[00:00:24] *** axisys has quit IRC
[00:03:10] *** d3vi1 has left #opensolaris
[00:03:47] *** stevel has joined #opensolaris
[00:03:49] *** ChanServ sets mode: +o stevel
[00:03:50] *** vmhobbes has quit IRC
[00:03:59] <richlowe> stevel: mornin'
[00:04:06] <Gman> hey steve
[00:06:05] <stevel> morning rich, glynn
[00:06:55] <richlowe> stevel: when can we expect the fixed onnv push?
[00:07:06] <stevel> hopefully today
[00:07:15] <stevel> sorry, i keep getting dragged to various presentations and talks
[00:07:16] <richlowe> cool.
[00:07:19] *** koolniczka has joined #opensolaris
[00:10:38] *** unixconsole has left #opensolaris
[00:11:50] *** RaD|Tz has joined #opensolaris
[00:11:57] *** koolniczka has quit IRC
[00:14:44] *** Cybernd has joined #opensolaris
[00:16:25] *** axisys has joined #opensolaris
[00:21:47] <twincest> richlowe> did you see my mystery file? :)
[00:22:36] <richlowe> Hm?
[00:23:25] <twincest> http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/cmd/man/src/util/instant.src/sqlnet.log
[00:23:44] <richlowe> tell me I didn't miss one :\
[00:23:58] <richlowe> if that's not an opengrok snafu, it's steve's fault ;)
[00:24:07] <twincest> it's not, it's in Hg too :)
[00:24:16] <richlowe> except src is down.
[00:24:23] <stevel> again???
[00:24:23] <twincest> again?
[00:24:24] <stevel> wtf
[00:24:24] *** Tpenta has joined #opensolaris
[00:24:27] <richlowe> twincest: for what it's worth, cmd/man/ is littered with detritus.
[00:24:39] <richlowe> twincest: we got several bits of it during the initial hg bundle stuff.
[00:24:40] <twincest> it was working after the previous downtime :(
[00:24:42] <richlowe> twincest: guess I missed bits. :(
[00:24:56] <twincest> rich: is that an artifact from hg, or are they really in onnv-gate?
[00:25:03] <richlowe> twincest: the others were really there.
[00:25:15] <Tpenta> gman: do you remember the null ptr deref bug in nimbus that was causing teh dual head issues?
[00:25:18] <richlowe> twincest: steve only killed those that outright made hg suffer, however.
[00:25:28] <Gman> Tpenta, yes
[00:25:29] <stevel> richlowe: that one is really there
[00:25:32] <richlowe> I guess I didn't go hunting for the just plain stupid :(
[00:25:35] <richlowe> My bad.
[00:25:38] * richlowe makes a note
[00:25:46] <Tpenta> have you got the cr num handy?
[00:25:52] *** hile_ has quit IRC
[00:25:55] <Gman> Tpenta, lemme check
[00:25:56] <richlowe> actually.
[00:26:02] <sommerfeld> sqlnet.log: what the heck?!?!!?!
[00:26:07] <richlowe> twincest: you signed the SCA now, findunref and patience, clean it out :)
[00:26:10] <stevel> goddamn glassfish crashes way too often
[00:26:15] <richlowe> sommerfeld: see above, src/man is just littered in junk.
[00:26:16] <twincest> rich: i filed 6490919 ;-P
[00:26:19] <richlowe> sorry, cmd/man
[00:26:39] <Gman> Tpenta, 6446921 ?
[00:26:40] <richlowe> sommerfeld: until they made hg unhappy, it had generated files under TW control too.
[00:26:45] <richlowe> sommerfeld: and other, similar, junk.
[00:27:03] <Tpenta> i just had a mail from rob thurlow asking me about why i cant run nimbus on dual headed sparc :)
[00:27:39] <Gman> Tpenta, heh
[00:27:46] <Gman> who's rob thurlow?
[00:27:47] <Tpenta> nah thats not it
[00:27:59] <Tpenta> its teh one where we get all the panels restarting every 20 seconds
[00:28:30] *** ChanServ sets mode: +o Tpenta
[00:28:47] * Gman queries again
[00:29:11] <Tpenta> becaus enimbus does a null ptr deref
[00:29:52] * richlowe runs findunref with no exceptions and waits
[00:30:08] <Tpenta> found it
[00:30:19] <Tpenta> i think it was 6452649 but reading to check
[00:31:13] <Tpenta> ok, that's a dup of 6452649
[00:32:01] <Tpenta> yup, that was it; hmmm still only dispatched
[00:32:27] <richlowe> twincest: I like the .eps logos in there, too.
[00:33:09] <Tpenta> i mean a dup of 6453340
[00:34:04] <Gman> Tpenta, looks like a very similar trace to 6446921
[00:34:22] <Tpenta> in which case, the bugs should be linked
[00:34:44] <sommerfeld> sqlnet.log is definitely in the findunref exception list...
[00:35:00] <Tpenta> it does doesnt it
[00:35:16] <Tpenta> i'll add that as a see also in 6453340
[00:35:26] <Gman> Tpenta, yeah, unfortunately not all engineers are so tuned in
[00:35:36] <Tpenta> so when was that delivered?
[00:35:58] <sommerfeld>  97/10/07 16:35:08
[00:36:03] <stevel> The "ON Tools Hg Conversion Project" is now open.
[00:36:04] <stevel> woot!
[00:36:06] <sommerfeld> (according to the SCCS delta)
[00:36:12] <Tpenta> vermillino46 shoudl be in b51, right? it's not fixed then
[00:36:14] <Gman> Tpenta, should be for b53 i think
[00:36:17] <Tpenta> ahh
[00:36:20] <Tpenta> ok
[00:36:26] <Gman> at least, the nimbus module was bumped then
[00:36:41] <Gman> [it may not be fixed though, in which case i'd start getting angsty]
[00:37:02] <Tpenta> I won't quite close it as a dupe yet, or should I?
[00:37:21] <Tpenta> I'll drop something in the evaluation. the wierd thing is that both bugs have the same RE
[00:38:18] <richlowe> stevel: yeah.
[00:38:56] <alanc> hmm, nv_53 WOS build is 11/20 - release will probably be delayed a little by the US holiday that week...should be out first part of December...an early Christmas present for Solaris Desktop fans?   (GNOME 2.14, FireFox 2.0, DRI for Intel 845-915 chips, and a few other goodies)
[00:39:24] <richlowe> 2.16, one would hope.
[00:39:26] <Tpenta> 2.14? I thought that woiuld be 2.16
[00:39:35] <Gman> Tpenta, yeah, make sure they're both assigned to erwann
[00:39:40] <Tpenta> they both are
[00:39:41] <alanc> yes, 2.16
[00:39:44] <alanc> stupid typo
[00:39:49] <Tpenta> :)
[00:40:17] <alanc> X gate closed today though, so Xorg 7.2 won't make that build - just the X dtrace probes
[00:40:38] *** jmcp has quit IRC
[00:41:19] <Gman> stevel, how possible is it for me to get an account on grommit for pos?
[00:41:31] <schily__> good night!
[00:41:36] *** schily__ is now known as schily
[00:41:42] <stevel> pos. lol... took me a sec to figure out what you mean
[00:41:45] <stevel> meant rather
[00:41:49] <stevel> gimme a minute
[00:41:58] <stevel> can you email me a public key for ssh?
[00:42:04] <kleppari> alanc, getting Xorg probes is nice, though :)
[00:42:05] <Gman> sure
[00:44:03] <Tpenta> stevel: nightly drop day?
[00:44:25] <stevel> tpenta: so it is
[00:44:30] <stevel> sorry, time difference threw me off
[00:44:33] <Tpenta> :)
[00:44:42] <twincest> nightly drops?  how two weeks ago
[00:44:44] <Tpenta> i'm just transferring my onnv-cloen tarball
[00:45:07] <stevel> tpenta: how much would you hate me if i started pushing closed-bins every day?
[00:49:23] <richlowe> he wouldn't!
[00:49:45] <richlowe> there's no requirement there for the -nd to be done daily, anyway.
[00:49:51] <richlowe> it'd be nice, sure, but not *necessary*
[00:52:17] <boyd> stevel: I assume you're going to -announce when the newly remade repo is back?
[00:52:34] <stevel> boyd: nah, tools-discuss, onnv-notify, and opensolaris-code probably
[00:52:41] <richlowe> not. notify.
[00:52:46] * richlowe makes a stand! ;)
[00:52:56] <stevel> fair enough
[00:53:03] <richlowe> request-sponsor, and *-notify shouldn't be used for discussion, really.
[00:53:09] <twincest> i wonder if i should be reading tools-discuss or -code
[00:53:10] <stevel> i'll replace onnv-notify with 10,000 spams to richlowe ;-)
[00:53:17] <richlowe> stevel: works for me.
[00:53:27] <richlowe> if it keeps junk off -notify, hit it as hard as you like :)
[00:53:30] <boyd> richlowe: I agree, but I don't think this is discussion
[00:53:31] * richlowe takes one for the team
[00:53:34] <Gman> haha
[00:54:06] <richlowe> boyd: how likely is it someone reading -notify wouldn't be reading tools-discuss or -code?
[00:54:26] <twincest> rich: me!
[00:54:28] <boyd> richlowe: I am (at least regularly)
[00:54:33] <Gman> richlowe, i bet there are hardly any people on notify still
[00:54:57] <boyd> It seems like -announce would be ok as a place
[00:55:03] <richlowe> Yeah, that works.
[00:55:07] <Gman> 25 ;)
[00:55:28] <Gman> real good mix of sun and non-sun people though
[00:55:31] *** uebayasi has joined #opensolaris
[00:55:32] <boyd> It's a pretty central announcement
[00:55:45] <richlowe> Gman: Not sure why sun folk would be on -notify.
[00:55:53] <richlowe> unless they can't get on the internal equivalents.
[00:55:56] <Gman> which sort of indicates that either a) not many people care, or b) not many people know about it
[00:56:09] <boyd> I've been reading the RSS from my onnv clone more actively lately
[00:56:11] <Gman> richlowe, danek's on notify, he cares! :)
[00:56:12] <Error_404> hmm, that's odd
[00:56:17] <boyd> Gman: I suspect B
[00:56:20] <richlowe> it's a beta, don't make statistical judgements until it's actually 'real'
[00:56:22] <Gman> boyd, yeah
[00:56:51] <Error_404> my laptop boots solaris, i get the (c) Sun Microsystems junk, then it shuts down immediately.
[00:56:52] <boyd> I've pointed a few people at onnv-notify here that have said "cool! I didn't know that existed"
[00:57:02] <twincest> Gman: there are only 50 contributors at all, i'm going with #1 ;-)
[00:57:10] <richlowe> twincest: 51!
[00:57:11] <Error_404> twincest: 51
[00:57:16] <boyd> Error_404: try booting with -kd and see the message
[00:57:17] <twincest> #51 doesn't count
[00:57:18] <Gman> heh
[00:57:21] <richlowe> don't round when the number is so small ;)
[00:57:25] <Error_404> twincest: hey... jerk
[00:57:29] <alanc> it's pretty trivial to get on the ongate internal putback notifications - happens automatically if you do a bringover from the gate even
[00:57:53] * boyd goes back to teaching DTrace
[00:57:55] <Gman> everyone who commits back to jds, *has* to join the notify alias
[00:58:10] <richlowe> I believe that's the intent with the hg stuff.
[00:58:13] <richlowe> similar, anyway.
[00:58:17] <Gman> cool
[00:58:34] <richlowe> assuming I'm right in that add-gateling does that, anyway.
[00:58:39] <Gman> i think it's important too for people to have an interest :)
[00:58:48] <sommerfeld> yeah, that's what add-gateling is for
[00:59:02] <alanc> "The synchronization process is expected to take approximately one day to complete." - apparently our bug db resync is slower than even I realized
[00:59:19] <boyd> Hehe
[01:00:17] <richlowe> stevel: I don't think it's a problem if you guys push out of sync.
[01:00:24] <richlowe> stevel: given the -nd bits are kind of a happy extra, right now.
[01:01:18] *** alanc has quit IRC
[01:01:38] <Error_404> ahh, I see it can't mount the root path
[01:01:40] <Error_404> i wonder why
[01:01:46] <boyd> Hey, stevel How did you make your complete bundles? against an empty repo?
[01:01:54] <stevel> boyd: yup
[01:02:00] <boyd> 'kay
[01:03:02] *** axisys has quit IRC
[01:03:05] <Kitty>  /42
[01:03:56] <stevel> ugh
[01:04:01] <stevel> i can't generate the changelog because bugster is down
[01:04:39] <Gman> it's working here
[01:04:50] <stevel> monaco queries don't work
[01:04:51] <Gman> sure you're not trying to use monaco or some other cli?
[01:04:55] <Gman> right
[01:05:05] <stevel> sorry, i'll rephrase my sentence
[01:05:17] <stevel> i can't generate the changelog because bugster is a piece of shit
[01:05:26] <Gman> i can do a query here, and export to a file
[01:06:57] <mlh> the solaris10/solaris express pages are about as clear as mud
[01:07:21] <mlh> there's links for solaris10 6/06, and there's solaris express 10/06
[01:07:32] *** lacaAFK is now known as laca
[01:07:57] <stevel> /home/stevel/ws/tonic/deliver.sh $fflag -p /net/donuthole/export/opensolaris/$DATE
[01:07:58] <mlh> it's unclear what the express 10/06 means; is the 10 for october or solaris10?  is the 06 for June or 2006?
[01:08:00] <stevel> whoops
[01:08:01] <stevel> wrong window
[01:08:29] <twincest> mlh: solaris express october 2006
[01:09:05] *** solarisjon has quit IRC
[01:09:40] *** Cyl has quit IRC
[01:09:57] <Error_404> hmm... solaris requires booting off the active partition?
[01:10:25] <dlg> anyone here familiar with writing drivers for pci devices?
[01:10:50] *** alanc has joined #opensolaris
[01:14:52] *** stevel changes topic to "Latest builds: SXCR: 51a (sparc) / 51 (x86) | ON build: 51 | ON nightly: 200611"
[01:15:08] *** stevel changes topic to "Latest builds: SXCR: 51a (sparc) / 51 (x86) | ON build: 51 | ON nightly: 20061106"
[01:23:22] *** Cybernd has quit IRC
[01:24:25] <mlh> twincest: thanks.
[01:25:21] <mlh> can I suggest to however is in charge that they make it clearer?  At least give release dates, but better would be to call express "solaris10 express 10/2006'
[01:28:29] <alanc> express isn't Solaris 10, it's the version that will likely be called 11
[01:29:10] <mlh> yeah I sorta realised that but I had to point out the ambiguity anyway
[01:29:25] <twincest> Solaris 11 0.2 ;-)
[01:29:29] <mlh> also; it's variously referred to as "Solaris Express 10/06 Software" and " Software Express for Solaris 10/06 Software"
[01:29:57] <mlh> or just Solaris Express Oct 2006
[01:30:00] <mlh> would be good
[01:31:31] <mlh> or follow industry conventions (gasp!) and call it Solaris beta
[01:32:43] <alanc> there's a separate Beta program when it's closer to actually being done
[01:33:01] <alanc> Express is "extended alpha"
[01:33:59] <mlh> ok, thanks
[01:34:53] <delewis> and it's still far more stable than real releases of certain operating systems. :-)
[01:35:25] *** Cyl has joined #opensolaris
[01:38:03] *** jafari has joined #opensolaris
[01:38:49] *** hile_ has joined #opensolaris
[01:38:58] * twincest downloads 335MB of fedora updates
[01:39:05] <twincest> their slogan should be, "Not for people who pay for Internet usage"
[01:40:31] <hile_> ugh
[01:40:42] <hile_> that also makes one wonder why you're using shitora in the first place
[01:41:02] <twincest> mainly because it includes very up to date kernels :)
[01:41:39] <twincest> (helpful for hardware support)
[01:42:36] *** klocze1 has quit IRC
[01:43:14] <hile_> The only place that I have linux right now is an old VirtualPC instance which is only used for NIMOL for AIX
[01:43:39] *** yarihm has quit IRC
[01:43:47] *** axisys has joined #opensolaris
[01:44:21] <axisys> anyone worked on traffic manager multipath? i want to setup a T2000 w/ 3510 and like to take advantage of it
[01:44:30] * twincest is planning on migrating to solaris once it has a couple more features
[01:44:50] *** bondolo has quit IRC
[01:44:59] <axisys> best would be someone already setup like that and has a url :-)
[01:44:59] <hile_> use whatever works best :-)
[01:45:17] <hile_> though i'd be hard pressed not to go with solaris or AIX somewhere that needed to be solid
[01:45:44] <twincest> hile: well, solaris works better.  it just doesn't work in all the areas i need :(
[01:46:28] *** kloczek has joined #opensolaris
[01:46:47] <hile_> ah
[01:47:22] <hile_> if my desktop were other than a SunRay used mostly for xchat/gaim/firefox/mathematica/matlab/etc, i might use something other than Solaris on the desktop
[01:48:20] <dlg> twincest: a working diff command?
[01:48:32] <twincest> gdiff
[01:48:39] <hile_> a working diff?
[01:48:40] <twincest> well, normal diff works fine for me
[01:48:41] <hile_> diff works fine
[01:48:50] <twincest> but you can use gnu diff if you like
[01:49:05] <axisys> twincest: i like tkdiff best
[01:49:09] <mlh> "shitora" <-- clever
[01:49:20] <axisys> dlg: for you that is.. not twincest :P
[01:49:41] <axisys> it color codes the diff.. very easy to detect diff
[01:50:32] <dlg> axisys: im fine with normal diffs
[01:50:32] <axisys> so anyone used Traffic Multipath w/ 3510 and T2000 or something close ?
[01:51:05] <axisys> like to cheat.. need to setup one
[01:54:08] <Tpenta> +alanc: congratulations on the integration of teh probes (just saw your blog)
[01:54:20] <kleppari> http://slashdot.org/article.pl?sid=06/11/07/2235233&from=rss
[01:54:23] <kleppari> vista is skynet
[01:54:24] <OnkelSchorsch_> hm. anyone have an ultra 20? mine seems to be quite loud.
[01:54:25] <alanc> thanks
[01:55:28] <jamesd> OnkelSchorsch_, mine only seems loud when the cpus are busy over 60% or more.. anything less than it just purs
[01:56:13] <OnkelSchorsch_> jamesd, so mine is probably borked :(
[01:56:28] <jamesd> my microwave almost 20 feet away makes more noise than my u20
[01:56:35] *** RaD|Tz has quit IRC
[01:56:37] <jamesd> call up sun they have a 2 year warranty
[01:57:02] <OnkelSchorsch_> I can call sun, just because it's too noisy?
[01:57:11] <richlowe> Hm.
[01:57:17] <axisys> OnkelSchorsch_: mine is pretty quite .. but like jamesd's one i never raised the cpu usage that high yet
[01:57:18] <alanc> isn't there an option in the BIOS to turn the fans down from "agressive cooling"?
[01:57:19] <jamesd> call up and see what tehy say.
[01:57:33] <alanc> or was that the previous models?
[01:57:46] <jamesd> i never looked at the bios in that much detail...
[01:57:57] <OnkelSchorsch_> I get it on the try and buy. I didn't get any contact info. or anything.
[01:58:11] <Tpenta> alac: how did you add the tags cloud to your blog?
[01:58:22] <jamesd> OnkelSchorsch_, grab the latest  driver disk from sun and patch the bios there was a few fixes that have to do with fan noise
[01:58:40] <OnkelSchorsch_> thanks. but I already installed those :(
[01:59:05] <OnkelSchorsch_> It was much more quiet before I installed those though :(
[01:59:14] <richlowe> Gman: you around?
[01:59:23] <Gman> yep, but heading out shortly
[01:59:37] <jamesd> OnkelSchorsch_, call sun.  complain about the noise if they don't have an anwser tell them you are conserned that the fans aren't cooling properly could be a bad bearing, they can swap the fan now or swap the cpu in 6 months...
[02:00:05] <OnkelSchorsch_> ok, thanks jamesd. I'll do that
[02:02:54] <stevel> hrm
[02:02:58] <stevel> what to do about notifications
[02:03:07] <stevel> pushing to ssh://stevel at hg dot opensolaris.org/hg/onnv/onnv-gate
[02:03:09] <stevel> searching for changes
[02:03:09] <stevel> remote: adding changesets
[02:03:09] <stevel> ^Cinterrupted!
[02:03:16] * stevel almost just spammed onnv-notify with another 1500 emails
[02:03:18] <stevel> :-P
[02:03:21] <Tpenta> stevel: i just started my builds
[02:03:32] *** nwf has joined #opensolaris
[02:04:07] <Gman> stevel, temporarily add a filter to onnv-notify?
[02:04:19] <stevel> gman: yeah, i think that's easiest
[02:04:26] <richlowe> stevel: mv
[02:04:36] <richlowe> (import to another gate, then switch them manually)
[02:04:37] <Gman> ok, later guys will be back in a couple of hours
[02:04:38] *** Gman is now known as GmanAFK
[02:04:46] <GmanAFK> stevel, thanks again for pos - it's looking fab ;)
[02:04:47] <richlowe> don't know if that's doable though.
[02:04:58] <richlowe> GmanAFK: you need a better acronym.
[02:05:09] <GmanAFK> richlowe, :)
[02:06:12] <stevel> richlowe: well, i want notifications sent for the ones i've missed in the time the bridge was down
[02:06:17] <stevel> i've thrown onnv-notify into emergency notification
[02:06:25] <stevel> i'll discard all the notifications except for the new ones
[02:06:34] <richlowe> Yeah, that makes it annoying.
[02:06:35] <richlowe> sorry.
[02:06:57] <stevel> damn i'm late for rehearsal
[02:09:37] <Tpenta> alanc: don't worry, I found it (and put it in)
[02:10:09] <alanc> Tpenta: oh, missed the query since you typoed my name
[02:10:28] <alanc> I stole the code that Dave Johnson sent to bloggers after the 3.1 rollout on b.s.c
[02:10:51] *** OnkelSchorsch_ has quit IRC
[02:10:54] <alanc> only change I made was to tweak the styles a little so they change color as well as size
[02:12:31] <Error_404> wtf? "WPA is only supported in the ath-0.3 driver" why... uhhh, not 0.4?
[02:12:36] <twincest> alanc: i suspect i know the answer to this, but is there any chance proprietary ATi drivers might appear? :)
[02:12:37] *** dlg has quit IRC
[02:12:48] <alanc> twincest: your guess is as good as mine
[02:12:57] <alanc> always a chance I suppose
[02:13:02] <twincest> i was hoping the AMd/ATi thing might affect it
[02:13:34] <alanc> it might, but if so, our VP's have to work it all out with their VP's long before us grunts get told what's going to happen
[02:13:55] *** dlg has joined #opensolaris
[02:13:57] *** lin has joined #opensolaris
[02:15:30] <Tpenta> alanc: yea i just lifted it and placed it in my title block
[02:17:11] *** miffe_ has joined #opensolaris
[02:17:39] *** dwc- has joined #opensolaris
[02:17:52] <stevel> damn
[02:17:56] <stevel> i fucked up the notifications anyway
[02:17:58] <stevel> oh well
[02:18:04] <stevel> onnv-gate is restored
[02:18:09] <stevel> pull/clone away
[02:18:14] <gisburn> mhhh
[02:18:28] <gisburn> stevel: any idea how the offspring will look like ?
[02:18:43] <stevel> you might possibly grow a new head
[02:18:54] <gisburn> why ?
[02:19:54] *** bengtf has quit IRC
[02:20:09] <gisburn> stevel: you didn't chopp off the old one (yet... uhm...)
[02:20:10] * gisburn hides
[02:20:25] * gisburn imagines stevel with a large sword seeking gisburn ...
[02:21:26] *** Kernel86_ has quit IRC
[02:21:52] *** Somethingelse has joined #opensolaris
[02:22:00] *** Kernel86|Laptop has joined #OpenSolaris
[02:22:19] <twincest> stevel: yay
[02:22:38] <twincest> okay, because i'm an hg newbie... to do the equivalent to "cvs up", i have to hg pull && hg update?
[02:23:11] *** Somethingelse has quit IRC
[02:24:07] <gisburn> stevel: will opengrok get something like bonsai's "annotate" functionality in the future ?
[02:25:21] <delewis> twincest: yes
[02:25:28] <delewis> hg pull -u also accomplishes that, too
[02:25:37] <twincest> ah
[02:26:03] <delewis> I was a bit confused about that, too, initially.
[02:26:08] <delewis> coming from the cvs and svn world
[02:26:37] *** Somethingelse has joined #opensolaris
[02:28:17] <stevel> fuck
[02:29:17] <twincest> oh dear, should i stop my pull? :)
[02:29:32] <stevel> no, that repo is fine
[02:29:36] <stevel> my child opengrok repo grew a head
[02:29:39] * stevel sighs
[02:32:13] <richlowe> gisburn: I'd kinda like that too, so RFE it. :)
[02:32:29] <gisburn> richlowe: what ?
[02:32:34] <richlowe> blame in opengrok.
[02:32:37] <richlowe> sorry, annotate.
[02:32:40] <richlowe> whatever term you want to use :)
[02:33:17] <gisburn> richlowe: well, I would opt for a paralle installation of bonsai that BOTH opengrok and bonsai can show it's strengths.
[02:33:27] <twincest> oh god, not this agian
[02:33:30] <gisburn> s/paralle/parallel/
[02:33:41] <gisburn> twincest: why ? it would save lots of time
[02:33:57] <twincest> set up your own bonsai if you want to show how good it is
[02:34:03] <gisburn> twincest: e.g. there would be no need to develop something, just install it.
[02:34:17] <gisburn> twincest: I can, but only internally which is useless for outside people.
[02:35:10] <richlowe> Yeah, let's make more work for the tonic folk.
[02:35:13] <boyd> It certainly would be kinda handy to be able to prune the changeset tree to remove bad heads
[02:35:14] <richlowe> they don't have *nearly* enough to do.
[02:35:40] <twincest> richlowe> i'd be happy to do it if they paid me ;-)
[02:35:41] * boyd starts his pull
[02:35:54] *** stevel has quit IRC
[02:36:13] <richlowe> ... I sure hope the sarcasm there was obvious.
[02:36:23] <gisburn> umpf
[02:36:29] <gisburn> richlowe: thanks you... ;-(
[02:36:49] <richlowe> there's a thousand worthwhile things that need to be done.
[02:36:55] <richlowe> why waste effort duplicating stuff?
[02:37:01] <twincest> #1) rewrite Hg in ksh93
[02:37:02] <gisburn> richlowe: you're talking with a person (myself) who detects sarcasm only if you prinnt it out and beat me with that paper
[02:37:05] <boyd> Haha
[02:37:36] <boyd> #2) rewite the bug interface in ksh93. It would *have* to be an improvement :)
[02:37:42] <richlowe> twincest: I think that's surprisingly possible.
[02:37:50] <richlowe> not necessarly *sensible*, but I bet you could do it...
[02:38:00] <boyd> I'm sure you could
[02:38:01] <richlowe> boyd: you don't want to go there.
[02:38:12] <gisburn> boyd: 2) may be a nice idea... ksh93 can act as webserver
[02:38:33] * boyd echo's richlowe's last comment :)(
[02:38:50] <twincest> bash can act as a webserver too, why don't we just ln -s bash /usr/bin/ksh93  ?
[02:39:00] <boyd> (I don't know what that smiley was)
[02:39:09] <gisburn> twincest: bash lacks some features
[02:39:12] * boyd dicks and covers
[02:39:17] <boyd> err ducks, I meant ducks
[02:39:28] <clee> hahahaha
[02:39:30] <twincest> gisburn: no-one uses those features anyway
[02:39:33] * clee submits to bash.
[02:39:48] <gisburn> twincest: e.g. no socket/service support, floating-point support, associative arrays etc.
[02:40:08] <boyd> Does anyone know anything about this flag day: http://www.opensolaris.org/os/community/on/flag-days/pages/2006110701/ ?
[02:40:47] * boyd has never missed socket support in a shell.
[02:40:49] <gisburn> twincest: http://svn.genunix.org/repos/on/branches/ksh93/gisburn/scripts/rssread.ksh has TCP support to generate plain text output for RSS
[02:40:52] <gisburn> s/for/from/
[02:41:24] <twincest> boyd: i can't read that page, opera formats it badly (the long lines wrap off the edge without a scrollbar)
[02:41:41] <boyd> Firefox too
[02:41:47] <boyd> but I have a wider screen I think :)
[02:42:03] <twincest> i only have one monitor :P
[02:43:01] <boyd> I was wondering if the Solaris_11/Tools/Installers/liveupgrade option was still going to work
[02:43:12] <gisburn> no
[02:43:25] <gisburn> at least not from B48 to B51
[02:43:44] <gisburn> liveupgrade is currently broken
[02:43:45] <boyd> So you mean it hasn't been updated to deal with it's own build?
[02:43:52] <twincest> hmm, i wonder if i should submit my scary patch for 6302757
[02:43:54] <richlowe> no it isn't.
[02:44:02] <twincest> the only way i found to do it was to link dc with bc
[02:44:04] <richlowe> actually, scratch that, I guess it might be to 51
[02:44:07] <richlowe> but it sure worked to 50.
[02:44:10] <boyd> I didn't get the impression that it was broken
[02:44:23] <gisburn> http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6487393
[02:44:27] <richlowe> it worked 50->51 too, though I haven't rebooted yet.
[02:44:27] <boyd> .. not from the flag-day anyway
[02:44:40] <boyd> (which applies to 52)
[02:44:51] <richlowe> gisburn: oh, jeeze.
[02:44:54] <richlowe> gisburn: how the hell did I make it past that?
[02:45:08] <twincest> is there a bc test suite in ON? :)
[02:45:16] *** jmcp has joined #opensolaris
[02:45:16] <clee> somebody should add 'pre { white-space: pre-wrap; }' to the CSS file for opensolaris.org
[02:45:32] <twincest> clee: i prefer overflow:scroll; or whatever it is
[02:45:46] <boyd> Yes, that one please
[02:45:55] <boyd> (any website folk here?)
[02:45:56] <richlowe> clee: i talked with someone or other about similar stuff a while ago.
[02:46:02] <richlowe> (regarding the heads-up bits)
[02:46:11] <twincest> boyd: don't count on it, i filed a trivial bug against opensolaris:website last year and it still hasn't been fixed
[02:46:24] * boyd sighs again
[02:46:27] <clee> twincest: 'overflow: scroll' sucks. horizontal scrollbars = lose
[02:46:35] <twincest> clee: not know where the lines ends = lose
[02:46:47] <clee> if you copy & paste from firefox, it preserves the proper line endings
[02:46:52] <clee> at least, it does here
[02:46:53] <boyd> twincest: +1
[02:46:55] <richlowe> twincest: it's still worth filing them, however.
[02:46:56] <clee> I don't know about your crazy moon browsers
[02:46:57] <twincest> i don't want to copy+paste just to see the text
[02:47:32] <twincest> oh, it wasn't last year, it was april
[02:47:35] <twincest> but still
[02:47:36] <boyd> It would be nice if FF could mark the line-wraps that it added as well as that copy/past trick
[02:47:52] *** sahafeez has quit IRC
[02:47:55] *** OnkelSchorsch has joined #opensolaris
[02:48:05] <boyd> twincest: That's last financial year (at least for us here in Aus)
[02:48:19] <twincest> boyd: oh, you fancy people with gainfl employment :P
[02:48:28] <twincest> added 1697 changesets with 638 changes to 13789 files (+1 heads)
[02:48:29] <twincest> not updating, since new heads added
[02:48:29] <twincest> (run 'hg heads' to see heads, 'hg merge' to merge)
[02:48:31] <twincest> ??!?!?
[02:48:45] <boyd> Aaahh...
[02:48:47] <richlowe> you didn't strip the repo back to the position it was trimmed in.
[02:48:49] * boyd stops the pull
[02:48:58] <twincest> he never said i had to do that!
[02:49:03] <richlowe> Yes, he did.
[02:49:04] <twincest> can i fix this?
[02:49:04] * boyd doesn' stop the pull
[02:49:06] <richlowe> read osol-code.
[02:49:09] <richlowe> twincest: Yes.
[02:49:11] <twincest> i don't read osol-code
[02:49:18] <twincest> i read onnv-notify :P
[02:49:28] <boyd> twincest: It was in -notify
[02:49:30] <richlowe> twincest: you have two options, one is a clone -r of that repo specifying the correct head.
[02:49:41] <richlowe> (which should ignore the branch ending on the old head)
[02:49:51] <richlowe> the other is to strip the old branch out of the existing workspace.
[02:49:52] <twincest> boyd: i thought that didn't apply now it had been fixed
[02:50:05] <richlowe> No, the applied *because* it was being fixed :)
[02:50:08] <boyd> No, that was to get ready for the fixed repo
[02:50:09] <twincest> rich: do either of those work if i have local changes?
[02:50:18] *** delewis has quit IRC
[02:50:21] <richlowe> that's slightly more painful.
[02:50:30] <twincest> maybe i should diff and rm -rf
[02:50:31] <richlowe> you'll need to move your local changes over onto the new head, either way.
[02:50:45] <richlowe> I would either export or diff for your local changes.
[02:50:50] <richlowe> then kill off the branch ending on the old head.
[02:51:00] <richlowe> (the new head is 3231f91da5c8)
[02:51:35] <richlowe> 'hg clone -r <that> onnv-clone onnv-fixed-clone' then apply your changes back to onnv-fixed-clone
[02:51:43] <richlowe> I think that will ignore the branch, anyway. :)
[02:51:56] <boyd> You can do some nice stuff for moving local changes with the "mq" extension
[02:52:14] <richlowe> boyd: I figured twincest wasn't using it.
[02:52:19] <richlowe> boyd: though mq makes merging a pain in the ass.
[02:52:19] * twincest waits for diff
[02:52:22] <boyd> indeed
[02:52:28] * richlowe is in the process of doing the migration via mq
[02:52:33] <richlowe> certainly more joyful than the alternatives though.
[02:53:44] <twincest> is 'onnv-clone' in that command my local directory holding the gate?
[02:54:00] <richlowe> the one with two heads, yes.
[02:54:04] <twincest> ok
[02:54:16] * boyd goes for lunch
[02:54:51] * twincest isn't liking hg :P
[02:57:25] <twincest> i hope doing this at the same time as a fedora upgrade isn't going to run me entirely out of disk space..
[03:02:24] <twincest> so i desperately want to rewrite cmd/acct/ to not be the most disgusting source code ever seen
[03:02:36] <twincest> is that likely to be accepted, or is that legacy code that should be left to rot now?
[03:02:59] *** Cyl has quit IRC
[03:04:45] <richlowe> if nothing else, this will teach me to keep less workspaces laying around.
[03:04:52] <richlowe> tis may well take me considerably into tomorrow to fix up
[03:06:08] <boyd> At least you can pull all but the first locally
[03:06:19] * boyd unlunches
[03:06:27] <richlowe> well yeah, that's why I keep a clone.
[03:06:42] <richlowe> once I actually have all the .hg/patches bits stored off safe, I'm going to recreate all the others.
[03:06:47] <richlowe> it's that bit that's time consuming.
[03:06:57] * boyd nods
[03:07:22] <boyd> Yay! Only one head!
[03:07:29] <richlowe> at least it makes the zbringover hack seem worthwhile :)
[03:07:38] <boyd> indeed
[03:07:41] <richlowe> once I get done shoving .hg/patches directories off to one side, the rest should fly :)
[03:11:22] *** Somethingelse has quit IRC
[03:11:33] *** Somethingelse has joined #opensolaris
[03:13:09] *** Somethingelse has quit IRC
[03:14:09] *** Somethingelse has joined #opensolaris
[03:21:04] *** B|nTaRa has joined #opensolaris
[03:22:25] *** bobbyz has quit IRC
[03:25:53] *** jeff_ has joined #opensolaris
[03:28:17] <jeff_> test
[03:33:03] <richlowe> twincest: for what it's worth, it does seem I may have missed some gunk in src/cmd/man
[03:35:53] <twincest> great, back to a single head
[03:37:21] *** OnkelSchorsch has quit IRC
[03:41:26] *** halton has joined #opensolaris
[03:41:45] *** jeff_ has quit IRC
[03:42:44] *** rachel has quit IRC
[03:43:44] *** rachel has joined #opensolaris
[03:47:23] *** simford has quit IRC
[03:47:38] <twincest> goddamn i hate our ticket system
[03:47:53] <twincest> 300 spam mails, which i have to close manually, because the 'mass move' feature requires Javascript
[03:49:49] *** klocze3 has joined #opensolaris
[03:50:28] *** LeftWing_ has quit IRC
[03:50:33] *** LeftWing has joined #OpenSolaris
[03:51:12] <boyd> twincest: you could always enable javascript.... oh wait.. you're in links aren't you?
[03:51:37] <twincest> boyd: Opera 9.02
[03:51:52] <boyd> Doesn't it have javascript?
[03:51:55] <twincest> i refuse to enable Javascript until someone demonstrates an implementation that isn't riddled with security holes
[03:53:05] <jamesd> twincest, what good is security if no one is willing to store anything of yours, so there is nothing to worry about.
[03:53:16] <jbk> our ticketing system runs on java, but only on windows :)
[03:53:23] <jbk> at least the client
[03:53:58] <boyd> Sheesh
[03:54:16] <twincest> jamesd: i don't understand what that means
[03:54:35] <boyd> twincest: So make a zone just to run your javascript enabled browser that you point only at the site that needs it
[03:54:37] <jamesd> twincest, most  data entry  web apps require javascript.
[03:54:56] <jamesd> twincest, with more added every day.
[03:55:44] <twincest> well, some do, some don't
[03:55:55] <twincest> most of those that require it are private sites
[03:56:02] <twincest> our ticket system accepts public mail from anyone
[03:56:16] *** halton has quit IRC
[03:57:07] <boyd> You could even make a special label under TX that can only talk to the ticket system
[03:58:31] <twincest> that wouldn't help, a javascript XSS vulnerability could still talk to the ticket syste
[04:00:52] *** slowhog has joined #opensolaris
[04:05:40] *** delewis has joined #opensolaris
[04:08:32] *** OnkelSchorsch has joined #opensolaris
[04:10:14] <edp> are zones with autoboot=true supposed to autoboot when their root is on a ZFS filesystem?
[04:11:09] <alanc> X build 52 sources posted now
[04:11:12] <jbk> i would think it should mean they should start upon system boot
[04:11:19] <jbk> regardless of what filesystem they are using
[04:11:27] <alanc> http://opensolaris.org/os/community/x_win/changelogs/changelogs-nv_50/ has the changelog and links to tarballs
[04:11:28] <edp> that's what i thought, but it doesn't seem to work
[04:11:40] <alanc> (This source drop is brought to you by the letters "D", "R", and "I".)
[04:11:44] <edp> in fact they don't even show up in `zoneadm list -v`
[04:14:28] *** alanc is now known as alanc-away
[04:14:49] <Error_404> dri?
[04:15:02] <jbk> for a few chipsets at least (from what I saw)
[04:15:11] <Error_404> intel 915?
[04:15:52] <alanc-away> yes
[04:15:57] <Error_404> w00t
[04:16:18] <Error_404> the only machine i have an actual graphics card hooked up to is intel
[04:16:30] <alanc-away> http://www.opensolaris.org/os/community/on/flag-days/pages/2006102901/ has details on what's supported and how to get it
[04:17:37] <Error_404> xserver.sfbay is... not gonna work
[04:17:46] *** deedaw has quit IRC
[04:18:45] <gisburn> dduvall: ping!
[04:20:20] <jbk> now if there was a way to get 64-bit broadcom wireless drivers, I'd be set :)
[04:21:36] <richlowe> Error_404: wait out snv_53 then, I guess.
[04:34:40] <Error_404> speaking of wireless... this wpa_supplicant package seems not work
[04:39:33] *** jafari has quit IRC
[04:41:53] *** delewis has quit IRC
[04:48:01] *** johnlev has quit IRC
[04:48:08] *** GmanAFK is now known as Gman
[04:50:20] <hile_> wub Gman
[04:50:23] <Gman> hey hile
[04:55:18] <twincest> only 228 spam tickets left!
[04:55:38] <gisburn> twincest: which list ?
[04:55:47] <twincest> noc at wikimedia dot org
[04:55:52] <gisburn> bah
[04:56:49] <sommerfeld> fascinating.  when you "zpool replace" an element of a raidz group, zfs does a parity-based reconstruction of the disk being replaced and writes the reconstructed data to *both* old and new disk.
[04:57:58] <richlowe> intentionally?
[04:58:29] <sommerfeld> that i'm not sure about.
[04:58:44] * gisburn waits for the ZFS feature to checksum each file and send the ids to RIIA for evaluation
[04:58:47] <sommerfeld> but from looking at zpool iostat, that's clearly what happens
[04:59:00] <richlowe> how odd.
[04:59:10] <gisburn> sommerfeld: read or write ?
[04:59:11] <richlowe> a hedge against interruption, maybe?
[04:59:25] <gisburn> sommerfeld: maybe it does a compare against the original version ?
[04:59:26] <sommerfeld> it's writing to both
[04:59:27] * richlowe is out of other even tenuously reasonable guesses
[04:59:34] <sommerfeld> it's not reading from the outgoing drive at all
[04:59:53] <richlowe> if you're doing a replace, it's unreasonable to expect the original to necessarily be workable.
[04:59:54] <sommerfeld> i just sent a note to zfs-discuss
[04:59:56] <richlowe> isn't it?
[05:00:50] <sommerfeld> well, in this case the drive is definitely mostly workable.  just generates hard read errors on a weekly basis.
[05:02:36] <sommerfeld> (and this failure is consistent with Elling's idea that most disk failures are not total failures..)
[05:05:02] <Error_404> I think it would be great if you could throw 2 drives at a single drive zpool & it'd turn it in to raidz
[05:05:12] <Error_404> or hell, throw 1 drive at a mirror & turn that thing in to a raidz
[05:05:21] <Error_404> though the mechanics of that one escapes me
[05:06:27] <sommerfeld> reshaping raid z is .. hard
[05:07:06] <hile_> Error: you mean like VxVM online relayout?
[05:07:10] <hile_> i *heart* that feature
[05:07:21] <jbk> it's nice, if you have a few months to spare :)
[05:08:07] <jbk> we've found it's usually faster to just attach a new mirror laid out in the manner you want than to do online relayout
[05:08:30] <jbk> as it seems to (counter intuitively) require more temp space than the size of the volume
[05:14:59] *** Lippman has joined #opensolaris
[05:17:43] *** Somethingelse has quit IRC
[05:20:01] <Gman> !seen brendang
[05:20:02] <Drone> brendang (brendang!n=brendang at m208-55 dot dsl.rawbw.com) was last seen in #opensolaris on Tue 17 Oct 2006 02:42 GMT, saying 'gotta run for muni. I'll try to log in more often again :)'.
[05:20:14] <Gman> hrm, what's bgregg's nick again?
[05:20:19] <richlowe> brendang.
[05:20:30] <Gman> huh, that's what i thought
[05:20:43] <richlowe> he hasn't been around for quite a while, at this point.
[05:20:51] <Gman> yeah :(
[05:20:57] *** nwf has quit IRC
[05:21:06] <twincest> PPP 1: Recent failure      : Our CHAP login failed - 'This account has been suspended. '
[05:21:14] <twincest> having to pay for internet access sucks :(
[05:22:38] <richlowe> Gman: no css?
[05:22:50] <Gman> richlowe, which what where?
[05:24:35] *** nwf has joined #opensolaris
[05:31:10] <richlowe> Hm, I suspect I just hosed this completely.
[05:55:03] * twincest orders 350G external HD
[05:55:08] <twincest> hopefully that's enough to store ON
[05:56:18] <dlg> no driver developers here?
[05:57:50] <jmcp> dlg: maybe there are, maybe there aren't
[05:57:55] <jmcp> depends on what sort of driver you're talking about
[05:57:56] *** stevel has joined #opensolaris
[05:57:57] *** ChanServ sets mode: +o stevel
[05:58:04] <dlg> scsi hba on a pci bus
[05:58:11] <richlowe> stevel: how'd it go?
[05:58:18] <jmcp> dlg: I might be able to help
[05:58:23] <dlg> ooh
[05:58:30] <jmcp> no promises!
[05:58:32] <stevel> richlowe: not bad
[05:58:37] <jmcp> dlg: what questions do you have?
[05:58:38] <stevel> i was giving a demo of building the code
[05:58:45] <jmcp> stevel: is avalon there with you?
[05:58:48] <stevel> and managed to screwup a printf statement
[05:58:52] * stevel is a wicked cool coder
[05:58:54] <richlowe> stevel: 5 minutes work, and several hours of waiting?
[05:58:55] <stevel> jmcp: yup
[05:59:01] <dlg> jmcp: you prefer here or /query?
[05:59:12] <jmcp> stevel: please say hi from me
[05:59:15] <stevel> richlowe: basically :) well, 50 minute preso - and several hours of waiting
[05:59:19] *** dakrone has joined #opensolaris
[05:59:20] <jmcp> dlg: is it a long question? :)
[05:59:28] <dlg> no
[05:59:31] <dlg> its just clues
[05:59:35] <stevel> jmcp: will do. dunno where he is right now, musta ducked out real quick
[05:59:48] <jmcp> dlg: fire away
[05:59:52] <stevel> richlowe: have you pulled from the new repo yet?  does it work?
[05:59:53] <jmcp> stevel: ta much
[06:00:02] <richlowe> stevel: myself, boyd and twincest all have.
[06:00:05] <dlg> im used to locking using interrupt masking
[06:00:10] <stevel> richlowe: and it worked?
[06:00:15] <richlowe> I think twincest botched the migration, but it worked for me at least.
[06:00:27] <stevel> good
[06:00:29] <dlg> i need some suggestions on how to approach locking
[06:00:32] <richlowe> beyond the pain of dancing all my workspaces around, and dealing with a merge that snuck up on me.
[06:00:33] <twincest> the instructions were incomplete :(
[06:00:38] <jmcp> dlg: carefully, for starters
[06:00:41] <richlowe> twincest: mine were not.
[06:00:46] <jmcp> dlg: are you writing a new hba driver?
[06:00:46] <richlowe> twincest: mine were *exactly* what I just did :)
[06:00:50] <stevel> twincest: how so?
[06:01:02] <dlg> jmcp: yes.
[06:01:04] <richlowe> and since I just did it 16 times, they better have worked ;)
[06:01:09] <jmcp> dlg: which chip?
[06:01:10] <twincest> stevel: i got the impression that a fixed onnv-gate would remove the need to do anything
[06:01:17] <dlg> megaraid sas from lsi logic
[06:01:18] <twincest> stevel: so i cloned it and ended up with two heads
[06:01:25] <richlowe> twincest: the clone shouldn't have given you two heads.
[06:01:27] <jmcp> dlg: oh, right
[06:01:29] <richlowe> a *pull* would have.
[06:01:32] <stevel> you cloned it from scratch?  that shouldn't have given you two heads
[06:01:35] <richlowe> a clone also didn't give me two heads.
[06:01:36] <twincest> eh, pull, whatever
[06:01:45] <twincest> stevel: pull, not clone
[06:01:45] <stevel> if you pulled onto an existing repo, it's possible
[06:01:46] <richlowe> oh, wait, I didn't actually clone.
[06:02:12] <twincest> i had a repo pulled from 'arp nits'
[06:02:19] <twincest> i pulled it again and had two heads
[06:02:34] <dlg> jmcp: i already figured the "very carefully" bit out
[06:02:39] <jmcp> dlg: :)
[06:02:41] <stevel> hg log -l 1 showed only the 'arp nits' at the top?
[06:02:52] <twincest> stevel: yes
[06:02:59] <richlowe> twincest: did what I said fix that for you?
[06:03:03] *** killing-joke has joined #opensolaris
[06:03:14] <richlowe> twincest: if not, I can maybe come up with better, I was just aiming for 'simple'
[06:03:17] <twincest> richlowe: yes
[06:03:32] <stevel> simple is: rm -rf onnv-gate; hg clone ssh://anon at hg dot opensolaris.org/hg/onnv/onnv-gate
[06:03:35] <stevel> :)
[06:03:50] <twincest> rm -rf was no good, i had local changes
[06:04:01] <stevel> ah
[06:04:04] <jmcp> dlg: so what part of the code are you looking at the locking for?
[06:04:05] <twincest> i did hg diff, hg clone, gpatch
[06:04:11] <dlg> maybe i should leave interrupts till later
[06:04:30] *** OnkelSchorsch_ has joined #opensolaris
[06:04:31] <dlg> jmcp: the interrupt handler
[06:04:44] <dlg> the hilevel, msi or whatever stuff is confusing to start with
[06:04:47] <dlg> so im copying amr, which uses a hilevel interrupt
[06:04:58] <dlg> whihc presumably wakes up something to do the actual work
[06:05:29] <killing-joke> i have a blank hard drive, a new motherboard, and a T1. where can i find some bootstrap code?
[06:05:42] <killing-joke> i would like to install opensolaris x86.
[06:05:43] <jmcp> dlg: you want to lock as finely as possible for as little as possible, and except for special cases, yes, you want to hand off to a worker taskq
[06:05:53] <twincest> killing-joke: T1 is a SPARC CPU.  it can't run Solaris x86
[06:06:02] <jmcp> note, *not* a thread. taskqs are the preferred way of doing things like this
[06:06:16] <jbk> or does he mean T1 the leased line
[06:06:17] <dlg> jmcp: whats a taskq?
[06:06:29] <killing-joke> actually, the T1 is my network bandwidth. i want to install x86 on a spare dual-opteron here at work.
[06:07:12] <twincest> killing-joke: then download Solaris 10 (www.sun.com/software/solaris/get.jsp) and install it, assuming you want a production-quality OS
[06:07:54] <killing-joke> it's for development. the engineers i am working with have specified opensolaris > 46.
[06:08:03] <jmcp> dlg: a more lightweight form of an lwp, iirc
[06:08:07] <twincest> killing-joke: then use http://www.opensolaris.org/os/downloads/on/ to install solaris express
[06:08:33] <dlg> jmcp: maybe i'll print the amr source and have a coffee
[06:08:57] <twincest> killing-joke: you probably want 3b (on the right), unless you want to do ON development
[06:09:01] <dlg> at least solaris has a sane coding style
[06:09:02] <killing-joke> twincest: thanks for the url. looks like it wants a registration id.
[06:09:05] <jmcp> dlg: good idea. figure out exactly what you must, unequivocally, do in interrupt context, and outsource the rest
[06:09:10] <dakrone> is an NFS server included out of the box in Opensolaris? I don't remember
[06:09:11] <twincest> yes, you have to register to download SX
[06:09:21] <jmcp> dakrone: yes
[06:09:28] <dakrone> jmcp: okay, thanks
[06:09:43] <dlg> jmcp: i have a fair idea of how the hardware works
[06:09:47] <dlg> it uses the same iop as the areaca
[06:09:49] <dlg> areca
[06:09:55] <jmcp> s/c// ?
[06:10:12] <dlg> ?
[06:10:25] <killing-joke> hmmm. i'm puzzled. does this mean that there is no downloadable install ISO for opensolaris?
[06:10:38] <jbk> there is
[06:10:43] <killing-joke> ... that does not require creating an account?
[06:10:58] <jbk> oh, i think you may be stuck there
[06:11:13] <twincest> 'opensolaris' is not an installable OS
[06:11:20] <twincest> it's a collection of source code
[06:11:49] <Error_404> like linux vs gnu/linux
[06:11:51] <stevel> killing-joke: nexenta, belenix, and schillix can all be installed
[06:11:52] <Error_404> heh
[06:12:21] <killing-joke> ah HAH!
[06:12:31] *** alanc_away is now known as alanc
[06:12:54] <killing-joke> thanks very much for the pointers. i do appreciate your help!
[06:12:58] <Triskelios> it's solaris express that requires a sun account
[06:13:13] <alanc> and it's not like "mickey mouse" and "clark kent" haven't downloaded hundreds of copies of Solaris - it's not like the registration asks for social security number and two forms of id
[06:13:28] <Error_404> that's what the liquor store does
[06:13:38] *** OnkelSchorsch has quit IRC
[06:13:55] <Triskelios> I still say that's a big deterrent...
[06:14:21] <richlowe> it's annoying, but unavoidable.
[06:14:33] <alanc> Error_404: I know xserver.sfbay won't work - if you want DRI, you can either build the source drop I released today or wait for nv_52 ISO's to come out in about 2 weeks
[06:14:49] <Error_404> my machine won't build ON
[06:14:50] <gisburn> alanc: I think Osamba BunkLaden has an account, too.
[06:14:52] <boyd> 6.0e+6 downloads seems to imply it's not too much of a deterrent
[06:14:55] <Error_404> it's broken
[06:15:08] <richlowe> boyd: well, it doesn't email you every time at least.
[06:15:13] <boyd> haha!
[06:15:14] * richlowe curses devpro
[06:15:16] <killing-joke> gisburn: as long as Osama can respond to the activation email, he's golden.
[06:15:35] <alanc> ON side is in build 51 binaries that are already available - it's only the X side that isn't in until build 52
[06:15:36] *** bobbyz has joined #opensolaris
[06:15:55] <alanc> and building X requires far less setup and machine mangling than ON does
[06:16:14] <gisburn> killing-joke: *shrug* ... not my account... :-)
[06:16:23] *** alanc is now known as alanc_away
[06:16:37] * alanc_away goes to eat his pizza before it gets cold
[06:16:54] <boyd> Yeah, do that
[06:18:43] <killing-joke> "SchilliX-0.5.2.iso.bz2  317289 KB  4/7/2006"  ...  310MB. i will be up and running in an hour.
[06:18:47] <killing-joke> thanks again!
[06:20:10] <Error_404> i wonder what counts as redistribution
[06:20:26] <Error_404> eg, can i take my sx:cr cd to a friend & install it on his computer?
[06:20:32] <Error_404> can i take it to work?
[06:21:06] <boyd> Technically I'd say no... but IANAL
[06:21:23] <jbk> i have a feeling they're not going to come knocking at your door verifying compliance
[06:21:24] <Error_404> can i make a vmware disk image & redistribute that?
[06:21:35] <jbk> but i don't work for sun..
[06:22:16] <boyd> jbk: I agree... an I don't either
[06:22:17] <killing-joke> if the player launchs something that is likely to advance the cause of Solaris, i would imagine the response would be very gentle.
[06:22:34] <killing-joke> however, if it was a super multithreaded pr0n downloader, ...
[06:23:48] <boyd> hehe
[06:23:49] <jbk> i suspect a lot of it is just to cover sun's butt if say someone hits a bug that causes an outage, etc.
[06:23:56] <richlowe> that wouldn't advance the cause?
[06:23:58] <richlowe> drat!
[06:24:19] <gisburn> decade = 10 years, right ?
[06:25:12] <killing-joke> gisburn: yes, deci-  ==  a "ten" prefix.
[06:25:28] <Error_404> way to thinKMetric
[06:25:39] <Error_404> as the signs @ the border leading in put it
[06:26:17] <killing-joke> here in silicon valley, the gas gauges in the giant SUVs display gallons per mile.
[06:27:10] <Error_404> what a kick in the groin that must be
[06:27:16] <jbk> are most of the digits displayed to the left of the decimal point? :)
[06:27:20] <jbk> err right
[06:27:25] * jbk just screwed that joke up..
[06:27:26] <Error_404> $/mile would be better
[06:27:41] <dwc-> what's sorta funny killing-joke is that in a lot of places, kilometers per litre is the standard measurement
[06:27:44] <Error_404> $2/mile... hooray
[06:27:57] <dwc-> rather than liters per kilometer
[06:28:09] <killing-joke> it should go up (in $units) like a cab fare.
[06:28:27] <killing-joke> push a big button at the end of the month to reset.
[06:29:40] <Tpenta> stevel: have you changed anything in bindrop since the weekend?
[06:29:43] *** deather_ has joined #opensolaris
[06:30:49] *** alobbs has joined #opensolaris
[06:31:26] *** kaiwai has joined #opensolaris
[06:31:29] <kaiwai> hi :)
[06:31:40] <kaiwai> one know the status of opensolaris and Intel 3945abg support?
[06:33:27] <boyd> dwc-: It's liters/100km here
[06:33:47] <kaiwai> hmm, don't you mean litres? :P
[06:34:01] <boyd> yes
[06:34:27] <boyd> But I couldn't be bothered fixing it
[06:35:16] *** Xh4 has joined #opensolaris
[06:35:39] * OnkelSchorsch_ litters the channel with some euro signs ?????? :D
[06:35:40] <sommerfeld> the toyota hybrid firmware lets you pick either mode in its real time gas usage videogame
[06:36:27] <boyd> Does it let you pick tetris? Ya gotta have tetris :)
[06:38:08] <Gman> stevel, is there any easy way to do up database looks ups on id's on opensolaris.org?
[06:38:26] <Gman> nm, figured out the url
[06:38:27] <boyd> bug lookups?
[06:39:06] <killing-joke> hunting BunkLaden
[06:41:26] <Gman> boyd, no user id's
[06:41:32] <Gman> !seen rteer
[06:41:35] <Drone> I've never seen rteer talk in #opensolaris.
[06:41:40] <Gman> !seen teer
[06:41:41] <Drone> I've never seen teer talk in #opensolaris.
[06:41:44] <Gman> hrm :)
[06:41:45] <boyd> I think it's richteer
[06:41:48] <jmcp> !seen richteer
[06:41:50] <Drone> RichTeer (RichTeer!n=chatzill at S0106080020b2d64f dot ok.shawcable.net) was last seen in #opensolaris on Sun 18 Jun 2006 23:27 GMT, saying 'Coool, looks like Tron.'.
[06:42:03] <Gman> ahh, not what i thought ;)
[06:42:04] * boyd wonders ehat looks like tron
[06:42:05] <kaiwai> !seen kaiwai
[06:42:06] <Drone> kaiwai is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:44 GMT, saying '!seen kaiwai'.
[06:42:15] <kaiwai> hmm
[06:42:16] <killing-joke> cute :)
[06:42:32] * boyd wonders about typing seen Drone but he things the universe may implode
[06:42:38] <killing-joke> lol!
[06:42:45] <richlowe> boyd: it does exactly what you'd expect.
[06:42:50] <richlowe> and oh boy is it ever annoying.
[06:42:53] <killing-joke> it kills all of us?
[06:42:55] <boyd> hehe
[06:43:09] <boyd> Glad I said it rather than trying it...
[06:43:24] <boyd> does that mean you've tried richlowe
[06:43:25] <boyd> ?
[06:43:32] <richlowe> no, gisburn has done it.
[06:43:35] <Tpenta> !seen drone
[06:43:37] <Drone> Drone is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:44 GMT, saying 'kaiwai is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:44 GMT, saying '!seen kaiwai'.'.
[06:43:44] <Tpenta> as you would expect :-D
[06:43:49] <Tpenta> now it gets annoying
[06:43:50] * killing-joke has always wondered what the knowledge would look like that *should* be suppressed
[06:43:59] <Tpenta> !seen drone
[06:44:01] <Drone> Drone is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:45 GMT, saying 'Drone is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:44 GMT, saying 'kaiwai is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:44 GMT, saying '!seen kaiwai'.'.'.
[06:44:04] <Tpenta> !seen drone
[06:44:06] <Drone> Drone is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:46 GMT, saying 'Drone is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:45 GMT, saying 'Drone is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:44 GMT, saying 'kaiwai is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:44 GMT, saying '!seen kaiwai'.'.'.'.
[06:44:07] <Tpenta> !seen drone
[06:44:08] <gisburn> !seen drone
[06:44:09] <gisburn> !seen drone
[06:44:11] <Drone> Drone is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:46 GMT, saying 'Drone is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:46 GMT, saying 'Drone is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:45 GMT, saying 'Drone is currently online in #opensolaris and last spoke on Wed 08 Nov 2006 05:44 GMT, saying 'kaiwai is currently online in #opensolaris and last spoke on Wed 08 Nov 200
[06:44:11] <Tpenta> ad nausium
[06:44:12] <gisburn> !seen drone
[06:44:20] <jmcp> will you guys stop that please!!!!
[06:44:25] <gisburn> jmcp: no.
[06:44:27] <boyd> I was wondering if Drone would see the "!seen Drone" in it's output and respond
[06:44:28] <killing-joke> lololo!
[06:44:35] <gisburn> jmcp: we want to torture you.
[06:44:45] <richlowe> jmcp: the easiest solution is to ignore the bot. :)
[06:44:49] * jmcp considers going off in a huff
[06:44:50] <boyd> fortunately not
[06:44:53] <richlowe> jmcp: it doesn't do anything /whois and seenserv won't :)
[06:45:06] * LeftWing throws a spare huff to jmcp.
[06:45:11] * killing-joke wipes eyes
[06:45:14] * jmcp catches LeftWing's spare huff
[06:45:19] <killing-joke> omg you guys a f'in funny
[06:45:23] <jmcp> richlowe: stop putting logic in... spoils a good huff
[06:45:24] <gisburn> dumb question: when /bin is a softlink to /usr/bin and "cd /bin" - what does an external application see as pwd - "/bin" or "/usr/bin" ?
[06:45:25] <killing-joke> ^are
[06:45:30] <hile_> rich: i don't know which is more productive: to ignore gisburn or ignore the  bot
[06:45:31] <kaiwai> *prods jmcp*
[06:45:32] <richlowe> jmcp: you don't huff right.
[06:45:40] <jmcp> hmmmph!
[06:45:43] * jmcp sulks
[06:46:01] * gisburn smacks hile_
[06:46:14] *** Lippman has quit IRC
[06:46:22] * gisburn feeds some cookies to jmcp
[06:46:25] * jmcp thwaps gisburn for thwapping hile_
[06:46:32] <gisburn> good cookies
[06:46:37] *** deather has quit IRC
[06:46:39] <LeftWing> gisburn: I'm under the impression it gets "/bin" as the pwd...
[06:46:54] <gisburn> Homemade by my little niece and myself!
[06:46:57] <gisburn> LeftWing: yeah
[06:47:03] * LeftWing writes a few lines of C.
[06:47:10] <boyd> gisburn: I think it depends on your shell
[06:48:35] <kaiwai> sooo
[06:48:42] <kaiwai> whats the story with Opensolaris 3945abg support?
[06:48:57] <jmcp> kaiwai: probably non-existant right now. have you filed an RFE on b.o.o. for it?
[06:49:15] <kaiwai> na :( I assume that it would automatically be supported due to wpi being ported
[06:49:24] <kaiwai> wpi and 3945abg exist on openbsd
[06:49:28] <jmcp> don't assume
[06:49:33] <kaiwai> I know, I know
[06:49:38] <jmcp> besides, if there's no RFE then there's no need, right?
[06:49:41] *** uebayasi has quit IRC
[06:49:54] <Tpenta> 20061106 nd binaries are up
[06:49:55] <jmcp> kaiwai: remember, *managers* need to buy in... and they want to see measurable demand
[06:49:57] <kaiwai> not really, if there is no RFE, there are people like me who are too lazy :)
[06:50:14] <kaiwai> so we bugger off to Linux *shudder*
[06:50:19] <kaiwai> or worse.....
[06:51:30] <boyd> Hey jmcp Congrates
[06:51:38] <jmcp> boyd: thankyou :)
[06:51:40] <kaiwai> whats happened to jmcp?
[06:51:43] <kaiwai> won lotto?
[06:51:48] <jmcp> not exactly
[06:51:51] <jmcp> see my blog
[06:51:54] * jmcp self-promotes
[06:52:07] <jmcp> kaiwai: http://www.jmcp.homeunix.com/roller/page/jmcp?entry=gainful_employment_part_2
[06:52:22] <Tpenta> ahhh you announced it
[06:52:29] <boyd> jmcp: How's your mandarin?
[06:52:34] <jmcp> boyd: pitiful
[06:52:37] * Tpenta had lunch with jmcp today
[06:52:57] <richlowe> Tpenta: you poor soul, you.
[06:53:02] <boyd> Hehe
[06:53:11] <Tpenta> james an i have worked together for years
[06:53:18] <richlowe> Tpenta: exactly :)
[06:53:25] <kaiwai> trying to find it :(
[06:53:32] <kaiwai> thank you :)
[06:53:45] <boyd> Hey, any more SOSUG's planned?
[06:54:10] <jmcp> richlowe: oi! I resemble that :)
[06:54:11] <Tpenta> i have been so flat out that I have been slack in that regard, I need to do one before xmas
[06:54:15] <jmcp> boyd: probably
[06:54:38] <kaiwai> *gives jmcp a hug and kiss* congratulations :)
[06:54:44] *** uebayasi has joined #opensolaris
[06:54:49] <Gman> boyd, organize one! ;)
[06:54:55] <boyd> Well, I'm gonna be in Syd on the week of 28th Nov (probably). Thought we may be able to catch up... jmcp won't be there I think
[06:55:16] <jmcp> kaiwai: thanks mate
[06:55:23] <jmcp> boyd: I'm back on the 2nd of Dec
[06:55:32] <boyd> Gman: I should organise one in Melbourne
[06:55:33] <kaiwai> well, I'm happy, next year, university, then become a secondary school teacher
[06:55:39] <kaiwai> bring on the pain <evil smile>
[06:55:46] <boyd> kaiwai: and then no more happy :)
[06:56:06] <kaiwai> well, I'm in a crappy job now; working in a supermarket, in a freezer all day
[06:56:31] <boyd> ... and of course, brendan's disappeared from the face of the earth, like the rest of the DTrace team
[06:56:42] <richlowe> yup
[06:56:50] *** |tsoome| has quit IRC
[06:57:02] <killing-joke> "connection to www.jmcp.homeunix.com (59.167.244.18) could not be established."
[06:57:03] <boyd> Suspicious, that
[06:57:06] <killing-joke> slashdotted?
[06:57:07] <jmcp> boyd: that's cos they're in the Lab ivory tower
[06:57:12] <jmcp> killing-joke: bugger
[06:57:21] <jmcp> anybody else getting that problem?
[06:57:26] <boyd> jmcp: slashdotted? wow!
[06:57:33] <Gman> mike was a little more active on the smf discuss alias this week
[06:57:34] <jmcp> boyd: I kinda doubt it
[06:57:39] <LeftWing> I can get to it.
[06:57:42] <boyd> No, mines working
[06:57:52] <killing-joke> alas, new york.
[06:58:06] <boyd> jmcp: But your front page is V slow to load the image
[06:58:11] <jmcp> hm
[06:58:24] <boyd> ... which is quite strange, I might add
[06:58:26] <jmcp> could be because last night I killed my swap allocation with a stupid script :|
[06:59:14] <boyd> Your hg repos link is broken too
[06:59:14] <kaiwai> hmm, apparently 3945abg is being tested internally
[06:59:42] <jmcp> kaiwai: that must hurt
[06:59:51] <kaiwai> hmm, if they don't release it, I'll internally test them ;) <evil laugh and smile>
[06:59:54] <jmcp> boyd: bugger. I think I might have forgotten to copy it into the zone
[07:01:08] <jmcp> boyd: I've removed the link for the time being
[07:01:19] * boyd nods
[07:01:20] <LeftWing> gisburn: http://pastebin.ca/243217
[07:01:47] <boyd> jmcp: That picture is quite uninsual
[07:01:50] <boyd> uniunusal
[07:01:54] <boyd> argh
[07:01:59] <boyd> strange
[07:02:12] <LeftWing> Strange?
[07:02:17] <jmcp> yeah, that too
[07:02:31] <jmcp> boyd: it's courtesy of the gimp with the oil painting mod
[07:02:36] <boyd> :)
[07:02:48] <gisburn> LeftWing: yeah, I just used |int main(int ac, char **av) { char buf[PATH_MAX+1] ; puts(getwd(buf)); return 0 ; }
[07:03:17] *** lin has quit IRC
[07:03:26] <LeftWing> Hehe, puts().
[07:03:45] <killing-joke> front page of /. says  "Sun To Choose GPL For Open-Sourcing Java"
[07:03:55] <LeftWing> Oh really? :O
[07:03:57] <boyd> Really!
[07:04:07] <LeftWing> GPL. :(
[07:04:11] <gisburn> LeftWing: yes, old but pretty in this case.
[07:04:12] <Tpenta> interesting, and we still get the slashdot mob hammering us
[07:04:37] *** gisburn is now known as slashmob
[07:04:44] <Tpenta> the story is from a journo i dont recognise. i do not yet see a sun press release. The journo claims that ME and SE will be GPL, but EE will be CDDL
[07:04:47] * jmcp hammers gisburn^Wslashmob
[07:05:01] <slashmob> Tpenta: HAMMMMMER!! CDDL is ILLEGAL!! HAMMMMMER!!HAMMMMMER!!HAMMMMMER!!HAMMMMMER!!HAMMMMMER!! CDDL is sooo evil! HAMMMMMER!!HAMMMMMER!!
[07:05:13] *** slashmob is now known as gisburn
[07:05:23] <killing-joke> stfu slashmob
[07:05:28] <killing-joke> always wanted to say that.   :)
[07:05:33] <jbk> i always like the term slashole
[07:05:37] <Tpenta>   /kick slashmob :-D
[07:05:39] <gisburn> heh
[07:06:03] <gisburn> Tpenta: I can add him to the channel banlist if you want...
[07:06:08] <killing-joke> hah!
[07:06:16] <gisburn> yessssss
[07:06:23] <gisburn> maximum punishmenski!
[07:06:46] <LeftWing> "I mean, I know some people have a mad on against the GPL, but it ain't what you'd call restrictive. The only thing it does is mandate that all derivitve works also have to be GPLed."
[07:06:57] <LeftWing> Not restrictive?  Sheesh.
[07:08:20] <Tpenta> lots of "I'll beleive it when I see it", or "they should have done this 5 years ago"
[07:08:26] <LeftWing> Heh.
[07:08:41] <jmcp> Tpenta: the usual fuckwits
[07:08:58] <boyd> I can't believe you guys are reading those idiots
[07:09:01] * gisburn learns new words from jmcp
[07:09:32] <LeftWing> boyd: I can't help it.  I know there are people saying that Java is slow and that desktop Java is "over" on there.  I have to read.  I have to be irritated.  It's part of my psyccheeee.
[07:09:35] * gisburn canns the pwd fasttrack
[07:10:05] <killing-joke> "RIP Java  (score: 1)  Now that the dirty hippies have access to the code - it is theirs to destroy much like the numerous number of projects they already have."
[07:10:07] <boyd> LeftWing: Hehe... don't do it... just think of them all as 12yo kids and walk away
[07:10:10] *** alobbs has quit IRC
[07:10:16] <LeftWing> Must... beat... the children...
[07:10:17] <LeftWing> ;P
[07:10:22] <boyd> hah
[07:10:34] <dlg> killing-joke: the source code is already available
[07:10:49] <dlg> the current license restricts distribution of it
[07:10:53] <jbk> it has been for a while
[07:11:08] <killing-joke> dlg: i tried to build it for FreeBSD 5.3 recently, but ran into a few problems.
[07:11:09] * jbk remembers tracking down why i/o sucked in older jdks
[07:11:23] <dlg> killing-joke: is that freebsds fault or javas?
[07:11:28] <jbk> turns out, all file i/o was being done a byte a time :)
[07:11:36] <killing-joke> dlg: mine, i'm sure.   :)
[07:12:49] <killing-joke> while you can replace the pmake-style make on fbsd with gnu gmake, that doesn't mean that everything will work correctly when you kick it off.
[07:13:25] <killing-joke> i decided to wait for jschwartz to open up the code
[07:13:32] <killing-joke> i didn;t have to wait very long!
[07:14:01] <dlg> a new license doesnt fix the build system
[07:14:17] <dlg> its still going to suck, its just going to be easier to get
[07:14:26] <killing-joke> oooh! sigged.
[07:16:08] <boyd> jbk: I found the same thing with dtrace a few months back
[07:16:21] * Gman hopes it'll be gpl for the vm
[07:16:30] <Gman> but not sure about gpl for some of the libraries
[07:16:31] *** bougie has joined #opensolaris
[07:16:33] <jbk> this was prior to dtrace (would have been nice)
[07:16:58] <jbk> i was writing a utility to process showrev -p output and put it into a nice spreadsheet
[07:18:00] <jbk> ended up reimplementing my own buffered i/o classes ontop of the File class (I think, been a while) as it seemed to be a thin wrapper around the standard unix i/o operations, so you could get more than 1 byte i/os
[07:18:17] <jbk> i'm hoping that's been fixed in newer jvms, but haven't looked
[07:19:11] <LeftWing> I get some reasonable buffered I/O out of 1.5
[07:19:24] *** lin has joined #opensolaris
[07:20:02] <Error_404> why on earth would sun go gpl, considering they own the cddl?
[07:20:09] <jbk> i'd still love to know the motivation for the original design
[07:20:49] *** lin has left #opensolaris
[07:21:19] <kaiwai> hmm, democrats won the lower house, interesting if they win the upper
[07:21:21] <jbk> i mean i know premature optimization is the root of all evil, etc, but essentially doing (for i=0; i < buf.len; ++i) { write(fd, buf[i], 1); } in the jvm implementation just.. well..
[07:21:30] *** linma has joined #opensolaris
[07:25:29] *** |tsoome| has joined #opensolaris
[07:26:28] <Error_404> kaiwai: american foreign policy will stay the same, the only difference is CNN's map of congress will be >50% blue
[07:27:21] <boyd> jbk: I think it's better... when I last looked... also, you have nio now
[07:27:21] <kaiwai> hmm, true, true, and the simple folk, people of the land, you know, morons, will continue to dictate the direction of the nation
[07:27:34] <kaiwai> aka "homo bashers, church goers and xenophobes united"
[07:28:01] <jmcp> kaiwai: not all church goers are unbalanced rightwing idiots
[07:28:02] *** coolvibe has quit IRC
[07:28:12] * jmcp is not a churchgoer
[07:28:35] <jmcp> hey, Eliot Spitzer got the NY Governorship
[07:28:39] <jmcp> *interesting* !!
[07:28:48] <kaiwai> hmm, explain the anti-gay constitutional ammendment
[07:29:03] <jmcp> which particular one?
[07:29:25] <hile_> it wasn't last time i looked your problem since you're neither an American nor living in the US
[07:29:26] <kaiwai> the states that have ammended their constitution, and the push by 'churchgoers' to get it made a federal change
[07:29:40] <kaiwai> but ignorance like that spreads to other relms
[07:30:12] <kaiwai> if they're not bashing 'mo's, they're persuing a 'we know whats best for you' agenda in regards to foreign policy
[07:30:15] <hile_> and i know very few republicans who *don't* want ot tell the religious right ot go fuck themselves
[07:30:22] *** dclarke has joined #opensolaris
[07:30:37] <dclarke> gday ..
[07:31:06] *** dakrone has quit IRC
[07:31:13] <kaiwai> hile_: true; and to think, Republicans used to be about freedom of choice, personal responsibility - aka, the 'new right'
[07:31:25] * jmcp guffaws - District of San Francisco had Prop J "Call for Bush/Cheney Impeachment" on the ballot. with 88% of precincts reporting, it's ahead 59-41%
[07:31:48] <kaiwai> hmm, like its going to make a different in the grand sceme of things
[07:31:58] <dclarke> is Bush being impeached ?
[07:32:01] <hile_> no.
[07:32:08] <dclarke> no .. it will make no diff
[07:32:15] <dclarke> never mind
[07:32:15] <jmcp> unlikely .... unless the Dems get control of the Senate too
[07:32:16] * kaiwai proposes dismatling the US system and replace it with a parliamentry democracy and Queen Elizabeth II as head of state
[07:32:23] <kaiwai> :)
[07:32:31] <dclarke> welcome to Canada
[07:32:48] <kaiwai> who would you prefer; GWB or Elizabeth?
[07:33:02] <kaiwai> dear I sing, "rule britannia' in the most obnocious voice :P
[07:33:09] <dclarke> ah .. well at least the Queen has grace and education
[07:33:15] <kaiwai> true, and class
[07:33:24] <kaiwai> not some slack jawed hick like GWB
[07:33:43] <kaiwai> damn, GWB needs some alocution classes to tone down that awful accent
[07:33:44] <dclarke> back to Solaris talk .. I think I have Firefox 2.0 ready to roll out
[07:33:55] <kaiwai> cool :)
[07:34:02] <hile_> that's been out for some time :)
[07:34:10] <dclarke> I threw the ultimate browser acid test at it
[07:34:23] <dclarke> its been out for Solaris ?
[07:34:31] <kaiwai> anyway, back to solaris, are Belkin PCMCIA wireless cards supported?
[07:34:44] <hile_> yes
[07:34:51] <hile_> pub/firefox/releases/conrib
[07:34:54] <hile_> er, contrib
[07:35:01] <dclarke> oh crap ... there goes my UPS
[07:35:09] <hile_> you need a larger UPS
[07:35:13] <dclarke> ah .. in the dark here
[07:35:16] <jmcp> dclarke: how is your ups the ultimate browser acid test?
[07:35:34] <dclarke> I tossed Lotus Domino Web Access at it
[07:35:39] <dclarke> its bloated and huge
[07:35:46] <dclarke> tons of script, CSS and dhtml
[07:35:49] <kaiwai> ewww, lotus notes :(
[07:35:57] <kaiwai> bloated, buggy and, bloated
[07:36:18] <dclarke> see pictures here : http://www.blastwave.org/dclarke/Lotus/
[07:36:34] <dclarke> kaiwai : ever worked with it ?
[07:36:34] <kaiwai> whats the story with wine currently? (on solaris)
[07:36:43] <kaiwai> oh yes, the client is awful
[07:36:51] *** laca has quit IRC
[07:36:52] <dclarke> kaiwai: really ?
[07:36:53] <kaiwai> just plain cruel and unusual punishment
[07:36:59] <dclarke> kaiwai: its not that bad
[07:37:16] <Error_404> oh, hey dclarke
[07:37:16] <kaiwai> give me evolution or LookOut! anyday
[07:37:19] <Error_404> didn't see you come in
[07:37:20] <dclarke> kaiwai: the server is abusive to setup but once its done right .. it runs and runs
[07:37:23] <kaiwai> heck, even Groupwise <shudder>
[07:37:39] * jmcp thwaps kaiwai for mentioning gropewise
[07:37:50] <richlowe> hah.
[07:37:50] <dclarke> I'd rather eat your socks after a long rainy day than run Groupwise
[07:38:04] <richlowe> jmcp: could be worse, could be notes.
[07:38:10] * richlowe just wants to see dclarke's reaction.
[07:38:37] <dclarke> in any case both Firefox 2.0 and GNOME 2.16.1 will roll out tomorrow or so
[07:38:37] <kaiwai> hmm, have Sun thought about bribing IBM to get Notes Client on Solaris x86?
[07:38:46] *** halton has joined #opensolaris
[07:38:51] <kaiwai> bribing and/or blackmail
[07:38:56] <dclarke> kaiwai: you mean to get Notes client on Sol x86 again
[07:38:57] <kaiwai> I prefer the later
[07:39:07] <dclarke> it was on Sol x86 back in the day
[07:39:16] <kaiwai> true, well, there is always wine
[07:39:23] <kaiwai> but wine seems to have a suckage factor on Solaris
[07:39:46] <dclarke> see wine at www.blastwave.org/wine
[07:40:05] <kaiwai> why haven't the wine maintainers accepted the solaris related patches?
[07:40:12] *** halton has quit IRC
[07:40:38] <dclarke> okay .. gotta go deal with flaky power
[07:40:41] <dclarke> ta ..
[07:40:49] <kaiwai> see ya
[07:40:50] *** dclarke has left #opensolaris
[07:45:57] <richlowe> someday I'll learn my lesson and not $q from kmdb.
[07:47:24] <dlg> richlowe: do you know how to modify a variable in the kernel using mdb?
[07:47:42] <jmcp>  variablename/W newvalue
[07:47:45] <richlowe> mdb -kw
[07:47:48] <richlowe> variable/W <value>
[07:47:49] <jmcp> if you've started mdb with "-kw"
[07:49:06] *** bougie has quit IRC
[07:49:51] <dlg> it cant find the symbol name :(
[07:50:09] <dlg> i want to change zfs_nocacheflush to 1
[07:50:19] *** halton has joined #opensolaris
[07:51:11] <richlowe> that's probably not such a great idea, but zfs`zfs_no_cacheflush/W 1 I think should do it.
[07:51:43] <richlowe> but without that second underscore.
[07:51:57] <dlg> no joy:
[07:51:57] <dlg> (
[07:52:35] <dlg> i think its a good idea in this case
[07:54:30] <richlowe> you're most likely running a build prior to that tunable being added.
[07:54:51] <dlg> solaris 6/06?
[07:55:33] *** kaiwai has quit IRC
[07:56:09] <richlowe> that tunable was added October 26th.
[07:56:16] <dlg> pwned
[07:57:15] <dlg> and the mpt driver is closed source (for no good reason) so i cant intercept the scsi command either
[07:59:08] <jmcp> dlg: mpt is closed source for a *very* good reason, namely, that $vendor (not Sun) won't let it be released
[07:59:20] <jmcp> dlg: you could use dtrace
[07:59:40] <dlg> im not smart enough to use dtrace
[07:59:59] <jmcp> dlg: you need to get out of that self-deprecating mode
[08:00:07] <jmcp> *everybody* is smart enough to use dtrace
[08:00:22] <dlg> how can you cache syncing from zfs with dtrace then?
[08:00:29] *** delewis has joined #opensolaris
[08:00:34] <dlg> either in the fs or in the sd or hba driver?
[08:00:54] <jmcp> you can watch them all, if you want
[08:01:09] <jmcp> www.opensolaris.org/os/community/dtrace is a good place to start.
[08:01:25] <jmcp> and read Brendan Gregg's stuff, then pull down a copy of the DTraceToolkit and play with it
[08:01:52] <dlg> mmm
[08:01:56] *** steleman has quit IRC
[08:02:03] <dlg> i'll come back to it after i finish this thing
[08:02:04] <richlowe> mpt doesn't appear to have any kind of obvious toggle regarding flushes
[08:02:09] <richlowe> at least, nothing obvious in the symbol table, anyway.
[08:02:17] <jmcp> dlg: the ROI on learning to use dtrace is *very* fast/good
[08:02:23] <dlg> of course not, it just passes the cdb striaght onto the device
[08:02:56] <jmcp> mpt_flush_hba and mpt_flush_target not what you're looking for? (found with strings /kernel/drv//mpt on nevada)
[08:03:03] *** |tsoome| has quit IRC
[08:03:18] <richlowe> well, I'm kinda hunting backward sure.
[08:03:28] <richlowe> but I was hoping for a toggle as specific to the device as I could find...
[08:03:33] <dlg> that cleans outstanding commands off the device or bus
[08:04:13] <jmcp> what was the original question? I've missed it....
[08:04:36] <dlg> id need to add code in the path that takes a command from the midlayer and puts it on the device
[08:05:02] <dlg> it would have to pull the cdb apart and drop SYNC CACHE requests
[08:05:33] <dlg> jmcp: i originally wanted to stop zfs from issuing sync cache commands a lot
[08:05:50] <jmcp> for performance?
[08:05:56] <dlg> reliability
[08:06:09] <dlg> i think the firmware on the disks freaks out, and they reset
[08:06:21] <jmcp> oh
[08:06:22] <dlg> which causes a disconnect from the bus, and then you lose ios
[08:06:30] *** |tsoome| has joined #opensolaris
[08:06:33] <dlg> and then the computer gets upset
[08:06:33] <jmcp> well, mpt_scsi_init_pkt is where the pkt and cdb are created, from what I remember
[08:06:51] <dlg> cdb should originate in sd
[08:06:58] <dlg> its only allocated in mpt
[08:07:09] <jmcp> .... and the structure elements are stashed correctly in mpt
[08:07:12] <dlg> yes
[08:07:23] <jmcp> yes, sd creates the cdb, but it has to be translated into a scsi_pkt for transport
[08:07:45] <dlg> this is moot though, since i dont have mpt source on solaris
[08:08:34] <jmcp> yeah
[08:12:05] <dlg> if i had the source i could intercept it when mpt gets a scsi_pkt and translates it into an mpt io request
[08:15:33] <richlowe> You maybe able to make zfs cease requesting via some combination of zil_disable, and vdev_t->vdev_nowritecache on the vdevs.
[08:16:50] *** robj has joined #opensolaris
[08:20:04] *** gisburn has left #opensolaris
[08:21:23] <dlg> how easy is that?
[08:22:45] * jmcp wanders off
[08:23:11] *** lisppaste3 has quit IRC
[08:24:05] <dlg> hrm, i have sd source
[08:25:52] *** lisppaste3 has joined #opensolaris
[08:28:02] <asyd> \_o<
[08:35:11] *** miffe_ has left #opensolaris
[08:36:47] <Triskelios> any idea what would cause a network interface suffix to be incremented to 1 when only one device exists?
[08:42:26] <richlowe> moving the device to a different slot.
[08:42:37] <richlowe> maybe other things, but that'd be my first guess.
[08:51:18] <Triskelios> ok, that might've been it (this is a laptop)
[08:51:47] <Triskelios> how do I reset the number? I tried reconfiguring
[08:52:17] *** PosixC has joined #opensolaris
[08:52:21] *** PosixCompliant has joined #opensolaris
[08:52:43] *** PosixCompliant has quit IRC
[08:52:43] *** PosixC has quit IRC
[08:57:39] *** sommerfeld has quit IRC
[09:00:52] *** steleman has joined #opensolaris
[09:05:45] *** Cyrille has joined #opensolaris
[09:21:25] <stevel> richlowe: i've started generating new build snapshots
[09:21:33] <richlowe> stevel: cool.
[09:22:20] <richlowe> stevel: shouldn't you be out and about doing stuff though?
[09:22:37] <stevel> sitting in tech days listening to a talk
[09:22:59] <Gman> how are they going?
[09:23:10] <stevel> so so
[09:25:01] <Gman> in quality of talks? or attendance?
[09:26:25] *** slowhog has quit IRC
[09:27:49] <Gman> stevel, thinking of announcing pso, even with the lack of hackergotchis
[09:27:56] <Gman> my gimp skills are sucking badly
[09:29:33] <stevel> gman: quality of talks
[09:29:49] <stevel> gman: sure, do you want to try and find a better URL for it? or are you okay with it living off of grommit?
[09:30:03] * stevel doesn't think we can get an official .os.o subdomain
[09:30:12] <richlowe> or a better acronym? :)
[09:30:26] <yusufg> Gman: you live in NZ right ?
[09:30:29] <stevel> but maybe we can buy a .org
[09:30:32] <stevel> richlowe: hehe
[09:30:38] <Gman> stevel, that would be a sucky solution though
[09:30:40] <Gman> yusufg, yeah
[09:30:59] <yusufg> Gman: came across this cool software developed by a NZ company
[09:31:19] <yusufg> artrage http://www.artrage.com/artrage.html
[09:31:53] *** Risky has quit IRC
[09:31:54] <Gman> cool
[09:32:30] <Gman> richlowe, though, the fact that it's a sucky acronym might encourage an official .os.o subdomain :)
[09:33:26] <Gman> http://en.wikipedia.org/wiki/Hackergotchi - pity the video is non-english :(
[09:34:52] *** deather_ is now known as deather
[09:36:13] *** heffnerd has quit IRC
[09:38:21] <Triskelios> can someone give me a copy of the cardbus driver from x86 b51? it was apparently already there when I installed the cardbus-0.3 driver which overwrote it..
[09:38:41] <Triskelios> /kernel/misc/cardbus
[09:39:08] *** heffnerd has joined #opensolaris
[09:41:16] <richlowe> stevel: sucky that this is so much work, though.
[09:44:09] *** PosixC has joined #opensolaris
[09:46:50] *** kmaraas has quit IRC
[09:47:17] *** PosixCompliant has joined #opensolaris
[09:48:01] *** PosixCompliant has left #opensolaris
[09:48:04] *** PosixC has quit IRC
[09:48:56] *** PosixC has joined #opensolaris
[09:49:11] *** PosixC has left #opensolaris
[09:50:18] *** pseudo_ has joined #opensolaris
[09:59:51] *** Xh4 has quit IRC
[10:02:47] *** bunker_work has joined #opensolaris
[10:03:11] *** neoxed_ has joined #OpenSolaris
[10:03:45] <Gman> stevel, is grommit ok to be hit with a few requests? :)
[10:10:06] *** Gman has quit IRC
[10:10:07] *** alanc_away has quit IRC
[10:10:39] *** robj has quit IRC
[10:10:51] *** alanc_away has joined #opensolaris
[10:11:00] *** Gman has joined #opensolaris
[10:12:34] *** miffe has quit IRC
[10:14:10] *** miffe has joined #opensolaris
[10:15:29] <richlowe> stevel: what's with onnv/f/?
[10:18:07] *** kmaraas has joined #opensolaris
[10:18:41] *** udos has quit IRC
[10:19:21] *** neoxed has quit IRC
[10:24:58] *** schily_ has joined #opensolaris
[10:29:13] *** Dar_HOME is now known as Dar
[10:29:14] *** calumb has joined #opensolaris
[10:33:09] *** schily has quit IRC
[10:33:32] *** uebayasi has quit IRC
[10:41:40] <raph_ael> hello
[10:43:21] *** pseudo_ has quit IRC
[10:51:43] *** lplatypus has quit IRC
[10:55:15] *** Snake007uk has joined #opensolaris
[10:57:28] *** coo has joined #opensolaris
[11:08:50] *** gnu2it2 has quit IRC
[11:10:51] *** jnvilo has joined #opensolaris
[11:11:23] <jnvilo> where can i download opensolaris cd's? it's my first time looking into it and it seems I have to download the normal solaris first?
[11:11:54] <Gman> only if you want to build the kernel
[11:12:08] <Gman> otherwise you can look at distributions like nexenta, bellanix, schillix
[11:15:37] <Gr|ffous> jnvilo, you might want to look here: http://www.opensolaris.org/os/about/distributions/ The opensolaris code is released in several forms, I run Solaris Express, Community release
[11:16:31] <jnvilo> thanks!
[11:18:35] <jnvilo> Gman: coming from a 10 year lnux background,, RHCE, submit bugs, patches, which distro should i be looking it?
[11:19:08] <Gman> possibly nexenta
[11:23:22] <jnvilo> Gman: yeps looking at it and it looks like it's GNU with opensolaris kernel :)
[11:23:30] <Gman> yeah
[11:23:38] <jnvilo> i will try them all out and come back here
[11:23:41] <timeless> it works nicely in a zone
[11:24:03] <timeless> i use it to run apt on an entirely unrelated repository to retrieve junk safely
[11:26:51] *** Odin- has quit IRC
[11:30:09] *** Odin- has joined #opensolaris
[11:36:37] *** coo has left #opensolaris
[11:40:02] <Gr|ffous> Gman, is gnome 2.16 still looking on track for snv53?
[11:40:13] <Gman> hoping to be, yeah
[11:40:28] <Gr|ffous> wicked :)
[11:40:42] <timeless> ooh
[11:40:43] <timeless> hrm
[11:40:50] <timeless> is that gnome2.16.9 or what? :)
[11:42:21] <timeless> arg
[11:43:01] <timeless> :* but the bug i reported against gnome is fixed for 2.18
[11:43:56] <Gman> they did?
[11:45:17] <Gr|ffous> where can I find the nevada release schedule?
[11:45:47] <PerterB> http://opensolaris.org/os/community/on/onnv_schedule.txt
[11:45:48] <timeless> http://bugzilla.gnome.org/show_bug.cgi?id=364901
[11:46:25] <Gr|ffous> thanks PerterB
[11:46:44] <Gman> timeless, ahh, right
[11:46:58] <Gman> timeless, they couldn't commit back to the stable branch because there was a string freeze
[11:47:06] <Gman> timeless, which lasts for the entire 2.16.x series
[11:47:20] <Gman> next available time scale is therefore 2.18
[11:47:51] <Gman> if you're desperate to try it out, i could create a package for you
[11:47:55] <Gman> [but not until tomorrow]
[11:48:13] <timeless> that'd be cool
[11:48:22] <timeless> i'll probably be traveling for the next 2 weeks
[11:48:28] <timeless> so take your time,
[11:48:33] <timeless> unless you happen to be in mountain view
[11:48:40] <timeless> in which case, can i pick it up in person? :)
[11:49:01] <Gman> i'm in .nz
[11:49:08] <Gman> so unlikely to get there before next year
[11:49:13] <timeless> hrm
[11:49:38] <timeless> are you likely near other people i know in .nz? :)
[11:49:48] <timeless> hrm, i don't ave business cards :(
[11:50:00] <Gman> heh, no idea
[11:50:04] <timeless> last time i met someone cycling and he delivered a business card to a guy a few states away
[11:50:06] <Gman> i'm in christchurch, mid way down the south island
[11:50:10] *** mega has joined #opensolaris
[11:51:21] * Gr|ffous is a proud .nz'er too!
[11:51:35] <Gr|ffous> though I don't have the good fortune of working for Sun.
[11:51:36] <timeless> http://www.nzherald.co.nz/search/story.cfm?storyid=0003D3AF-EFBA-1525-9BD383027AF100AD
[11:51:44] <timeless> i have no idea how to figure out if that's near anything
[11:52:28] *** jnvilo has quit IRC
[11:53:50] <Doc> how many pairs of velcro gloves do you own?
[11:54:18] <timeless> for cycling + skiing?
[11:54:29] <timeless> probably half a dozen or a dozen
[11:54:44] <timeless> (not that i ski, but my snow gloves do work for skiing)
[11:55:02] <Doc> you're not from NZ, are you... :)
[11:55:26] <timeless> haven't had the pleasure of visiting :(
[11:56:31] <Gman> timeless, not a clue ;)
[11:56:50] <Gman> ok, time to sleep for me, later all
[11:56:51] *** Gman is now known as GmanAFK
[12:00:06] <PerterB> you a fairly serious cyclist then, timeless?
[12:06:09] * Gr|ffous is atm, I have a 160km race on the 25th. It's a fun-race, but 160km still qualifies as serious! ;)
[12:07:21] <PerterB> cool... hilly?
[12:08:14] <PerterB> (I normally do about 120km most weekends but suffering from a sprained knee at the minute)
[12:08:45] <Berny> .oO(crazy you're all crazy)
[12:09:04] <Berny> .oO(i do 500km each weekend... in my car :>)
[12:09:14] <Gr|ffous> Yeah, it has some killer hills apparently. I haven't actually been on the course, it's down south
[12:09:29] <Gr|ffous> why would you do 120 each weekend - that's no fun!
[12:10:02] <Berny> .oO(too much time?)
[12:10:26] <PerterB> training :) it's only about 5 hours and not all that hilly... in fact the reason it's 120km is that Box Hill, the nearest halfway decent hill to me, is 60km away :)
[12:11:35] <Gr|ffous> do you have a big race coming up?
[12:12:16] <PerterB> not for a while, planning to do the Cape Argus early next year and want to put in a better time this year
[12:12:35] <Gr|ffous> I'm not familiar with that one sorry, what are the details?
[12:13:03] <PerterB> biggest actual race in the world... 30,000 cyclists in Cape Town, it's big fun
[12:13:05] *** Risky has joined #opensolaris
[12:13:16] <PerterB> www.cycletour.co.za
[12:13:42] <Gr|ffous> yikes
[12:14:17] <Gr|ffous> 109...
[12:14:25] <PerterB> of course, the carbon cost of a bunch of us flying out there is probably more than Berny's 500 car miles per weekend :(
[12:15:31] *** Tpenta has quit IRC
[12:17:42] *** luxh has joined #opensolaris
[12:21:00] *** lplatypus has joined #opensolaris
[12:22:47] <Berny> yeah i never understand how you can spend so much money on a bike
[12:24:44] <PerterB> because a high end racing bike costs an awful lot less than a racing car... unless you're a serious athlete though you run into dimishing returns pretty quick
[12:25:13] <Berny> oh well
[12:25:34] <Berny> i mean here in germany you see loads of hobby cyclist with damn expensive bikes who never do a race
[12:26:10] <PerterB> right... and how many ferrari owners race? :)
[12:26:16] <Berny> they just spend huge amounts of money and believe the streets are theirs (although there are excellent cycling paths right next to the street)
[12:26:34] *** tarzeau has joined #opensolaris
[12:26:37] <Berny> i don't own a ferrari (except for the notebook)
[12:27:56] <Berny> i just have a middle class car (which by the doesn't have a electronic max speed regulator - therefor i can pretty easily "kill" quite a number of highly expensive cars which just stop at a certain speed)
[12:29:19] <Berny> but thats happening 2-3 times a year when i'm in the mood to have that kind of fun... rest of the year it's just relaxed getting from a to b without pedalling all the time :-)
[12:30:04] <Berny> .oO(i know i'm a lazy bastard when it comes to move myself around by using my feet :>)
[12:34:04] <Stric> PerterB: I've been in a race with about 16k people.. biking 300km on one day..
[12:34:30] *** Andrew____ has joined #opensolaris
[12:34:42] <PerterB> that's a serious day out... don't think I've ever done more than about 220
[12:35:10] <Stric> I was a bit tired after :)
[12:35:15] <PerterB> :)
[12:35:41] <luxh> Stric: vattennrundan?
[12:35:46] <PerterB> I did do the Etape du Tour last year though which was 185km but over 3 Alps
[12:35:53] <PerterB> that hurt :)
[12:36:23] <luxh> Stric: oh yeah, spelling. meant vatternrundan :)
[12:36:50] <Stric> luxh: yeah.. and some people got hit by a car (moron who biked without being in the race then got into the car and fell asleep by the wheel) and died a few minutes before us..
[12:37:13] <PerterB> yow :(
[12:37:22] <Stric> we had stopped just before that for a few minutes.. if we hadn't, it could have been me..
[12:37:52] <luxh> shit :(
[12:38:16] <LeftWing> Probably the most reasonable post I've seen on Slashdot in quite a while:  http://developers.slashdot.org/comments.pl?sid=205415&threshold=1&commentsort=0&mode=thread&pid=16762895#16763573
[12:38:57] <Stric> the people who arrange it demands that person who has competed -must- sleep at least 6h before getting into the car, but that guy ignored it with not so good results..
[12:39:17] <Berny> can one get zfs list to print out sizes in bytes (or at least all out in the same unit?)
[12:47:18] *** halton has left #opensolaris
[12:48:35] *** bougie has joined #opensolaris
[12:49:57] 
[12:49:58] *** kloczek has quit IRC
[12:50:03] <Berny> why the hack does the blastwave pkg for ntop require glib and all it's dependencies?
[12:50:27] <bougie> hello :p
[12:51:15] *** danielinu has joined #opensolaris
[12:57:08] <quasi> Berny: I think because of the xmldumpPlugin in ntop which uses gdome2 and glib-2
[12:58:18] <Berny> yeah and it's running a webserver... christ...
[12:58:44] <quasi> the default config is pretty scary
[13:00:30] <quasi> I went for jnettop last time I wanted something like that
[13:01:13] <Berny> let's google for that one
[13:01:29] <quasi> jnettop.kubs.info
[13:02:08] <Berny> cheers
[13:03:55] <quasi> at least it can be built without gui
[13:05:16] <Berny> though ntop just told me what i suspected
[13:05:45] <Snake007uk> we have 2.2TB of data which is made up of millions of 100k size files. we need to transfer this data from one DC to another, and currently we use DD to copy raw disk to another RAW disk and transport over
[13:06:41] <quasi> Snake007uk: sounds like a candidate for zfs ;)
[13:07:35] <Snake007uk> quasi but how would you transport
[13:07:38] <Snake007uk> over network
[13:07:48] <Snake007uk> to different DC
[13:08:22] <Snake007uk> we were thinking of tarring up but not sure if tar can transfer at full bandwitdh
[13:08:28] <Snake007uk> but considering the number of files i am not sure
[13:08:53] <Stric> tar is about as simple as it can get..
[13:09:12] <Berny> tar cpf - | ssh host "tar xpf -"
[13:09:19] <Berny> or something along that line ;-)
[13:09:22] <Snake007uk> well ls takes to display all listing
[13:09:30] <quasi> Snake007uk: zfs send
[13:09:35] *** User755 has joined #opensolaris
[13:09:40] <Snake007uk> zfs send?
[13:09:47] <Doc> zfs send or rsync
[13:09:48] <Stric> ls is probably stat():ing everything and then sorting, then displaying..
[13:09:53] *** Risky has quit IRC
[13:09:58] <lplatypus> tar may be less efficient than dd because it doesn't read the disk sequentially
[13:10:12] *** Patte has joined #opensolaris
[13:10:16] <Doc> to use dd you'll need to umount the filesystem first tho
[13:10:33] <Doc> or at least use some form of snapshotting
[13:11:17] <Snake007uk> rsync sync usaging memeory and the machine runs out of memory very quickly
[13:11:20] <Stric> the preferred method depends on a bunch of factors.. what is most important; least downtime, downtime-but-fastest-transfer etc
[13:11:48] <quasi> Snake007uk: "Example 10: Remotely Replicating ZFS Data" in man zfs
[13:12:09] <Doc> snake: so do it in groups
[13:12:20] <Doc> eg, one sub-directory at a time
[13:12:25] <Snake007uk> quasi,  u may have somehting here
[13:13:29] <quasi> Snake007uk: ymmv on the example numbering - I see it is example 12 in U3-beta
[13:14:08] <Snake007uk> ill write up a proper document for the problem, and see what people come up with.
[13:14:46] <Stric> Snake007uk: are you using zfs or something else?
[13:15:31] <Snake007uk> ext3 this is linux, im trying to make a move to ZFS
[13:15:48] <Stric> so you can't use zfs send.
[13:16:11] <Berny> .oO(i should get those quad gbit adapters into my box soon!)
[13:16:21] <Stric> and dd is probably a pretty bad idea too if you're changing filesystem
[13:16:32] <Snake007uk> if we could
[13:16:36] <Snake007uk> move to zfs
[13:16:37] <Stric> so you have to use some online tool like tar, scp, rsync
[13:18:06] <Stric> tar and scp should work fine for a "one shot" sync.. rsync if stuff is changing while you copy
[13:18:06] <Patte> Is there a way to reset/clear the bootmemory (bios in PC) in an Ultra 10?
[13:18:58] <Stric> hold down L1 and 'n' while powering up I think
[13:18:59] <dlg> can you get to the prom?
[13:20:06] <Patte> What is L1? Yes I get the ok prompt.
[13:20:22] *** Risky has joined #opensolaris
[13:20:28] <dlg> i think you can go reset-all on the ok prompt
[13:20:37] <Stric> reset-all doesn't clear anything
[13:20:46] <Stric> it's just like a "reboot"
[13:22:01] <Patte> Is there a button called L1 on the keyboard?
[13:22:15] <Stric> you have a key named 'Stop' or similar?
[13:23:12] <Stric> Patte: http://docs.sun.com/source/819-2946-10/oboot.html
[13:23:36] <Stric> it's the "Reset NVRAM" thing you want if you want to clear settings (permanently)
[13:24:22] <Patte> Yes I have 'Stop'.
[13:25:06] <Stric> turn on the power, hold down stop and 'n' until keyboard leds flash
[13:29:43] <Patte> Ok, thank you. I tried to install SXRC 51a, but it just hangs during the install. Is my Ultra 10 too old perhaps?
[13:31:26] <lasseoe> No it'll be fine
[13:32:04] <lasseoe> dunno why it hangs though, could be any number of things
[13:32:37] <dlg> u10s dont like really recent pci devices
[13:33:25] <Patte> My riser is empty.
[13:34:40] <dlg> oops, sorry
[13:37:35] <Patte> The screen is really strange The text is too big so the lines are too long.
[13:42:00] <Snake007uk> need help with this probelm :)
[13:42:04] <Snake007uk> http://www.unixcraft.co.uk/?cat=1
[13:53:53] *** calumb is now known as calLNCH
[13:54:04] <Berny> oehm
[13:54:25] <Berny> how can i get a throughput of 105.6Mbps on a 100Mbit nic? 8-)
[13:56:54] <quasi> Snake007uk: commented
[13:57:03] <quasi> Berny: compression
[13:57:32] <Berny> hmm, not that i would know how i configured that though
[13:58:03] *** Elendal has joined #opensolaris
[13:58:48] <quasi> probably not at the nic level, but at the protocol level instead
[13:58:57] <Berny> not that i know of
[13:59:14] <Berny> 99.9999% is good old nfs right now
[13:59:35] <quasi> ah, then maybe just a calculation error? ;)
[13:59:52] <Berny> probably ;-)
[14:00:13] <Berny> 90 more gigs and all the stuff is in a zfs 8-)
[14:01:44] <Berny> i should have installed the gbit nics before but than again our fab network guys only have 100mbit in this building :-\
[14:02:23] <dlg> gbic nics are better designed
[14:02:28] <dlg> its worth using them even on 100mbps networks
[14:04:19] <Berny> got two boxes waiting to get quad gbit nics :-)
[14:04:27] <Berny> next scheduled downtime
[14:04:54] <Berny> and then lets pray the damn switches work ok with link aggregation
[14:08:32] *** bengtf has joined #opensolaris
[14:09:49] *** logic has joined #opensolaris
[14:12:17] *** vmhobbes has joined #OpenSolaris
[14:15:58] *** uebayasi has joined #opensolaris
[14:16:53] *** Patte has quit IRC
[14:21:22] *** logic_ has quit IRC
[14:27:11] *** jcsmith has joined #opensolaris
[14:35:35] <Snake007uk> quasi, ??
[14:35:46] *** Cass has joined #opensolaris
[14:36:30] *** solarisjon has joined #opensolaris
[14:37:07] <quasi> Snake007uk: 13:42 < Snake007uk> http://www.unixcraft.co.uk/?cat=1
[14:37:17] <Snake007uk> yeah what about it ?
[14:37:34] <Snake007uk> oh u commented ok
[14:37:34] <quasi> Snake007uk: 13:56 < quasi> Snake007uk: commented
[14:37:45] <Snake007uk> ok let me read, and thanks :)
[14:38:20] <asyd> tar | nc rulez
[14:38:59] <quasi> ufsdump
[14:40:54] <Snake007uk> nc ? never heard of that
[14:41:09] <Snake007uk> is there something like ufsdump for ext3
[14:41:34] <quasi> netcat
[14:42:03] <quasi> of course ext3 is really a crappy fs when you've got many files
[14:42:05] <Snake007uk> ok
[14:42:10] <Snake007uk> yep
[14:42:22] <Snake007uk> this is a new issue for me :) and a good one too :)
[14:42:51] <quasi> not that I'd advocate reiser in general, but it is one of the things their btree indexes tries to help
[14:46:09] <Snake007uk> yeah thats why we are moving to resier, they really dont want to move to solaris, the linux peeps here are adament about that (considering i am linux person - i am opened minded and use whatever does the job) and in this case Solaris with ZFS does the job :)
[14:53:14] *** kmaraas has quit IRC
[14:53:24] *** bunker_work has quit IRC
[14:54:16] *** calLNCH has quit IRC
[15:00:24] *** TruBl_ has joined #opensolaris
[15:00:27] <Snake007uk> is there anything on gnome on linux that i cant have in gnome on solaris
[15:01:05] *** mega has quit IRC
[15:01:11] <TruBl_> Can anyone point me to some introductory docs for using DTrace ?
[15:08:57] *** |tsoome| has quit IRC
[15:12:05] <tarzeau> Snake007uk: yes
[15:12:24] <Snake007uk> tarzeau, i just asked in #solaris they said no
[15:12:45] <tarzeau> Snake007uk: it was not the same question
[15:12:46] <Cyrille> actually tarzeau said no...
[15:12:53] <tarzeau> 15:08 < Snake007uk> guys is all software which is available for linux/gnome available for solaris U2
[15:12:54] *** |tsoome| has joined #opensolaris
[15:13:01] <Cyrille> and true enough, it wasn't the same question...
[15:13:07] <Snake007uk> oh
[15:13:11] <Snake007uk> sorry
[15:13:15] <Snake007uk> well i get the idea :)
[15:13:28] <tarzeau> you want to find out what you really want to ask
[15:15:34] <Cyrille> Snake007uk: for instance, on that site you mentioned, they list vmware workstation, which doesn't support s10 as a host OS, and since it's not an open source product, there's little chance of your being able to compile your own version on that platform.
[15:18:05] *** kmaraas has joined #opensolaris
[15:18:59] <quasi> TruBl_: http://opensolaris.org/os/community/dtrace/ is a good starting point
[15:20:00] *** Vanuatoo has quit IRC
[15:20:46] <Snake007uk> Cyrille, i realised i asked a very silly question, i was just thinking of installing solaris 10 U2 on my IBM t42 and do work on solaris, as i could do what ever i wanted to on linux and know where everything is. the last thing i need to do  is keep looking around for software or there equivlent.
[15:21:33] *** |tsoome| has quit IRC
[15:22:18] *** |tsoome| has joined #opensolaris
[15:24:01] *** Vanuatoo has joined #opensolaris
[15:27:09] *** uebayasi has quit IRC
[15:28:06] *** Elendal has quit IRC
[15:28:40] *** Elendal has joined #opensolaris
[15:29:16] *** axisys has quit IRC
[15:32:27] *** eboutilier has joined #opensolaris
[15:33:05] *** Somethingelse has joined #opensolaris
[15:36:05] *** damienc has joined #opensolaris
[15:40:27] *** Xh4 has joined #opensolaris
[15:40:58] <TruBl_> Device Question: Why is the root of the device tree output from prtconf called "nexus" ?
[15:46:47] <oxygene> TruBl_: nexus is a term used for hub/bus drivers
[15:47:28] <TruBl_> oxygene: I know that ... but why use the word "nexus" ?
[15:47:33] *** lulf has quit IRC
[15:48:30] *** Snake007uk has quit IRC
[15:49:45] *** LeftWing has quit IRC
[15:49:53] *** LeftWing has joined #OpenSolaris
[15:50:33] *** axisys has joined #opensolaris
[15:52:02] *** nwf has quit IRC
[15:57:37] *** yippi has joined #opensolaris
[16:04:24] <Error_404> TruBl_: because that's the meaning of the word?
[16:04:45] *** |tsoome| has quit IRC
[16:05:41] *** laca has joined #opensolaris
[16:06:38] <Error_404> "why do we use the word blue to describe the color of the clear sky" is an equally absurd question
[16:08:39] *** yippi has quit IRC
[16:10:56] *** RaD|Tz has joined #opensolaris
[16:12:04] *** TruBl_ has quit IRC
[16:12:47] *** RaD|Tz has quit IRC
[16:13:09] *** ndroux has quit IRC
[16:13:16] *** RaD|Tz has joined #opensolaris
[16:13:26] *** RaD|Tz has quit IRC
[16:13:49] *** RaD|Tz has joined #opensolaris
[16:25:53] *** yippi has joined #opensolaris
[16:33:07] *** nwf has joined #opensolaris
[16:33:48] *** dunc_ has joined #opensolaris
[16:36:40] *** sommerfeld has joined #opensolaris
[16:36:41] *** ChanServ sets mode: +o sommerfeld
[16:37:04] *** darrenm has joined #opensolaris
[16:44:53] *** |tsoome| has joined #opensolaris
[16:51:00] *** logic has quit IRC
[16:51:01] *** logic_ has joined #opensolaris
[16:57:46] *** LinuXo has joined #opensolaris
[16:58:40] <elektronkind> today's lesson: Don't misspell Bill Sommerfeld's name.
[16:59:11] *** RaD|Tz has quit IRC
[16:59:25] *** zarathustra has quit IRC
[16:59:28] *** zarathus1ra has joined #opensolaris
[17:01:48] <Error_404> elektronkind: ?
[17:09:15] *** gdamore has joined #opensolaris
[17:09:25] <gdamore> good morning!
[17:20:57] <axisys> hmm.. mpathadm does not seem to be available on sol 10 u2
[17:21:06] <jbk> i heard u3
[17:22:06] <axisys> jbk: ok
[17:22:27] <quasi> sounds plausible (according to man on a box near here ;)
[17:22:35] <jbk> that at least i believe what they were targeting...
[17:23:07] *** slowhog has joined #opensolaris
[17:24:05] <jbk> it'd be nice if that piece ever gets opened up, would be curious to see how they balance i/o's across the paths
[17:25:33] <axisys> looks like sol 10 u2 does the multipath automatically http://www.rafb.net/paste/results/he9d0q83.html
[17:25:37] <axisys> nice
[17:26:03] <jbk> yes, but there really isn't any way to admin it
[17:26:12] <jbk> take paths offline/online, etc.
[17:26:19] <axisys> this is a v490 attached to 3510
[17:26:22] <jbk> that should be coming though
[17:26:54] *** Anil has joined #opensolaris
[17:27:06] <axisys> jbk: stmsboot -d (disable)
[17:27:17] <axisys> jbk: stmsboot -e (enable)
[17:27:33] <jbk> that's all or nothing
[17:27:39] <axisys> i c
[17:27:40] <jbk> and not an online thing
[17:27:41] <Anil> Hi, does the OpenSolaris Community provide an estimate date for Solaris 11 release?
[17:31:36] <darrenm> Anil: that isn't for the community to choose
[17:31:41] *** mrdeviant has joined #opensolaris
[17:32:03] <darrenm> just like the community wouldn't choose when the next rev of Nextenta or Belenix etc comes out
[17:32:12] <darrenm> Solaris is Sun's distribution based on OpenSolaris.
[17:32:19] <darrenm> When Solaris 11 comes out is for Sun to choose
[17:32:34] <darrenm> When uname changes from 5.11 to something else would be a community choice though.
[17:32:41] <darrenm> Do you get the difference Anil ?
[17:33:12] <hile_> darrenm: quick question -- with BSM is there any way to say "audit/log everything that is done in a profiled shell"?
[17:33:49] <darrenm> hile_: the profile shell writes audit records with event AUE_prof for every privileged exec it does (well actually pfexec(1) does that).
[17:34:07] <darrenm> so it depends what you mean by everything
[17:34:42] <jbk> heh.. we have management that wants to log all keystrokes for anything done as root..
[17:34:45] <Anil> Oh
[17:34:46] <jbk> still not sure how to do that
[17:34:56] <hile_> ah.  ..
[17:35:11] <hile_> I need to run out; i'l lthink more on that later
[17:35:22] <hile_> my employer does that
[17:35:40] <hile_> as much as it pains me, root's shell appears to be a custom ksh build that logs like hell
[17:35:44] <Stric> jbk: dtrace? :)
[17:35:54] <jbk> we still have lots of solaris 8 stuff :)
[17:35:58] <jbk> needs to work there too
[17:36:13] <axisys> we have bash w/ logging patched
[17:36:26] <hile_> why the fuck would you want bash?
[17:36:34] <axisys> i think was it bofh that we used to patch
[17:36:59] <axisys> not for root but users
[17:37:24] <axisys> root uses /sbin/sh
[17:37:29] <axisys> on < sol 10
[17:38:40] <jbk> i mean i suppose we can log all exec and file modifications, but then space concerns come into play...
[17:39:18] <darrenm> jbk: I seriously doubt you actually want all keystrokes
[17:39:24] <darrenm> what are you going to do with that data
[17:39:42] <jbk> witchhunt by management :)
[17:39:43] <Auralis> why not just use the stuff intended for such a purpose, audit?
[17:39:45] <darrenm> do you really want to know how good a typist the people with root access are and how many times a day the hit ^H ?
[17:40:07] <darrenm> I think what you probably want to do is explain that the more useful thing is a list of all the commands run with privilege.
[17:40:24] <darrenm> make root a role, and log all the execs (with arguments and environment variables) that it does.
[17:40:31] <jbk> basically we're encountering more work errors because they're forcing us to do more work in less time, and their solution is to scrutinize everything being done when things break :)
[17:40:37] <darrenm> sure that doesn't get you what they did to /etc/hosts but it is actually readbale.
[17:40:40] <jbk> i didn't say it was a valid reason :)
[17:41:11] <darrenm> jbk: the only valid reason for keystroke logging in my opinion is password sniffing :-)  Everything else is management looking at the problem wrongly.
[17:41:30] <jbk> i agree :)
[17:42:14] <jbk> i've looked at making root a role before, however the biggest issue is that you have to assign roles to every individual user, it'd be a lot nicer if it could be done based on group membership as well
[17:42:44] <sickness> is there anyone using belenix usb script?
[17:42:49] *** zarathus1ra is now known as zarathustra
[17:44:31] <darrenm> jbk: so all you are doing is moving the problem from assigning a role to each person to putting each person in a given group.  How does this help you  ?
[17:45:09] *** vectra has joined #opensolaris
[17:45:10] <vectra> re
[17:46:03] <jbk> because we are a heterogeneous environment and a dedicated group of operators handles user account management, they aren't very technical, but managing group membership is something they already do, and is largely uniform across hpux, aix, linux, etc.
[17:46:25] <jbk> having to special case all sun boxes poses issues
[17:47:43] <darrenm> jbk: Sun's Identity Manager will help you with that ;-)
[17:48:09] <darrenm> You aren't the first to ask about role membership being based on group membership and I doubt you will be the last
[17:48:22] <jbk> you know though
[17:48:43] <jbk> every bit of docs i can find on that on docs.sun.com don't tell you crap about what it can actually do if you're not talking to a web-based j2ee app
[17:49:12] <jbk> so it's not even an option at this point, as we are already using some M$ based solution for identity management
[17:49:48] *** johnlev has joined #opensolaris
[17:50:11] <darrenm> is there a reason why your identity management solution can't have the SolarisRBAC schema added to it and deploy that data too ?
[17:51:06] <darrenm> that BTW IIRC is how Sun's Identity Manager deals with some of the RBAC data.
[17:51:15] <jbk> heh...
[17:51:18] <jbk> yes and no
[17:51:40] <jbk> one thing they're still trying to decide is how to link everything
[17:52:01] <jbk> we started to consolidate all the unix enviornments into ldap
[17:52:10] <jbk> however, hp-ux & aix have issues
[17:52:55] <jbk> and the AD people don't like sun's AD sync tool, and want us to use products that are ad site aware instead
[17:53:08] <jbk> (don't like it cause it's not AD site aware)
[17:53:50] <jbk> it's a mess :)
[17:56:07] *** laca is now known as lacaAFK
[17:56:08] <darrenm> so do you have a solution other than AD itself, is there a provisioning tool or are you just using AD ?
[17:56:53] *** ndroux has joined #opensolaris
[17:57:42] *** Lippman has joined #opensolaris
[17:58:16] <jbk> most are still using a home-grown provisioning tool that's cross platform, but doesn't understand rbac (and the code is hideous), there are some boxes that are now under ldap, but are still undecided how to sync data with AD
[18:00:00] <PerterB> it's not exactly RBAC but sudo can give out root privs based on group (or netgroup) membership and do logging... although there would be the problem of people doing sudo /some/sh to evade logging
[18:00:05] *** darrenm has quit IRC
[18:01:11] <jbk> well we use sudo today.. and keep being told we will replace it with powerbroker 'someday', but most sysadmins still su - when they need to do stuff as root
[18:01:34] <PerterB> there's also commercial alternatives like UPM
[18:01:39] <elektronkind> I'd run a NIS server for project and *_attr maps that's populated with data from AD
[18:01:49] <jbk> but i'm still not sure that they will settle for anything but full keystroke logging :)
[18:01:57] <PerterB> I think UPM does that
[18:02:29] <PerterB> no idea how much it costs these days though
[18:04:54] <mrdeviant> is cvs.opensolaris.org timing out for anyone else ?
[18:05:32] <elektronkind> yeah, it is for me.
[18:05:53] *** danielinu has quit IRC
[18:06:00] <PerterB> yup
[18:06:41] <axisys> jbk: looks like you enable multipath per port as well in /kernel/drv/fp.conf
[18:06:48] <axisys> you can*
[18:06:56] <elektronkind> yeah, that's where you do that
[18:07:05] <jbk> axisys: that's not dynamic though -- i think with mpathadm though it should be able to
[18:07:15] <elektronkind> turning it on in scsi_vhci.conf will activate it globally
[18:07:58] *** Snake007uk has joined #opensolaris
[18:08:14] <axisys> elektronkind: thats for third party .. for sun u dont need to touch it.. just fp.conf
[18:08:49] <elektronkind> for anything you need to set mpxio-disable=no
[18:09:02] <elektronkind> sun devices are detected automatically if that's on
[18:09:15] <elektronkind> for others you need to make a definition in scsi_vhci.conf
[18:09:25] <elektronkind> (eg: Apple Xserve RAID)
[18:09:35] <axisys> elektronkind: yep
[18:09:38] <jbk> they could come up with a better interface for that too.. i love counting spaces and trying to remember the lengths of the vendor and product fields from an inquiry :)
[18:10:17] <axisys> jbk: u could change there are run stmsboot -u
[18:10:31] <axisys> jbk: u r right that requires reboot
[18:10:50] <jbk> it looks like mpathadm (when delivered) will address that though..
[18:10:59] <axisys> jbk: agree
[18:11:33] <jbk> we use vxdmp, and the biggest thing is that sometimes paths get stuck and never come back online automatically
[18:12:03] <elektronkind> dmp sucks. I've never been happy with it.
[18:12:04] <jbk> so whenever we have to do any sort of san work (switch or storage array), we have to touch every impacted box and verify all the paths are up and if not, bring them online manually
[18:12:47] <jbk> i think once we can do that with mpxio, we'd probably look at it if only to see one device for a given lun
[18:13:18] <jbk> we use clariions and using those with dmp is annoying due to all the clutter in /dev/dsk
[18:14:32] *** Darwin has joined #opensolaris
[18:15:41] <PerterB> there's always Powerpath ;)  (*ducks and runs*)
[18:15:56] <jbk> hahah
[18:16:22] <PerterB> it's big and horrible and sucky and expensive, but at least it usually brings paths back online
[18:16:25] <jbk> a friend at another company that uses powerpath on solaris has told me horror stories :)
[18:16:46] <jbk> apparently one time it decided to spontaneously change all teh device names
[18:16:57] <PerterB> yeah, it sometimes does that
[18:17:11] <jbk> they managed to catch it in time, but it almost completely corrupted all their oracle databases on that system
[18:17:15] <Stric> Hm. Anyone good at DNLC stuff? I have zfs with 550k inodes, serving it over nfs.. I've set ncsize to 600k and rebooted.. kstat -n dnlcstats  gives a whole bunch of stuff in pick_heuristic and pick_last which according to docs means "I had to get rid of stuff in dnlc, it was full" ..?
[18:17:16] <PerterB> we used it a lot at my last place.... $10k per server, I think there must have been some kind of golf course deal there
[18:17:57] <jbk> hahahahaha
[18:19:25] *** Cyrille has quit IRC
[18:21:46] <jbk> of course i think 'golf course deal' could describe just about any sort of business transaction involving emc
[18:22:15] <jbk> at least they would make sense then...
[18:23:01] <PerterB> yeah... that and the fact they flew most of the stakeholders out to Ireland by private jet for the weekend so they could land the original Symmetrix order
[18:23:51] *** LinuXo is now known as RaD|Tz
[18:26:22] *** eboutilier has quit IRC
[18:41:25] *** kmaraas has quit IRC
[18:53:05] *** Tpenta has joined #opensolaris
[18:54:14] *** ChanServ sets mode: +o Tpenta
[18:58:18] *** nwf has quit IRC
[19:01:41] *** nwf has joined #opensolaris
[19:07:16] *** bondolo has joined #opensolaris
[19:14:04] <dunc_> PerterB: i've had success with my SATA disks on another 32bit box by the way, so it's something specific to that box
[19:14:21] <PerterB> bios thang, maybe
[19:15:27] <dunc_> i'm stumped as to what though have messed with everything i know, and more besides :)
[19:17:50] *** Jiko_ has joined #opensolaris
[19:23:59] *** damienc has quit IRC
[19:25:07] *** digijohn has joined #opensolaris
[19:25:12] *** Cybernd has joined #opensolaris
[19:25:23] *** digijohn has left #opensolaris
[19:27:31] *** Jiko has quit IRC
[19:27:48] *** Xh4 has quit IRC
[19:30:07] *** paul has quit IRC
[19:30:12] *** paul has joined #opensolaris
[19:32:24] *** nwf has quit IRC
[19:33:08] *** Jiko has joined #opensolaris
[19:33:38] <bobbyz> Are there any utilities other than smpatch for solaris 10 that can do automatic patch downloads/installs?  I am trying to install a solaris system that is as reduced as possible so I used the reduced network core and added what I needed from there.  However, all of the dependencies required to get smpatch going has tripled the size of the system
[19:34:39] <dwc-> wget?
[19:35:24] <mrdeviant> pca ?
[19:35:50] *** gnu2it2 has joined #opensolaris
[19:35:53] <dwc-> http://sunsolve.sun.com/search/document.do?assetkey=1-9-82023-1
[19:36:15] <dwc-> grab the patch list from the readme
[19:36:24] <bobbyz> dwc-: I'm new to solaris.  I've been using linux for a long time, but not solaris.  I'm reaing through as much documentation as possible, but there's an ocean of it to read.  When you say wget, do you mean there is an easy way to retrieve the patches from a repository somewhere?
[19:36:27] <bobbyz> ahhh
[19:37:02] <jbk> do they still allow you to grab the recommended bundles without much hassle?
[19:37:25] <jbk> i haven't really kept too many tabs on all the changes to patches since they don't really affect us
[19:37:38] <quasi> jbk: I think as long as you register on sunsolve
[19:40:16] <dwc-> http://sunsolve.sun.com/pub-cgi/show.pl?target=patches/patch-access
[19:40:24] <dwc-> you can grab the patch lists from the readme in there
[19:40:54] <bobbyz> dwc-: Thanks, that helps a ton
[19:40:59] <dwc-> or... apparently now it wants me to sign in to see the readme
[19:42:04] *** Jiko_ has quit IRC
[19:43:09] <hile_> i have no hassle at all getting patches.
[19:43:34] <hile_> the wonders of having an SSE and/or D1 engineer on-site M-S 0700-0000 :)
[19:44:40] *** Somethingelse has quit IRC
[19:44:50] *** pikapika has joined #opensolaris
[19:45:09] <jbk> :)
[19:45:16] <jbk> we use tlp
[19:45:45] <bobbyz> tlp?
[19:45:50] <quasi> jbk: tlp is pretty handy
[19:45:55] <jbk> yes it is
[19:46:04] <jbk> if sun would ever pick a diagnosis enginge :)
[19:46:11] <bobbyz> what is that?
[19:46:20] <jbk> it's a tool for larger sites
[19:46:40] <jbk> for patching
[19:46:43] <quasi> but personally I'd probably be tempted to use pca if sun keeps insisting on insane pricing
[19:46:51] <jbk> pca?
[19:46:52] <quasi> "traffic light patch"
[19:47:05] <mrdeviant> how is sun's pricing insane?
[19:47:11] <quasi> http://www.par.univie.ac.at/solaris/pca/
[19:47:52] <quasi> mrdeviant: I can't tell you what $oldjob paid for tlp - but is sure as hell wasn't cheap
[19:48:11] <bobbyz> hey!  That pca is exactly what I'm looking for.  Thanks for mentioning it
[19:50:28] <hile_> I use pca a lot.
[19:51:04] <hile_> especially for my backups box at home (which runs 8) ... 10 boxes get updated when update releases come out
[19:52:14] *** Somethingelse has joined #opensolaris
[19:52:19] <pikapika> hello
[19:52:20] *** nwf has joined #opensolaris
[19:52:33] *** andersmo has quit IRC
[19:52:48] <axisys> hile_: even for production u auto update ?
[19:52:49] * quasi has taken to using smpatch on 10 - it tends to be pretty solid
[19:52:58] *** yippi has quit IRC
[19:53:04] <hile_> I never said I auto-update
[19:53:19] <hile_> I use pca for an easy "what the hell is out there since I last looked"
[19:53:31] <quasi> auto-update? that's like asking for trouble ;)
[19:53:33] <oxygene> hmm.. lots of patches (w/o support contract) for 10u2 in pca these days.. isn't that usually a sign that the next release is really near?
[19:53:47] <axisys> hile_: u said get updated.. so i assumed auto ;-)
[19:54:16] <hile_> I don't patch boxes more than about quarterly unless there's an  urgent issue that needs it
[19:54:25] <quasi> oxygene: prolly - you haven't heard u3 named 6/11? ;)
[19:55:00] <axisys> i never used tlp.. how does it perform compare to pca or smpatch?
[19:55:07] <oxygene> quasi: I know that
[19:55:21] <oxygene> quasi: 01/06 appeared 12/05 already, so I don't count too much on that
[19:56:22] <quasi> which reminds me that I should probably get some bugs filed towards u3-beta ;)
[19:57:33] *** _william_ has joined #opensolaris
[19:57:40] <axisys> "For random optimization, the cache block size is 32 Kbyte" where do setup the cache size for storage
[19:58:11] <axisys> in /etc/system ? what parameter is that?
[19:58:17] *** yippi has joined #opensolaris
[19:58:26] <_william_> hi all
[19:58:31] *** alanc-away is now known as alanc
[20:00:52] *** Lippman has quit IRC
[20:01:47] *** andersmo has joined #opensolaris
[20:03:18] <bobbyz> Hey, thanks for the information about pca et. al. guys.  I really do appreciate the help.  I am trying to read as much documentation as possible through google and sun.com but I've hit a couple of brick walls that you guys have really helped me past  :)
[20:03:48] *** andersmo_ has joined #opensolaris
[20:04:18] <axisys> i have two fiber chnl adapters for my T2000. can I just connect them to fc0 and fc1 to get multi path for sol 10 u2?
[20:04:23] *** andersmo_ has quit IRC
[20:04:30] * Symmetria dances
[20:04:50] <Symmetria> my solaris bos just set the record for fastest outbound traffic pushed by a single box outta .za in an hour long period
[20:05:07] <Symmetria> well, I can guarantee it was a record :p cause I've had ph one calls asking if I've lost my mind
[20:08:21] <sommerfeld> what was it pushing?
[20:08:37] <Symmetria> heh well over 200mbit
[20:08:45] <Symmetria> (yeah its not huge, but .za hasnt got hugte bandwidth)
[20:08:50] <Symmetria> and in .za bandwidth that is huge
[20:09:13] <Symmetria> 1255
[20:09:15] <Symmetria> heh
[20:09:38] <quasi> Symmetria: start pushing pr0n - then you'll easily use a lot more than that ;)
[20:13:05] <Error_404> and it's apparantly a good way to make a tidy profit
[20:14:43] <Error_404> i'd do it but i don't know any girls willing to get naked on the internet
[20:14:49] <sommerfeld> and Josh Berkus misses an opportunity to say "lawyers, guns, and money"...
[20:15:31] *** Azureus1 has joined #opensolaris
[20:15:44] <Error_404> ?
[20:17:23] <sommerfeld> post to osol-discuss
[20:17:29] <sommerfeld> " I'd think it more likely that those
[20:17:30] <sommerfeld> components have large amounts of code licensed to Sun from third parties.  So
[20:17:30] <sommerfeld> fixing it would require both lawyers and money."
[20:17:54] <Error_404> i don't pay attention to osol-discuss
[20:19:12] <alanc> osol-discuss mail delivery seems to be weird - never got rich's original, but see josh & harpster's responses
[20:19:52] <alanc> ooh, nevada build 52 ISO's posted internally...
[20:20:15] <Error_404> opensolaris.org is down
[20:20:29] <Error_404> The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later.
[20:20:44] <mrdeviant> opensolaris.org goes down a lot, it seems
[20:20:46] <alanc> I thought only Java had been slashdotted
[20:21:01] <Error_404> mrdeviant: more than a drunk sorority chick
[20:34:22] *** lacaAFK is now known as laca
[20:35:23] *** RaD|Tz has quit IRC
[20:38:50] *** Kronuz has joined #opensolaris
[20:38:53] *** sahafeez has joined #opensolaris
[20:40:00] <Kronuz> hmm... what is the difference between the channel #solaris and #opensolaris ?
[20:40:11] <twincest> we're nicer than they are
[20:40:18] <Kronuz> lol :P
[20:41:29] <jamesd> #opensolaris is more professional, and #solaris is much more relaxed, and not constricted by protocol
[20:41:48] <mrdeviant> we're professional ?
[20:41:51] <sahafeez> we are professinal
[20:42:02] <sahafeez> jink
[20:49:20] <Kronuz> oh
[20:50:14] <Kronuz> hey, I'm planning on using Solaris in our new servers... (I'm just not very fond of Linux for some reason)
[20:50:31] <Error_404> because it's shit?
[20:50:34] <oxygene> jamesd: #solaris is not even constricted by etiquette
[20:50:48] <Kronuz> that's what I've always thought
[20:50:51] <Kronuz> (honestly)
[20:50:55] <Kronuz> but I wanted to talk with someone who has use Solaris
[20:50:58] <jamesd> true
[20:51:11] <sahafeez> good choice. forget everything you know about *nix (linux, bsd) and you will have an easier time. bigadmin is your friend
[20:51:19] <Kronuz> 'cause I'm more of a BSD guy... but support on servers is not that good for those
[20:52:06] <Kronuz> (though I'm haven't used much of those either... you could say I'm rather new to the *nix world)
[20:52:23] <Kronuz> and Solaris sounds great, now that's open source and "free"
[20:52:42] <sahafeez> well bsd means you have a better understanding then some of the linux people. :)
[20:52:59] <Kronuz> yeah, I like to think so :)
[20:53:09] <delewis> Kronuz: it's been "free" for quite some tiem.
[20:53:13] <delewis> s/tiem/time/
[20:53:13] <Kronuz> but I was wondering... what are the plans of Sun? will Solaris always be free from now on?
[20:53:14] <delewis> since Solaris 8
[20:53:17] <Kronuz> hmm
[20:53:26] <Kronuz> I thought it was free since 10
[20:53:28] <delewis> no
[20:53:31] <Kronuz> oh
[20:53:39] <delewis> it was free if you had under 8 CPUs for Solaris 8
[20:53:46] <delewis> and under 2 CPUs for 9
[20:53:55] <twincest> "under 2" :)
[20:54:01] <delewis> inclusive. :-)
[20:54:10] <Kronuz> and for 10?
[20:54:23] <Kronuz> is free for any number of processors?
[20:54:28] <Auralis> yes
[20:54:33] *** bougie has quit IRC
[20:55:01] <Error_404> as per the license, solaris 11 will always be "free" (in the stallman sense)
[20:55:01] <Kronuz> so the plan is to keep it free, 'cause honestly I'm always a bit affraid Sun changes its mind and starts charging money on the next version :P
[20:55:10] <delewis> Kronuz: why?
[20:55:15] <Kronuz> dunno
[20:55:20] <delewis> Sun hasn't changed it's mind in 6 years regarding the cost of Solaris
[20:55:21] <Kronuz> it might happen
[20:55:26] <delewis> (only licensing terms per processor count)
[20:55:41] <Error_404> if it did, someone'd fork the last free version a la Linux
[20:55:41] <delewis> that's a lot more than Red Hat can say
[20:55:43] <delewis> or IBM
[20:55:48] <Auralis> it is opensource, there are distors out there that are not from sun
[20:55:53] *** bougie has joined #opensolaris
[20:55:56] <delewis> which change their cost of media/contract costs every year or more
[20:56:09] <Kronuz> hmm... really are there distros?
[20:56:18] <oxygene> Kronuz: lookup nexenta, belenix, schillix, martux
[20:56:19] <delewis> Kronuz: yes
[20:56:31] <Kronuz> interesting
[20:56:48] <Auralis> http://www.genunix.org/
[20:57:32] *** nwf has quit IRC
[20:58:05] <Kronuz> what about the hardware? is UltraSPARC really much better than Opteron ?
[20:58:13] <delewis> why does it matter?
[20:58:26] <delewis> Solaris 10 runs on Opteron as well as it does UltraSPARC.
[20:58:31] <delewis> as Sun is a huge Opteron vendor.
[20:58:38] <Kronuz> I'm just wondering, 'cause we're buying servers
[20:58:43] <delewis> but in some cases, yes, UltraSPARC might out-perform Opteron
[20:58:48] <delewis> though, more likely
[20:58:53] <Kronuz> and if it's worth we might consider SPARC
[20:58:54] <delewis> Opteron will blow UltraSPARC out the window.
[20:58:56] <sahafeez> sun makes the best servers
[20:58:58] <delewis> you may also consider Niagra
[20:59:04] <delewis> if you have a heavily threaded workload
[20:59:28] <oxygene> (heavily threaded without fpu needs, that is)
[20:59:31] <Kronuz> what about Xeon sustems?
[20:59:39] <Error_404> space heaters
[20:59:40] <delewis> Kronuz: go SPARC if you want elegance, scalability, and reliability over performance.
[20:59:55] <delewis> Kronuz: Xeons are AMD64, nowadays, so it doesn't really matter.
[21:00:02] <mrdeviant> uh, no.
[21:00:11] <delewis> I'd get my Opteron gear from Sun, regardless. :-)
[21:00:27] <mrdeviant> xeons may implement the amd64 isa, but the system architecture is very different
[21:00:29] <Kronuz> I've read Intel systems can get a bit wormer
[21:00:51] <delewis> mrdeviant: for better or worse?
[21:00:58] <delewis> I've only read the ISA specifications
[21:01:08] <Error_404> Kronuz: plus FSB is an outdated crap technology
[21:01:09] <mrdeviant> amd's is much more scalable.
[21:01:20] <Kronuz> hey, what would you recommend to get me started with Solaris?
[21:01:24] <mrdeviant> the opteron design is numa.
[21:01:25] <delewis> Intel never gets scalability right. :-)
[21:01:28] <delewis> well, I knew that.
[21:01:41] <Error_404> Kronuz: download a copy & install it on a machine you have kicking around
[21:02:18] <Kronuz> but I meant literature (for guys comming from the BSD world... or at least linux world)
[21:02:34] <Error_404> bigadmin.com
[21:02:47] <Auralis> docs.sun.com has more then you ever need or want
[21:02:56] <Error_404> them too
[21:03:27] <Kronuz> I wanted to use FreeBSD, but the support from manufacturers is bad... and the philosophy FreeBSD team has about corporate world is that "they don't need them at all" (or so they say, tho' I find that statement even stupid :P)
[21:03:55] *** zsakr has joined #opensolaris
[21:05:13] <Kronuz> hey, about ZFS... is that one designed in any way for a SAN or was it created with something else in mind?
[21:05:47] <Auralis> ZFs is a next gen general purpose fiulesystem
[21:07:01] <hali> i can't really think of many scenarios where ZFS is note suited
[21:07:04] <hali> not*
[21:07:13] <hali> clustering .. but thats in the pipeline
[21:08:06] <Kronuz> what about Virtualization? .... oh, it's just that I have a bunch of questions, and there's so much information about everything from everyone that one can get lost
[21:08:14] *** GmanAFK is now known as Gman
[21:08:31] <Gman> hey all
[21:08:43] <Kronuz> hi Gman
[21:09:14] <hali> ZFS is great for Virtualization.. you can assign a zone a ZFS and the zone admin can create more filesystems under that
[21:09:41] <hali> great if you rent out containers... easy to assign containers more space
[21:09:49] <Kronuz> hali: you make it sound so easy :P
[21:09:58] <Error_404> it *IS* easy
[21:10:13] <hali> Kronuz: the whole zfs setup is two commands .. more or less... it's as simple as it gets
[21:10:31] <hali> you have the zpool command to build the "pool" and zfs to manage whats in the pool
[21:11:12] <Kronuz> so the pool can take up the whole volume (or generally it should)... or what?
[21:11:25] <Error_404> it can
[21:11:25] <Kronuz> what about a SAN?
[21:11:41] <hali> you can give zfs pretty much anything... a disk, a slice, even a file on a ufs filesystem
[21:11:49] <mrcranky_> you can have a pool in the san that uses one or more whole LUNs
[21:11:50] <Error_404> zfs set exportnfs=on pool/fs
[21:11:56] <hali> you can give ZFS a luns as well.. works great
[21:12:24] <Kronuz> anyway, back to virtualization... is virtualization just used for security (as if you rent containers, as hail said) or also to improve performance by increasing CPU usage in some way?
[21:12:39] <hali> both yes and yes
[21:12:53] <Kronuz> hmm.. so it does increase CPU usage
[21:13:18] <hali> usage in the sense that you have 0.5-1% overhead
[21:13:31] <Kronuz> that sound's weird... wouldn't more processes or threads do the same?
[21:13:34] <noyb> timeless: did you ever get your kernel module built ?   I hope you found the help you needed.
[21:13:36] <hali> usage in the sense that you can have more applications on one box and hence get better utilization
[21:13:43] <Kronuz> without the need of virtualization? (that's something I don't get)
[21:14:24] <kleppari> is it impossible to compile Solaris with Sun Studio 11?
[21:14:34] <delewis> kleppari: no..
[21:14:37] <delewis> I've been doing it for weeks now
[21:14:46] <Kronuz> Sun Studio?
[21:14:47] <Kronuz> hmm
[21:15:00] <oxygene> Kronuz: zones work by adding a flag to processes.. every zone has its own flag, and given that flag on a process, the kernel decides what the process can see.
[21:15:09] <hali> Kronuz: if you don't need virtualization, you don't need it... it's just there for free if you some day you decide you do
[21:15:20] <oxygene> Kronuz: ie. one kernel, lots of system instances - no virtualization in the sense on vmware and friends
[21:15:35] <hali> i usually use containers to resource limit in a clean way...
[21:15:40] <kleppari> delewis, oh, great :)
[21:17:06] <Kronuz> yeah, but I was wondering about the CPU usage increase via virtualization, 'cause it got me confused when could virtualization be needed to increase performance if creating threads or process should do the same
[21:17:21] <Kronuz> or not?
[21:19:22] <sickness> heh, I'd like vmware for solaris host SO much :'
[21:19:28] <Error_404> say if your database server (for example) is important
[21:19:42] <Error_404> you can put it in a zone where it gets exclusive rights over a processor or four
[21:19:51] *** Sugarat has joined #opensolaris
[21:20:03] <delewis> Error_404: you can do that with projects, as well.
[21:20:07] <delewis> (which is what zones use anyway)
[21:20:34] <Error_404> I know, but it was the only real "performance increase" for zones i could think of
[21:20:59] <Kronuz> so I suppose virtualization is more about organizing structure and security than increased CPU usage then
[21:21:13] <Error_404> yeah
[21:21:29] <Kronuz> (never got that about the increased CPU usage and it doesn't sound very logical either)
[21:21:45] <Sugarat> I've been trying to install Solaris 10 on my PC for hours now and its just not working. After selecting the disk to install on and making the slice its sat there saying - preparing the system for solaris installation
[21:21:52] <Error_404> though if you have an app that requires linux for some stupid reason, you can either run 2 machines (one solaris, one linux), or run the app in a branded zone
[21:22:13] <Kronuz> lol, should be a really stupid reason
[21:22:40] <oxygene> Kronuz: increased cpu usage as in "I have a machine running here with 20% cpu usage, and I'd _love_ to run something else on it, but then the support for that weird app running on it blames everything on the additional installations"
[21:23:50] *** rachel_ has joined #opensolaris
[21:23:51] *** rachel has quit IRC
[21:23:51] *** laca has quit IRC
[21:23:52] *** laca_ has joined #opensolaris
[21:23:53] <Kronuz> Error_404: I'd probably choose windows 95 over linux anyday... thought that might, just might, be overstating a bit
[21:24:43] * Gman wonders whether he should dial into the pteam today
[21:24:45] <Sugarat> on what PC's should Solaris run? the install fails at various points on this machine
[21:25:08] <delewis> Sugarat: peecees that have components listed on the HCL at bigadmin
[21:25:26] <Kronuz> I mean, if you want to increase CPU usage you add more processes or threads, unless something like oxygene says happens (if there's an incompatibility between installed apps), on which case you'd want to virtualize, if I understood correctly
[21:25:45] <Sugarat> this is a 500Mgz Pentium 3 so I guess it should run
[21:25:57] <delewis> zones just give you a chance to logically separate your services running on a single system.
[21:26:12] <delewis> and they're nice for development/testing environments
[21:26:35] <delewis> want a quick DNS setup to play around with? create four or five zones for root and slave servers
[21:26:40] <delewis> same goes for NIS, LDAP, Kerberos
[21:27:01] <Kronuz> I tried installing Solaris in a Dual Core AMD 64 some time ago, but it just freezed at boot time after a couple times I just stopped trying
[21:27:19] <delewis> Sugarat: the processor has little to do with it
[21:27:23] <Kronuz> I have a DFI mobo with nForce4
[21:27:32] <delewis> the most important component is a supported disk controller
[21:27:37] <delewis> next you'll want a supported NIC and so fourth
[21:28:13] <Kronuz> it was bad I couldn't get it working that time... :( otherwise I'd know more now about Solaris
[21:28:24] <Kronuz> :P
[21:28:26] <Sugarat> well the disks are detected fine so I guess it works
[21:28:33] <Error_404> don't the u20/u40 run nforce4?
[21:28:35] <Sugarat> except that it crashes when trying to install
[21:28:48] <Kronuz> u20/u40 ?
[21:29:00] <delewis> Ultra 20 and Ultra 40
[21:29:04] <Error_404> sun Ultra 20  or Ultra 40 workstations
[21:29:06] <Kronuz> it was shortly after booting when it crashed (or freezed)
[21:29:12] <delewis> Sun's Opteron-based workstation systems
[21:29:21] <delewis> the 25 and 45 are SPARC-based
[21:29:48] <Kronuz> hey, I'll be back in a few minutes...
[21:29:51] <Kronuz> thanks for the help :)
[21:30:00] *** Kronuz is now known as KronuzAFK
[21:30:34] <Sugarat> if the solaris 10 cd's from Sun dont want to install, do you think I would have more luck with OpenSolaris ?
[21:30:47] <Error_404> Sugarat: try it
[21:30:58] <Error_404> bunch more drivers in SX
[21:33:38] <Sugarat> ok cool
[21:35:40] <delewis> OpenSolaris != Solaris Express
[21:37:51] <axisys> could this be a zfs bug?
[21:37:52] <axisys> http://www.rafb.net/paste/results/EG5n4Q16.html
[21:38:02] <Error_404> you're just really out to shoot down my simplifications today, aren't you delewis ?
[21:38:05] <Error_404> ;)
[21:38:10] <axisys> zpool status shows degraded .. but svm slices shows just fine
[21:39:10] <sommerfeld> c1t1d0s7  UNAVAIL      0     0     0  cannot open
[21:39:23] <sommerfeld> so it could be a transient failure to open
[21:39:33] <axisys> sommerfeld: c1t1d0s1 and c1t1d0s0 on svm shows no problem
[21:39:56] <axisys> replaced the disk also just in case.. still showing issue few months later
[21:40:09] <sommerfeld> and "zpool online c1t1d0s7" doesn't reopen it?
[21:40:47] <sommerfeld> what does prtvtoc /dev/rdsk/c1t1d0s2 report ?
[21:40:52] <axisys> sommerfeld: ran it couple times
[21:42:43] <axisys> sommerfeld: let me check
[21:43:35] <axisys> http://www.rafb.net/paste/results/NQJ3hj63.html
[21:44:28] <sommerfeld> looks fine.  do you see any difference between that and the same on the other disks?
[21:45:00] <axisys> sommerfeld: let me check.. btw.. running the zpool status online again.. http://www.rafb.net/paste/results/5pGjB086.html
[21:46:51] <axisys> sommerfeld: all four disks (raid z) look same http://www.rafb.net/paste/results/P3ypwQ58.html
[21:47:40] <axisys> time for a bug report ;-)
[21:47:48] <sommerfeld> yes, definitely.
[21:50:02] <axisys> browsing sunsolve bug report to make sure it is not duplicate
[21:50:47] *** Somethingelse has quit IRC
[21:51:36] <axisys> sommerfeld: last zpool online completed .. http://www.rafb.net/paste/results/NDzgQh36.html :-(
[21:51:45] <axisys> still same status
[21:53:23] <sommerfeld> hmm.  i'd go mucking around with dtrace fbt to see exactly what part of the stack is getting in the way.
[21:54:22] <axisys> sommerfeld: except i am not familiar w/ dtrace yet..
[21:54:41] <axisys> signed up for a dtrace class in arlington, va in last week december :P
[21:55:04] <axisys> hopefully it wont get cancelled on me for the third time
[21:55:19] <axisys> since last year
[21:55:52] <axisys> i do have the DTT downloaded.. any particular script there could help?!
[21:57:50] *** mrcranky_ has quit IRC
[21:59:50] *** osdummy has joined #opensolaris
[22:05:31] *** Sugarat has left #opensolaris
[22:08:15] *** yippi has quit IRC
[22:09:20] *** KronuzAFK is now known as Kronuz
[22:09:23] <Kronuz> hey
[22:09:32] <Kronuz> are there ports in Solaris?
[22:09:38] <Kronuz> (or something like that)
[22:09:55] <Kronuz> also, how do you upgrade the system?
[22:10:28] <Kronuz> is there a way to download the stuff and compile it to upgrade? (as in the *BSD systems)
[22:10:50] <axisys> can i use a FC adapter on T2000 to coonect to fc0 and fc1 on 3510?
[22:11:03] <axisys> to achieve mpxio
[22:11:04] <twincest> you can compile the ON source and install it if you like the bleeding edge, but there's no reason to do so
[22:11:19] <twincest> (unless you want to modify it)
[22:11:49] *** hspaans has joined #opensolaris
[22:11:56] <hspaans> g'day to all
[22:12:12] <timeless> presumably he's more interested in upgrading additional apps like a browser or editor
[22:14:21] *** Burana has joined #opensolaris
[22:14:25] <hspaans> people working with sol10 11/06?
[22:14:54] <Burana> When is it coming out?
[22:15:20] <hspaans> this month, but there seems something with zoneadm
[22:15:39] <hspaans> don't pin me on the date btw
[22:16:22] <Burana> I hope it comes this month. I want to setup a couple of new servers
[22:16:52] <hspaans> and you can't with 06/06?
[22:16:59] *** Symmetria has quit IRC
[22:17:16] <Burana> I just started at a new company and are setting up a new solaris build
[22:17:23] *** alchemist3 has joined #opensolaris
[22:17:40] <Burana> the company isn't using sol10 yet
[22:17:59] <Burana> and 11/06 has a couple of zfs patches included
[22:18:10] <hspaans> raidz2 yes
[22:18:56] <Burana> there's a lot of work for me waiting, but with the new solaris features, it's fun :-)
[22:19:19] <Azureus1> I could mount ext3 linux from solaris using 2 packages from belenix site ; is it possible to mount resierfs from solaris ?
[22:19:40] <axisys> i just attached a 3510 to the T2000 and rebooted.. it shows wrong magic number
[22:19:48] <sahafeez> no
[22:20:01] <Auralis> Azureus1: if you write the driver, then yes
[22:20:02] <axisys> http://www.rafb.net/paste/results/hQc1ML17.html
[22:20:10] *** bougie has quit IRC
[22:20:15] <hspaans> axisys: start format and label them
[22:20:34] <hspaans> Burana: test your applications on sol10 before you go live
[22:20:48] <axisys> hspaans: let me try
[22:21:19] <Burana> hspaans: we have mainly oracle and sybase databases. Nothing too complex.
[22:21:26] <hspaans> Burana: I already encountered some apps that started to barf on sol10 that worked without problem on sol8 and 9
[22:21:45] *** Darwin has quit IRC
[22:22:06] <Burana> hspaans: if we migrate we will start with the development system, and then go forward to test and prod.
[22:22:11] *** Azureus1 has quit IRC
[22:22:47] <hspaans> Burana: just sharing my experiences ;-)
[22:23:18] <Burana> hspaans: my philosophy is (where possible) to have a global zone with no applications. And a local zone containing the application (even if there is only one local zone totally)
[22:23:43] <hspaans> but no sun god arround for some zoneadm question?
[22:23:49] <Burana> hspaans: the zonemigration feature should come handy in the future... what do you think?
[22:24:06] <Burana> hspaans: what is your question?
[22:24:19] <hspaans> Burana: check sun cluster docs is has some good hints
[22:24:37] *** bengtf has quit IRC
[22:24:56] <Burana> hspaans: what do you think about keeping the global zone free from applications?
[22:25:06] <hspaans> Burana: 11/06 has in the man page that you can use -N when you detach a zone but zoneadm doesn't support that option
[22:25:31] *** bengtf has joined #opensolaris
[22:25:34] *** alchemist3 has quit IRC
[22:25:35] <hspaans> Burana: nice but don't do it to has zones running
[22:25:42] <hspaans> s/has/have/
[22:26:11] <hspaans> but when its in a zone you can migrate it without too much problems
[22:27:01] <Burana> The documentation says, you can only migrate to the same os release/patch level on a new host. But what if the new host is never?
[22:27:17] <hspaans> your fucked
[22:28:09] <Burana> maybe, maybe not
[22:29:12] <lasseoe> unsupported though, even if it does work :)
[22:29:18] <hspaans> currently you can't do a live upgrade when you have zones so
[22:29:21] <axisys> i see 175G on one disk and 175G on the second
[22:29:46] <axisys> one connected to fc0 and one fc1
[22:30:00] <axisys> is that mean i am seeing the same 175G or all 350G?
[22:30:21] <lasseoe> depends how your storage is configured
[22:30:23] <Kronuz> twincest: but say you install Solaris today, in a few months or even a year, you'll want to upgrade to the next stable Solaris, right? well, how hard or easy that is?
[22:30:29] <Kronuz> (that's what I meant)
[22:30:46] <Kronuz> (also about the other apps like postgres or apache)
[22:30:49] <twincest> kronuz: insert DVD, boot, upgrade
[22:30:58] <hspaans> axisys: check mpxio and thet 3510 if you offer two luns or one over two paths
[22:31:05] <twincest> upgrading will upgrade the bundled postgres and apache, as they are part of solaris
[22:31:33] <Kronuz> twincest: I see, but can't that be done via Internet instead of a DVD?
[22:31:39] <sommerfeld> axisys: to recycle something a CS prof once said: "paint one, see if the other one changes color"
[22:31:54] <hspaans> Kronuz: I upgrade every two weeks my sxrc box with live upgrade
[22:31:55] <Kronuz> like keeping the new packages ready as they come out (like in sync)
[22:32:15] <Kronuz> oh... live upgrade... hmm
[22:32:33] *** miffe has quit IRC
[22:32:42] <hspaans> Kronuz: turn it on before going to the shop and is finished when I'm back
[22:32:43] *** lloy0076 has joined #opensolaris
[22:32:54] <Kronuz> are those things like Live Upgrade and Sun Studio free with the Solaris?
[22:33:03] <twincest> kronuz: production solaris doesn't work like that, it doesn't integrate new things the moment they're available.  any updates will be available via patches
[22:33:07] <delewis> Live Upgrade, yes
[22:33:13] <lloy0076> I just downloaded the DVD parts for on_51 and the unzip program has the license terms in French.
[22:33:14] <delewis> Sun Studio, no, but it is freely available as a separate download.
[22:33:26] <Kronuz> what's the catch? with the free stuff?
[22:33:31] <delewis> Kronuz: there's no cathc.
[22:33:33] <delewis> catch*
[22:33:38] <delewis> it just isn't included in Solaris
[22:33:47] <Kronuz> it's just that this almost sounds too good to be true :P
[22:33:48] <delewis> Sun makes money off support and consulting for Sun Studio
[22:33:54] <delewis> Kronuz: why? gcc is free
[22:33:57] <twincest> kronuz: the free stuff doesn't include commercial support.  (in particular, you don't get all solaris patches)
[22:34:37] <Kronuz> oh, there's the catch, if you don't get the support you don't get the patches, right?
[22:34:38] <Burana> Sun Studio 11 is officialy free only for OpenSolaris Community members....
[22:34:43] <delewis> Kronuz: no
[22:34:43] <twincest> that's not what i said
[22:34:45] <delewis> you get patches
[22:34:48] <twincest> you don't get _all_ patches
[22:34:48] <Burana> ...or at least it was in the beginning
[22:34:49] <Kronuz> o_O
[22:34:50] <delewis> Burana: *wrong*
[22:34:54] <delewis> Sun Studio 11 is free for *anyone*
[22:34:58] <delewis> commercial or not
[22:35:05] <delewis> Sun Studio 10 was free only for OpenSolaris users
[22:35:17] <hspaans> Kronuz: you get all the security patchen and important patches that are needed
[22:35:17] <Burana> delewis: ok then it was like that
[22:35:25] <delewis> and you are entitled to Sun Studio patches if you don't have a support contract
[22:35:35] <delewis> without a support contract for Solaris 10, you are only entitled to security patches
[22:35:38] <delewis> and some recommended patches
[22:35:39] <hspaans> Kronuz: new feature stuf you need to buy or with for the next update release
[22:35:43] <Kronuz> what exactly patches consist of?
[22:35:45] <twincest> most of the patches you don't get are new feature patches that will be available for free in the next update
[22:35:48] <hspaans> s/with/wait/
[22:35:57] <twincest> although the ZFS patch was available to free users, which was nice
[22:36:13] <Kronuz> are those in binary form (the patches)
[22:36:16] <twincest> kronuz: patches are updates to a released version of solaris
[22:36:19] <twincest> yes, they're binaries
[22:36:24] <Kronuz> oh, I see
[22:36:29] <hspaans> twincest: and why you think that is? ;-) pushing zfs ;-)
[22:36:41] <Kronuz> and Live Update does it for you, if I understood correctly
[22:36:46] <delewis> no
[22:36:54] <delewis> patchadd, and friends install patches for you
[22:36:55] <twincest> live upgrade is not for patches, it's for new releases
[22:36:56] *** _dreams_ has joined #opensolaris
[22:36:57] *** jcea has left #opensolaris
[22:37:04] <Kronuz> oh
[22:37:05] <delewis> Live Upgrade just gives you the ability to upgrade to another Solaris release with minimum downtime
[22:37:20] <delewis> (it can also be used as a backup mechanism to ensure you have a working Solaris boot environment somewhere)
[22:37:52] <delewis> you upgrade a separate instance while you're still up, and the only downtime is the downtime it takes to reboot your system into the new boot environment.
[22:38:03] <delewis> and if things are fucked
[22:38:08] <delewis> you just reboot back into the old boot environment
[22:38:11] <delewis> and things are as they were
[22:38:13] *** zsakr has quit IRC
[22:38:13] <Kronuz> hspaans: you said you upgrade every two weeks, that's not really the "official" Solaris then, is it?
[22:38:23] <twincest> kronuz: he was talking about solaris express
[22:38:26] <delewis> Kronuz: you're confusing Solaris Express with Solaris with OpenSolaris
[22:38:29] <twincest> the beta version of solaris 11
[22:38:30] <Kronuz> oh, I see
[22:38:31] <delewis> Solaris Express is evolving into Solaris 11
[22:38:35] <delewis> Solaris 10 is current stable release
[22:38:38] <twincest> that does not get patches, instead there's a new release every two weeks
[22:38:42] <delewis> that is really not based off of OpenSolaris
[22:38:45] <Burana> does anybody know when solaris 11 is out?
[22:38:49] <Kronuz> I'm probably confusing all the three: Solaris, Solaris Express and OpenSolaris
[22:38:56] <delewis> they share separate code bases and Solaris 10 occasionally gets things backported to it
[22:38:59] <Burana> ...or how many sol10 updates are planned?
[22:39:00] <Auralis> Burana: october 2007
[22:39:01] <jbk> when it is done :)
[22:39:08] <delewis> Solaris 11 *will* be an OpenSolaris-distribution
[22:39:15] <delewis> meaning it uses the various OpenSolaris bits, like ON, JDS, etc.
[22:39:43] <hspaans> Burana: expected on every once in a will until 2009 orso
[22:39:49] <Burana> otherway arround, which opensolaris features will never make it into solaris? :-)
[22:39:57] <delewis> Burana: huh?
[22:40:04] <delewis> all releases after Solaris 10 will be OpenSolaris-based
[22:40:09] <delewis> so Solaris will receive all OpenSolaris features
[22:40:12] <delewis> that
[22:40:13] <hspaans> Burana: its not if, but when
[22:40:13] <delewis> er
[22:40:26] <Kronuz> so Solaris Express is the "current" branch of the official Solaris OS
[22:40:29] <delewis> that's like asking which Linux kernel features a specific Linux distribution won't be receiving
[22:40:36] <delewis> Kronuz: yes
[22:40:40] <alanc> October 2007 is a placeholder date in the motd, not an actual expected ship date at this point
[22:40:45] <Kronuz> Solaris is the released version (Solaris 10) and there's just one of those
[22:40:49] <Burana> I mean, theoretically there can be opensolaris distributions that have features that will not be in solaris
[22:40:54] <delewis> Kronuz: correct
[22:40:57] <Kronuz> and OpenSolaris is?
[22:41:12] <delewis> OpenSolaris is merely various bits that are optional or necessary to create an OpenSolaris-distribution
[22:41:14] <twincest> opensolaris is an umbrella for a lot of solaris-releated opensource projects
[22:41:16] <Kronuz> the testbead for Solaris future developments?
[22:41:17] <hspaans> Kronuz: opensolaris is just the code and community
[22:41:28] <delewis> bits are like ON (which is the kernel+userland) and JDS (modified-Gnome), and a few more, like NWS (Networking and Storage)
[22:41:31] <delewis> some of these bits are optional
[22:41:34] <delewis> like JDS
[22:41:38] <delewis> some of them are required, like ON
[22:41:43] <delewis> and even some of ON is optional
[22:41:44] <delewis> like the userland
[22:41:49] <alanc> OpenSolaris isn't just a testbed - it's the actual next release - testbeds are kept in separate branches
[22:41:51] <delewis> which Nexenta has taken and replaced with a GNU userland
[22:41:54] <Kronuz> so the work being done by OpenSolaris is there at Solaris Express?
[22:41:59] <delewis> Kronuz: yes
[22:42:00] <jmcp> delewis: nit, NWS == NetWork Storage .... the networking part is already in ON
[22:42:05] * jmcp pedants
[22:42:08] <delewis> Solaris Express is continually receiving the latest ON bits with each release
[22:42:12] <Kronuz> oh, I think I understand now :)
[22:42:15] <delewis> jmcp: :-)
[22:42:20] <Burana> ...yes that makes it clear
[22:42:24] <Kronuz> so, Solaris Express is not for production, I suppose
[22:42:27] *** rockz has joined #opensolaris
[22:42:28] <delewis> Kronuz: no
[22:42:31] <delewis> it's basically an Alpha
[22:42:40] <Burana> but I read that you can buy support for solaris express?
[22:42:41] <jmcp> Kronuz: some people run SX in production
[22:42:45] * jmcp looks around for benr
[22:42:47] <delewis> when Solaris 11 gets ready to be released customers will be able to pre-evalauate it before release
[22:42:50] <delewis> and then it will be a beta
[22:42:55] <jmcp> Burana: no, you can't
[22:43:09] <jmcp> Burana: since it's not an actual "product" per se, Sun doesn't sell support for it
[22:43:12] <delewis> you can buy a support contract that's pretty worthless.
[22:43:15] <Kronuz> but how stable is it? 'cause some "not for production" software is actually really good for the real thing
[22:43:19] <delewis> it basically gives you access to the Solaris Express bugdb
[22:43:24] <delewis> which I assume is similar to bugs.opensolaris.org
[22:43:29] <Kronuz> jmcp: SX is just the short for Solaris Express, I suppose
[22:43:33] <jmcp> Kronuz: yup
[22:43:36] <Kronuz> :)
[22:43:37] <delewis> Kronuz: more stable than many actual releases of other OSes
[22:43:37] <rockz> I want to make a raid-0 for 2 of my local disks.  Should I use zfs or the old SVM  ?
[22:43:47] <jmcp> Kronuz: and re stability ... http://www.scalewithrails.com/
[22:43:48] <delewis> Solaris development continually requires "FCS quality all the time"
[22:43:52] <Kronuz> delewis: that's good to know :)
[22:43:55] <twincest> jmcp: i thought you can buy an SX license for $99 that includes some form of support
[22:44:05] <delewis> twincest: read up to what I said
[22:44:09] <jmcp> nah, that's a right to use
[22:44:15] <twincest> ah
[22:44:17] <hile_> mornin jmcp
[22:44:24] <jbk> i recall there was some program, but i'm guessing it's not going to be the same as a regular support contract
[22:44:29] <jmcp> hi hile_, one-n-all
[22:44:38] <jbk> which for strictly solaris support really isn't that expensive anyway
[22:44:39] <Kronuz> 'cause some other systems' are really really ugly in non-stable state (and in stable state too) .... *caugh* loonix *cough*
[22:44:43] <jbk> if you want to get it
[22:45:03] <hspaans> jbk: its the basic plan for 125 dollar a year?
[22:45:08] <delewis> Kronuz: there's some rough edges in Solaris Express occasionally, but nothing that effects stability.
[22:45:17] <delewis> you might get some annoying dmesg output, but that's about it usually
[22:45:22] <delewis> or have to symlink a missing lib
[22:45:23] <jmcp> Kronuz: I've been running Solaris Express builds for years, and I've come across very few bugs which make the system unstable
[22:45:31] <rockz> we're getting linux support from Sun too :)
[22:45:32] <jbk> that sounds right, though i'd suggest looking on sun's website..
[22:45:48] <jbk> heh
[22:46:02] <Kronuz> jmcp: so I can just forget about the patches and just use LiveUpdate and use SX... I guess
[22:46:05] <Burana> Solaris Support is actually quite cheap compared to redhat support and oracle unbreakable
[22:46:21] <jbk> you should put quotes around 'un' :P
[22:46:22] <delewis> *very* cheap actually.
[22:46:23] <jmcp> Kronuz: there aren't patches for SX. I would use LiveUpgrade myself
[22:46:25] <delewis> and you get so much more
[22:46:32] <delewis> like pushed patch media
[22:46:34] <hspaans> yes solaris is cheap compared to redhat support. then you all for that
[22:46:39] <Kronuz> jmcp: exactly, that's what I meant ;)
[22:46:40] <hspaans> s/then/thank/
[22:46:41] <delewis> (whereas Sun sends you CDs containing patches as they're released)
[22:46:46] <jmcp> Kronuz: ok, glad we're on the same page
[22:46:54] <delewis> reminds me of HP's HP-UX contracts
[22:47:03] <Kronuz> 'cause SX already's patched
[22:47:07] <Kronuz> (I would guess)
[22:47:10] <delewis> (most sites that of HP-UX systems have drawers and drawers full of patch CDs)
[22:47:13] <axisys> hspaans, sommerfeld well i have a fc adapter with two pairs.. one pair goes to fc0 and other pair to fc1 of the 3500.. is that mean i am looking at the same side of the disks or both.. should i move the fc1 to fc4 to make sure i am looking at all 12 disks
[22:47:28] <Kronuz> and LiveUpdate upgrades from the net, right?
[22:47:29] <axisys> (sorry about the long question) i am not storage expert
[22:47:32] <delewis> Kronuz: no
[22:47:37] <delewis> there is no "upgrading from the net"
[22:47:40] <Kronuz> oh
[22:47:41] <hspaans> Kronuz: no from cd images
[22:47:55] <Kronuz> why is there no upgrade from the net again?
[22:47:56] <twincest> but you can download the DVD image and upgrade from it without burning it
[22:47:58] <delewis> with live upgrade, you either upgrade from a flash archive on the network or disk or mounted ISOs (which can be loopback mounted)
[22:47:59] <Burana> There was a blog entry of some guy that compared support prices, ok to post link?
[22:48:00] <Kronuz> just wondering...
[22:48:03] <twincest> so it's basically from the net, just indirectly
[22:48:04] <Kronuz> I'm used to it
[22:48:28] <hspaans> Burana: tinyurl maybe?
[22:48:29] <delewis> Kronuz: the closest thing Solaris has to "net installs" is a Jumpstart install
[22:48:42] <delewis> where you have a Jumpstart server on your network where the systems boot off of and do installs.
[22:48:45] <Kronuz> hey, so where do I get the Solaris Express?
[22:48:48] <jmcp> delewis: closest thing? gee :)
[22:49:06] <axisys> i have 12 146G disks.. so i am definitely not seeing all disks
[22:49:08] <Burana> http://tinyurl.com/wwfo9
[22:49:09] <delewis> jmcp: :-)
[22:49:20] <jmcp> delewis: I'm having a real pedantic morning
[22:49:23] <twincest> kronuz: btw, you probably want Solaris Express: Community Release rather than Solaris Express
[22:49:31] <Burana> TinyUrl is nice, but the blog has a nice url (www.c0t0d0s0.org :-)
[22:49:34] <delewis> jmcp: well, I'm that way in the morning, too
[22:49:34] <hspaans> axisys: log on to the 3510 before you continue and check you config IN the 3510
[22:49:37] <twincest> kronuz: SX is only released every coupe of months (but has more testing)
[22:49:38] <Kronuz> and to upgrade you have to get the whole DVD every time, right?
[22:49:39] <delewis> usually by the afternoon, I'm fairly relaxed.
[22:49:43] <jmcp> Burana: or for the rest of us, http://www.opensolaris.org/os/downloads/on/
[22:49:50] <delewis> Kronuz: correct
[22:50:01] <delewis> but after you download it, you can just loopback mount the ISO
[22:50:04] <delewis> and do a Live Upgrade
[22:50:08] <delewis> (which is what I do)
[22:50:13] <Kronuz> (that's why I wanted a net upgrade: to save bandwidth) :P
[22:50:14] * delewis tries to remember the last time he actually burnt a DVD
[22:50:19] <delewis> Kronuz: this is 2006. :-)
[22:50:26] <Kronuz> hehe
[22:50:30] <Kronuz> still
[22:50:34] <jmcp> Kronuz: remember, "the network has no latency" ....
[22:50:39] <sommerfeld> axisys: i am not a storage expert either.  and haven't had a chance to play with a 3500 yet.
[22:51:04] <Kronuz> twincest: what's that about two Solaris Express releases ? O_O
[22:51:11] <Kronuz> the community and the other
[22:51:14] *** nwf has joined #opensolaris
[22:51:15] <Kronuz> o_O
[22:51:27] <Kronuz> (just when I thought I understood)
[22:51:31] <axisys> hspaans: hmm.. how do i login to 3510.. i did know i could do that
[22:51:37] <Auralis> community is needed for building opensolaris and is normaly one step newer then normal express
[22:51:53] <jmcp> axisys: telnet to it. rtfm to figure out the default ip address
[22:51:59] <twincest> kronuz: there's SX and SX:CR.  SX:CR is released every two weeks.  every couple of months, an SX:CR release is declared 'more stable' and rebranded as SX
[22:52:15] <twincest> that's it
[22:52:21] <hspaans> axisys: you can use serial (speed is 38400) or network and sccli is also a limited option (for password reset for example)
[22:52:25] <jmcp> though Sun tries for every month with SX
[22:52:34] <Kronuz> oh.. so it goes Solaris 10 -> Solaris 10 w/patches -> Solaris Express -> Solaris Express Community Release (last two being part of OpenSolaris)
[22:53:00] <hspaans> jmcp: SX is just SXCR from a few months ago when I downloaded it
[22:53:02] <twincest> there's no need to build ON as root, right?
[22:53:16] <Burana> You forgto Solaris 10 update releases
[22:53:28] <jmcp> hspaans: yes, but the release cycle for SX is monthly .... mostly
[22:53:33] <jmcp> twincest: correct
[22:53:34] <Kronuz> (sorry about the bunch of questions, but it's a bit confusing when you're not used to it)
[22:53:44] <jmcp> Kronuz: that's one of the reasons we've got #opensolaris
[22:53:51] <Kronuz> :)
[22:53:54] <hspaans> ow
[22:54:14] <Kronuz> do any of you work at Sun as official programmers of Solaris?
[22:54:32] <rockz> Question:  I want to make a raid-0 for 2 of my local disks.  Should I use zfs or the old SVM  ?
[22:54:41] <jmcp> Kronuz: I will (again) from tuesday next week
[22:54:47] <twincest> rockz: what are you going to put on the disks?
[22:54:47] <Burana> rockz: what do you run on it?
[22:54:52] <twincest> jmcp: heh, consulting?
[22:54:55] *** gisburn has joined #opensolaris
[22:54:58] <Auralis> rockz: i would currently use svm for the system disks and zfs for the rest
[22:54:59] <jmcp> twincest: not as such, contracting
[22:55:07] <gisburn> Good moring everyone!
[22:55:09] <Kronuz> jmcp: oh, cool :)
[22:55:15] <hspaans> contracting? ow nice I need you ;-)
[22:55:20] <Kronuz> g'morning gisburn
[22:55:22] <Gman> hi gisburn
[22:55:27] <rockz> twincest: just storage for oracle dump files that will be exported from the oracle DB nightly , just for extra backup, not worthy
[22:55:27] <jmcp> Kronuz: I used to work on the NWS (fibrechannel) consolidation
[22:55:35] <twincest> rockz: i'd use zfs then
[22:55:35] <jmcp> hspaans: can you afford me though, that's the question :)
[22:55:45] <Burana> rockz: that could be a good thing to get familiar with zfs
[22:55:52] <rockz> Burana: ahh okay
[22:55:56] <hspaans> jmcp: no problem platinum++ ;-)
[22:55:59] <Kronuz> jmcp: FC consolidation? hmm.. cool
[22:56:08] <jmcp> hspaans: as long as I get all the dough ... nowurries :)
[22:56:15] <rockz> Burana: I installed Sol10 sparc when zfs was not out.  Can I now install it ?
[22:56:30] <twincest> rockz: ZFS is available if you have current patches installed
[22:56:45] <Burana> rockz: you can upgrade with the solaris 10 update 2 release
[22:56:49] <twincest> (i forget the exact patch #s you need)
[22:56:54] <Burana> rockz: should be easy
[22:57:02] <hspaans> jmcp: *g*
[22:57:03] <Kronuz> jmcp: by the way, you must know this one then... is FC the "best" thing to connect storage to a server (I'm still trying to figure that out, against other technologies such as iSCSI)
[22:57:15] <jmcp> Kronuz: it really does depend on what requirements you have
[22:57:17] <hspaans> axisys: can you connect?
[22:57:26] <rockz> twincest: I did a "smpatch update -L"  , would zfs come with it ?
[22:57:33] <jmcp> Kronuz: in true consultant style, the answer is "It Depends"
[22:57:34] <jmcp> (tm)
[22:57:40] <twincest> rockz: should do, do you have the 'zfs' command?
[22:57:52] <Burana> kronuz: it depends how pays and how much :-)
[22:57:53] <jmcp> rockz: do you get a positive response for "pkginfo -l SUNWzfs" ?
[22:57:59] <Kronuz> jmcp: ... I was hoping to finally unveil the mistery :P
[22:58:02] * hspaans wonders btw what is going to come from the Microsoft/Linux allianences
[22:58:21] <Kronuz> Burana: like if you have the money it's the best, or what?
[22:58:22] <hspaans> rockz: please do a live upgrade or you will break things for sure
[22:58:25] <oxygene> hspaans: more uncertainty for those linux users who are not novell customers, probably
[22:58:31] <jmcp> Kronuz: there is no single best way, it really does depend on your environment and what your needs are
[22:58:36] *** lloy0076 has left #opensolaris
[22:58:47] <rockz> jmcp: hmm, i don't
[22:58:50] <Kronuz> jmcp: if you need speed, for instance
[22:58:53] <Burana> Kronuz: iSCSI is cheap, FC is expensive.... iSCSI has latency, FC not
[22:58:55] <delewis> hspaans: nothing of relevance, I'm sure.
[22:59:03] <hspaans> rockz: many tools are now linked against ZFS since 06/06
[22:59:04] <twincest> okay, hg is confusing.  now it says 'aborting update spanning branches!' when i try to hg pull -u.  what do i do?
[22:59:07] <rockz> it says  "SUNWzfs not found"
[22:59:16] <jmcp> rockz: you need to have that package - integrated with Solaris 10 update 2
[22:59:16] <Kronuz> (not much of integrity... like for just reading files hadly writting stuff)
[22:59:43] <rockz> ahh i see
[22:59:45] <hspaans> delewis: I'm not that sure about that
[22:59:56] <rockz> can I get Solaris 10 update 2 with smpatch   ?
[23:00:03] *** mrdeviant has quit IRC
[23:00:21] <delewis> hspaans: sure I am. that part of the market does not effect me or interest me in anyway.
[23:00:21] <Kronuz> 'cause I'll need to have both things but it could be in different volumes and technologies (reliable writting and consistency for a database and really fast reading access)
[23:00:27] <delewis> and I wish more people would take this point of view, too.
[23:00:30] <jmcp> rockz: dunno. I suspect not
[23:00:36] <delewis> Linux users are too obsessed with "what Microsoft is doing"
[23:00:41] <delewis> screw 'em. forget about it. move on.
[23:00:52] <Kronuz> and I was thinking FC would be great for the first, but for speed, I'm not sure if there's something better
[23:00:59] <hspaans> delewis: I have to deal with managers who think Linux is the next best thing. sorry
[23:01:18] <Burana> kronuz: FC still has less latency, and this is important for databases, not so for fileservers.
[23:01:19] <delewis> hspaans: :-)
[23:01:34] <jmcp> Kronuz: do you mind if your reads and writes for storage to go via the network stack (iSCSI, nfs, nas) or do you want them to go via their own hardcore transport ?
[23:01:43] <jmcp> Burana: depends on your fileservers really
[23:01:45] <rockz> hspaans: what's the command I can use for live upgrade to obtain solaris update 2 ?
[23:02:00] <delewis> obtain?
[23:02:10] <hspaans> start with Sun Download Manager ;-)
[23:02:11] <delewis> you can't obtain Solaris 10 Update 2 with Live Upgrade
[23:02:12] <Kronuz> jmcp: for the fast access thing, it would be better to have it as fast as possible no matter how
[23:02:19] <jmcp> of course
[23:02:20] <Kronuz> (like for serving files)
[23:02:21] <delewis> you download the ISO parts, cat them, loopback mount them
[23:02:26] <delewis> and *then* you Live Upgrade
[23:02:35] <jmcp> Kronuz: my bias is for fc and direct-attached scsi.
[23:02:44] <jmcp> that was true even before I started working on the NWS consolidation
[23:02:51] <Kronuz> so, FC must be really fast then
[23:02:56] <delewis> Kronuz: it is :-)
[23:03:00] <Kronuz> :-)
[23:03:08] <jmcp> NFS, NAS and iSCSI definitely have their place, but they're not top of my list
[23:03:11] <Burana> kronuz: do you have san already?
[23:03:19] <twincest> is FC faster than iSCSI over 10GE?
[23:03:20] <hspaans> Kronuz: not that fast, only 1, 2 or 4 Gbit/s
[23:03:21] <Kronuz> nope, I'm still planning what to buy
[23:03:22] <delewis> it's also a sane protocol
[23:03:25] <rockz> hspaans: Sun Download Manager is the easy way then?  I don't have to download the ISO, cat , loopback mount like delewis described ?
[23:03:29] <delewis> that does what it needs to do and nothing more
[23:03:32] *** bobbyz has quit IRC
[23:03:40] <jmcp> twincest: at that level you really need to ask about costs too
[23:03:45] <delewis> rockz: SDM is just a glorified Java app to download it fo ryou
[23:03:51] <delewis> you still need to follow the steps I just laid-out.
[23:03:52] <Kronuz> Burana: I'm even still deciding the brand :P
[23:04:05] <hspaans> delewis: don't spoil the fun for rockz ;-)
[23:04:12] <Kronuz> first I was thinking EMC for the SAN then IBM now I dont' knwo :P
[23:04:25] <delewis> Kronuz: originally, it all started with Brocade :-)
[23:04:26] <hspaans> HDS? ;-)
[23:04:30] <richlowe> evenin' all.
[23:04:31] <rockz> delewis: ohh
[23:04:40] <hspaans> g'day richlowe
[23:04:56] <Kronuz> hspaans: hmm.. HDS
[23:05:10] <rockz> is there a step by step procedure for this upgrade documented by Sun ?
[23:05:12] <Kronuz> does Sun have SAN solutions?
[23:05:21] <jmcp> Kronuz: when you're purchasing big-end SAN gear, make sure you get professional services included with the sale
[23:05:23] <jmcp> Kronuz: yes
[23:05:23] <Burana> Kronuz: Sun sells HDS
[23:05:49] <jbk> as well as their own gear for mid-tier type stuff (though don't know if they just slap their name on it or what)
[23:05:54] <hspaans> rockz: yes type liveupgrade on sunsolve
[23:05:55] <Kronuz> jmcp: really? are those that difficult to handle?
[23:05:56] <Doc> we also bought a little company you might have heard of call StorageTek
[23:06:07] <hspaans> Storagewho?
[23:06:14] <jmcp> Kronuz: if you don't really have any experience in the area, getting professionals in is damned good idea
[23:06:23] <jbk> heh..
[23:06:24] *** oxygene has quit IRC
[23:06:25] <jmcp> Kronuz: too many people / customers / companies completely fuck it up
[23:06:26] <Kronuz> :P
[23:06:43] <jbk> a friend of mine that deals with a decent size san environment actually is very happy with the mid-tier storagetek stuff
[23:06:45] <Burana> Kronuz: I know the HDS UPS100. Reliable, expensive, management software is not the best
[23:06:46] * jmcp used to work in Sun's support group doing technical escalation management
[23:06:57] <rockz> jmcp: is it even safe to upgrade to update 2 or it's recommended to install the new solaris from scratch ?
[23:06:58] <jbk> says it kicks clariion's butt in terms of $/performance
[23:07:12] <rockz> from the original solaris 10
[23:07:13] <jbk> and this guy is hard to please :)
[23:07:22] <jmcp> rockz: an upgrade should be fine. do you really think Sun would have released an update which wasn't upgrade-safe? :)
[23:07:22] <Kronuz> wat do you say about the IBM DS4700, if you've worked with it
[23:07:31] <rockz> oh, hhehe
[23:07:32] <hspaans> jbk: Dell sell clariion's so?
[23:07:33] <Kronuz> or the EMC CX3-20
[23:07:37] <rockz> just have to hear it from you , that's all :)
[23:07:54] <twincest> richlowe: how do i keep local changes when hg updating without my repository growing two heads?  it refused to update unless i removed my local changes
[23:08:09] <delewis> Kronuz: I've done a lot of work with IBM storage
[23:08:13] <delewis> rock-solid.
[23:08:22] <richlowe> twincest: either commit then, and merge, or use Mq.
[23:08:25] <delewis> mostly with the ESS800 (Shark, which is high-end), and the DS4500 and DS4800
[23:08:34] <delewis> which are mid-range.
[23:08:34] <richlowe> 'commit them'
[23:08:35] <Burana> Kronuz: I recommend that you buy a storage array that virtualizes the disks. Then you can make LUN's of any size...
[23:08:42] <twincest> i did commit them, then my repository grew the extra head
[23:08:53] <richlowe> ... and then merge :)
[23:08:58] <jmcp> delewis: as it happens, as of next week my Director will be the bloke who used to manage the DS/ES storage series @ IBM
[23:09:06] <delewis> jmcp: nice
[23:09:07] <twincest> this is confusing.  :P
[23:09:16] <jmcp> delewis: yup
[23:09:21] <delewis> the ES models actually have a pSeries in the cabinet (and two if you go with the 800 series)
[23:09:24] <delewis> :-)
[23:09:31] <Kronuz> Burana: Don't all do that?
[23:09:44] <delewis> the DS4000 series have glorified xSeries, usually.
[23:09:45] <Kronuz> Burana: virtualized disk
[23:10:09] <Burana> Kronuz: Not all.
[23:10:26] <delewis> personally, I'm still waiting for my Photon to arrive :-(
[23:10:33] <Burana> Kronuz: If the support would be better in europe, I would consider this http://www.compellent.com/
[23:10:40] <Kronuz> I suppose I'll go for the DS4700 as you say IBM is not too bad, and since I already checked it the most
[23:10:45] <hspaans> rockz: http://docs.sun.com/app/docs/doc/817-5505?q=live+upgrade
[23:10:56] <Burana> Kronuz: ben rockwood wrote a little bit about it in his blog
[23:11:08] <Kronuz> ben rockwood?
[23:11:11] <jmcp> Kronuz: under the hood the IBM midrange series has a lot of similarities to the STK stuff
[23:11:18] <Burana> Kronuz: You don't know ben rockwood?
[23:11:23] <delewis> and IBM stuff is being really OEM'd, nowadays
[23:11:24] <Kronuz> <.<
[23:11:25] <Kronuz> >.>
[23:11:30] <Burana> Kronuz: he is a legend :-) www.cuddletech.com/blogs
[23:11:31] <twincest> he's new ;-)
[23:11:33] 
[23:11:34] <rockz> hspaans: thanks alot, going there
[23:11:35] <delewis> I can't remember the name (Eugenia) or something like that
[23:11:42] <delewis> which sells re-branded DS-class storage solutions
[23:11:47] <delewis> (or IBM buys their DS-class series from them)
[23:11:49] <hspaans> cuddletech rules ;-)
[23:12:10] <Burana> hspaans: yes, but he is busy at the moment :-(
[23:12:11] <delewis> I haven't quite figured it out, yet
[23:12:13] <gisburn> cuddlerules tech!
[23:12:23] <jmcp> delewis: Engenio
[23:12:28] <delewis> jmcp: bingo
[23:12:32] <delewis> what's the deal with that?
[23:12:42] <delewis> is IBM buying their storage solutions from them and re-branding them?
[23:12:43] <jmcp> delewis: the name?  who tf knows, not me
[23:12:46] <delewis> or is the other way around
[23:12:48] <delewis> ah
[23:13:03] <jmcp> the first
[23:13:29] <Burana> I'm going to bed. Bye!
[23:13:37] <hspaans> bye
[23:13:50] <Kronuz> Burana: what is he famous for?
[23:13:51] *** Burana has left #opensolaris
[23:13:53] *** lloy0076 has joined #opensolaris
[23:14:18] * hspaans loves zfs snapshots
[23:14:28] <Kronuz> (I feel ignorant now... even more)
[23:14:44] <lloy0076> What's the easiest way to burn DVD images in Open Solaris? I have ON_50...
[23:14:51] <twincest> lloy: cdrw
[23:14:58] <lloy0076> I've tried using Nautilus but it won't recognise that the disc I've put in is actually a blank disc
[23:15:00] *** eboutilier has joined #opensolaris
[23:15:04] <hspaans> Kronuz: we all did when we started so nothing new and you'll learn about it ;-)
[23:15:18] <Kronuz> :-)
[23:15:24] <Error_404> i feel ignorant on a near constant basis
[23:15:37] <Error_404> oh, wait
[23:15:38] <Kronuz> hey, what about load balancing under Solaris
[23:15:39] <Error_404> s/feel/am
[23:15:47] <Kronuz> (both link balancing and servers load balancing)
[23:15:49] <hspaans> *g*
[23:15:52] <jmcp> Kronuz: define load balancing
[23:16:01] <twincest> by link balancing you mean port aggregation?
[23:16:03] <jmcp> Kronuz: ipqos-sorta thing?
[23:16:10] <twincest> 0/river@nvdev:~/onnv-gate>hg merge 57ca8d0bec7f
[23:16:10] <twincest>  local changed usr/src/cmd/man/src/util/nsgmls.src/lib/entmgr_inst.cxx which remote deleted
[23:16:14] <twincest> huh? i didn't change that file
[23:16:25] <Kronuz> well, after much research in the web, I think I found basically two types of network load balancing
[23:16:37] <Kronuz> servers load balancing and link load balancing
[23:16:46] *** bobbyz has joined #opensolaris
[23:17:01] <Kronuz> link load balancing mening basically either a dynamic intalligent DNS
[23:17:02] <hspaans> with IPMP you can connect a server redundant to your network
[23:17:20] <richlowe> twincest: how'd you end up with that file in general?
[23:17:23] <hspaans> DNS loadbalancing is a lie
[23:17:24] <Error_404> gated ?
[23:17:26] <Kronuz> or some other method to balance the load among internet links
[23:17:33] <richlowe> twincest: that's part of the man junk that we clobbered a few months back.
[23:17:35] <twincest> richlowe: i don't know.  shouldn't i have it?
[23:17:39] <richlowe> twincest: it (was) versioned, yet also generated.
[23:17:45] <richlowe> it should no longer be versioned...
[23:17:54] <Stric> hspaans: it's not a lie, but it's not even.. the ones with highest and lowest ip get more traffic than the ones in the middle..
[23:18:03] <richlowe> twincest: how far behind was the workspace before you went trying to update it?
[23:18:24] <twincest> oh crud, i'm in the wrong directory
[23:18:28] <twincest> onnv-gate is the unfixed version
[23:18:33] <richlowe> maybe you aren't.
[23:18:43] <richlowe> I suspect that bloody thing is still there :(
[23:18:52] <richlowe> ah, no, it's not.
[23:18:56] <twincest> no, i am, but i'll let you know if it's still there when i do it in the right directory :)
[23:18:57] <Kronuz> ... the other load balancing (the servers load balancing) would be to have a connection go to one of several servers int a pool that basically do the same thing
[23:19:21] <Kronuz> you've seen the F5 solutions? (BIG-IP and the like)
[23:19:27] *** Cybernd has quit IRC
[23:19:32] <OnkelSchorsch_> hm. on my sparc machines I can use prtpicl -vc temperature-sensor and prtpicl -vc fan-speed but how can I find out how hot the machine is and the fan speed on an x86 machine?
[23:19:50] <Kronuz> but I just thougth those are too expensive and that there should be software doing that.. and there is, but for linux
[23:19:53] <jmcp> OnkelSchorsch_: at this point, you can't
[23:19:58] <OnkelSchorsch_> :(
[23:19:58] <Kronuz> so I was wondering what's there for Solaris
[23:19:59] <hspaans> Stric: both round robin and global directors depend on resolvers and applications that most of the time are broken
[23:20:16] <Error_404> Kronuz: there's probably something
[23:20:35] <osdummy> How do I add some static routes (not default route) in Solaris 10? My NFS client mounts depend on some static routes being there so do I need to write a SMF Service?
[23:20:36] <Stric> hspaans: yep. the one with the lowest ip gets the most.. then the highest ip.. then kinda evenly spread in the middle
[23:20:38] <twincest> hspaans: we did dns load balancing for a long time and it was functional.  (not great, but it works)
[23:21:03] <Kronuz> 'cause I'd much rather spend the money on a SAN or an extra server than on a load balancer that could be done by sofwaare
[23:21:22] *** lloy0076 has quit IRC
[23:21:40] <Stric> Kronuz: it all depends on what you're trying to achieve.. if it's important high availability thingie, then dns won't do the work..
[23:21:56] <Stric> if it's "I want to add more capacity to this not too important service", then dns will do fine
[23:22:03] <Error_404> doesn't gated do what you're looking for?
[23:22:17] <twincest> i thought gated was a BGP implementation
[23:22:19] <rockz> besides the zfs software, is there any other great benefit for upgrading from u1 to u2  ?
[23:22:30] <Kronuz> Stric: what are the solutions for high availability?
[23:22:46] <Kronuz> (other than the "intelligent DNS")
[23:22:48] <hspaans> twincest: try to use it with a mainframe ;-)
[23:22:53] <Stric> intelligent dns isn't HA
[23:23:09] <delewis> HA is Sun Cluster, VCS, or HACMP :-)
[23:23:12] <Kronuz> hmm.. it sounded like that when they tried selling it to me :P
[23:23:17] <delewis> and lots of multi-pathing.
[23:23:18] <twincest> hspaans: i didn't say it's suitable for all applications
[23:23:49] <Stric> Kronuz: what if a client caches the dns results and end up on a broken server? then your "high availability" isn't worth shit..
[23:24:14] <hspaans> twincest: in theory it should work and most people it will work, but when you're more demanding it doesn't add up
[23:24:23] <Kronuz> Stric: supposedly the zones have a 30 seconds life span
[23:24:31] <sommerfeld> Stric: if the client caches the dns results for longer than the DNS TTL the client is broken
[23:24:42] <twincest> sommerfeld: clients are broken
[23:24:43] <Stric> sommerfeld: welcome to the world we live in.
[23:24:44] *** rockz has quit IRC
[23:25:04] <kleppari> is anyone familiar with a bug in sunstudio 11 in regard to window drawing?
[23:25:12] <twincest> sadly "your client is broken" is not a good enough answer most of the time :)
[23:25:19] <kleppari> new windows take ~10 sec. to get drawn
[23:25:28] <kleppari> unless I manually move or resize that window
[23:25:29] <gdamore> a lot of clients cache longer in an attempt to deal with misconfigured zones that basically disable caching
[23:25:39] <Stric> Kronuz: if it's within a controlled (non-broken) environment, dns could work.. if you involve internet, no.
[23:26:28] * hspaans agrees with Stric
[23:26:31] <Kronuz> Stric: So the best is to have a cluster there at that level
[23:26:48] <Stric> Kronuz: well, it depends on what you're trying to do
[23:27:00] <Stric> "I want load balancing" is kind of vague
[23:27:26] <Kronuz> (I've heard a lot about clusters, but I still have to demystify the term)
[23:28:01] <hspaans> Kronuz: sun cluster documentation explains a lot in the beginning and it free to download and read
[23:28:14] <Kronuz> Stric: well, it's probably just that I still have vage ideas of what it is (I know the basics but I suppose it's more complicated than I think it is)
[23:28:26] *** laca_ is now known as lacaAFK
[23:29:09] <Stric> Kronuz: then you need to tell what you're trying to achieve from this.. just added capacity without added security, just added security, added security & capacity, .. etc
[23:29:44] <Stric> (raid0, raid1, raid1+0 ;)
[23:30:30] <Kronuz> Stric: Suppose I need to things: first, I have several internet links (say several T1 or T3) for a company that needs to serv a large website, now the thing would be to balance the load among the links as not just a few get all the traffic
[23:30:34] <Kronuz> how is that called?
[23:30:39] <twincest> richlowe: now i have three heads
[23:30:47] <Kronuz> two things* (that was the first one)
[23:30:51] <richlowe> twincest: hang on.
[23:31:15] <hspaans> Kronuz: OSPF and BGP are going to fix that for you
[23:32:21] <Kronuz> the second one is I have a pool of several servers that are basically mirrors (do the same) and I want to either round robin or something better the incomming connections to each of those servers in the pool (including perhaps SSL decryption beforehand, as to not have each server doing that)
[23:32:24] <jmcp> Kronuz: that's called "design for availability"
[23:32:45] <Kronuz> jmcp: the first one, or both ?
[23:33:02] <twincest> i think i'll just start again again
[23:33:09] <jmcp> Kronuz: both
[23:33:16] <Kronuz> it's just that "design for availability" seems to sound a litte bit more vage than link and load balancing :P
[23:33:30] <jmcp> Kronuz: because it's a higher-level concept
[23:33:33] <Kronuz> yep
[23:33:38] <Stric> the first is link load balancing (and a bit of high availability), the second is high availability
[23:34:12] <Kronuz> Stric: isn't the second load balancing too?
[23:34:20] <Kronuz> (I would guess so)
[23:34:22] *** bengtf has quit IRC
[23:34:38] <Stric> well.. kinda.. yeah..
[23:34:46] <Kronuz> :-)
[23:34:55] <hspaans> that can be a side effect yes ;-)
[23:34:58] <jmcp> Kronuz: for me, a serious clusterware-based solution is required for both your scenarios
[23:35:15] <Kronuz> hmm
[23:35:23] <Kronuz> I really need to start reading more about clusters
[23:35:54] <Kronuz> what exactly would be the difference between a cluster and what I'm suggesting?
[23:36:03] <jmcp> Kronuz: go to sun.com/blueprints
[23:36:04] <hspaans> Kronuz: "building internet firewalls" may also be a good start
[23:36:08] <Stric> if it's important, go with some clusterware.. if it's not too important (like opensource software distribution, nothing you lose money over) then regular roundrobin will do fine
[23:36:10] <Kronuz> do clusters run a single instance of an operating system?
[23:36:18] <jmcp> Kronuz: nope, multiple
[23:36:28] <jmcp> but they provide you with the high availability that your scenarios need
[23:36:30] <jmcp> s/need/require/
[23:36:50] <Kronuz> so clusters are basically more thougth out solutions of what I've just described?
[23:36:59] <jmcp> yes
[23:37:02] <Kronuz> hmm
[23:37:07] <jmcp> if you want to completely over-simplify
[23:37:11] <Kronuz> :P
[23:39:48] <jmcp> Kronuz: do you want to do it right, or do you want to play around at the edges?
[23:40:04] <jmcp> that's the difference between an "enterprise" solution and "poor man's cluster"
[23:40:16] <jmcp> PMC tends to get you yelled at when (not if) it breaks
[23:40:17] <Doc> doing it right sounds like too much work - i'd go for the other one
[23:40:34] <jmcp> both by your management and by the vendor you log support calls with
[23:40:35] <hspaans> Doc: no hire someone ;-)
[23:40:38] <jmcp> Doc: you're a wuss
[23:41:38] <Kronuz> well, I want to get it right... but I'm still lacking of a lot of knowledge
[23:42:09] <hspaans> Kronuz: take courses in that case
[23:42:10] <Kronuz> I have the basic idea and solid foundations but I'm not sure about current technology and best practices
[23:42:14] <jmcp> Kronuz: look at the Sun Blueprints
[23:44:22] <Kronuz> are creating clusters a difficult thing?
[23:44:31] <Kronuz> or is it the hard work in the learning how?
[23:44:37] <twincest> grah, now what? remote: abort: repository hg/onnv/onnv-gate not found!
[23:45:11] <Kronuz> (I'm starting to have a headache)
[23:45:30] <delewis> Kronuz: proper planning is the most time consuming task
[23:45:30] <Kronuz> ... I hope I'm not causing you one too :P
[23:45:38] <delewis> implementation is fairly straight-forward, provided you've got a plan.
[23:46:19] <Kronuz> yep, I'm at the planning stage, and I want to get it right from the start... that's why I'm reading a lot... and that's when I got to the point I needed Solaris ;-)
[23:47:37] <twincest> ah, the repo is broken again.  fun
[23:48:13] <delewis> clustering software isn't really that difficult to understand -- you've got a set of resources that need to be transferred to that node when this node fails, and each node is checking the other's heartbeat to see if it's still alive.
[23:48:26] <delewis> that's basically what all, sane clustering software does and is capable of doing.
[23:49:03] <delewis> things get slightly more complicated when you need to make sure when one system fails the common application between them picks up on the other node without a hitch.
[23:49:15] <gisburn> sickness: ping!
[23:49:19] <delewis> but the clustering software should handle all of that, anyway.
[23:49:24] <Kronuz> delewis: so it's like installing Solaris in differnet machines, and then installing some clustering software to check the status of eachother and do something if someone fails?
[23:49:24] <gisburn> sickness: is hal-9000 offline ?
[23:49:35] *** sch0 has joined #opensolaris
[23:49:36] <delewis> Kronuz: yes
[23:50:00] <delewis> and in the case something fails, the node that's still alive can inherit all the sources *cleanly* of the dead node (disks, IPs, etc.)
[23:50:05] <Kronuz>  I thought one needed some sort of special connections between processors or stuff
[23:50:06] <jmcp> Kronuz:  you install the machines identically (here's where jumpstart helps)
[23:50:15] <delewis> which is why you want shared storage *cough* SAN *cough* between the nodes
[23:50:28] *** axisys has quit IRC
[23:50:31] <delewis> s/sources/resources
[23:50:33] <jmcp> Kronuz: you need ethernet or myrinet for heartbeats between the nodes
[23:50:50] <Kronuz> O_O myrinet :S
[23:50:57] <delewis> ethernet does the task fine
[23:50:59] <jmcp> you have to have shared storage too ... which is why SANs are a really good idea. dual-pathed scsi is acceptable though
[23:51:12] <jmcp> delewis: I know
[23:51:29] <jmcp> just wanted to point out that there are other options... including infiniband
[23:51:32] <Kronuz> and a "grid" means just the way the ethernet is interconected among servers, right?
[23:51:41] <delewis> no
[23:51:48] <Kronuz> :-(
[23:52:24] <Kronuz> infiniband would be the ideal thing for clusters or what? (other than ethernet)
[23:52:25] <delewis> usually a "grid" implies there's some workload being distributed across the nodes
[23:52:28] <delewis> which isn't really HA
[23:52:43] <delewis> *pure* HA is where all systems are identical, and all systems have the same workload.
[23:52:53] <delewis> if one node fails, the workload isn't affected.
[23:53:23] <delewis> with grid computing, you've got one node doing something, another node doing something else, and if that node fails *poof* so much for availability.
[23:53:53] <delewis> usually, when you talk about clustering people tend to get confused, because there's two meanings -- grid computing and HA.
[23:53:56] <Kronuz> delewis: but I meant Cluster Grids
[23:54:21] <Kronuz> at any rate, I suppose it's not just the way they're connected
[23:54:38] <Kronuz> but anyway, how do the incomming network connections work in a cluster
[23:54:44] *** eboutilier has quit IRC
[23:54:45] <hspaans> delewis: you mean like a tandem? ;-)
[23:54:46] <Kronuz> they come to a single node always?
[23:54:47] <delewis> Kronuz: the biggest goal in an HA environment is to make sure that any connections between your nodes have multiple-paths.
[23:54:58] <jmcp> Kronuz: go and read the blueprints, they explain all this stuff really well
[23:55:01] <delewis> that either means the medium that's being used for heart-beat
[23:55:02] <delewis> network
[23:55:05] <delewis> or storage
[23:55:23] <delewis> you want paths paths, preferably isolated from one another (like on different switches, etc.)
[23:55:33] <Kronuz> jmcp: I went there, but there are a bunch of files :S
[23:55:38] <Kronuz> let me check again
[23:56:00] <delewis> Kronuz: this may sound silly, but you may want to check out the IBM Redbooks, too
[23:56:03] <delewis> pertaining to HACMP
[23:56:06] <Kronuz> hehe... can't find the window... I have like 30 windows open with my browser... :P
[23:56:13] <delewis> the first two chapters of every HACMP guide are concepts
[23:56:15] <Kronuz> oh, here it is :)
[23:56:24] <delewis> which is pretty much how *any* clustering suite -- HACMP, VCS, or Sun Cluster work
[23:56:27] <delewis> and proper practices
[23:57:34] <Kronuz> I never thought of actually getting into clusters (I had read about those, but they sounded way too complicated and too "big" for mere mortals like me)
[23:57:53] <sahafeez> ugh! $600 p/night for a hotel in DC
[23:57:54] <delewis> if you're in any sort of real, production environment, you need to know clustering :-)
[23:58:21] <delewis> and clustering and SAN sort of go hand-in-hand anyway
[23:58:28] <delewis> both have the same practices
[23:58:34] <delewis> paths. paths. and more paths.
[23:58:43] <richlowe> sahafeez: look through northern VA too.
[23:58:46] <Kronuz> ...and as it turns out it seems a cluster is exactly what I was ending up doing from scratch (and probably not a very good one as I was designing the whole thing myself) :P
[23:58:53] <richlowe> sahafeez: everything tends to be more expensive actually within DC.
[23:59:02] <hspaans> sahafeez: is that included flatfree roomservice?
[23:59:18] <sahafeez> my new office is 1 block from the white house - do not want to drive ;)
[23:59:46] <Kronuz> okay, there's this BigAdmin, BluePrints and SunDocs... what else?
[23:59:49] <richlowe> sahafeez: metro.





top