:25] <jengelh> oh, c++
[18:42:40] <jengelh> well in that case, you should be using static_cast<char *> or new char[] anyhow ;)
[18:42:57] <timeless> new throws exceptions, the code doesn't need those
[18:43:00] <timeless> it already crashes :)
[18:43:09] <richlowe> movement: the simplest way to do the xen stuff would probably be to do what onnv did.
[18:43:10] <timeless> besides this was a sample
[18:43:26] <jengelh> C++, just for a sample? Sounds like megalomania.
[18:43:35] *** yippi has joined #opensolaris
[18:43:40] <richlowe> movement: (get a project called xen, or xennv, or matrix, or whatever for the repo)
[18:43:40] <timeless> i try to write moderately correct samples
[18:44:03] <axisys> i got the solaris express dvd downloaded and burned using cdrw -i image on a dvd-R .. but ultra 20 does not seem to read it.. i used the same ultra 20 to burn the image.. so i know ultra 20 recognize the dvd
[18:44:24] <axisys> ultra 20 cannt be booted from the dvd
[18:44:43] *** RaD|Tz has joined #opensolaris
[18:44:51] <axisys> i did a lofiadm and mount it and that went fine.. so image is not corrupted
[18:45:29] <axisys> i tried to boot using a sol 10 u2 and point to network filesystem.. but it says the image is not sol 10.. so network install wont work either
[18:45:33] <axisys> what a dilemma
[18:45:40] <Symm> hrm
[18:46:01] <Symm> vfstab entry for swap, is just swap - /swapfilename - - - right?
[18:46:04] *** sparc-kly has joined #opensolaris
[18:46:05] *** ChanServ sets mode: +o sparc-kly
[18:46:56] <timeless>                 t1 = (char *) my_malloc(strlen(outname) + 2);
[18:46:57] <timeless>                 strcpy(t1, outname);
[18:47:07] <timeless> is the real glimpse code that crashed, complete w/ the cast :)
[18:48:13] <dunc> folks, if i use zfs to mirror a pair of disks, and then once i've filled em, can i just add another pair to my pool and double the size
[18:50:53] *** axxl has quit IRC
[18:51:29] <sommerfeld> dunc: yup.
[18:51:30] * timeless wonders how to build this thing
[18:52:23] *** elflord has quit IRC
[18:52:42] <dunc> thanks sommerfeld
[18:52:50] *** calumb has joined #opensolaris
[18:53:23] *** sparc-kly__ has quit IRC
[18:53:40] *** elflord has joined #opensolaris
[18:53:50] <axisys> i not sure why but after rebootign the ultra 20 couple times now I see the solaris 11.. eehaa
[18:53:55] <axisys> i am not..
[18:53:59] <axisys> cool
[18:54:19] <axisys> installing sol 11 now
[18:55:01] *** dcFood has quit IRC
[18:56:22] <trygvis> root@www4 # zoneadm -z mail boot
[18:56:22] <trygvis> zoneadm: zone 'mail': /usr/sbin/devfsadm unexpectedly terminated due to signal 10
[18:56:25] <trygvis> zoneadm: zone 'mail': call to zoneadmd failed
[18:56:28] <trygvis> wtf?
[18:57:42] <axisys> trygvis: zlogin -C mail .. see what error it gives in console
[18:57:49] <timeless>      SIGBUS           10      Core       Bus Error
[18:58:30] <trygvis> no outputt
[18:58:41] <trygvis> aha
[18:58:45] <trygvis> /zones is full
[18:59:15] <timeless> man, gimp's start screen is animated
[18:59:50] *** coffman has quit IRC
[19:00:29] *** deather__ has quit IRC
[19:00:31] *** deather__ has joined #opensolaris
[19:03:49] *** dunc has quit IRC
[19:04:03] *** sparc-kly_ has quit IRC
[19:06:13] *** LordKing has quit IRC
[19:07:53] *** alanc-away is now known as alanc
[19:08:24] *** rachel has quit IRC
[19:09:07] *** rachel has joined #opensolaris
[19:09:13] *** LordKing has joined #opensolaris
[19:11:41] <richlowe> alanc_away: way to reply to that, and get it past my filter :(
[19:12:04] <alanc> the ksh93 flame from Felix?
[19:12:25] <alanc> I know, we should ignore him until he goes away or nrubsig beats him down
[19:13:33] <alanc> just keep reading all the blogs of developers dropping out of projects like Debian because they're nothing but flamefests and hoping we can stop that from starting here
[19:14:38] *** logic__ has joined #opensolaris
[19:16:57] *** deather__ is now known as deather
[19:16:58] <trygvis> hmm .. what's the correct way to lofs /pool0/zones/myzone/local to /zones/myzone/root/local
[19:17:19] *** packetjunkie has joined #opensolaris
[19:19:10] <packetjunkie> I have a Sun Blade 100. The previous owner had installed BSD on it. Searching Google showed that this caused issues the disklabel which is why the Solaris installer cannot detect the disks. Problem is that the format utility won't handle the IDE drive. I tried using a Gentoo CD to create a Sun disklabel, but that didn't work. Any ideas what I can do?
[19:19:41] <|tsoome|> google solaris EFI SMI label
[19:20:00] <|tsoome|> and short answer is format -e and then label command
[19:20:06] *** logic_ has quit IRC
[19:21:37] <packetjunkie> I had tried that but I got an error writing VTOC.
[19:22:00] <zarathustra> hello
[19:22:01] <packetjunkie> I'll check that web page; maybe after reading it I'll figure out how to get around that. Thanks.
[19:22:06] <jamesd_> hi
[19:22:31] <|tsoome|> that -e is important one;)
[19:22:43] <zarathustra> has anybody tried to install munin ?
[19:22:58] <packetjunkie> |tsoome| I used -e ;(
[19:23:53] *** |tsoome| is now known as tsoome
[19:24:20] <dwc-> isn't EFI an x86 thing?
[19:24:50] <sickness> zarathustra: boot a bsd installcd, and issue a dd if=/dev/zero of=/dev/yourdisk
[19:24:59] <sickness> should set the disk like new...
[19:25:07] *** Godsey has left #OpenSolaris
[19:25:22] <richlowe> dwc-: EFI-the-label-format, not EFI-the-firmware.
[19:25:45] <zarathustra> sickness: :P
[19:27:54] <axisys> zarathustra: i like hobbit better.. here is ademo page maintained by the author http://hswn.dk/hobbit/
[19:28:13] <axisys> zarathustra: it is super fast..
[19:28:47] <packetjunkie> tsoome: Any idea about the error? "error writing VTOC." "no backup labels"
[19:29:01] <packetjunkie> Google shows others with the problem, but no solution posted.
[19:29:20] <tsoome> what is your drive geometry?
[19:29:52] <packetjunkie> cyl 3880 hd 16 sec 135
[19:30:25] <tsoome> I think there is/was also some bug related how solaris interprets efi label....
[19:30:26] <axisys> zarathustra: i do 3425 test in 2.82 secs in avg on a Netra T1 200, 1 500Mhz cpu
[19:30:58] <richlowe> http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6358153
[19:31:08] <richlowe> alanc: could you tell me what the 8th word there should be? :)
[19:31:22] * richlowe can guess, but had thought that was fixed
[19:31:56] <alanc> you mean it's not supposed to be "xxxxx" ?
[19:32:31] <richlowe> b.o.o used to mask some variations of "Sun", "SUN", "SUNW", etc.
[19:32:34] <alanc> SMI
[19:32:35] <richlowe> as I said, I thought it had ceased to do that.
[19:32:40] <richlowe> ... I guess not.
[19:32:41] <alanc> why the hell is that masked?
[19:32:44] <richlowe> other CRs it works just fine.
[19:33:02] <alanc> (SMI == "Sun Microsystems, Inc.")
[19:33:06] <richlowe> http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6337932
[19:33:09] <richlowe> but there, it works just greate.
[19:33:12] <richlowe> I know how SMI expands :)
[19:33:22] <alanc> SMI isn't masked later in the same bug!
[19:33:40] <alanc> only in the first note, not the later ones
[19:34:07] *** Symmetria has joined #opensolaris
[19:34:41] <packetjunkie> tsoome/richlowe: Thanks for the help/info. :)
[19:34:41] <alanc> the masking of the device path strings is bogus too, but that's expected b.o.o brain damage
[19:35:20] *** mage2 has joined #opensolaris
[19:35:29] <mage2> can you still download 01/06?
[19:35:49] *** LordKing has quit IRC
[19:35:57] <richlowe> alanc: I (had) my suspicions regarding the SMI/Sun-ish masking, but I can't confirm though.
[19:36:07] <richlowe> it only doing one of them, however, suggests I'm wrong.
[19:36:38] *** packetjunkie has left #opensolaris
[19:39:07] *** OnkelSchorsch has quit IRC
[19:39:19] <mage2> ?
[19:39:38] <mage2> does anyone know if Solaris 01/06 is still downloadable?
[19:40:00] *** ericr has joined #opensolaris
[19:41:00] *** Symm has quit IRC
[19:41:21] <richlowe> sigh.
[19:41:27] <richlowe> alanc: bug report, and mail to website-discuss will be forthcoming.
[19:41:32] <richlowe> though I totally don't know why I even bother.
[19:44:03] *** nachox has quit IRC
[19:46:21] <gdamore> random thought:  core group is < 500MB.  I wonder if it would be worthwhile to consider trying to run Solaris from CF for a NAS device.
[19:46:30] <gdamore> (or other embedded application?)
[19:47:01] <gdamore> the biggest problem being, I think, whether /var/ needs to be writeable or not.
[19:47:06] <richlowe> there was someone, I forget who, who put together a cut-down thumb drive with a basic Solaris on it.
[19:47:19] <richlowe> I think eboutillier has done it, and someone else who I think was sending it to zfs-discuss?
[19:47:21] <mrdeviant> how do the live cds handle the issue? tmpfs for /var ?
[19:49:59] <gdamore> one the reasons i'm keenly interested in this is that CF takes ~0 watts compared to a real disk. :-)
[19:50:42] <jengelh> cf degrades faster and is not as cheap :(
[19:51:24] <andersmo> doesn't read-mostly cf last pretty long?
[19:51:54] *** koolniczka has joined #opensolaris
[19:52:40] <andersmo> (assuming you could run from it in a livecd-like fashion without any swap or transient writes - only permenent configuration changes are written to flash?)
[19:53:00] <richlowe> of course, just to finish of this barrel of fun.
[19:53:13] <richlowe> alanc: could you find the CR ID of the CR I just filed regarding this for me?
[19:53:44] <richlowe> synopsis should be "bugs.opensolaris.org information masking is far too overzelous", or words to that effect.
[19:53:48] <richlowe> filed less than a minute ago.
[19:54:08] <gdamore> read-only cf lasts pretty well.
[19:54:09] *** lacaAFK is now known as laca
[19:54:17] <alanc> don't see it in the triage queue yet - usually takes at least a few minutes
[19:54:57] <gdamore> and as far as "cheap", I can buy a 4GB CF pretty darn cheap right now.
[19:55:14] <gdamore> and 1GB even cheaper.  if Solaris core will work in 500MB. :-)
[19:55:43] <gdamore> of course, putting swap on flash would be an incredibly bad idea... but I'd run without swap.
[19:56:30] <timeless> swap in flash is bad
[19:56:36] <timeless> but swap on an mmc is better
[19:56:43] * timeless grumbles about small devices
[19:57:03] <quasi> writing much on flash is bad
[19:57:59] <andersmo> Heh. Swapping to flash is a bad idea - isn't plain old RAM almost as cheap as flash anyway? =)
[19:58:30] <andersmo> OK, maybe the flash price has dropped faster.
[19:58:51] <andersmo> Still. Not expensive in the quantities you can buy flash drives. =)
[19:58:59] <timeless> andersmo: all i know is that there's a limit to how much ram vendors will put in
[19:59:05] <richlowe> I'm not sure I'd cut the system down brutally *and* have no swap.
[19:59:08] <timeless> although they don't really like growing flash either
[19:59:13] <richlowe> I think that's a situation I'd be very careful to make sure it could dump...
[19:59:15] *** Tpenta has quit IRC
[19:59:27] <timeless> richlowe: we make sure you can dump to your mmc
[19:59:31] <timeless> it works well enough
[19:59:45] <timeless> we finally even have a tool to automatically analyze the cores, at least, i think
[19:59:54] <andersmo> I haven't followed the discussion until I jumped in, I'm probably missing some context. =)
[19:59:57] * timeless causes lots of cores
[20:04:23] *** svoboda has quit IRC
[20:07:44] <richlowe> stevel: Thanks, man.
[20:07:55] <stevel> richlowe: np
[20:07:58] <richlowe> stevel: thanks for not pasting the priority justification, too ;)
[20:08:04] * richlowe is not in a great mood
[20:08:06] <stevel> :-D
[20:08:24] <stevel> i got a good laugh out of it
[20:08:35] <stevel> nit: "maybe" should have been split up into two words
[20:08:44] <stevel> "may be" as opposed to "maybe"
[20:08:49] *** calumb has quit IRC
[20:09:19] <alanc> here it is: CR 6488418 Created P3 opensolaris/triage-queue bugs.opensolaris.org should never mask instances of Sun Microsystem's name.
[20:10:18] <alanc> I noticed your bug fails to mention the current masking implementation is broken, and misses masking some 8-)
[20:10:39] <richlowe> alanc: Well yes.
[20:10:44] <richlowe> I don't mention bugs that are beneficial to us.
[20:10:49] <richlowe> because I *KNOW* some fool would give them priority and 'fix them'
[20:10:51] *** bunker has joined #opensolaris
[20:11:22] <richlowe> I value my temper as a precious resource, I'll go to great lengths not to lose it...
[20:11:49] * stevel ponders going to great lengths to provoke it
[20:12:14] <richlowe> stevel: no need, when I find appropriate words, I'm filing the interest list CR too.
[20:12:16] *** svoboda has joined #opensolaris
[20:12:21] <richlowe> that may well require a thesaurus.
[20:12:39] * stevel eagerly awaits the bugs-triage-queue mail
[20:13:03] *** alobbs has quit IRC
[20:13:04] *** Doc has quit IRC
[20:13:20] *** Doc has joined #opensolaris
[20:15:15] *** sparc-kly_ has joined #opensolaris
[20:19:27] *** yippi has quit IRC
[20:19:39] <axisys> my second disk looks f***d
[20:19:41] <axisys> http://www.rafb.net/paste/results/iVIPlS93.html
[20:20:02] <axisys> they are exact same size disk
[20:20:11] *** sparc-kly has quit IRC
[20:20:40] <axisys> any way to fix that?
[20:20:47] <quasi> axisys: anything interesting from iostat -E
[20:20:54] <quasi> axisys: get a new disk ;)
[20:21:18] <axisys> quasi: i guess.. the guy who shipped the disk prob'ly f**d it up
[20:21:22] <axisys> that bastard
[20:21:57] *** opSuse has quit IRC
[20:22:23] <axisys> iostat -E gives this http://www.rafb.net/paste/results/U4ptjJ69.html
[20:23:23] *** laca has quit IRC
[20:23:34] <esproul> having trouble with ucd-snmp on a 6/06 machine.  seems a value (ssRawContexts) that is supposed to be a counter, is actually moving backward
[20:23:45] <esproul> anyone seen anything like this?
[20:24:38] <quasi> esproul: I've seen several odd things out of snmp on 6/06 with sun x86 machines
[20:24:59] <esproul> quasi: any workaround?
[20:25:48] <quasi> esproul: I had to do something else and never got back to the problem
[20:25:50] *** kupfer has joined #opensolaris
[20:25:51] *** ChanServ sets mode: +o kupfer
[20:26:22] <esproul> quasi: thanks, maybe i'll post to a mailing list
[20:26:25] *** sahafeez has joined #opensolaris
[20:27:26] <quasi> esproul: sounds like a plan
[20:28:11] <trygvis> do I have to do something special to get console access when giving "platform console" on the SP to a v40z?
[20:28:50] <quasi> trygvis: yes
[20:29:03] <quasi> trygvis: at least if it hasn't been set up before
[20:29:18] <trygvis> it haven't, I've set it up :)
[20:29:47] <quasi> trygvis: first you have to point the serial at the sp - that's described in the sp doc
[20:30:07] <trygvis> I've done that, I can the bios output
[20:30:15] <quasi> trygvis: but you also need to change speed to 9600 (even if they do suggest 19200)
[20:30:30] <quasi> on an install?
[20:30:43] <trygvis> when rebooting
[20:30:52] <quasi> then you see the bios?
[20:30:59] <trygvis> yep
[20:31:02] <zarathustra> axisys: thanks, Ill take a look to hobbit
[20:31:03] <quasi> that's really all there should be to it
[20:31:05] *** LordKing has joined #opensolaris
[20:31:29] <trygvis> weir
[20:31:30] <trygvis> d
[20:32:04] <quasi> trygvis: but I wonder if you set the full command that they recommend in the doc - pointing console to the sp
[20:32:25] <axisys> zarathustra: u can also subscribe to the hobbit mailing list hobbit-subscribe at hswn dot dk
[20:32:34] <trygvis> this is the line I used: platform set console -s sp -e -S 9600
[20:32:35] <zarathustra> axisys: it looks like nagios
[20:32:43] <axisys> axisys: it does..
[20:32:48] <quasi> trygvis: that looks about right
[20:32:49] <axisys> zarathustra: it does
[20:32:59] <axisys> IMing to myself ;-)
[20:33:10] *** dduvall__ has joined #opensolaris
[20:33:23] *** dduvall__ is now known as dduvall
[20:33:29] <axisys> zarathustra: i used nagios and bigbrother before hobbit came about..
[20:33:52] <zarathustra> axisys: :O
[20:33:59] <axisys> zarathustra: it is all written w/ C and use mostly memory to do the work
[20:34:08] <axisys> zarathustra: hence super fast
[20:34:12] <zarathustra> axisys: have you tried munin ?
[20:34:19] <axisys> zarathustra: nawp
[20:35:08] <trygvis> quasi: should I give it any special commands to get the login prompt?
[20:35:49] <quasi> trygvis: nope, pressing enter works for me
[20:36:01] <trygvis> grr
[20:36:18] <zarathustra> axisys: Ill try hobbit
[20:36:32] <axisys> zarathustra: all graphs just work
[20:36:35] <quasi> trygvis: I know you can do something to redirect the console on the os side ... perhaps check the system handbook for infodocs
[20:37:48] <zarathustra> axisys: but the colors are not so good for me, is it easy to change them ?
[20:37:58] <trygvis> I wonder .. I've also used the single external serial port to talk to the disk array, I wonder if that has messed up something?
[20:41:55] <zarathustra> axisys: does it work like client/server ?
[20:42:37] *** dunc has joined #opensolaris
[20:43:25] <axisys> zarathustra: yes.. also whenever there is a new update client picks up the new version from server
[20:43:40] <zarathustra> :O
[20:43:55] <axisys> axisys: there are few contributions on that.. couple gif files
[20:44:05] <razrX> the hobbit interface looks remarkebly similar to bb ;)
[20:44:49] *** RaD|Tz has quit IRC
[20:46:13] <axisys> razrX: because it is next gen bb w/ major major change
[20:46:32] <axisys> razrX: the author said in his page so as well
[20:46:45] <razrX> at work we use munin and nagios but I'll look into hobbit
[20:47:06] <zarathustra> razrX: have you installed munin in a solaris 8 box ?
[20:47:23] <razrX> no, munin runs on a debian box unfortunately
[20:47:44] <zarathustra> razrX: ok, because I am having problems :(
[20:47:46] <razrX> we do have the munin-client running on various solaris 8 boxes on SPARC though I think
[20:48:02] <razrX> the munin-server is definitely a debian box
[20:48:19] <zarathustra> razrXdo you know where to find good documentation about munin ?
[20:48:35] *** mage2 has quit IRC
[20:48:56] <razrX> not really zarathustra, it's a lot of perl so you can write the plugins yourself
[20:49:07] <dunc> ok folks, i'm about to create a fileserver, i've used evms on linux in the past but i'm thinking i might give ZFS a go, but I don't think I'm up for the  zfs on root thing just yet.
[20:49:21] <zarathustra> razrX: ok
[20:49:27] <axisys> razrX: i get about 3500 test in less than 3 secs.. hard to beat performance like that
[20:49:30] <razrX> I can't remember setting up the munin-client was all that difficult
[20:49:43] <dunc> but i'd still like mirrored root, so what's the preferred method of std mirroring nowadays? i've used disksuite before
[20:49:58] <Stric> dunc: same. it's just called svm nowadays
[20:50:05] *** bondolo has joined #opensolaris
[20:50:20] <dunc> still md.tab and metadb action?
[20:50:30] <Stric> metadb, yes.. md.tab no
[20:50:36] <dunc> ok
[20:50:45] <dunc> well i'll be mostly at home anyway
[20:50:55] *** alobbs has joined #opensolaris
[20:51:06] <zarathustra> alobbs !
[20:51:13] *** Godsey has joined #opensolaris
[20:51:13] <dunc> is it ok to chop off a partition from each disk and then mirror them, and still use all the rest for my zpool?
[20:51:16] <Godsey> http://forum.sun.com/jive/thread.jspa?threadID=109908&tstart=0
[20:51:20] <axisys> razrX: author has a live page to see how it performs, check these two links out
[20:51:24] <axisys> razrX: http://hswn.dk/hobbit-cgi/bb-hostsvc.sh?HOST=voodoo.hswn.dk&SERVICE=bbtest
[20:51:26] <Godsey> I found this, same raid controller that I have
[20:51:29] <dunc> and will it be ok to add whole disks onto that at a later date when i've filled itt?
[20:51:33] <alobbs> ey zarathustra.. you're everywhere man! ;)
[20:51:34] <dunc> s/itt/it/
[20:51:38] <axisys> razrX: http://hswn.dk/hobbit-cgi/bb-hostsvc.sh?HOST=voodoo.hswn.dk&SERVICE=bbgen
[20:51:40] <zarathustra> alobbs: yeah !
[20:51:55] <Godsey> I wonder if it's not finding the 64bit drivers
[20:53:00] <Godsey> can I set my bios to lock processor to 32bit, install and then later change to 64bit?
[20:53:11] <trygvis> hmm .. does anyone happen to have a SMF descriptor for Qmail?
[20:53:12] <Godsey> (provided I'm able to install w/ 32bit)
[20:53:52] *** Ireul has joined #opensolaris
[20:55:13] <sickness> trygvis: you know, you were supposed to use daemontools... so maybe an SMF descriptor for the daemontools alone could suffice...
[20:55:51] <quasi> trygvis: it might have
[20:55:58] <trygvis> heh, but there's no need for daemontools
[20:56:15] <quasi> who need daemontools when you have smf?
[20:56:20] <trygvis> exactlyy
[20:56:37] <sickness> well, I don't know, not that I'm a fan of daemontools...
[20:58:55] <quasi> trygvis: http://sunsolve.sun.com/search/document.do?assetkey=1-25-79754-1
[20:59:19] *** Ireul has left #opensolaris
[21:00:15] *** Gman has joined #opensolaris
[21:00:32] <trygvis> hmm .. can I use any other sun login to access that?
[21:00:45] <axisys> sickness: i use daemontools on < sol 10
[21:01:10] <quasi> trygvis: you might - otherwise you can just register and get access
[21:03:16] <quasi> trygvis: there's another one as well that says you need to do what you did on the sp, set it up in the bios as well and eeprom console=ttya on the os
[21:03:51] <quasi> then it should work after the next init 6
[21:04:31] *** sparc-kly has joined #opensolaris
[21:04:32] *** ChanServ sets mode: +o sparc-kly
[21:05:41] <trygvis> seems like it's set to: console=text
[21:05:46] <sickness> axisys: I used daemontools on openbsd, freebsd and linux, but I'm not so fond of them...
[21:05:54] *** kupfer has quit IRC
[21:06:45] <axisys> axisys: it just works once i set it up.. i have been using it on solaris.. some people like Gerrit Pape's runit pkg.. similar in design but may be w/ more features
[21:06:52] <axisys> sickness: that was for u
[21:07:00] <axisys> i keep IMing myself
[21:07:03] <sickness> axisys: I never had a problem with "old" bsd-style way of starting things, and if a daemon crashes a lot, there's something to fix, trying to dumbly restart it every time sound to me as a microsoftish way of doing things, not a real unix way, the unix way should be to fix it and make it run rock solid the next time...
[21:07:40] <axisys> *shrug*
[21:08:17] <quasi> trygvis: Com Port Address:  On-board COM A Console connection: Direct Baud Rate: 9600 Flow Control: NONE Console Type: ANSI  in the bios according to http://sunsolve.sun.com/search/document.do?assetkey=1-9-82083-1
[21:08:26] <sickness> my only concern about SMF is initial "lack" of manifests/methods, beyond that I found it to be ok
[21:08:27] <axisys> sickness: i am sure u saw this http://cr.yp.to/daemontools/faq/create.html#why
[21:08:45] <sickness> exactly like daemontools, once you get used to it, no prob, you can go fine
[21:09:00] <axisys> sickness: u could use another service as template
[21:09:15] <axisys> sickness: may be sendmail
[21:09:38] <stevel> lol @ richlowe's newest bug
[21:09:51] <stevel> "b.o.o interest-list notifications are worse than useless."
[21:10:07] <sickness> axisys: yeah I read all that documentation, but I don't have all that "impossible" requirements, djb thinks that all the people should be smart asses like him, it turns out that a lot of people is really more stupid than him so maybe prefers stupid things to work with instead of his "mean" tools ;)
[21:10:33] <trygvis> quasi: seems like I have to buy a servicee plan to be able to activate the account
[21:11:20] *** sparc-kly_ has quit IRC
[21:11:34] <Gman> Submitter wants to work on bug
[21:11:35] <Gman>    No
[21:11:39] <Gman> richlowe, ;)
[21:11:40] <axisys> sickness: hehe .. he is famous for that
[21:12:02] <Gman> richlowe, you suck though, apparently you're running snv01 :)
[21:12:12] <stevel> yeah. maybe it's fixed in snv_51
[21:12:13] <stevel> :-P
[21:12:21] *** nwf has quit IRC
[21:12:24] <richlowe> Gman: I was going to file that, too, but I really can't do 3 in a day.
[21:12:25] <sickness> imho I'd choose rc.local over every other solution, it's deadly simple and I'm used to it, if in a hurry I can fire up a console and fix things even without docs, even drunk in a disco from midpssh on my cellphone!, with daemontools or smf you have to climb a learning curve... that said I was fine with daemontools in production.
[21:12:33] <sickness> :)
[21:12:42] <richlowe> Gman: what's the point of showing me a list of snv builds, for OpenSolaris/website and similar?
[21:12:55] <asyd> 1
[21:12:58] <asyd> raa
[21:13:12] <richlowe> Gman: if I start filing *all* of them, I'll be at it until Friday, I'm being choosy :)
[21:14:24] <Gman> :)
[21:14:31] *** cheatersrealm has joined #opensolaris
[21:15:33] <richlowe> (though if anyone else feels like joining in...)
[21:16:39] <cheatersrealm> I want to make a fileserver, I've got 2x40gb drives and 4x320gb drives, what are some good ideas on how to set it up?
[21:17:16] <hile_> use the 40GB drives as your rootdisks, obviously
[21:17:19] <oxygene> hmm.. svm mirrored 2x40gb for system, 4x320 in zfs (somehow, ask the ones with more zfs experience) for data
[21:17:22] <cheatersrealm> yeah
[21:17:29] <cheatersrealm> svm?
[21:17:30] <sickness> 2x40gb: ufs raid1 root disk, 4x320gb: raidz pool :P
[21:17:34] <jamesd_> 2x40   mirrored for now changed to the root pool in time  when zfs boot is supported... the  4x320 either as mirrored or raidz depending on speed and reduncancy needed.
[21:17:54] <cheatersrealm> I'd be using raid-z/zfs
[21:17:57] <cheatersrealm> (on the 320gb)
[21:18:07] <cheatersrealm> what about the boot into upgraded environment
[21:18:12] <sickness> most important thing: be sure to fill them with pr0n after all! :P
[21:18:26] <cheatersrealm> 2x20gb (mirrored) root's?
[21:18:30] <richlowe> raidz, raidz2, or striped across two mirrors depending on needs.
[21:18:42] <dunc> is it ok to chop off a few gig for system off some huge disks, and mirror the remaining space with zfs, and then add 2 whole disks (to mirror also) to the pool at a later date?
[21:19:02] <richlowe> mirror the 2 40's for the whole system, not just /, stick everything non-data on there.
[21:19:06] <quasi> trygvis: nope. I log in without a service plan
[21:19:20] <jamesd_> root pool will need to be seperate when zfs boot ==
[21:19:24] <cheatersrealm> right
[21:19:43] <quasi> trygvis: I just created an account on the page that asks for it
[21:20:00] *** nwf has joined #opensolaris
[21:20:00] <richlowe> jamesd_: I don't believe it has to be separate, there's just limits on its shape.
[21:20:04] <richlowe> jamesd_: or do you see something I don't?
[21:20:11] <hile_> chartersrealm: Does your backup vendor support backing up zfs?
[21:20:22] <cheatersrealm> hile_: back up?
[21:20:34] <hile_> yes.
[21:20:35] <richlowe> I think that's a No. :)
[21:20:38] <dunc> lol
[21:20:43] <hile_> lol richlowe
[21:20:44] <jamesd_> richlowe, last i read it either have to a single drive or mirrored in the root pool... no other single drives or raidz allowed in it.
[21:21:06] <cheatersrealm> ok, so last time I tried to install solaris, it didn't seem to support my raid card
[21:21:09] <cheatersrealm> well sata
[21:21:12] <cheatersrealm> it's only sort of raid
[21:21:13] <richlowe> jamesd_: yeah, but that's not necessarily *separate*, just of the right shape (there has to be a better word than shape...)
[21:21:26] <richlowe> all I can come up with are synonyms, so screw it. :)
[21:21:29] <sickness> richlowe: type?
[21:21:31] <hile_> chartersrealm: also do you forsee a need to deport the data diskgroups to another system which might be running a solaris version older than s10u2?
[21:21:50] <cheatersrealm> hile_: no, this is going to be my only solaris system
[21:21:57] <richlowe> or if you're running snv, to s10u2 even.
[21:22:08] <richlowe> since that's zpool v. 2, not v3 as in snv.
[21:22:14] *** laca has joined #opensolaris
[21:22:24] <cheatersrealm> well I don't plan on back compatibility
[21:22:50] *** sparc-kly_ has joined #opensolaris
[21:23:11] <axisys> sickness: and there is upstart that ubuntu started
[21:24:01] *** coolvibe has quit IRC
[21:25:07] <sickness> axisys: oh, didn't know it, what's that?
[21:25:26] <sickness> axisys: wait, there's also the "make" thing, I think some distro has that, maybe archlinux?
[21:27:08] *** _william_ has joined #opensolaris
[21:28:25] <axisys> sickness: it is a replacement of rc script
[21:28:28] <_william_> hi all
[21:30:53] <cheatersrealm> so what is the deal with the syba sata (sil3114) cards?  would there have been support for that in solaris express 3/06 ?
[21:31:53] *** sparc-kly has quit IRC
[21:33:32] <axisys> looking for "how to install opensolaris".. just installed solaris express latest (48)
[21:36:23] <axisys> got it..
[21:38:08] <trygvis> quasi: I made a new account, but I still can access that site
[21:39:50] <quasi> trygvis: odd - it worked for me - I have no contracts tid to my account
[21:39:57] *** spackest has left #opensolaris
[21:40:11] <quasi> (well, that is none that I know of ;)
[21:46:11] *** elflord has quit IRC
[21:53:39] *** saablover has joined #opensolaris
[21:53:52] *** hile_ has quit IRC
[21:53:55] <saablover> anyone from the sun ray team online ?
[21:54:00] *** Cybernd has joined #opensolaris
[21:54:41] <cheatersrealm> sun rays are awesome
[21:54:45] <cheatersrealm> my friend got a few
[21:54:58] <saablover> I know
[21:55:12] <saablover> but video playback doesn't work
[21:55:35] <saablover> from the browser
[21:56:49] <boyd> I wouldn't expect good performance from video playback on a sunray
[21:57:12] <saablover> I don't expect full screen video playback
[21:57:22] <saablover> just basic video files
[21:57:27] <saablover> in a small video
[21:57:32] <saablover> window that is
[21:58:11] <quasi> boyd: no? damn, then I'm going to have to give up on my plan to use a sunray 2FS to drive the tv
[21:58:36] * boyd shrugs... I'm no expert....
[21:58:58] <saablover> when video playback would be directed to the onboard video controller
[21:59:00] *** LordKing has quit IRC
[21:59:01] <saablover> it would work
[22:01:11] <boyd> There is not much intelligence in one of them... what would you send to the terminal over the net?
[22:03:37] <Peanut> boyd: actually, with the video chipset, it should be able to do color-space conversion and hardware scaling in the SunRay, and some pieces of software make use of that capability.
[22:04:42] <saablover> I think a great market is lost with streaming media, it would be so cool to see several sun rays playing video from a central system
[22:05:04] <cheatersrealm> thanks for the help, guys
[22:05:05] *** cheatersrealm has left #opensolaris
[22:05:56] *** Bart_M has quit IRC
[22:07:57] <boyd> Peanut: Sure, colorspace and scaling, but you still have to get the stream to the terminal
[22:08:06] <boyd> (after decoding)
[22:08:44] <Peanut> The colorspace and scaling are a big part of the decoding I think.
[22:09:39] <boyd> Hmm... maybe... I dont really know. I'd have thought that decompression and frame re-ordering, etc would be up there.
[22:10:09] <saablover> on a well designed network it should work without a problem sending the video to the terminals
[22:10:16] <Peanut> Don't know exactly how the division of labour is between server and client.
[22:10:27] <saablover> but this breaks the low bandwidth thing of sunray
[22:11:12] <boyd> I don't think low bandwidth has been a big target... the original release *required* a dedicated 100Mbps network
[22:11:53] <boyd> I think the main target has been TCO for office installations
[22:12:03] <lasseoe> yup
[22:12:09] <boyd> .. for which video has not been a priority
[22:12:12] *** Tpenta has joined #opensolaris
[22:12:50] *** ChanServ sets mode: +o Tpenta
[22:12:57] <boyd> Morning, Tpenta
[22:12:59] <Doc> alan
[22:13:15] <Tpenta> morning folks
[22:13:27] <Doc> well, afternoon
[22:13:58] <boyd> Hi Doc, NY?
[22:14:02] <PerterB> uh, evening
[22:14:05] <Doc> yup
[22:14:25] <boyd> Doc: Where you going tonight?
[22:14:40] <Doc> halloween parade
[22:14:51] <boyd> Cool
[22:15:00] <Doc> was in the audience for letterman last night, which was great
[22:15:13] <boyd> It has to be.... more convicing than the kids in my street last night :)
[22:15:59] <Tpenta> we had lots of kids coming around trick or treating last night, fotos on flikr later today. we made an effort on the house too :)
[22:16:49] <Doc> blah.. just what we need in australian - another warped american tradition
[22:17:12] <Tpenta> the kids have fun; that's the important thing
[22:17:31] <Tpenta> at least they don't egg the houses that don't have sweets
[22:17:33] <quasi> Doc: look at the bright side - more money for the shop owners ;)
[22:18:36] <boyd> Tell me about it... the 7/11 round the corner had pre-made gouge the sucker bags
[22:19:01] <Doc> which reminds me, i need to upload some more photos...  blah
[22:20:08] *** jmarcot has joined #opensolaris
[22:20:17] * quasi grmbls at sun slowly changing all newsletters to html format - now the time has come to blueprints :(
[22:21:07] <Doc> html rather than pdf?
[22:21:28] <Doc> or the blueprints notification newsletter
[22:21:40] <quasi> the newsletter
[22:22:29] <quasi> I like the blueprints being pdf, but having to find an html mailer to read the newsletter is annoying
[22:23:27] <Doc> what mail program do you use?
[22:23:50] <Doc> http://www.flickr.com/photos/docbert/272563811/
[22:23:59] <Doc> damn i love that photo :)
[22:24:02] <boyd> Geez, they should at least be multipart/alternate
[22:24:02] <quasi> mutt
[22:24:21] <boyd> Yeah, that's a good pic, Doc
[22:24:36] <quasi> Doc: very cool pic
[22:24:48] <Doc> quasi: setup lynx as a helper and use it to view the html
[22:25:22] <Doc> text/html;      lynx -force_html %s
[22:25:36] <quasi> Doc: doesn't make copying the urls for use in firefox any easier
[22:26:06] <Doc> so set it up to use filefox if your in x, lynx if your ot
[22:26:30] <Doc> easy to do, and pleanty of examples out there showing how
[22:26:42] *** FBdev has quit IRC
[22:27:08] <quasi> except I don't have ff on the server my mail runs on
[22:27:33] <Doc> so use imap
[22:28:11] <quasi> no, sun should just stop wasting electrons on crappy html ;)
[22:28:23] <Doc> and even so, copying urls out of lynx is no harder than out of a text based email
[22:29:07] <Doc> hmm.. share price dropped a bit
[22:33:54] *** hspaans has joined #opensolaris
[22:34:29] *** FBdev has joined #opensolaris
[22:43:48] <Tpenta> doc, send the url to david smith to send to his dad
[22:44:07] <Doc> which url?
[22:44:19] <Tpenta> for the photo
[22:44:37] <Doc> i sent all of the sams the main url
[22:44:54] <Tpenta> :) you did know that john breeds alpacas didn't you?
[22:45:09] <Doc> ahh.. no, i didnt
[22:46:01] <Doc> anyway, off too do some shopping
[22:46:02] <Tpenta> he does
[22:46:09] *** glagasse has quit IRC
[22:50:38] *** nwf has quit IRC
[22:50:49] *** saablover has quit IRC
[22:54:25] *** nwf has joined #opensolaris
[22:54:38] <delewis> anyone familiar with QFEs? they do have four, *separate* interfaces, right? (meaning I could assign each interface a separate IP, etc.)
[22:55:08] *** slowhog has quit IRC
[22:56:23] <lasseoe> delewis, yes
[22:56:31] <delewis> lasseoe: cool
[22:56:58] <delewis> will I have any trouble using them in Solaris 10 or Nevada?
[22:57:06] <lasseoe> delewis, nope should just work.
[22:57:08] * delewis is just making sure before he purchases one for his router
[22:57:12] <delewis> lasseoe: excellent.
[22:57:20] <lasseoe> uhm
[22:57:31] <lasseoe> afaik QFE cards use the system MAC address though
[22:58:13] <boyd> That can be changed
[22:58:16] <delewis> I don't think that would really be a problem, as I'll only be using two interfaces (one to the to the DSL modem and one to the local switch)
[22:58:20] <boyd> use-local-mac? = false
[22:58:21] <delewis> so two, completely separate subnets
[22:58:46] <lasseoe> boyd, pretty sure some or all qfe cards don't have local MACs
[22:58:53] <Symmetria> hrm
[22:58:54] *** solarisjon has left #opensolaris
[22:58:59] <delewis> lasseoe: probably the older, SBus ones, right?
[22:59:08] <delewis> which is what I'll be purchasing for an Ultra 2
[22:59:10] <lasseoe> delewis, yeah
[22:59:21] <Symmetria> anyone here involved or know anyone who is involved with opensolaris.org? I wanna know what the deal with mirroring it is, if Im allowed to, and who I can speak to about it
[22:59:24] <delewis> hmm, well, I don't *think* it will be a problem.
[22:59:29] <delewis> like I said, two subnets
[22:59:38] <delewis> so identical mac addresses shouldn't matter
[22:59:41] <boyd> lasseoe: All the ones I've used are no prob... but they're PCI
[23:00:51] <delewis> what does SBus become saturdated at about, anyway? gigabit speeds I would presume.
[23:01:02] <delewis> which might become an issue if I ever decide to start using the other two interfaces
[23:01:17] <axisys> any preferred one to install opensolaris out of this list http://www.opensolaris.org/os/community/documentation/doc_index/install/
[23:01:33] <axisys> i already got the solaris express b48 installed
[23:02:07] <lasseoe> boyd: ah I might have remembered wrong, think it was one of my old hme SBus cards that didn't have a local mac
[23:02:09] <lasseoe> http://sunsolve.sun.com/search/document.do?assetkey=1-9-16733-1
[23:05:20] *** laca is now known as lacaAFK
[23:06:07] * delewis wishes SunSolve worked well with Opera
[23:06:22] * boyd wishes SunSolve worked well.
[23:06:27] <delewis> haha :-)
[23:06:28] <delewis> true.
[23:06:45] <delewis> boyd: I don't think it's worse than docs.sun.com, though
[23:06:50] <delewis> which seems to go down once about every week
[23:07:23] <delewis> it's also been *slow* lately.
[23:07:49] *** mrdeviant has quit IRC
[23:07:58] <boyd> My pet peeve with d.s.c is the disappearing search targets. Search for something, see the relevance indicators, go to one of the indicated places... they're all gone. Sucker!
[23:08:11] <delewis> yes, I've noticed that, too
[23:09:24] <delewis> it just seems like it takes forever to access some docs
[23:09:44] <boyd> Indeed.. I fetch PDFs and read/search locally these days
[23:09:57] <delewis> yeah, I always fetch the PDFs unless I'm just searching for something quickly
[23:09:58] <richlowe> docs.sun.com hasn't been particularly speedy for several years.
[23:10:02] <delewis> I can't *stand* to read HTMl documentations.
[23:10:15] <delewis> richlowe: i think it's detiorated even more, recently.
[23:12:37] *** jmarcot has quit IRC
[23:13:11] *** axisys has quit IRC
[23:23:27] *** fik has joined #opensolaris
[23:31:13] *** EdLin has left #opensolaris
[23:31:18] *** slowhog has joined #opensolaris
[23:32:08] *** inaddy has joined #opensolaris
[23:44:07] <Tpenta> http://www.flickr.com/photos/43071033@N00/tags/halloween2006/  :-D
[23:44:45] <Symmetria> hrm
[23:44:51] <Symmetria> sunfreeware is bigger than I thought it would be
[23:46:23] *** fik_ has joined #opensolaris
[23:53:17] <Error_404> Symmetria: it sure is
[23:54:12] *** jlc has joined #opensolaris
[23:54:40] <hspaans> Tpenta: nice cat ;-)
[23:54:46] <jlc> If I add a network interface after the install what steps should I take to get the system to recognize it?
[23:54:49] <jlc> touch /reconfigure
[23:54:54] <jlc> i did a sys-unconfig
[23:54:59] <jlc> and it still doesn't show up
[23:55:09] <jlc> 3com 3c509 i think it is
[23:55:39] <Tpenta> hspaans: :-D
[23:55:51] <hspaans> jlc: check with prtconf and prtdiag
[23:56:38] *** jsubl2 has joined #opensolaris
[23:57:33] <hspaans> jlc: you should see a driver for you card
[23:58:47] <jlc> hurm
[23:58:52] *** fik has quit IRC
[23:59:52] <jlc> touch /reconfigure would have made it shown up in there right?





top