[00:03:32] *** andersenep has quit IRC
[00:03:33] *** psychicist has quit IRC
[00:03:52] <TommyTheKid> for what its worth, if someone else says something about not being able to ssh out, getting stuck on "next method: public key" .. they may need to logout of gnome and login again. My session was not happy because when I logged in, I was "nobody" due to NFSv4 issues, and the ssh-agent stuff was not able to prompt me properly.
[00:05:18] <lewellyn> well, yeah.
[00:05:23] <lewellyn> you never gave that piece of data :P
[00:05:23] *** Macer has left #opensolaris
[00:08:03] *** dijenerate has quit IRC
[00:10:16] *** |Inthewings| has quit IRC
[00:10:17] <TommyTheKid> it would sure be nice if the fancy gdm login screen would tell me what my username is when its asking for a password
[00:10:47] <TommyTheKid> is that field available such that I could customize some file to make it appear before the hostname?
[00:11:18] *** Jondice has quit IRC
[00:14:48] *** victori has quit IRC
[00:16:10] *** Mdx4 has quit IRC
[00:16:24] <TommyTheKid> hah, I only just connected the dots today after a nice sleep. I could ssh as any user BUT me :)
[00:16:45] <LaidBack_01> does anyone know what mod_apreq is?
[00:17:48] <TommyTheKid> http://tinyurl.com/2c8l5b5
[00:18:09] <LaidBack_01> trying to install perl modules that play nice with the packaged version of apache is a nightmare... it looks like i need to install a whole new version of apache.
[00:18:47] <TommyTheKid> you might need to use Sun's compiler
[00:19:14] <lewellyn> s/might//
[00:19:21] <LaidBack_01> TommyTheKid, while that's cool and all, and I've used it as well, it's not what I meant. I phrased the question incorerctly.  anyway, it's frustration settling in as to just how difficult changing from something familiar to something new can be...
[00:19:45] *** jlliagre has left #opensolaris
[00:19:49] <lewellyn> LaidBack_01: if you use packaged software, you need to use the packaged bundles. otherwise you end up with a mess, or rolling your own packageset.
[00:20:01] <lewellyn> in that case, you may be better off with a third-party packageset
[00:20:08] <TommyTheKid> as lewellyn said, LaidBack_01 you will need to get the Sun compiler if you wish to compile perl modules against the existing perl, or apache modules against the packaged apache
[00:20:08] <LaidBack_01> lewellyn, that'd be really cool, but there is no package for mod_apreq
[00:20:21] <lewellyn> alternatively, update the sfw stuff to use the perl version you need ;)
[00:20:31] <lewellyn> LaidBack_01: write a spec file :)
[00:20:37] <lewellyn> there are examples in sfe
[00:21:06] <LaidBack_01> that might be the best bet.  Getting tired of the dependency hell that this is.
[00:21:11] <easyE> http://www.graphviz.org/Gallery/directed/fsm.html looks pretty much like what you would want.
[00:21:34] <easyE> Which has this dot representation http://www.graphviz.org/Gallery/directed/fsm.gv.txt
[00:21:37] *** olivier_54 has joined #opensolaris
[00:21:40] <lewellyn> easyE: wrong channel?
[00:21:44] *** dijenerate has joined #opensolaris
[00:21:58] <easyE> oops, sorry.
[00:22:03] <lewellyn> :)
[00:22:25] <LaidBack_01> naw, that's just the dep hell that osol is ...  just a graphical representation.
[00:22:36] *** samc has quit IRC
[00:23:21] <lewellyn> LaidBack_01: create a spec file. submit it to jucr and/or sfe. let everyone benefit. :)
[00:23:40] <lewellyn> who knows, it may even get pulled into the main distro eventually if it turns out to be popular in contrib :)
[00:24:11] <LaidBack_01> I think no one else is stupid enough to try to run bricolage on osol.  It seems like a great idea at first, but it's just an exercise in frustration so far.
[00:24:35] <LaidBack_01> that said, i'll try to write my first spec file.
[00:24:48] *** hrist has quit IRC
[00:25:12] <lewellyn> LaidBack_01: if you need help, there are people ;)
[00:25:31] *** xapiens has quit IRC
[00:25:36] <LaidBack_01> yep, thanks.
[00:27:18] *** hunter has joined #opensolaris
[00:28:02] *** dover has joined #opensolaris
[00:28:10] *** xapiens has joined #opensolaris
[00:28:46] *** _setuid_H has joined #opensolaris
[00:28:54] *** bobbyz has quit IRC
[00:28:55] <TommyTheKid> sfe looks interesting
[00:29:11] *** moepmoep has quit IRC
[00:29:40] *** niq has quit IRC
[00:30:44] *** xapiens has quit IRC
[00:30:56] *** hrist has joined #opensolaris
[00:31:12] <TommyTheKid> especially the SFEfreeradius .spec ;)
[00:31:13] <hunter> I think I've horked this opensolaris box beyond salvage, but damn if I know how I did it.
[00:31:39] *** jdberkeley has joined #opensolaris
[00:32:04] <hunter> Where does opensolaris store the ipaddress associated with a given port?
[00:32:24] *** victori has joined #opensolaris
[00:32:31] <TommyTheKid> hunter: by "port" do you mean network interface?
[00:32:42] <hunter> I assigned the wrong IP to this machine, discovered my mistake, and have corrected the /etc/hostname.igb0 file. But on a reboot, its still trying to use the wrong IP I put in at first.
[00:32:54] <hunter> yes, network interface.
[00:33:12] *** moepmoep has joined #opensolaris
[00:33:22] <jdberkeley> does anybody know when opensolaris 2010 will be released?
[00:33:35] <TommyTheKid> hunter: make sure, if you are using /etc/hostname.INTERFACE files that you have network/physical:nwam disabled and network/physical:default enabled
[00:33:54] <hunter> I do
[00:33:54] *** andersenep has joined #opensolaris
[00:33:57] <TommyTheKid> yes, someone does
[00:34:07] <LaidBack_01> jdberkeley, the topic shows the correct answer
[00:34:13] *** benr has joined #opensolaris
[00:34:13] *** ChanServ sets mode: +o benr
[00:34:25] <TommyTheKid> someone with divination capabilities ;)
[00:34:42] <hunter> TommyTheKid: I enable phys:default and disable the wam before I create that file.
[00:35:07] <jdberkeley> so nobody knows... frustrating indeed
[00:35:21] <TommyTheKid> it will be ready when it s ready
[00:36:03] <jdberkeley> I hope oracle doesnt screw up opensolaris
[00:36:06] <TommyTheKid> we don't need a fleet of extermantor trucks circulating on release day due to a bunch of bugs
[00:36:12] *** hrist has quit IRC
[00:36:35] *** Aria has joined #opensolaris
[00:36:41] *** stukreit has joined #opensolaris
[00:36:51] <CosmicDJ> jdberkeley: http://www.c0t0d0s0.org/archives/6510-Future-of-OpenSolaris.html
[00:36:56] <LaidBack_01> okay, so after building a zone and doing the config via zlogin -C, I'd like to drop back to the global zone.  Sending ~. logs me out entirely.  Anyone else experience this?
[00:37:14] <TommyTheKid> hunter: pastebin the following... ifconfig -a; ls /etc/hostname*; ls /etc/dhcp*; cat /etc/inet/hosts; cat /etc/hostname*; svcs -x network/physical
[00:37:21] <tsoome> you can redefine escape char
[00:37:30] <tsoome> man zlogin
[00:37:40] <TommyTheKid> LaidBack_01: yo can re-define escape key or do ~~.
[00:37:47] <hunter> hmmm.
[00:38:03] <hunter> TommyTheKid: I'll have to pick this up tomorrow - looks like the console server just fell over
[00:38:06] *** jdberkeley has left #opensolaris
[00:38:06] <TommyTheKid> you may have to check the syntax of those commands, I just rattled them off
[00:38:09] <TommyTheKid> oops
[00:38:28] <hunter> I'll save them somewhere and look at those in the morning.
[00:38:54] <hunter> I also am now missing the /rpool/boot directory, and /export/home/$USER refuses to mount.
[00:38:58] <hunter> Something bad wrong.
[00:38:59] *** rulereric has quit IRC
[00:39:01] *** victori has quit IRC
[00:39:08] <hunter> And I'm not going to look at it anymore.
[00:39:19] <TommyTheKid> LaidBack_01: for SSH (assuming you used ssh to login to your server with ssh, the escape key is "~" ... so "~." would close the connection, (~? lists other options) ... one of the interesting ones is "~~" which sends an actual ~ to the other side
[00:39:24] <LaidBack_01> so it works for you guys?  I have a snv_134 desktop, snv_134 server, and then the zones of course are the same, when I type that I always end up back at my desktop - that doesn't happen to you?
[00:39:37] <TommyTheKid> hunter: its good to know when to walk away, I usually end up throwing things before I give up :)
[00:39:48] *** ivo_ has quit IRC
[00:40:38] <TommyTheKid> LaidBack_01: ... so use ~~. to pass the first tilde to the zlogin command (instead of the ssh from your desktop) and it will close out of the zlogin
[00:40:48] <LaidBack_01> ah, yep, ~~ does do it
[00:40:55] <LaidBack_01> cool, thanks!
[00:40:59] *** sponix has joined #opensolaris
[00:41:00] <TommyTheKid> welcome
[00:41:23] <TommyTheKid> LaidBack_01: it doesn't get fun till yo have to do 3 or 4 tildes to get to the right place
[00:42:05] <LaidBack_01> heh, thank God for screen.
[00:42:10] <TommyTheKid> ... at that point, it might be worth looking at "man zlogin" ( specifically the -e c) to set the secape caracter to something else
[00:43:13] *** lucalenardi has joined #opensolaris
[00:43:18] *** Calibellus has joined #opensolaris
[00:44:42] <lewellyn> LaidBack_01: ^Aaaaaaaa
[00:44:43] <lewellyn> :)
[00:44:58] *** jpd1234 has quit IRC
[00:45:39] <LaidBack_01> yeah, that sort of implies taht zones can contain zones.  I didn't think that was possible.
[00:45:50] <TommyTheKid> no
[00:45:53] <lewellyn> i was referring to nested screens
[00:46:15] <TommyTheKid> but you can ssh from a laptop to a bastion host, then to a server in a lab, then to another server then zlogin :)
[00:46:17] <LaidBack_01> so... when would you use nested zlogins?
[00:46:29] <LaidBack_01> Oh, okay, i see what you mean.
[00:46:38] <LaidBack_01> yeah, I do that a lot too.
[00:46:49] <lucalenardi> Hi. Anybody knows how to start opensolaris in console mode? My BE doesn't start anymore and I cannot see what is heppening because I'm stacked beyond the blue splash screen.
[00:46:58] <LaidBack_01> course I'm still one of those guys that doesn't use ^D or anything - I just type exit.
[00:47:11] <LaidBack_01> but now with the ~~ bit that'll be easier.
[00:47:21] <LaidBack_01> but I can see what you mean anyway...
[00:48:00] <LaidBack_01> press "esc" and the blue splash screen should go away.
[00:48:15] <LaidBack_01> if you can ssh to the box in question you can disable gdm and then reboot it.
[00:48:20] <LaidBack_01> then it won't boot X.
[00:48:27] <LaidBack_01> er... start X
[00:48:34] <TommyTheKid> you need ~~. if you use zlogin -C (console mode)
[00:48:42] *** Wes- has joined #opensolaris
[00:48:53] <TommyTheKid> or do you mean your global zone
[00:49:35] <TommyTheKid> for a global zone, you edit the kernel$ line in grub to say console=text probably ?
[00:49:44] *** projectmoon has joined #opensolaris
[00:50:03] <TommyTheKid> you can edit /rpool/boot/grub/menu.lst to make the change permanent
[00:50:20] *** hrist has joined #opensolaris
[00:51:06] <benr> ya, -B console=text
[00:51:15] <benr> make sure you update the boot archive after doing so.
[00:51:15] *** falconn19 has joined #opensolaris
[00:51:42] <TommyTheKid> well, a reboot should do that ;)
[00:51:58] <benr> i NEVER take that for granted.
[00:52:03] <benr> always better to just do it.
[00:52:06] <benr> bootadm update-archive
[00:52:08] <benr> simply enough. ;)
[00:52:10] <TommyTheKid> indeed
[00:52:26] <TommyTheKid> oh and reboot -p :)
[00:52:40] <benr> thats right, good call. :)
[00:52:52] <TommyTheKid> fast reboot doesn't appear to read my console=ttya, so I doubt it reads the console=text either :)
[00:53:03] <TommyTheKid> although I consider that to be an issue :)
[00:53:06] *** mlisok_ has quit IRC
[00:53:24] *** victori has joined #opensolaris
[00:53:57] <benr> no, you need to re-engage grub.
[00:54:10] *** rv- has joined #opensolaris
[00:54:54] <rv-> anyone knows how to check who are the leaders for an osug ?
[00:55:05] <TommyTheKid> it reads the kernel line enough to know what file to boot, it could read the options :)
[00:55:44] * TommyTheKid of course has no clue what it really does, just from an end user perspective its mildly annoying, especially now that the live CD fast reboots ;)
[00:56:08] <TommyTheKid> ... that one has the function of confusing the ever loving hell out of me when I actually do the reboot -p several days later and the ISO is still mounted :p
[00:56:41] <TommyTheKid> (the remote console was still running on my sun ray session)
[00:57:04] <TommyTheKid> smrt: explain sfe
[00:57:04] <smrt>  ⌂ See spec-files-extra
[00:57:11] <TommyTheKid> smrt: explain spec-files-extra
[00:57:12] <smrt>  spec-files-extra (SFE) is a collection of popular software which is not yet in OpenSolaris, but should be. The project ports software and creates spec files for packages. You need pkgbuild to use SFE, see: explain pkgbuild , bootstrap-sfe-latest-os20nn , http://pkgbuild.sourceforge.net/spec-files-extra/ , http://pkgbuild.svn.sourceforge.net/viewvc/pkgbuild/spec-files-extra/trunk/ the wiki explains how to use the build environment and gives installati
[00:57:38] *** ewdafa has quit IRC
[00:57:51] <lewellyn> the livecd should eject the cd when it fast-reboots :P
[00:58:24] <Meths> lewellyn: Can smrt split long factoids?
[00:58:27] <TommyTheKid> it probably does, but would that cause the javaRconsole to stop mapping it so that the next time it reboot -p's it doesn't see it
[00:58:49] <TommyTheKid> smrt: explain pkgbuild
[00:58:50] <smrt>  pkgbuild builds software packages for OpenSolaris from recipes (spec files). It is used for JDS (desktop), Source Juicer, and spec-files-extra (SFE) packages. http://pkgbuild.wiki.sourceforge.net/pkgbuild+on+OpenSolaris explains how to use pkgbuild and related tools. #pkgbuild hosts pkgbuild and packaging related discussion. (See also: explain spec-files-extra, bootstrap-sfe-latest-os20nn)
[00:59:42] <lewellyn> Meths: not yet
[01:00:00] <lewellyn> Meths: i have to parse numerics at server join time for that to work effectively
[01:00:44] <lewellyn> Meths: it's better to have linked factlets though, imo.
[01:01:58] *** sergiusens has joined #opensolaris
[01:02:10] <TommyTheKid> i think smrt should take a "to NICK" on the explain so that we dont have to say SoAndSo: do what smrt said :p
[01:02:36] <TommyTheKid> i can make work for anyone :)
[01:02:42] <TommyTheKid> i should be a manager :)
[01:03:01] <eviljames> pft
[01:03:04] <eviljames> you should be in government!
[01:03:05] <lucalenardi> Ok! Thanks the system was hang waiting for the maintenance user. Then a bootadm update-archive solved my problem.
[01:03:12] <TommyTheKid> thats it!
[01:05:17] <lewellyn> TommyTheKid: it's on the todo list
[01:05:31] <TommyTheKid> :)
[01:06:14] <lewellyn> right now, i've gotta make sure bills get paid, then i have more interesting things to work on than smrt. if i got paid to work on smrt, there'd be more neat things in it ;)
[01:08:09] *** Odin- has joined #opensolaris
[01:08:11] <TommyTheKid> i know how that goes
[01:09:43] *** myosound has joined #opensolaris
[01:13:13] *** _setuid_H has quit IRC
[01:13:40] *** lucalenardi has quit IRC
[01:13:49] *** Shoggoth has joined #opensolaris
[01:15:04] *** xapiens has joined #opensolaris
[01:21:02] *** hunter has quit IRC
[01:23:13] <TommyTheKid> damnit
[01:23:29] <TommyTheKid> bootstrap dumped stuff in my home directory
[01:23:43] *** axisys has joined #opensolaris
[01:24:17] *** LaidBack_01 has quit IRC
[01:28:16] *** lattera has quit IRC
[01:29:45] *** ingenthr has joined #opensolaris
[01:30:28] *** ingenthr1 has quit IRC
[01:30:40] <lewellyn> it dumps it in whichever user you tell it to :)
[01:30:51] <TommyTheKid> i want it in /ws
[01:31:03] <TommyTheKid> I shouldn't have heeded the "end user adjustables" comment
[01:31:26] <TommyTheKid> /home/whatever is NFS mounted (aka slow)
[01:31:28] *** myosoun_ has joined #opensolaris
[01:34:05] *** sponix has quit IRC
[01:34:53] *** Laidback01 has joined #opensolaris
[01:35:03] *** myosound has quit IRC
[01:36:03] *** myosoun_ has quit IRC
[01:36:43] *** ingenthr1 has joined #opensolaris
[01:37:02] <Laidback01> okay, so there is no maven in the default package repos of opensolaris?
[01:38:26] <lewellyn> probably not
[01:38:30] *** ingenthr has quit IRC
[01:38:37] <lewellyn> i don't think i've seen it in webstack
[01:38:42] <lewellyn> and i don't think it's in sfw
[01:38:51] <lewellyn> and i don't think i've seen it in sfe or jucr
[01:39:16] <Laidback01> I feel I've made a grevious mistake putting osol as a desktop.  time to go back to something with java sanity.  you'd think that with the group that wrote java they'd have  included the more standard build systems... I understand no hudson... I guess it's good ant is there...
[01:39:37] <lewellyn> i know nothing of maven :P
[01:39:48] <lewellyn> but you also have to keep in mind that solaris got kind of stagnant
[01:39:57] *** comay has quit IRC
[01:40:14] <Laidback01> I'm going home.  I've had enough of this for one day.  maybe tomorrow will be brighter...
[01:40:20] *** myosound has joined #opensolaris
[01:40:23] *** Laidback01 has quit IRC
[01:40:40] *** humpalla has quit IRC
[01:45:16] <TommyTheKid> lewellyn: and it tries to pfexec mkdir -p ${PACKAGESDIR}/SOURCES which failes because root cannot write to NFS home directories... I am going to try just overriding the USERHOMEDIR.. I think that will do the trick
[01:46:00] <TommyTheKid> i realize it says it wont work in all instances :)
[01:46:28] *** rv- has left #opensolaris
[01:49:43] <lewellyn> TommyTheKid: bug tomww ;)
[01:50:37] <lewellyn> ooh. rv isn't nat/sun anymore...
[01:50:59] *** piwi has quit IRC
[01:53:15] <alanc> must be visiting a Oracle building or using Oracle VPN then - MPK17 is still behind the NAT registered to sun.com
[01:55:03] *** Laidback01 has joined #opensolaris
[01:55:19] *** olivier_54 has quit IRC
[01:55:38] <Laidback01> okay, my bad, maven is super easy to install.  just always came preinstalled... sorry about being a bitch.  just tired and against tight deadlines.
[01:55:59] <lewellyn> understandable :)
[01:56:05] <lewellyn> but yeah. make specs. get them into contrib ;)
[01:56:47] <Laidback01> I'll attempt it.  Guess it's the only way it's going to get done is if people just do that.
[01:57:53] <Laidback01> any ideas on cups?  It wants me to login as root.  the Sun/Oracle docs on "Getting started with printing" are no longer availalbe, and I've not found relevant information at opensolaris.  Would like to have a printer...
[01:59:01] <TommyTheKid> assuming you mean on the desktop side... System -> Administration -> Printers
[01:59:07] <TommyTheKid> sorry
[01:59:32] <TommyTheKid> apparently its broken?
[02:00:09] <TommyTheKid> https://localhost:631 I guess
[02:00:52] <TommyTheKid> if you need to login as root, try "pfexec su -"
[02:01:58] <Laidback01> naw it's all good.  Have to add a group called cups, add myself to that group, add a Require @CUPS to the cupsd.conf, restart the scheduler, and I'm good.
[02:02:35] <TommyTheKid> for what its worth, the root password appears to be set to the same password as your initial user you created during install too
[02:02:53] <TommyTheKid> i am being told its time to leave, lights went out
[02:03:09] <Laidback01> have fun
[02:03:17] <TommyTheKid> you have fun, I am going to eat dinner :)
[02:03:39] <TommyTheKid> there is a System -> Administration -> Print Manager that seems to work?
[02:03:45] *** Aria has quit IRC
[02:03:45] <TommyTheKid> does ask for root password (pathetic)
[02:03:48] <TommyTheKid> meh
[02:03:49] <TommyTheKid> nite
[02:04:12] <Shoggoth> has anyone else seen odd behaviour when doing a zpool scrub - I have one pool that was taking forever to scrub but after it hit about 33% it speed up by a factor of 2 or 3
[02:04:14] <Laidback01> yeah, but you can put in any name you want, and a password with hit.
[02:04:52] *** ingenthr has joined #opensolaris
[02:05:02] *** ingenthr1 has quit IRC
[02:06:55] *** Amse has quit IRC
[02:08:20] *** wdp_ has quit IRC
[02:10:24] *** gebi has joined #opensolaris
[02:11:50] *** balrog has quit IRC
[02:13:30] *** balrog has joined #opensolaris
[02:13:35] *** stevel has quit IRC
[02:21:42] <Andys^> Shoggoth: yes, well it goes through the database of blocks in a linear fashion. takes longer when it hits areas of the disk with high fragmentation rates or small blocksizes / small files
[02:22:58] *** ingenthr1 has joined #opensolaris
[02:22:58] *** ingenthr has quit IRC
[02:23:06] <Shoggoth> Andys^: ok... that makes some sense, altough I'm still a little surprised as the filesystem should have low fragmentation as it's a staging area that had data freshly zfs recv'd to it
[02:24:05] <Shoggoth> and small blocks/files I would have expected them to be somewhat evenly distributed... does zfs try to allocate files of similar sizes in the same area of disc?
[02:24:22] <Andys^> nope
[02:24:32] <Andys^> sorry when i say linear fashion
[02:24:37] <Andys^> i didnt mean in the order the files were created :)
[02:24:44] <Andys^> it does a traversal of the tree in its database
[02:25:02] <Andys^> the upside is, it finishes very quickly on empty disks
[02:30:05] <Shoggoth> Andys^: yes an empty disc would be pretty quick - I'm still a bit hazy on the other though... are you saying that the tree clumps together like-sized files>
[02:30:07] <Shoggoth> ?
[02:30:34] <Shoggoth> it would make sense...I've just never bothered digging into that level of detail wrt zfs
[02:31:02] <Andys^> i'm guessing it does a breadth-first traversal
[02:32:13] <Shoggoth> ok... so colour me thick but unless like-sized files are clumped together I don't understand why I would have a step in my performance graph
[02:32:43] <Shoggoth> apologies if I'm not seeing the bleeding obvious... I had about 3 hrs sleep last night :(
[02:34:03] <Andys^> to answer your question you need to look at the source code, i can't tell you any more
[02:34:09] <Shoggoth> fair nuff
[02:34:36] <Shoggoth> I think I'll need a lot of caffeine before I do that kind of spellunking
[02:34:49] <Andys^> heheh, yes
[02:34:53] <Andys^> and even if you find the answer
[02:34:56] <Andys^> will it really help you?
[02:35:03] <Shoggoth> well yes actually
[02:35:09] <Andys^> there's not much you can do about it except fiddle with the zfs_scrub_limit parameter to slow it down slightly
[02:35:15] *** Hunner has quit IRC
[02:35:22] <Shoggoth> my solaris box is actually running paravirtual on top of a linux dom-0
[02:35:39] <Shoggoth> I did that so I could have disc encryption (still waiting on zfs native encryption)
[02:36:10] <Andys^> right..?
[02:36:13] <Shoggoth> but now that lofi has encryption I converted one pool over to use it... I was planning on doing all my pools so I could get rid of the hypervisor and boot solaris on the bare metal
[02:36:26] <Andys^> ok..?
[02:36:40] <Shoggoth> but lofi is shockingly slower that linux device mapper encryption
[02:36:54] <Shoggoth> or at least it would seem from the scrub's I've performed on all my pools
[02:37:09] *** wdp_aao has quit IRC
[02:37:11] <Andys^> ah i see
[02:37:14] <Shoggoth> but now I'm wondering if the data I have in the pool in question is just an example of worst case
[02:37:18] <Andys^> but thats not a really fair test until you run it bare metal
[02:37:34] <Shoggoth> lofi is approx 2.5 to 3 times slower
[02:37:35] <Shoggoth> :(
[02:37:47] <Shoggoth> I ran that pool bare metal
[02:37:49] <Shoggoth> no difference
[02:37:53] <Andys^> oh :/
[02:38:02] <Shoggoth> yes... thus my dilemma
[02:38:08] <Andys^> same type of encryption in both cases?
[02:38:22] <Shoggoth> bare metal vs virtualised... yes
[02:38:38] <Andys^> i think testing speed with zpool scrub isn't a good test either
[02:38:48] <Shoggoth> probably true
[02:39:05] <Shoggoth> but the lofi encrypted pool is also half the size of the others
[02:39:13] <Shoggoth> which is somewhat worrying
[02:39:23] <Andys^> o_O
[02:39:25] *** achester has quit IRC
[02:39:26] <Shoggoth> so actually it's really about 4 to 5 times slower
[02:39:31] <Shoggoth> come to think of it
[02:39:33] <Andys^> its not a good test though
[02:39:39] <Shoggoth> true
[02:39:51] <Andys^> scrub is NOT a benchmark, and is especially not a benchmark if the filesystems have different contents!!
[02:39:52] <Shoggoth> I don't have the spare hardware to do a more scientific evaluation unfortunately
[02:40:04] <Shoggoth> _but_
[02:40:31] <Shoggoth> one of my clients has a few spare TB on one of their disc arrays... I'm hoping to get some time on it to do something more scientific
[02:40:44] *** leito has quit IRC
[02:43:14] <Shoggoth> Andys^: thanks... your comments have gone my gears meshing.... I think I'll hold off doing anymore on my setup until I've had a go with the clients array :)
[02:43:14] *** ingenthr1 has quit IRC
[02:43:18] *** ingenthr has joined #opensolaris
[02:43:26] <Shoggoth> gone = got*
[02:43:55] <Andys^> heheh
[02:43:59] <Andys^> "You know what really grinds my gears?"
[02:44:05] <Shoggoth> pray tell
[02:44:19] <Andys^> no, thats a quote from Family Guy
[02:44:25] <Shoggoth> whoosh
[02:47:04] *** sk0rd has joined #opensolaris
[02:51:42] *** infomomo has joined #opensolaris
[02:51:47] *** infomomo has joined #opensolaris
[02:53:33] *** achester has joined #opensolaris
[02:59:31] <jmcp> Andys^: "sand"
[02:59:55] <infomomo> :D
[03:00:00] *** freakazoid0223 has quit IRC
[03:00:18] <Andys^> heheh
[03:01:12] *** Norrin has joined #opensolaris
[03:01:23] *** cybernd has quit IRC
[03:02:28] * lewellyn bets lblume will show up soon...
[03:02:55] *** panos has joined #opensolaris
[03:05:24] *** sponix has joined #opensolaris
[03:06:46] <Norrin> he always comes after me? lol
[03:11:06] <lewellyn> it's a bit after 9am there
[03:11:38] *** stukreit has quit IRC
[03:13:21] <lewellyn> smrt: explain sfe-bootstrap-latest
[03:13:22] <smrt> Um. I seem to not know anything about sfe-bootstrap-latest...
[03:13:24] <lewellyn> grr
[03:13:48] <lewellyn> smrt: explain bootstrap-sfe-latest
[03:13:49] <smrt>  see bootstrap-sfe-latest-os20nn
[03:13:55] <Andys^> mm.. AMD's 6-core CPU is out
[03:14:04] <lewellyn> ok. i really gotta move smrt elsewhere and poke at new databases :P
[03:14:09] *** leito has joined #opensolaris
[03:14:10] <lewellyn> smrt: explain bootstrap-sfe-latest-os20nn
[03:14:11] <smrt>  a script which sets up a build environment to compile and assemble SVR4 and/or IPS Packages locally from a repository of RPM-style build recipes. It is best used in a separate zone for the first time. Needed tools like CBE, scripts, compilers, pkgbuild/pkgtool, on IPS/OS2009.nn additional packages are installed automaticly. For updates wget and re-run the script from http://tinyurl.com/bootstrap-sfe-latest-os20nn
[03:15:48] <Norrin> Andys^: it's a given the Intel one will be better?
[03:15:52] <lewellyn> tomww: this is the first time i'm running it since i sent my patch to you. far less spurious output. we should figure out how to reduce it further :D
[03:16:51] <Andys^> Norrin: yes, but the AMD one is cheaper. and the motherboards are cheaper too
[03:17:00] *** Syndrome71 has joined #opensolaris
[03:17:29] <Syndrome71> Bork.
[03:19:07] <lewellyn> Syndrome71: you making us meatballs? i'm hungry ;)
[03:19:48] <Syndrome71> Heh. No... But i guess i could...
[03:20:29] <Syndrome71> Love the topic. Says it all...
[03:20:31] <lewellyn> tomww: i think i figured out how to streamline the process more, too :D
[03:21:21] *** Brave has joined #opensolaris
[03:21:50] *** myosound has quit IRC
[03:22:37] *** Syndrome71 has quit IRC
[03:29:22] *** Funk-E has joined #opensolaris
[03:30:22] *** Brave has left #opensolaris
[03:33:51] *** Laidback01 has quit IRC
[03:36:18] *** falconn19 has quit IRC
[03:37:42] <Shoggoth> just read the zpool man page and I'm still not 100% clear - I've got a raidz2 pool and I'd like to take one of the spindles out and use it for something else temporarily and then add it back in... zpool remove, zpool detach, zpool replace - which one?
[03:38:13] <Andys^> detach
[03:38:16] <Shoggoth> ta
[03:38:38] <Andys^> sorry
[03:38:42] <Andys^> wrong :)
[03:38:43] <Andys^> detach is for mirrors
[03:38:44] <taemun> lol
[03:38:46] <Andys^> offline
[03:38:47] <taemun> yeah
[03:38:49] <taemun> haha
[03:38:49] <Andys^> is probably what you want
[03:39:15] <Andys^> remove is for hot spares and cache devices only
[03:39:16] <taemun> Shoggoth: tbh, you could probably just yank it
[03:39:19] <Andys^> detach is to break a mirror in half
[03:39:23] <Andys^> offline is the one
[03:39:30] <taemun> Andys^: detach is for breaking a replace as well
[03:39:32] <taemun> iirc
[03:39:45] <Andys^> then when you want to re-add a disk, use 'online' if the disk is identical to before, otherwise use 'replace'
[03:39:48] <taemun> (I probably dont)
[03:39:54] *** theron has quit IRC
[03:40:15] <Shoggoth> ok so once I've offlined it I can do a zpool create on it - do temporary things and then do a zpool replace and it will resilver...?
[03:41:19] <Andys^> yep
[03:41:27] <Shoggoth> ok.. tx
[03:42:38] *** vraa has quit IRC
[03:48:38] *** gebi has quit IRC
[03:50:31] *** sushiandbeer has joined #opensolaris
[03:52:11] <sushiandbeer> Anyone know what's up with the next release?  I do a news.google.com search everyday for opensolaris, hit opensolaris.com and .org and am unable to find anything posted lately.
[03:52:28] <alanc> only those who read /topic know
[03:53:08] <alanc> (well, they know there's not much we can tell you - it's late, that happens to software)
[03:54:09] <sushiandbeer> ok, thanks.
[03:55:26] <alanc> there were recent posts to opensolaris-discuss about progress, but not completion
[03:57:06] <Andys^> I heard that oracle is trying to delay it because they're evil </FUD> :)
[03:57:24] <sushiandbeer> gracias
[03:58:33] *** perlmongo has quit IRC
[03:59:18] <sushiandbeer> The whole thing is sad - our company is ditching Solaris and SPARC because of the whole uncertainty of it all -- really sad.  Sun/Oracle killed us by discontinuing the lower-end SPARC boxes, so we started down the OpenSolaris route, but now that's going to hell in a handbasket...  bleh
[03:59:36] <sushiandbeer> oh well
[04:00:57] <lewellyn> sushiandbeer: make sure your oracle sales rep knows this ;)
[04:03:51] <sushiandbeer> i would if he'd ever call me back...  since the take-over, they won't even return our calls or emails.  Nice.
[04:04:19] <sushiandbeer> we're not real small, but not huge.  Maybe 50 SPARC boxes ranging from old 4500 to T5240
[04:04:23] *** Funk-E has quit IRC
[04:04:36] <Andys^> sushiandbeer: that is really disappointing :(
[04:04:37] <lewellyn> i'm kind of hopeful that they got rid of some of the more ineffectual sales staff :P
[04:05:04] <lewellyn> sushiandbeer: oracle has "contact sales" buttons all over the site, btw.
[04:05:10] <sk0rd> sushiandbeer: we have less boxes, but they're super important. I don't mean to sound like the kid in the emporer has no clothes, but my sales rep has actually been really helpful
[04:05:15] <sk0rd> yeah, he wants to sell me stuff
[04:05:41] <sk0rd> but he's not retarded and is really willing to work with me
[04:06:00] *** CodeWar has joined #opensolaris
[04:06:48] <sk0rd> I've also found, by following the mailing lists, that thing that people should have been asking there SE's they just refuse to on principle
[04:07:46] <lewellyn> yeah. there's a lot of "we want certain things, but we won't tell the people who can make sure it gets tallied up as something which may make money!"
[04:07:54] <sk0rd> some guy copied nearly all the lists the other day saying "I'm going to buy this, but it's broken"
[04:08:09] <sk0rd> then call your freaking sales engineer and he'll get someone on it!
[04:08:14] <lewellyn> oh that too
[04:08:51] <sk0rd> I'm tried and true a solaris guy
[04:09:02] <sk0rd> I'm a sysadmin, I hate steaks and strippers
[04:09:10] <sk0rd> but there's a time to call on your SE
[04:10:02] <sushiandbeer> LOL -  steaks are nice.
[04:10:15] <CodeWar> as are strippers
[04:10:54] <sk0rd> they are. but I take my work seriously
[04:11:08] <sk0rd> my fat wife is on the couch snoring now
[04:11:08] <CodeWar> what do you guys use Solaris for btw
[04:11:15] <sk0rd> I love steaks and strippers, it's so not that
[04:11:59] <sk0rd> CodeWar: everything
[04:13:49] * alanc uses Solaris to make new Solaris
[04:13:50] <sk0rd> small workgroup stuff like wikis and samba shares to running millions of set top boxes
[04:15:04] <lewellyn> sk0rd: what kind of specs do your STBs have if they're running solaris?
[04:15:45] <CodeWar> alanc,  how recursive
[04:16:38] <alanc> buy all the Solaris you want, we'll make more!   and it won't make you fat like those potato chips that used the same slogan...
[04:17:12] * lewellyn watches alanc patent a perpetual solaris machine
[04:18:00] <alanc> unfortunately, it's lubricated with the tears of owners of old sparc hardware, which is why we make them cry so much
[04:18:32] * alanc suddenly realizes why management never sends him to meet with customers
[04:19:10] <CodeWar> farts?
[04:19:24] <lewellyn> alanc: and that's why you didn't get to man the install lounge in june? ;)
[04:19:57] <alanc> well, that also had to do with my plane not landing until like 3pm the day of communityone
[04:20:21] <lewellyn> excuses ;)
[04:20:27] <CodeWar> but seriously why bother selling Solaris when you can put it in a large refrigerator like box...
[04:20:32] <CodeWar> and sell the box for a million bucks
[04:20:42] <alanc> besides, I show up at those things, and people expect me to know how to make their laptop video drivers work, and man, that shit's hard and confusing...
[04:20:44] <lewellyn> CodeWar: to make people buy the boxes
[04:21:18] <lewellyn> alanc: i wouldn't have. my i945 was happy right then ;)
[04:21:59] <alanc> you know, we'll probably turn /dev back on just in time to push the build with the new KMS-enabled i915 driver, and you can find all sorts of fun new bugs!
[04:22:21] <lewellyn> except that i have other issues precluding me from using that laptop atm :P
[04:22:31] <lewellyn> like... i kinda stole its ram ;)
[04:22:35] * lewellyn should buy more
[04:22:46] <CodeWar> hahaha listen to wat Dr Phil ha to say "we dont see a fat lion .. a fat coyote because animals eat to survive:"
[04:22:50] <CodeWar> mans a genius
[04:22:51] <freetown2> is the OpenSolaris desktop dream over?
[04:23:02] <Andys^> freetown2: no, just the server dream ;)
[04:23:17] <alanc> we had a dream?
[04:23:23] <Andys^> i did :(
[04:23:24] * lewellyn refrains
[04:25:03] <freetown2> aw man. Where else am I going to get a half hourly/for 12 hrs per day/five days a week snapshot solution?
[04:25:35] <DerSaidin> is sigquit suppost to make programs coredump?
[04:25:35] <lewellyn> netapp! ;)
[04:26:08] <lewellyn> DerSaidin: man termio
[04:26:34] <lewellyn> (simple answer: usually)
[04:26:34] <freetown2> lewellyn, netapp can do that? I forgot the 'for a period of three months' too
[04:26:43] <DerSaidin> ah, cool
[04:26:49] <DerSaidin> thanks
[04:26:51] <lewellyn> freetown2: dunno. ask your oracle sales rep ;)
[04:27:07] <alanc> DerSaidin: yes, that's why it says "Core" as the default action in "man signal.h"
[04:27:40] <freetown2> lewellyn, last time I spoke to a Sun rep, i got ignored...you sure the Oracle side is any better?
[04:27:56] <freetown2> s/rep/sales/
[04:28:05] <freetown2> in case that gets misconstrued
[04:29:53] *** tosh has quit IRC
[04:30:23] <CodeWar> most companies in the valley I know of use netapp for their backend storage ...
[04:30:36] <CodeWar> havent heard of any that uses a Sun solutoin though would be delighted to hear that change
[04:31:03] *** steleman has joined #opensolaris
[04:31:58] *** mroconnor has joined #opensolaris
[04:32:50] <freetown2> CodeWar, well, i use samba for file serving and zfs's snapshot capabilities to provide file recovery due to end user mistakes for the last three months
[04:32:55] <freetown2> can netapp do that?
[04:32:55] *** steleman has left #opensolaris
[04:33:02] *** derekv has quit IRC
[04:33:24] <CodeWar> freetown2, I m no filesysteme xpert but snapshots I m pretty sure is part of netapp
[04:33:36] <jbk> yes, but there's a 255 snapshot limit/vol
[04:33:51] * freetown2 throws netapp out the window
[04:33:53] <jbk> and a 512(?) limit on # of volumes
[04:34:02] <Andys^> freetown: carefully, that could kill someone down below :)
[04:34:03] <jbk> or was it 256 for that as well
[04:34:04] <freetown2> ZFS rules!
[04:34:10] <jbk> and that's across a cluster if you're clustering
[04:34:54] <CodeWar> is there a site that does a comparison of a complete netapp solution with a complete hw+sw solution based on ZFS
[04:35:03] <Andys^> Netapp also doesn't support compression, and dedupe only works in background
[04:35:31] <freetown2> what's this dedup that has been surfacing lately?
[04:35:34] * alanc has an old Netapp filer at home - it's currently being used as a printer stand, because really - who needs all that noise for a raid array of 6 9gb disks?
[04:37:21] <CodeWar> its a great place to work i ve been told.. even pay you $9K each time your wife/significant other is pregnant (seriously)
[04:37:44] <freetown2> cor....dedup sounds useful...i wonder how many duplicates of video and picture files i have here...
[04:38:00] <Andys^> CodeWar; the australian government paid me $5k to have a baby ;)
[04:38:34] <freetown2> you guys just reminded me of a joke made on an Indian couple that ended up have over 40 children
[04:41:13] <sk0rd> ha
[04:41:51] <skullone_> the opensolaris dream is over? ouch
[04:42:04] <CodeWar> actually I m surprised to hear Sun is such a serious contender in storage.. the names I d heard so far as a part time storage spectator was EMC/netapp
[04:42:48] <skullone_> i find it hard to believe that sun has one of the better filesystems around, but just cant find a place for it in the market
[04:43:48] *** ryoohki has joined #opensolaris
[04:44:30] *** sushiandbeer has quit IRC
[04:45:38] <jbk> skullone_: don't underestimate the ineptness of sun marketing =]
[04:45:46] *** nikolam has quit IRC
[04:46:33] <skullone_> if they couldve just made money with their ineptness like microsoft did
[04:47:29] *** lattera has joined #opensolaris
[04:47:33] <CodeWar> microsoft sales /OEM handlers are anything but inept .. who said they are?
[04:47:43] <skullone_> im confused by compellent making a ZFS storage appliance
[04:49:07] <skullone_> wouldnt they be at a disadvantage by relying on sun to push out updates to zfs and in-kernel cifs/comstar?
[04:50:09] *** CodeWar has quit IRC
[04:53:09] *** nikolam has joined #opensolaris
[04:53:22] <sk0rd> see, there's nothing to compare strategy to, which conuses the crap out of everyone
[04:53:42] <sk0rd> sun HAD no way of pushing out updates to something awesome
[04:53:51] <sk0rd> because solaris is the tortoise
[04:53:59] *** het has quit IRC
[04:54:00] <sk0rd> and opensolaris was the hare
[04:54:22] *** het has joined #opensolaris
[04:55:06] <sk0rd> but, I LIKE the tortoise
[04:55:39] <sk0rd> because I like spending time with my kick ass kids
[04:55:53] <lewellyn> heh
[04:56:14] <sk0rd> the 3 yr old boy actually asks me to put on detroit style techno when in the car for christs sake
[04:56:22] <sk0rd> it's very shredded wheat
[04:56:32] <skullone_> what?  =
[04:56:34] <skullone_> =/
[04:56:36] <sk0rd> BUT FROSTED IS SOOOOO GOOD
[04:57:06] <zedrich> dont mind him - he's high
[04:57:17] *** lblume has joined #opensolaris
[04:59:06] *** Wild_Geese has quit IRC
[04:59:30] *** Wild_Geese has joined #opensolaris
[05:00:30] *** LaidBack_01 has joined #opensolaris
[05:01:09] <sk0rd> ha, I wish, I'm on call
[05:01:17] *** het has quit IRC
[05:02:03] <sk0rd> It's the crux really, there's tons of people out there that were like "I can leave this alone" and for everyone one of us there's a sales guy
[05:02:52] *** Spencer_tt has quit IRC
[05:04:16] *** infomomo1 has joined #opensolaris
[05:05:15] *** sk0rd has quit IRC
[05:05:35] *** infomomo has quit IRC
[05:08:37] *** Aria has joined #opensolaris
[05:19:26] *** Spencer_tt has joined #opensolaris
[05:20:16] <lewellyn> lblume: moo
[05:22:24] <stallion> Hah!
[05:22:28] <stallion> efe is *done*
[05:22:51] <lewellyn> it passed NICDRV?
[05:22:56] <stallion> Now to finish the documentation and get NICDRV cooking
[05:23:03] <lewellyn> heh
[05:23:15] *** sushiandbeer has joined #opensolaris
[05:23:20] <stallion> not yet - but its passed my own smoke testing; transferring > 2GB at wirespeed with no problems in a tight loop
[05:23:24] <stallion> so NICDRV should be cake
[05:23:35] <lewellyn> famous last words
[05:23:43] <stallion> heh, NICDRV and I are old friends ;)
[05:23:44] <lewellyn> watch it work too well for nicdrv or something ;)
[05:24:00] <stallion> trust me - after getting dnet to pass NICDRV, I'm pretty familiar with the f'n thing
[05:24:45] * lewellyn actually doesn't know much about NICDRV
[05:24:52] <lewellyn> does it actually need the hardware installed?
[05:24:53] <stallion> I'll have to start an integration branch so I can get the spwr renaming in
[05:24:59] <stallion> on the client side, yes
[05:26:03] <lewellyn> ah
[05:27:14] *** soro has joined #opensolaris
[05:32:25] <stallion> I've not tried NICDRV on Indiana yet, so this should be interesting
[05:32:57] *** CodeWar has joined #opensolaris
[05:34:08] <CodeWar> pm-updatemanager seems to think I have an update pending on package dircproxy but when I proceed to update it .. it fails saying it was already updated.
[05:34:17] <CodeWar> who can help me shut  updatemanager up
[05:34:25] <lewellyn> CodeWar: is it in contrib?
[05:34:53] <CodeWar> lewellyn, /pending actually
[05:35:10] <CodeWar> pkg://pending/dircproxy at 1 dot 2.0.1,5.11-0.111:20100423T054958Z
[05:35:28] <lewellyn> heh. pending is "you get to keep the pieces" ;)
[05:36:01] *** CodeWar has quit IRC
[05:36:10] *** CodeWar has joined #opensolaris
[05:37:42] * lewellyn grumbles
[05:37:47] <lewellyn> how do i get info on that? :P
[05:37:56] <lewellyn> pkg: 'dircproxy at 1 dot 2.0.1-0.111' matches multiple packages
[05:38:26] <CodeWar> lewellyn, funny thing is I didnt even have that package installed ... I just installed it from /pending dont know why updatemanager thought I deserved this package
[05:39:17] <lewellyn> i blame roboporter
[05:40:26] <lewellyn> that timestamp makes no sense
[05:40:31] <lewellyn> http://jucr.opensolaris.org/review/packages/1690/
[05:40:38] <lewellyn> dircproxy at 1 dot 2.0.1,5.11-0.111:20091022T004800Z
[05:40:47] <lewellyn> that's not the timestamp you gave
[05:41:10] <lewellyn> maybe they're rebuilding everything so 2010.1H will work, finally :P
[05:42:15] <CodeWar> just remembered I m not using anything from /pending might as well remove it
[05:42:27] <lewellyn> i forget how to verify :P
[05:45:11] <lewellyn> what are SUNWkvm and SUNWcar again?
[05:45:45] *** notdk has joined #opensolaris
[05:45:59] *** projectmoon has quit IRC
[05:46:09] *** kukhuvud has quit IRC
[05:46:39] *** notdk has quit IRC
[05:46:46] <CodeWar> time to go watch Lost
[05:47:00] <lewellyn> i already am
[05:47:18] <CodeWar> they play it early in canada?
[05:47:26] <lewellyn> i'm already lost
[05:47:32] <lewellyn> i don't watch tv. no time ;)
[05:47:33] <CodeWar> :-)
[05:47:40] <Norrin> what time does http://abc.go.com post the new V episode?
[05:47:43] <Norrin> that's what i'm waiting on
[05:47:55] <lewellyn> and how would i know if it's early in canada? :P
[05:47:57] <CodeWar> V"? oh that crazy thing after Lost?
[05:48:19] <CodeWar> 10pm PST is when it goes live
[05:48:25] <Norrin> V & caprica only currently running shows i watch
[05:48:43] *** lattera has quit IRC
[05:48:44] *** vraa has joined #opensolaris
[05:48:52] <CodeWar> used to watch prison break .... that used to air  an hour early in Canada
[05:49:46] * Norrin on episode 19 of X Files right now though (netflix)
[05:50:14] *** LaidBack_01 has quit IRC
[05:50:23] <CodeWar> actually its a good idea to watch it when its all done rather than wait a week and see
[05:50:41] <TommyTheKid> i wait to download Lost from eztv, no commercials,  x264, 24" monitor.. done and done
[05:51:46] * lewellyn wonders what SUNWcar got renamed to
[05:51:55] <vraa> do you watch x264 on opensolaris?
[05:52:26] <TommyTheKid> um, no :)
[05:52:58] <lewellyn> ok... SUNWcar became system/library/processor
[05:53:16] <lewellyn> so that explains that one... i still wonder what SUNWkvm is
[05:53:21] <stallion> ahh
[05:53:24] <stallion> code complete. love it.
[05:54:25] <lewellyn> oh well. i guess i'll never know.
[05:55:24] *** relling has quit IRC
[05:56:28] <stallion> kvm?
[05:56:29] <stallion> one sec
[05:56:50] <TommyTheKid> cant you find that with pkg search ?
[05:56:52] *** kukhuvud has joined #opensolaris
[05:56:54] <stallion> Core Architecture, (Kvm)
[05:57:01] <lewellyn> yeah. i got that much. ;)
[05:57:10] <stallion> pkginfo -x :P
[05:57:18] <lewellyn> SUNWcar was Core Architecture, (Root)
[05:57:55] <lewellyn> i did try the basic legwork before muttering on irc :)
[05:58:52] <lewellyn> hm. graphicsmagick isn't in any repo that i'm finding immediately. that's somewhat unfortunate
[06:01:07] <lewellyn> oh well. off with me! :D
[06:01:16] <stallion> lewellyn: feel up to some testing?
[06:01:30] <lewellyn> of?
[06:01:35] <stallion> efe
[06:01:50] <lewellyn> what hardware does that require again?
[06:02:04] <stallion> SMC EtherPower II
[06:02:11] <lewellyn> yeah. i don't have that lying around :P
[06:02:15] <stallion> gotcha
[06:02:23] <stallion> oh well, I need to get a webrev ready anyway
[06:02:27] *** Aria has quit IRC
[06:02:33] <stallion> not to mention update onnv-clone :/
[06:02:36] <lewellyn> i've got lots of dnet cards around it seems
[06:02:41] <lewellyn> and rge :P
[06:03:01] <alanc> graphicsmagick?  that's different from imagemagick?
[06:03:02] *** sushiandbeer has quit IRC
[06:03:54] *** LaidBack_01 has joined #opensolaris
[06:04:49] <lewellyn> alanc: yes. :)
[06:05:03] <stallion> lewellyn: heh. doesn't help me much ;)
[06:05:10] <lewellyn> http://www.graphicsmagick.org/FAQ.html#how-does-graphicsmagick-differ-from-imagemagick
[06:05:11] <alanc> oh, since there's plenty of imagemagick in the repo
[06:05:20] <stallion> I need to do the spwr rename anyway or gdamore will probably smack me around a bit before filing the PSARC case ;)
[06:05:42] <alanc> ah, I see
[06:05:53] <lewellyn> GraphicsMagick is usually faster than ImageMagick. The baseline execution overhead for simple commands is much lower, and GraphicsMagick is also more efficient at dealing with large images.
[06:05:59] <lewellyn> that's why i need it :P
[06:06:06] <alanc> don't tell joerg, since he already thinks imagemagick should be banned for stealing his convert command name
[06:06:10] <lewellyn> imagemagick pukes on 300 meg tiffs for some reason :P
[06:06:12] <stallion> jmcp: around?
[06:07:23] *** het has joined #opensolaris
[06:09:20] <lewellyn> now. i'm really gone. hopefully gm will be easy to write a spec for tonight :)
[06:11:19] *** achester has quit IRC
[06:14:09] <gosx> should i generally be offloading checksums etc in my ethernet adapter?
[06:14:14] <gosx> (bnx)
[06:14:27] <gosx> if my CPUs aren't loaded, does it matter?
[06:17:52] *** sushiandbeer has joined #opensolaris
[06:18:30] *** paddy has quit IRC
[06:20:54] *** radsy has quit IRC
[06:21:53] *** CodeWar has quit IRC
[06:22:34] <stallion> alanc: could you triage a bug for me?
[06:22:43] <alanc> sure
[06:22:49] <stallion> just submitted one for spwr
[06:23:23] <alanc> ah, was going to guess it wasn't 6947728 JM363 AHCI - fails to configure connected disks
[06:23:41] <alanc> 6947888 Provide an open source implementation of spwr(7D) seemed more likely
[06:23:44] *** derchris has quit IRC
[06:23:49] *** derchris has joined #opensolaris
[06:24:01] <stallion> yup
[06:24:12] <alanc> should I just assign it to gdamore or leave it to let him think he's volunteering?
[06:24:27] <stallion> go for it, I think hes expecting it by now ;)
[06:24:53] <alanc> done
[06:24:58] <stallion> excellent. tyvm.
[06:26:46] *** lateau has joined #opensolaris
[06:26:55] *** LaidBack_01 has quit IRC
[06:27:09] <alanc> theres a whole 3 other spwr bugs open - wonder how many of those you'll fix as side effects
[06:29:27] <alanc> one is just a driver_aliases addition for a new model
[06:29:50] <alanc> well, it was new in 1999, when the bug was filed
[06:30:06] <alanc> 4232450 spwr: need to add driver_aliases entries for SMC9432TX/MP
[06:34:55] <stallion> interesting
[06:35:20] <stallion> alanc: do you recall which build integrated the IPS bits into ON proper?
[06:35:25] <alanc> 136
[06:35:26] <stallion> I'm not seeing anything interesting in snv_134
[06:35:29] <stallion> figures ;)
[06:36:06] <alanc> since I just had that flag day page open to grab the SVR4 packages from for bootstrapping a SXCE machine into IPS
[06:36:27] <alanc> yeah, after it slipped past 131, it was pretty much going after 2010.03 branched
[06:37:04] <skullone_> anyone know whats up with: CR 6924824   destroying a dedup-enabled dataset bricks system
[06:37:19] <alanc> you can tell which it is if your gate has a pkg subdir instead of pkgdefs
[06:39:03] <stallion> ahh gotcha
[06:39:36] <alanc> http://static.opensolaris.org/on/flagdays/pages/20100302184109.html has more details
[06:39:39] <stallion> this is going to be a pain - GLDv3 was integrated in 136 (along with the IPS bits), and I can't get a damned workspace built without doing it from scratch
[06:39:45] <stallion> ty
[06:44:21] *** freakazoid0223 has joined #opensolaris
[06:46:00] <sickness> alanc: omg 1999? solaris 2.6? =)
[06:46:01] <sickness> ghgh
[06:46:45] <alanc> bug says Solaris 8 actually - must have been during S8 development
[06:47:06] *** xapiens has quit IRC
[06:47:07] <alanc> but still, quite moldy
[06:47:28] <sickness> eheh
[06:47:56] <stallion> alanc: looks good: http://cr.opensolaris.org/~stallion/6947888/
[06:48:16] <alanc> stallion's work does fix the root cause blocking that issue - "Need to work with vendor for update"
[06:48:42] <stallion> from what I understand, the original spwr was pretty nasty
[06:48:54] *** olivier_54 has joined #opensolaris
[06:48:59] <stallion> this driver is actually pretty clean - even performs well (no issues w/ wirespeed)
[06:49:01] <alanc> umm, you will need to change the copyright headers, but to do that properly, you need your sponsor's help
[06:49:12] <stallion> which headers?
[06:49:33] *** RoyK has quit IRC
[06:49:50] <alanc> the ones that say "Copyright 2010 Sun...."
[06:49:56] <alanc> http://static.opensolaris.org/on/flagdays/pages/20100324114440.html
[06:50:08] *** hemantha has quit IRC
[06:50:32] <stallion> bah
[06:50:32] <alanc> since oracle uses first & last year of edits format, someone with access to the full internal history has to look up the dates for the first year
[06:50:42] <stallion> thats going to be a PITA
[06:50:46] <alanc> casper has a magic script to make it happen
[06:50:50] *** hemantha has joined #opensolaris
[06:50:51] <stallion> thats handy
[06:51:04] <stallion> fortunately my webrev is against snv_134! :D
[06:51:24] *** sushiandbeer has left #opensolaris
[06:51:30] <stallion> I'll have to bug Garrett later though - getting this through C-Team is going to be difficult until IPS builds resume
[06:51:34] <alanc> besides, you only have 3 files that aren't brand new - your life is much easier than when Olga hit this on the ksh93 upgrade in snv_138
[06:51:42] <stallion> ooph I bet
[06:52:38] *** het has quit IRC
[06:53:47] <lewellyn> poor olga :P
[06:54:03] <lewellyn> are they still doing giant batches of patches? :(
[06:54:40] <lewellyn> also, does that mean every file in every consolidation now has a © 19xx-20yy header?
[06:55:49] <stallion> lewellyn: as they get modified, yes.
[06:56:55] *** Zauberpo1y has joined #opensolaris
[06:57:09] *** hemantha has quit IRC
[06:57:21] <lewellyn> ah good. i was afraid someone had to go through and make the larry happy with copyright on every extant file :P
[06:57:46] <stallion> well, you can't really do that and be in compliance with the law
[06:58:01] <lewellyn> no?
[06:58:25] <stallion> IANAL, but I believe in order to change/extend the copyright, you have to edit the work
[06:58:26] *** ryoohki has quit IRC
[06:58:50] *** Zauberpony has quit IRC
[06:58:57] <lewellyn> i thought it was legit to clarify the copyright, however.
[06:59:24] <jmcp> muttermuttermutter
[06:59:30] <lewellyn> ianal, either, though. and i'm not paying one to ask about this ;)
[06:59:52] *** hemantha has joined #opensolaris
[07:00:21] <alanc> lewellyn: each consolidation has the choice of doing a mass update or doing it as each file is updated
[07:00:43] <lewellyn> and both approaches would suck in their own ways :)
[07:00:56] <TommyTheKid> wow, the bootstrap latest has like major chaos going on, its like someone had a copy/paste error in an editor
[07:00:57] *** hemantha has joined #opensolaris
[07:01:02] <stallion> jmcp: efe is done ;)
[07:01:05] <lewellyn> TommyTheKid: chaos how?
[07:01:18] <lewellyn> TommyTheKid: you're using the actual latest one, too?
[07:01:18] <stallion> 1.6KLOC ... not too far off from what I expected.
[07:01:27] <alanc> JDS is doing the mass update, I may do it for X, ON said "OMG!   Are you freaking kidding me!   15000 files!" and chose to make each engineer be responsible for updating the files they touch
[07:01:29] <jmcp> stallion: w00t!
[07:01:31] * stallion was hoping for 1.5K
[07:01:42] <jmcp> what hardware does it need?
[07:01:46] <TommyTheKid> it says bootstrap-sfe-latest-os20nn
[07:01:48] <stallion> SMC EtherPower II
[07:01:51] <TommyTheKid> fiik :)
[07:01:53] <stallion> great chips, just not overly popular
[07:02:16] * alanc is considering the mass update so he doesn't have to answer the questions of what to do on every putback, since he doesn't have a team to delegate those to
[07:02:18] <lewellyn> stallion: are they readily available on craigslist and such? :)
[07:02:28] <stallion> lewellyn: I had to dig on eBay for mine
[07:02:37] <lewellyn> :(
[07:03:15] <TommyTheKid> lewellyn: the verswion named in the file says ... os200905-20100325-1122 .. is that latest?
[07:03:24] <stallion> jmcp: out of curiosity, how the hell am I going to be able to get this integrated without IPS being updated to at least snv_136?
[07:03:36] <stallion> There are some GLDv3 changes that will be made, and the new packaging crap will have to be dealt with
[07:03:41] <lewellyn> echo "Version:  bootstrap-sfe-latest-os200906-20100325-1122"
[07:03:46] <lewellyn> that's the one i just installed
[07:03:53] <TommyTheKid> it seems to sortof startover again around line 200
[07:03:56] <jmcp> stallion: your sponsor inside the firewall will have to do the legwork
[07:04:14] <lewellyn> TommyTheKid: it's fine here. download it again? :P
[07:04:15] <TommyTheKid> (thats approximate cause I have been making some changes as I go thorugh
[07:04:22] <stallion> jmcp: sounds like I'm about to owe Garrett a shitload of beer.
[07:04:31] <TommyTheKid> well its different logic in thhis secttion, but its re-setting things again
[07:04:34] <jmcp> he might settle for a slab, rather than a shitload ....
[07:04:37] <jmcp> :-)
[07:04:46] <stallion> well
[07:05:04] <lewellyn> TommyTheKid: the script could use some cleanup, but hopefully soon it won't need as much hackiness in the first place.
[07:05:08] <stallion> I can at least make the necessary code changes - I can only run NICDRV against snv_134 though
[07:05:24] <lewellyn> at least now (aside from one issue i saw) it deals properly with 80x24 screens better :)
[07:05:52] <alanc> you can always build later bits yourself or even *cough* apply the bfu archives to at least get onnv_137 - does NICDRV depend on much outside ON?
[07:06:02] *** ryoohki has joined #opensolaris
[07:06:07] <stallion> not really
[07:06:19] <stallion> cap-eye-install isn't my idea of a good time though
[07:06:36] <lewellyn> oh, you can still bfu 137?
[07:06:48] * lewellyn should look over the flag days...
[07:06:48] *** cmbntr has joined #opensolaris
[07:06:51] * alanc feels people fainting as he actually recommends BFU - though for it's intended purpose of testing an ON build, not as an upgrade tool
[07:07:10] <stallion> alanc: I don't think i've actually done a buf since snv_36
[07:07:13] <stallion> bfu rather
[07:07:14] <lewellyn> meh. sxce is dead anyhow. i brick the box, i move to indiana sooner ;)
[07:07:22] <alanc> lewellyn: bfu still works, it's  just more hardcore now, since acr is gone
[07:07:23] *** hemantha has quit IRC
[07:07:30] <stallion> ooph
[07:07:31] <lewellyn> meh. that's fine :)
[07:07:34] <stallion> I rather liked acr
[07:07:41] *** hemantha has joined #opensolaris
[07:07:53] <alanc> and you can't have 138 archives until they fix uploads to dlc post-data-center-move
[07:07:57] <lewellyn> alanc: you wouldn't happen to have semi-recent svr4 packages of x, would you? ;)
[07:08:16] <lewellyn> maybe i'll learn how to actually build sfw and have an almost-consistent system.
[07:08:23] <alanc> internally yes - haven't posted externally lately
[07:08:55] <alanc> probably will get some Xorg 1.8 test packages out at some point soonish though, if we're going to make our claimed integration of 141
[07:09:19] *** hemantha has quit IRC
[07:09:29] <alanc> I could post 137 packages externally, if push to dlc was working...
[07:09:52] <lewellyn> well, i'm not tackling that tonight anyhow. i have other things to do before passing out :)
[07:10:06] *** hemantha has joined #opensolaris
[07:10:12] <alanc> 138 ones even for x86 (hopefully 138 sparc packages will be built by tomorrow, now that I've fixed the damage I did to the buildmachine)
[07:13:37] <lewellyn> any pkgtool users about? :)
[07:15:31] *** cmbntr has quit IRC
[07:15:54] <stallion> somewhat
[07:16:05] <alanc> only an occasionaly user, and only of the SVR4 output-ing version
[07:16:24] <lewellyn> yeah. i'm trying to set up my pkgtoolrc for ips use
[07:16:36] <stallion> oh?
[07:16:47] <lewellyn> WARNING: /export/home/osol/.pkgtoolrc: Unknown option "ips"
[07:16:48] <lewellyn> :(
[07:16:49] * alanc saw you ask in #pkgbuild, so knew how to escape....
[07:16:49] <stallion> All I ever set is .pkgbuildmacros
[07:16:54] <lewellyn> alanc: :)
[07:16:58] <lewellyn> INFO: IPS packages will be installed by default from http://localhost:10000/
[07:17:03] <lewellyn> that's also incorrect
[07:17:17] *** sysspoof_ has joined #opensolaris
[07:19:55] *** DukeOfMilan has joined #opensolaris
[07:20:51] *** merzo has quit IRC
[07:21:35] *** ryoohki has quit IRC
[07:22:29] * jmcp realises where the stmsboot -L bug went off the rails
[07:22:32] * jmcp gets annoyed
[07:23:50] <jmcp> seriously, folks, if you turn on MPxIO and *then* add targets to a system, DO NOT expect to see a non-MPxIO path in stmsboot -L -- it *never* existed in the first palce
[07:23:51] <jmcp> place
[07:24:15] <TommyTheKid> heh
[07:24:22] <TommyTheKid> maybe its bed time for you too jmcp  ;P
[07:24:42] <jmcp> TommyTheKid: it's only 15:30 for me
[07:24:48] <jmcp> plenty of time left in the day
[07:25:01] * jmcp gets out a tool d'thwapping
[07:26:39] *** het has joined #opensolaris
[07:27:03] <stallion> alright gents - time for bed
[07:27:05] <stallion> night all
[07:27:06] <jmcp> fnite
[07:27:07] * stallion &
[07:27:07] <jmcp> gnite
[07:29:25] *** DukeOfMilan has left #opensolaris
[07:29:38] <TommyTheKid> oh you need to take a break for afternoon treat (cookies)
[07:29:58] *** ingenthr has quit IRC
[07:30:19] <jmcp> TommyTheKid: seriously mate, this is something that's been derailed by people who do not understand how that !#$!@#$!@#$ing utility works
[07:30:29] <TommyTheKid> yes :)
[07:31:00] <TommyTheKid> ok well its like 11:36PM here, so I am going to sleep
[07:31:02] <TommyTheKid> nite!
[07:31:05] <jmcp> gnite
[07:33:24] * lewellyn puts off building things with pkgtool since he can't find docs on how to tell it which kind of package to emit :P
[07:34:08] <jmcp> lewellyn: do you think that this bullet point in a presentation would get the point across:        Check your speeling, and you'r grammer
[07:34:34] <jmcp> I'm going to be talking to a bunch of 1st year software eng students tomorrow about how to do it right
[07:35:00] <lewellyn> Cheque you're
[07:35:08] <jmcp> heh
[07:35:22] <jmcp> well, they will be 1st years, so I might pause for effect
[07:35:59] <lewellyn> cheque doesn't work so well here. but i imagine it'd get the point across there ;)
[07:36:28] <jmcp> actually, not sure it will here either - people pretty much don't have cheque accounts any more, or write cheques
[07:36:43] <jmcp> I've got one, because the facility came free with my mortgage account
[07:37:12] *** nickw has quit IRC
[07:37:13] <lewellyn> but the word isn't from a foreign language, there :)
[07:38:27] <jmcp> yes it is
[07:38:33] <jmcp> it's just less foreign to us
[07:38:44] <lewellyn> heh
[07:39:09] <lewellyn> ok. Czech then
[07:39:33] <Andys^> Chek ur speeling?
[07:40:26] <lewellyn> too obvious. my idea is properly-spelled words which are not the correct words ;)
[07:41:00] <lewellyn> four the wen
[07:43:25] <lblume> It's "chèque", you accent-disabled people :-P
[07:44:46] <lewellyn> è
[07:44:58] <lewellyn> i can do accents fine. but we don't use them here ;)
[07:45:17] <lewellyn> rèsumè becomes resume here
[07:45:38] *** nickw has joined #opensolaris
[07:46:40] <lblume> You miss all the beauty of a finely accented letter :-D
[07:47:18] <lewellyn> trochej can provide those
[07:47:42] <lewellyn> łż etc
[07:47:50] <jmcp> ç
[07:47:51] *** EugeneKay has joined #opensolaris
[07:47:55] * EugeneKay waves
[07:48:04] *** lateau has quit IRC
[07:48:13] <lewellyn> i still need to figure out how to type CCCP
[07:49:11] *** nickw has quit IRC
[07:49:53] *** jamesd2 has quit IRC
[07:50:23] <balrog> hi lewellyn
[07:50:29] *** jamesd2 has joined #opensolaris
[07:50:30] <balrog> CCCP as in? :P
[07:50:31] <lewellyn> hoila
[07:50:53] <lewellyn> <Compose> <C> <C> <C> <P>
[07:51:03] <balrog> I don't get it
[07:51:12] <balrog> :[
[07:51:12] *** jamesd2 has quit IRC
[07:51:28] <lewellyn> that's the key sequence to draw a hammer & sickle
[07:51:50] *** jamesd2 has joined #opensolaris
[07:51:58] <lewellyn> just like ł is <Compose> <l> </>
[07:52:05] <balrog> now that I'm very familiar with
[07:52:14] <lblume> But it doesn't work in gnome :'(
[07:52:16] *** hajma has quit IRC
[07:52:19] <balrog> well the meaning of CCCP, lol
[07:52:21] <balrog> :(
[07:52:22] <lewellyn> lblume: is that the reason?
[07:52:38] <lewellyn> alanc: who do we file a bug against? :D
[07:52:51] <lblume> Yes, alanc said some time ago that it's in X composition tables, but gnome uses different ones
[07:52:52] <freetown2> OpenSolaris. You need KDE
[07:53:11] <lewellyn> smrt: explain kde
[07:53:11] <smrt>  KDE 4 is not yet officially available for OpenSolaris. If you feel enterprising, you may wish to check out http://solaris.kde.org/ and http://techbase.kde.org/Projects/KDE_on_Solaris for experimental packages or to see if you are able to help the porting effort! (See also: korona)
[07:53:17] <lewellyn> freetown2: "patches welcome"
[07:53:55] *** CodeWar has joined #opensolaris
[07:54:22] <freetown2> just raising the hand in case nobody thinks people want KDE :P
[07:54:29] *** jamesd2 has quit IRC
[07:54:45] <lewellyn> there are gnomes busily working away at it ;)
[07:54:58] <lewellyn> *rimshot*
[07:55:00] *** jamesd2 has joined #opensolaris
[07:55:47] *** jamesd2 has quit IRC
[07:56:20] *** jamesd2 has joined #opensolaris
[07:56:59] <freetown2> we need some kute dwarven engineers then
[07:57:08] * freetown2 coughs
[07:57:14] <lewellyn> wouldn't they be qute?
[07:57:21] *** jamesd2 has quit IRC
[07:57:26] <lewellyn> or is that too trollish? ;)
[07:57:42] <EugeneKay> I'm looking into building a high-performance fileserver for something less than the $100k NetApp wants. I'll be storing primarily .vmdk's for ESX. If I read the recommended best practices correctly, my best bet will be to use a mix of SSDs and 10 or 15k RPM SAS drives, with the SSDs setup as "ZFS Intent Log" devices. Are these assumptions sound?
[07:57:50] *** jamesd2 has joined #opensolaris
[07:57:55] <lblume> What's the ISO639 code for Dwarvish?
[07:58:05] <lewellyn> 42
[07:58:11] <jmcp> lblume: .tk ?
[07:58:25] *** nickw has joined #opensolaris
[07:58:47] <jmcp> nope
[07:58:48] *** jamesd2 has quit IRC
[07:58:54] <jmcp> .dv :-)
[07:59:14] *** Mace_N900 has joined #opensolaris
[07:59:14] <EugeneKay> Mounting the volumes to ESX would be over NFS and vanilla-grade GbE(matching bandwidth to that of the disks, roughly)
[07:59:22] *** jamesd2 has joined #opensolaris
[07:59:38] *** jamesd2 has quit IRC
[07:59:54] <Andys^> EugeneKay: i've done something similar
[08:00:27] <Andys^> EugeneKay: the important thing to note is that in that type of setup, *all* incoming writes are treated as syncronous writes which get dumped to the ZIL (intent log) device
[08:00:45] <EugeneKay> So I've read. Not entirely sure what that means(still reading)
[08:00:46] <Andys^> so it really hammers the ZIL devices, so you need good ones. needs to be resourced correctly
[08:01:04] <lewellyn> and make sure you have spares around just in case :)
[08:01:06] <Andys^> it means any time a VM writes anything to its disk, its gets put onto the ZIL device first before it gets written to the main pool of disks
[08:01:10] <EugeneKay> IE, that long discussion in #vmware earlier about "quality" SSDs
[08:01:13] <Andys^> yeah
[08:01:23] <EugeneKay> That's what I was afraid of.
[08:01:26] <Andys^> it depends how much your cluster writes
[08:01:33] <Andys^> you need to measure it
[08:01:41] <EugeneKay> Not much - it's not applications.
[08:01:51] <Andys^> you can actually build it using two hard disks as the ZIL, and see how much writes you're getting, and then switch to SSD
[08:01:53] <EugeneKay> My bottlenecks are CPU and memory, on the ESX hsot side.
[08:02:11] <Andys^> ok - if memory is a bottleneck then if VMs start swapping that'll also use alot of writes up
[08:02:50] <EugeneKay> I would like to configurethe setup such that ESX swaps to "local" SSDs, rather than bothering to hit the file server.
[08:03:03] <Andys^> thats fine, as long as you're aware of the limitations of that..
[08:03:07] <EugeneKay> I am. :-p
[08:03:28] *** victori has quit IRC
[08:03:40] <EugeneKay> What about the feasibility of using a ramdisk as my ZIL drive?
[08:03:51] <EugeneKay> IE, spec-ing a server out with some ludicrous amount of DDR2
[08:03:54] <Andys^> most corporate type setups don't "need" the ZIL SSDs in that it'll run fine off hard disks, just slower
[08:03:57] <Andys^> ZIL lowers the latency nicely
[08:04:05] <Andys^> well, ZIL must not be kept safe
[08:04:06] <Andys^> err
[08:04:11] <Andys^> the data in ZIL must be kept safe
[08:04:16] * EugeneKay nods
[08:04:18] <Andys^> it had better be battery-backed ram
[08:04:21] <Andys^> otherwise you're wasting time
[08:04:37] <EugeneKay> Well, a RAMSan is within the realm of possibility
[08:04:42] <Andys^> now you're talking :)
[08:04:53] <Andys^> thats the extreme step though
[08:05:02] <Andys^> before that you'd look at FusionIO PCIe cards or STEC SSDs
[08:05:22] <EugeneKay> Like I said, I'm looking to go from a $100,000 estimate to a $10k one
[08:05:34] <Andys^> ah. $10k. you're looking at Intel X25-E then ;)
[08:05:46] <Andys^> they're rated for 100GB/day for 5 years
[08:05:53] <Andys^> of writes
[08:06:10] <EugeneKay> I'm a 3-year-SLA kinda guy.
[08:06:17] <Andys^> RAMsan costs more than a $100k netapp so... ..
[08:06:26] <EugeneKay> There's that.
[08:06:38] <Andys^> how much space do you need?
[08:07:03] <Andys^> with $10k you probably will be forced to use 7200rpm disks
[08:07:04] <jmcp> all of it, duh
[08:07:04] * EugeneKay does the back-of-envelope calculations
[08:07:15] <EugeneKay> My application is VERY clone-able
[08:07:26] <Andys^> so check out eg. Supermicro chassis with 16 or 24 SAS bays, and 16 to 24 x Seagate Constellation 2TB
[08:07:29] <EugeneKay> IE, one "team" of VMs, cloned out 20 times.
[08:07:42] <Andys^> in that case maybe you can use smaller 10k disks
[08:07:52] <Andys^> and load up the fileserver with as much ram as possible and use zfs dedupe... :)
[08:07:57] <Andys^> but this is a little untested/unstable right now
[08:08:07] <EugeneKay> That's my thinking.
[08:08:25] <tsoome> dedupe is not good for speed:P
[08:08:33] <lewellyn> have there been any credible reports of zfs compression actually HURTING things?
[08:08:39] <Andys^> he's implying this application doesn't have much writing though
[08:08:40] <lewellyn> (with the default settings)
[08:08:40] <bdha> lewellyn: gzip9. :)
[08:08:43] <freetown2> are supermicro 36 disk chassis not two good?
[08:08:44] <bdha> boo.
[08:08:46] <lewellyn> bdha: ;)
[08:08:59] *** RoyK has joined #opensolaris
[08:09:01] * bdha has yet to see compression cause problems.
[08:09:03] <Andys^> freetown2: no one said that. but he didnt say how much space he needs yet
[08:09:12] * EugeneKay is trying to figure it
[08:09:13] <lewellyn> bdha: likewise.
[08:09:18] <DesiJat> how do u handle chassis/server failure?
[08:09:24] <lewellyn> i'm starting to wonder why compression isn't the default ;)
[08:09:27] <Andys^> you don't, for $10k ;)
[08:09:31] <lewellyn> especially on rpool
[08:09:35] <Andys^> lewellyn: it is, for metadata
[08:09:39] <lewellyn> i've seen some nice performance gains for rpool
[08:09:44] <EugeneKay> The pain is that people will be creating VMWare Snapshots, each of which will eat 4GB of disk space as it dumps memory to disk
[08:09:49] <freetown2> Andys^, oh, just that i rarely see the 36 disk chassis being mentioned but the 24 does get plenty of it
[08:09:55] <lewellyn> Andys^: yes. i'm not talking about metadata. but rather data :)
[08:10:19] <DesiJat> i've been looking at Supermicro chassis * 2 and doing replication via Linux + DRBD
[08:10:20] <Andys^> compression hurts performance on low-end PCs/notebooks, with opensolaris's default 30 second lag for writing
[08:10:22] <EugeneKay> I would say no more than 5TB of "uncompressed" per fileserver
[08:10:25] <lewellyn> EugeneKay: how are you planning on handling bandwidth?
[08:10:43] <lewellyn> i can tell you that you're going to likely outperform a gigE nic
[08:10:49] <EugeneKay> Parallel GbE
[08:10:53] <Andys^> DesiJat: i'm currently building a shared SAS disk setup (with dual port backplane), with two opensolaris heads. i don't yet know how to automate failover, but the principle is sound
[08:11:03] <lewellyn> and aggregation isn't a guarantee by any means that clients will get even distribution
[08:11:14] <Mace_N900> hi
[08:11:19] <EugeneKay> Or, I can spring for a 10gb card
[08:11:26] <EugeneKay> But that gets expensive quick
[08:11:32] <Andys^> EugeneKay: with NFS it can only spread load across >1 link if you have >1 datastore. with iSCSI it can only do it if you have >1 LUN
[08:11:33] <DesiJat> i dunno what the equivalent of DRBD is on Solaris :/
[08:11:40] <Andys^> DesiJat: AVS
[08:11:41] <EugeneKay> I will have 1 datastore per "Team" of VMs
[08:11:53] <lewellyn> that'd be what i'd recommend. a 10GbE nic for the server on a switch with the consumers all on gigabit
[08:12:01] <Andys^> how many consumers are there, though?
[08:12:07] *** olivier_54 has quit IRC
[08:12:12] <EugeneKay> Right now? ~20 teams, 4 VMs each
[08:12:21] <Andys^> how many ESX boxes?
[08:12:24] <EugeneKay> 2.
[08:12:38] <Andys^> right.. so 10G is probably uncalled for depending on the application
[08:12:40] <Andys^> what ARE the applications?
[08:12:42] <Mace_N900> ugh. i hated esx :)
[08:12:54] <EugeneKay> In a nutshell? Running windows desktops.
[08:12:54] <lewellyn> yeah. 10gig is probably not the right choice for 3 boxes
[08:12:56] <EugeneKay> Lots of them.
[08:13:06] <madwizard> Coffee
[08:13:10] <Mace_N900> something about a windows only client
[08:13:12] <Andys^> 80 windows VMs isn't very much, 1G will be fine for that
[08:13:12] <DesiJat> Andys^: hrrm....
[08:13:19] <lewellyn> madwizard: we need more accented characters :D
[08:13:25] <EugeneKay> I will be expanding to include more vHosts down the line
[08:13:33] <lewellyn> 40 per gigabit nic, at that
[08:13:38] <freetown2> does opensolaris support umem cards?
[08:13:44] <Andys^> EugeneKay: in that case i'd use 7200rpm SATA disks with X25-E for ZIL, and as much ram as possible in the server.. like... 48Gb would be ideal
[08:13:51] <EugeneKay> IE, when I put this fileserver in, replacing what I have now, I'll be bringing my vHost count up to 8
[08:14:10] *** e1kg has joined #opensolaris
[08:14:14] <EugeneKay> I'm eyeing one of those fancy 18-DIMM boards, loaded up with 4GB per slot.
[08:14:22] <Andys^> yes, that'd be excellent
[08:14:22] <lewellyn> and i'd probably go with a gigabit nic per esx box and perhaps a "management" nic
[08:14:23] <EugeneKay> Can only run at DDR2-667, but eh.
[08:14:27] <Andys^> that would give you heaps of RAM to perform fast deduping
[08:14:46] <Andys^> EugeneKay: umm.. the current Xeon platform takes registered DDR3
[08:14:54] <Andys^> if thats what you're referring to by 18 DIMMs..
[08:15:05] <EugeneKay> I haven't read the specs on the 56xx CPUs yet
[08:15:06] *** victori has joined #opensolaris
[08:15:15] <Andys^> dont worry much about the CPU for the fileserver
[08:15:25] <Andys^> just buy the basic E5620's and put all money into RAM
[08:15:31] <EugeneKay> I know that the 55xxs can only run at 667 when you have triple-channel
[08:15:36] <EugeneKay> and 3 DIMMS per channel
[08:15:39] <Andys^> with say 24 x 500gb seagate SAS
[08:15:40] *** Mace_N900 has quit IRC
[08:15:46] <Andys^> its not 667 dude, its 800
[08:15:57] <Andys^> and for 56xx  its 1066
[08:16:03] <EugeneKay> Nossir. Read the specs. Triple-channel, 3 DIMMs, 667 limit. Triple-channel 2-DIMMS is 800.
[08:16:10] <Andys^> but thats not a worry, there's stacks of ram bandwidth
[08:16:11] <CodeWar> and 12MB L2 as wel
[08:16:13] <Andys^> where'd you read that?
[08:16:23] <EugeneKay> Hang on, lemme grab the link
[08:16:41] <lewellyn> oh... madwizard... ;)
[08:16:46] <EugeneKay> It's the motherboard manual for the Supermicro board I like.
[08:16:47] <Andys^> Supermicro X8DTN+
[08:16:51] *** Norrin has quit IRC
[08:16:54] <Andys^> Supports up to 192 GB 1333 / 1066 / 800MHz DDR3 ECC Registered memory
[08:17:01] <EugeneKay> Yeah, but read the manual
[08:17:09] <Andys^> 667 DDR3 doesnt even exist!
[08:17:09] <EugeneKay> The faster speeds are only available with lesser DIMM counts
[08:17:19] <Andys^> the minimum speed for this platform is 800
[08:17:26] <Andys^> if the manual says 667 its a typo
[08:17:33] *** CodeWar has quit IRC
[08:17:34] <EugeneKay> I *may* be off by a step, then.
[08:17:48] <EugeneKay> But I do know that Triple channel, 3-DIMMS is slower than triple-channel 1 DIMM
[08:17:49] <EugeneKay> :-p
[08:17:52] <Andys^> yeah it is
[08:17:56] <Andys^> but its seriously no concern
[08:18:01] <gosx> we built our fileserver with an X25E as L2ARC, and for our workloads, we saturate gigE all the time
[08:18:02] <Andys^> its so much faster than the previous xeon platform its not funny
[08:18:04] <lewellyn> but you don't care about speed. you care about amount.
[08:18:06] <EugeneKay> Yeah. Bottleneck won't be there. :-p
[08:18:08] <gosx> it's only backed by 2 vdevs
[08:18:29] <gosx> our hot working set fits completely in L2ARC
[08:18:31] <lewellyn> gosx: i saturate gigE with a simple raidz2 with 6 disks and compression :P
[08:18:34] <EugeneKay> My current vHosts are 2xE5520s with 16GB of DDR2
[08:18:40] * RoyK += 0xc0ffee
[08:18:40] <gosx> lewellyn: yea, but we do in random IO
[08:18:47] <Andys^> EugeneKay: also if you care about power saving, the 56xx series supports 1.35V DIMMs
[08:18:51] <Andys^> DDR3, not DDR2
[08:18:52] <EugeneKay> They wouldn't let me have the money for the extra 16GB per box until I showed that this stuff works
[08:18:58] <lewellyn> as do they. but they have some mighty compressible data :)
[08:19:00] <EugeneKay> Sorry, yesh.
[08:19:23] <Andys^> so yeah.. i think what i just recommended should work out close to $10k
[08:19:29] <Andys^> assuming your $vendor isn't a ripoff merchant
[08:19:31] <EugeneKay> Sounds right.
[08:19:36] <lewellyn> one dataset is back up to almost 4x again
[08:19:38] <EugeneKay> I like Newegg.
[08:19:40] <EugeneKay> :-p
[08:19:44] <Andys^> you can use supermicro's LSI SAS cards too, to talk to the backplane
[08:20:06] <gosx> something like a dell T710 with the 56xx would be a good baseline, they include the 6Gbit/s LSI adapter now in their base model
[08:20:11] <EugeneKay> What large-disk-count chassis would you recommend?
[08:20:16] <Andys^> Supermicro 836E1
[08:20:18] <gosx> H200 they call it, but it's the 9211-8i
[08:20:32] <gosx> EugeneKay: does it have to be a single box?
[08:20:34] <Andys^> the difference between 3 and 6gbps SAS is basically zip for this application
[08:20:53] <EugeneKay> The less machines I have to manage, the better.
[08:21:09] <gosx> almost better off with a medium-beefy head unit and an attached JBOD for expansion
[08:21:27] <EugeneKay> The 2.5" JBODs I've seen are $2200 for JUST empty disk shelf
[08:21:42] <EugeneKay> Running off a single SAS expander
[08:21:43] <Andys^> EugeneKay: and use Seagate's "ES" enterprise disks.. say 16 x 1tb, and configure them with RAID10
[08:21:58] *** photon_chac has joined #opensolaris
[08:22:14] <EugeneKay> RAID10 in hardware, or let ZFS do it?
[08:22:18] <Andys^> zfs
[08:22:23] <Andys^> use mirroring, is my point :)
[08:22:36] <EugeneKay> Aye.
[08:22:53] <Andys^> what i've described is a very similar setup to what i've used in several different incarnations
[08:23:26] <gosx> striped vdevs each being a mirror works very well
[08:23:27] <Andys^> its alot faster than a $100k netapp, but less redundant
[08:23:29] <EugeneKay> How many X25-Es per 1TB drive, would you say?
[08:23:38] <FastJack> good morning
[08:23:39] <Andys^> EugeneKay: oh, you only need two X25-E
[08:23:50] <EugeneKay> Rather, how many 1TBs per X25E
[08:23:55] <EugeneKay> IE, 8:1
[08:24:03] <Andys^> that ratio isn't really important
[08:24:10] <Andys^> the X25-E can write faster than the incoming data
[08:24:18] * EugeneKay nods
[08:24:22] <gosx> X25E can write around 220MB/s
[08:24:29] <gosx> 220-240
[08:24:37] <EugeneKay> The big or the small one?
[08:24:42] <Andys^> small
[08:24:42] <gosx> both are about the same
[08:24:55] <EugeneKay> I mean, is there any benefit ot the big one, aside from the price tag being double?
[08:25:00] <Andys^> nope
[08:25:04] <gosx> not for SLOG
[08:25:18] <gosx> if it's slog, you'll only be using probably 1GB at MOST
[08:25:22] <gosx> probably less
[08:25:24] <gosx> a lot less
[08:25:33] * EugeneKay nods
[08:25:33] <Andys^> he's planning to have alot of system ram, so it could use alot.  but not many  writes from his apps, so probably not
[08:26:04] <gosx> I would strongly consider an L2ARC device for any large setup
[08:26:10] <gosx> it's huge how much it helps us
[08:26:17] <Andys^> again, he's planning to have alot of system RAM
[08:26:24] <gosx> didn't he say 16GB?
[08:26:27] <Andys^> 72GB
[08:26:31] <Andys^> 16gb was for the vmware host
[08:26:31] <EugeneKay> 18x4GB
[08:26:39] <gosx> oh
[08:26:43] <gosx> yea 72GB is a lot
[08:26:45] <EugeneKay> I'm gonna have 32GB per VMWare host
[08:26:59] <Andys^> i'm running >50 VMs off a box with 24GB and it very rarely touches the l2arc
[08:26:59] <EugeneKay> 48GB if I can ge them to let me have the R610s instead of the R410s
[08:27:12] <Andys^> but i'm not using dedupe.
[08:27:29] <Andys^> EugeneKay: you'll probably find,like me, that the compression works so damn well on VM disk images you won't care about dedupe
[08:27:31] <lewellyn> madwizard: what? did you just show up for a moment and leave us for hours? :)
[08:27:38] <EugeneKay> I could *not* use dedupe, but I don't see the point in taking a 100GB set of .vmdks and multiplying them by 20.
[08:27:59] <EugeneKay> When it can just read the common stuff to memory and be done with it
[08:28:04] <EugeneKay> (the fileserver, that is)
[08:28:08] <madwizard> lewellyn: ?
[08:28:14] <Andys^> i'll tell you why not: because dedupe isn't even available in a 'stable' release of opensolaris yet
[08:28:24] <EugeneKay> Hah.
[08:28:24] <lewellyn> madwizard: you didn't respond a bit ago... i have a question for you ;)
[08:28:29] <Andys^> and it may prove to be slower, and you'll have enough space not to use it
[08:28:36] <madwizard> lewellyn: Shoot
[08:28:38] <EugeneKay> Indeed.
[08:28:45] <madwizard> Kawusia
[08:28:47] <lewellyn> madwizard: you use pkgtool a bit, right? do you have a .pkgtoolrc?
[08:28:52] <madwizard> Żażółć gęślą jaźń
[08:29:10] *** MuffiXXL has joined #opensolaris
[08:29:14] <lewellyn> lblume: there's some accents :D
[08:29:22] <madwizard> lewellyn: Is it created by default? Cause I never made any myself. :)
[08:29:27] <lewellyn> no :P
[08:30:53] <lewellyn> i get the suspicion that the only people who use a pkgtoolrc are people who are using svr4 packages ;)
[08:31:04] <Andys^> EugeneKay: you might like to consider the X8DT3-F if you need to save money, it has 12 DIMM slots and builtin LSI SAS controller, so just plug it straight in to the backplane
[08:31:14] <lewellyn> madwizard: do you publish to a repo that's not localhost?
[08:31:20] <Andys^> 48GB is probably more than enough
[08:31:31] <EugeneKay> I hate saying "ought to be enough"
[08:31:47] <Andys^> as i said
[08:31:50] <Andys^> i'm running 24GB
[08:32:02] <Andys^> and its not even using all of that
[08:32:03] *** MuffiXXL has quit IRC
[08:32:23] <gosx> i like now the new xeon chipsets allow you to populate all the memory, even if not all CPU sockets are filled
[08:32:25] <lewellyn> how much does that 48gb board and cpu go for, decked out with full ram?
[08:32:37] <RoyK> Andys^: turn on dedup and lower block size to 16k, and you'll see your memory go sooner than you think :)
[08:32:37] <gosx> can stuff 1TB of DRAM in the latest dell boxes
[08:32:50] <EugeneKay> What was the figure, 1GbE should be plenty for a vHost?
[08:33:00] <EugeneKay> Running with ~50 VMs
[08:33:03] <Andys^> EugeneKay: yeah.
[08:33:19] <Andys^> RoyK: right..i'm recommending that he doesnt turn on dedupe
[08:33:30] <lewellyn> your disks will be faster than your nic though ;)
[08:33:43] <EugeneKay> Mmmkay. Each R610, with 2xE5620s and 48GB of memory will be running me about $5,000 by the time it's all buttoned up.
[08:33:44] <Andys^> 1gbps of random IO is pretty serious stuff
[08:33:46] <lewellyn> (not necessarily a bad thing)
[08:34:06] <EugeneKay> 8 of those per $10k storage appliance ought to be fine
[08:34:10] <lewellyn> Andys^: it all depends on the data, and how much compression helps
[08:34:23] <EugeneKay> IE, add a pair of 4-port GbE TOE cards to the motherboard's 2.
[08:34:25] <Andys^> EugeneKay: yep. you probably will need a plan for what happens if the box dies though?
[08:34:30] <madwizard> lewellyn: Nope
[08:34:34] <madwizard> lewellyn: I don't
[08:34:37] <lewellyn> madwizard: you're useless ;)
[08:34:45] <lewellyn> have some vokda in your coffee :)
[08:34:52] <madwizard> lewellyn: Yes, I know, my wife told me
[08:34:53] <EugeneKay> Andys^ - buy 3, and have a spare one with 16 vHosts, of course.
[08:35:00] <gosx> Andys^: do you enable ToE/checksum offload, and if so, how much CPU load did that save you?
[08:35:08] <gosx> and which NIC hardware are you using?
[08:35:13] <Andys^> EugeneKay: if you're building your own.. you're gonna need 2 boxes? unless your boss can handle days of downtime :)  and if you buy a dell, at least you have some kind of same-day replacement option
[08:35:23] <Andys^> gosx: i only have Intel NICs, and its on by default
[08:35:37] <EugeneKay> The data server(s) I'll be building myself. N+1 redundancy
[08:35:39] <gosx> k.  it's not on by default on the broadcom bnx driver
[08:35:43] <gosx> gonna have to measure
[08:35:48] <Andys^> broadcom :(
[08:36:02] <EugeneKay> IE, 20 vHosts, 3 data servers.
[08:36:05] <EugeneKay> If one fails, no biggie.
[08:36:22] <gosx> somehow our bnx hasn't wedged since we got it, unlike most T610 users apparently
[08:36:23] <EugeneKay> I just gotta re-provision the VMs
[08:36:31] <EugeneKay> And re-load them into ESX.
[08:36:32] <Andys^> EugeneKay: yep. but the primary option for duplicating data on opensolaris right now is ZFS send/receive, so the backup data box will be some measure of time out of sync with the primary
[08:36:39] <EugeneKay> That's OK.
[08:36:54] <EugeneKay> Nothing here is "mission critical"
[08:37:05] <Andys^> you don't even have to do that, if you shut down the ESX hosts, gives the primary storage IP to the secondary box, then reboot all the ESX hosts
[08:37:19] <EugeneKay> There's that option as well.
[08:37:20] <Andys^> then ESX picks up the datastore on the new box, using the same IP as before
[08:37:54] *** MuffiXXL has joined #opensolaris
[08:38:37] *** perlmongo has joined #opensolaris
[08:38:44] <EugeneKay> Hmmm.... so, a total solution for less than the previously quoted cost for just the fileserver
[08:38:47] * EugeneKay likes
[08:39:01] <EugeneKay> Now, to convince my "boss".... me.
[08:39:17] <Andys^> how much did netapp quote you?
[08:39:27] * EugeneKay digs through his email
[08:39:45] <Andys^> see the netapps come with dual redundant controllers. very minimal downtime. everything else is worse than ZFS though
[08:39:59] <EugeneKay> I have Ye Olde NetApp FAS 250 right now. :-p
[08:40:02] <Andys^> thats the reason for the price difference
[08:40:14] <EugeneKay> FAS2040 Active-Active with (12)600GB 15K SAS drives, iSCSI, FCP, NFS, FlexClone, MultiStore - $117,907 - Roughly 4.3TB useable.
[08:40:32] <Andys^> niiice  *gag*
[08:40:37] <EugeneKay> Pretty much.
[08:40:41] <gosx> $25K/TB seems pricey
[08:40:48] <EugeneKay> It's a NetApp.
[08:40:54] <Andys^> you can shave off money if you go for 20 x 300gb or 450gb, and no NFS license
[08:41:07] <Andys^> i'd say you can get that down to ~$50k-70k
[08:41:22] <EugeneKay> Could, but why bother when I can do it for 1/10th the price? :-p
[08:41:23] <Andys^> for something that is still slower. (but more reliable)
[08:41:36] <EugeneKay> Like I said, reliability is not #1 priorirty here, thankfully.
[08:41:41] <Andys^> well yeah, if its not mission critical, netapp's a complete waste of money
[08:41:57] <EugeneKay> I only have one bloody power bus to my rack anyway. One flipped breaker and I'm off the air.
[08:42:16] <Andys^> right..heheh
[08:42:23] <Andys^> got budget for a UPS? ;)
[08:42:30] <EugeneKay> Don't want one.
[08:42:37] *** merzo has joined #opensolaris
[08:42:43] <EugeneKay> An hour of downtime is acceptable. Everybody just goes to lunch.
[08:42:47] <Andys^> its not about downtime
[08:42:54] <Andys^> its about treating your server hardware nicely with a clean shutdown
[08:43:00] <EugeneKay> There's that.
[08:43:04] <Andys^> so that a power glitch doesn't fry the lot
[08:43:17] <Andys^> not saying you need a huge UPS, just... something from APC that'll keep it running for 5 mins
[08:43:17] <EugeneKay> For the full build-out I'll likely be changing data centers entirely
[08:43:32] <EugeneKay> So it's moot.
[08:47:29] *** raichoo has joined #opensolaris
[08:47:43] * EugeneKay does the math on Us.... 42U, minus 2x1 for gateways, 3x3 for fileservers, 3x1 for Switches, an 1 for a KVM leaves 27 for vHosts.
[08:48:32] *** CodeWar has joined #opensolaris
[08:48:50] <EugeneKay> Sigh. I'm going to have to write a goddamned investment prospectus for this project. >_<
[08:49:03] <Andys^> you also need to leave 1RU free per 3 servers for spacing :)
[08:49:12] *** jamesd2 has joined #opensolaris
[08:49:12] <Andys^> and maybe more for cable management
[08:49:31] <EugeneKay> Indeed.
[08:49:50] *** MuffiXXL has quit IRC
[08:50:18] *** Nardul has joined #opensolaris
[08:50:30] * EugeneKay wonders if he should order up one of those BackBlaze red cases and fill it up with 2TB drives, as long as he's at it.
[08:50:58] * freetown2 spurts out all his coffee
[08:51:07] <freetown2> EugeneKay, you owe me a new screen and keyboard
[08:51:27] <EugeneKay> I hope it was a cheap keyboard.
[08:51:42] <freetown2> it's gold plated
[08:52:10] *** |AbsyntH| has joined #opensolaris
[08:52:34] * EugeneKay shrugs
[08:52:59] <EugeneKay> A guy's gotta have somewhere to keep his "important data", you know?
[08:53:08] <freetown2> hahahaha
[08:53:56] <freetown2> so what's wrong with a Supermicro 36 bay chassis and SAS2?
[08:54:32] <freetown2> you could always add a 45 bay chassis later too...
[08:55:45] <EugeneKay> ? http://www.supermicro.com/products/chassis/4U/847/SC847E2-R1400U.cfm
[08:57:58] <EugeneKay> Actually, the 847A looks better. All Low-Profile cards :-p
[08:58:20] <EugeneKay> This one http://www.supermicro.com/products/chassis/4U/847/SC847A-R1400.cfm
[08:58:48] <EugeneKay> The only problem with that thing is the need for a forklift to install it.
[09:05:28] *** Nardul has quit IRC
[09:05:43] <freetown2> or get four burly men...
[09:06:16] <EugeneKay> Looks like hte manual says that thing has a mounting location inside, too
[09:06:21] <Andys^> EugeneKay: you shouldnt need 3 storage boxes should you? 2 should be enough
[09:06:27] <Andys^> or you mean down the track..
[09:06:37] <EugeneKay> Well, if I'm gonna do this, might as well get it all at once.
[09:07:19] <Andys^> financially you're better off getting two bigger ones than 3
[09:07:32] <EugeneKay> I like N=1
[09:07:34] <EugeneKay> N+1*
[09:07:39] <Andys^> 2 is N+1?
[09:07:46] <Andys^> you only need 1 box, so 2 is N+1 :P
[09:07:52] <EugeneKay> Not if I do a whole rack
[09:08:03] <Andys^> thats my next question
[09:08:13] <Andys^> why are you doing a whole rack? you should max out the RAM in the ESX boxes first
[09:08:24] <EugeneKay> Current ESX boxes are "development"
[09:08:26] <Andys^> you're definitely not CPU bound with windows desktops
[09:09:33] <EugeneKay> Once I have all my fancy management scripting good to go, and the customer is ready to move to production, I'm going to have a sudden case of OH GOD I NEED MORE SERVERS
[09:09:33] <Andys^> and don't forget the effect of page sharing in ESX with identical VMs.. you can fit more per GB of RAM
[09:11:08] *** twisti has joined #opensolaris
[09:13:53] <CodeWar> i m guessing in addition to onnv-gate there would be internal release branches that folks use internally for Solaris development.  uber question who does the integrations from mainline to release branches dev or a release guy?
[09:15:25] <lewellyn> i think that's what release engineering exists for
[09:16:47] <CodeWar> see trouble is release engineers will find it difficult to do the merges when theres a conflict
[09:17:20] <CodeWar> sure they can ping engineers but it er well..hmm .. lets see theres lots of things people do curious what oracle folks do internally
[09:17:56] <lewellyn> i think they're only technically oracle folk, at this point
[09:18:12] *** Nardul has joined #opensolaris
[09:18:15] <lewellyn> i haven't gotten the impression that the cultures have been attempted to be merged yet
[09:18:53] <CodeWar> seriously people that have work to do .. serious projects to deliver wouldnt care about culture/ atmosphere / temperature / mood and what not .. theres you theres your manager and theres a deliverable
[09:19:34] <CodeWar> problem starts when there are 3 teams doing the same thing and they start interfering with each other .. I dont get the sense thats going to happen with Oracle / Sun merger
[09:20:25] <CodeWar> well I guess hotspot/ jrockit folks might ego clash a bit
[09:22:09] *** Dagobert has joined #opensolaris
[09:23:02] *** sysspoof_ has quit IRC
[09:24:31] <CodeWar> in other news ubuntu is threatening to release 10.4 ... their 9.10 was terrible though
[09:25:22] <EugeneKay> Lynch me if you like, but I use Windows 7 on all computers with a mouse attached.
[09:25:22] *** kaktus__ has joined #opensolaris
[09:25:47] <CodeWar> use the best tool for the job .. nothing wrong with that
[09:25:56] *** Nardul has quit IRC
[09:25:56] *** perlmongo has quit IRC
[09:25:56] *** freakazoid0223 has quit IRC
[09:25:56] *** vraa has quit IRC
[09:25:57] *** Shoggoth has quit IRC
[09:25:57] *** andersenep has quit IRC
[09:25:57] *** p3rror has quit IRC
[09:25:57] *** mbreitba has quit IRC
[09:25:57] *** devians has quit IRC
[09:25:57] *** echobinary has quit IRC
[09:25:57] *** zedrich has quit IRC
[09:25:57] *** enderst has quit IRC
[09:25:57] *** Wes-- has quit IRC
[09:25:57] *** pksu has quit IRC
[09:25:57] *** lyarick has quit IRC
[09:25:57] *** MrWGW has quit IRC
[09:25:57] *** abyssmal_ has quit IRC
[09:25:57] *** djinni has quit IRC
[09:25:57] *** Thijssss has quit IRC
[09:25:57] *** oninoshiko has quit IRC
[09:25:57] *** jdparker has quit IRC
[09:25:57] *** jim_ec2 has quit IRC
[09:25:57] *** kaktus has quit IRC
[09:26:05] *** sysspoof_ has joined #opensolaris
[09:26:20] <EugeneKay> And Fedora works plenty good for my webserver..... CentOS is well and good, but all their packages are bloody old.
[09:26:37] <EugeneKay> 6-month lag on Firefox is just unacceptable.
[09:26:43] *** Bateau has quit IRC
[09:27:17] <CodeWar> why would the webserver need to run firefox though
[09:27:32] <EugeneKay> It doesn't, but it's a systemic thing.
[09:27:33] *** Wes-- has joined #opensolaris
[09:28:19] <CodeWar> where does canonical get its money from btw .. they re on time and they deliver good changes its a different thing I wont use it but heck ..
[09:28:27] <EugeneKay> Corporate support
[09:28:34] <CodeWar> ahh
[09:28:38] *** kaktus__ is now known as kaktus
[09:28:41] <EugeneKay> Same place Red Hat makes money on RHEL vs Fedora or CentOS
[09:28:42] *** Odin- has quit IRC
[09:28:53] <EugeneKay> See also MySQL
[09:29:03] <EugeneKay> Or Sun on Solaris ;-)
[09:29:05] <CodeWar> RHEL maeks its money from enterprise no surprises there but ubuntu? hmm really
[09:29:20] <jnss> opensolaris is very picky when it comes to suppoirting hardware
[09:29:30] <jnss> opensolaris is not open to suggestion
[09:30:13] <jnss> add more hardware drivers, please
[09:30:15] <EugeneKay> Yup. I don't believe they've turned a profit yet though
[09:30:25] <CosmicDJ> really? they're not open for "hey you f* include support for xy or I'll run linux"? ;)
[09:30:42] <jnss> That's it
[09:30:46] <jnss> you nailed it ;)
[09:30:48] <EugeneKay> (Canonical, that is)
[09:31:10] <CosmicDJ> I'm pretty sure they accept patches...
[09:31:12] *** phimic has joined #opensolaris
[09:31:28] <EugeneKay> Ubuntu is currently funded by Mark Shuttleworth's ridiculously deep pockets, because he likes the idea.
[09:31:39] <jnss> he went into space
[09:31:40] <CodeWar> why would you want to share a good IP/solution/patch with Solaris Oracle when you could contribute to the linux kernel ... more open more peer review
[09:31:50] <jnss> he has a large ego. he must think he's some kind of god
[09:31:53] *** Nardul has joined #opensolaris
[09:31:53] *** perlmongo has joined #opensolaris
[09:31:53] *** freakazoid0223 has joined #opensolaris
[09:31:53] *** Shoggoth has joined #opensolaris
[09:31:53] *** andersenep has joined #opensolaris
[09:31:53] *** p3rror has joined #opensolaris
[09:31:53] *** mbreitba has joined #opensolaris
[09:31:53] *** devians has joined #opensolaris
[09:31:53] *** echobinary has joined #opensolaris
[09:31:54] *** zedrich has joined #opensolaris
[09:31:54] *** enderst has joined #opensolaris
[09:31:54] *** MrWGW has joined #opensolaris
[09:31:54] *** pksu has joined #opensolaris
[09:31:54] *** lyarick has joined #opensolaris
[09:31:54] *** abyssmal_ has joined #opensolaris
[09:31:54] *** djinni has joined #opensolaris
[09:31:54] *** Thijssss has joined #opensolaris
[09:31:54] *** oninoshiko has joined #opensolaris
[09:31:54] *** jdparker has joined #opensolaris
[09:31:54] *** jim_ec2 has joined #opensolaris
[09:31:58] <EugeneKay> He made his money from Thawte, who he sold to Verisign for $575M in cash.
[09:32:05] <jnss> just imagine how selfish youd have to be
[09:32:16] <jnss> to spend millions for your own space flight
[09:32:29] <jnss> when 60,000 people starve to death every day
[09:32:39] <jnss> mostly in africa. shuttleworth is african. south african
[09:32:41] <CodeWar> so ubuntu is charity? guy has a lot of money and decides to blow it on a linux distro instead of marijuana .. good
[09:32:49] <jnss> he should try feeding his continent
[09:32:57] <EugeneKay> Well, it's "Charity, but it'd be nice if I could make some money on this"
[09:33:45] <jnss> he sure did  a good thing with ubuntu. i am jsut saying, try saving a life once in a while
[09:33:58] <joshua_> ah, the old fashioned science fallacy
[09:34:27] <CosmicDJ> jnss: do your homwork dude, he spent some of his money for charity
[09:34:32] <joshua_> "why are people doing research on <x>? many more people die of cancer!"
[09:34:36] <EugeneKay> There's worse things he could blow his money on. Like his $20M trip to the ISS.
[09:34:45] <EugeneKay> :v
[09:34:49] <CodeWar> joshua_,  well said
[09:35:06] *** tsoome has quit IRC
[09:35:15] <jnss> joshua_, hardly a fallacy
[09:35:37] <joshua_> EugeneKay, I'm not even convinced that's "blowing the money" either; if I were doing that, I would consider it a donation to / funding for the space program in question, while he gets a ride into space out of it
[09:35:37] <jnss> we all know cancer should have been cured by now
[09:35:56] <jnss> or at least the killer disease known as starvation
[09:36:11] <EugeneKay> Indeed. Encouragement of interest in space is a good thing.
[09:36:17] <joshua_> pouring money into it will not solve anything; you have to micro the money, or it will go warlords or whathaveyou
[09:36:31] <joshua_> and, in general, not towards people who will solve problems
[09:36:33] <jnss> joshua_, i relaly agree with that
[09:36:48] <joshua_> he has expertise in solving software problems, and that is what he likes to do
[09:37:10] <EugeneKay> A few years back, there was a bunch of publicity about the Gates foundation and something about helping women in Africa. The rheoteric was "Why doesn't he just buy a small African nation?"
[09:37:32] *** olivier_54 has joined #opensolaris
[09:37:56] <CodeWar> I dont believe in helping any country beyond a certain limit and that too only in the wake of a natural disaster..  a community/nation should fight for its own
[09:38:21] <jnss> hmm i guess you had to bring up the gates foundation
[09:38:31] <EugeneKay> There's a terribly politically incorrect one about the Haiti quake I know.
[09:39:09] <EugeneKay> Before the earthquake, there were 2.5 million starving Haitians. Afterwards, there were only 2 million, and they were actually getting the international food aid that they had been begging for.
[09:40:03] <CodeWar> ok lets not get so insensitive now .. we re better people than that
[09:40:16] <EugeneKay> I like to think so.
[09:40:40] <EugeneKay> But, the point gets at the whole "help only for the publicity" thing.
[09:42:33] <CodeWar> scout threads .. HTM I m dying of curiosity to know how Rock is doing (canned/ clipped/changed / on track?) read Dice's paper other day
[09:43:30] <CodeWar> one thing I like about Sun (old Sun) and SPARC is they come up with these large leap revolutionary changes in small timeframes and dont worry too much about whether its too early for the market or too much to deliver
[09:43:37] <CodeWar> compare with intel of the 90s..
[09:44:05] <EugeneKay> Truthfully, ZFS is genius.
[09:44:09] *** perlmongo has quit IRC
[09:44:24] <CodeWar> from what lil I understood of it .. yes its revolutionary
[09:44:35] <EugeneKay> It's a crying shame that there's no linux implementation, because, let's face it, that's where the *NIX scene is in the modern age
[09:45:23] <freetown2> EugeneKay, if you can wait for btrfs whatever that will be...
[09:45:36] <CodeWar> while I sympathize with Solaris if you made quality ZFS freely available onLinux there wont be much charm turning to solaris for most folks
[09:45:53] <freetown2> but I hope to stick with OpenSolaris and ZFS. Larry, please don't kill OSOL
[09:46:06] <EugeneKay> I likely will, once it's "good"
[09:47:05] <EugeneKay> Butter FS is supposed to be the successor to ext3/4, with a good bit of the ZFS magic ideas mixed it
[09:47:18] <EugeneKay> It's not as if ZFS is anything truly "new", see NetApp's WAFL.
[09:47:38] <CodeWar> whoever got there first can claim credit for something big .. dunno who did frankly
[09:47:48] <EugeneKay> It just happens to be a good implementation of the idea set, and not forced to come with a $100k+ device
[09:48:33] <joshua_> ggibson, whenever ZFS is brought up, likes to insist that NetApp did it firs t;P
[09:49:52] <CodeWar> Patterson asked a very pertinent question during one of the hotos conferences for which he was raped but it made sense .. most of the OS gliterrati like to focus on perf and in the process lose out on thinking out of the box by shining an old metal
[09:50:03] <EugeneKay> They did.... but they have all sorts of arbitrary limits
[09:50:33] *** Erwann has joined #opensolaris
[09:52:09] *** Tpenta has joined #opensolaris
[09:53:05] *** SYS64738 has quit IRC
[09:53:27] *** ChanServ sets mode: +o Tpenta
[09:53:32] *** SYS64738 has joined #opensolaris
[09:54:40] <CIA-33> vitezslav batrla - Sun Microsystems - Prague Czech Republic <Vitezslav.Batrla at Sun dot COM>: 6945319 pcieb_initchild() misinterprets return from ndi_merge_node(), issues bogus warning
[09:55:11] *** dover has quit IRC
[09:59:41] <CodeWar> 20 bucks in 1995 to 250 in mid 2000 ... looks like the golden time to join Sun was 95 .. by 2000 your stocks would have vested ripe just ready to sell
[09:59:47] <CodeWar> any oldtimers from then here?
[10:01:03] *** Nardul has quit IRC
[10:02:42] <palowoda> I joined Sun in 1995.
[10:03:44] <CodeWar> hopefully you cached out bro and didnt get too greedy waiting for it to rise even higher
[10:04:12] <palowoda> Don't worry.
[10:05:18] <bonsaikitten> cashed, not cached ;)
[10:05:53] <CodeWar> sorry ... was debugging a cache problem :-) waiting to recompile and looking at stocks of different companies
[10:05:56] <CodeWar> :-)
[10:06:08] *** Nardul has joined #opensolaris
[10:06:45] <palowoda> Investments are another ball of wax. There is more than just the high tech sector.
[10:06:55] *** tsoome has joined #opensolaris
[10:07:22] <CodeWar> true though stocks from your company is free money ... you dont invest your own money
[10:09:56] <palowoda> I've never considered money free, although investing your own money can give you heart attacks sometimes.
[10:10:09] <lblume> Those who got paid with the 2000 stock and couldn't legally cash them before a few years must not think exactly the same :-)
[10:10:20] *** jamesd2 has quit IRC
[10:10:37] <CodeWar> lblume, the feeling is that of drowning with hands tied behind your back
[10:11:01] <palowoda> Actually I should say trading as opposed to investing.
[10:11:03] <CodeWar> friend of mine at broadcom had it even worse.. couldnt sell their company blackwindow was on i m told
[10:11:56] *** Yu\2 has joined #opensolaris
[10:12:10] <lblume> Anyhow, interesting discussion, but go tto split to get closer, but not quite inside Sun's offices in 清华, bye all.
[10:12:28] <palowoda> Later lblume.
[10:14:08] <CodeWar> palowoda, you still with Sun
[10:14:57] <palowoda> No, I was with them (and I stand corrected) 94 to 2005.
[10:16:36] *** houst0n has joined #opensolaris
[10:16:55] <palowoda> I don't know if that makes me an old timer or has been. :)
[10:17:03] <CodeWar> makes you a rich bloke
[10:17:42] <madwizard> Coffee
[10:17:43] <palowoda> No I'm no different than the rest of you guys and gals.
[10:19:20] <tsoome>  nono, you are special:D
[10:20:11] <palowoda> tsoome: Thanks but I have 7 cats and 5 raccoons to feed which I found out can be expensive.
[10:20:14] *** ss1 has joined #opensolaris
[10:20:21] <tsoome> ouch!
[10:20:33] <ss1> hello all
[10:20:38] <houst0n> Cat / Raccoon stew is a delicacy here
[10:20:44] *** SYS64738 has quit IRC
[10:20:46] *** nsuperbus has joined #opensolaris
[10:20:53] <tsoome> lol
[10:21:41] <EugeneKay> I prefer a good Possum roast
[10:21:59] <palowoda> houst0n: You would not believe how stupid cats and raccoons try to get along.  But it can be entertaining.
[10:22:08] <houst0n> I generally dislike animals
[10:22:34] <houst0n> Robots, sure.. Animals? No.. Not unless I mistakenly impregnate my girl and she decides to pop one out
[10:22:43] <houst0n> Then they're probably useful
[10:22:50] <tsoome> .oO
[10:23:02] <tsoome> you better be careful then;)
[10:23:06] <houst0n> I do enjoy eating them, though
[10:23:44] <palowoda> When a 40lb raccoon walks into your office and ask for food you have to explain to it the options.
[10:25:27] <houst0n> I think the first question here would be, 'what is 40lb?!' We don't understand the concept of '5 pounds 17 ounces and 13 quibbles along 3 yards of wibblestick' in europe.. I think it's a bit under 20KG tho
[10:25:41] <EugeneKay> At times like that, I am glad that I work from home, because office co-workers might be upset if I were to draw my 9mm and unload a whole clip into the thing
[10:25:43] <houst0n> And we'd simply kill it with a big manual
[10:25:43] <houst0n> heh
[10:25:45] *** SYS64738 has joined #opensolaris
[10:25:53] <EugeneKay> houst0n - 2.2lb to the kg
[10:26:09] <houst0n> Okay, so 18
[10:26:46] *** randw has quit IRC
[10:26:51] *** randw has joined #opensolaris
[10:28:48] <palowoda> I'm on my fourth generation of raccoons they obey me like dogs.
[10:29:12] <palowoda> Of coarse I got the food.
[10:30:11] <EugeneKay> Heh.
[10:30:13] <CodeWar> didnt think of raccoons as social animals before
[10:30:14] <houst0n> Why would you keep a raccoon? Do you make something out of their fur? Or do you like the idea of your kids catching rabies or something?
[10:30:29] <EugeneKay> Kids can get rabies from dogs just as easily.
[10:30:45] <CodeWar> or from other kids
[10:30:54] <EugeneKay> Breeding across multiple generations like that is how dogs were domesticated
[10:31:03] <palowoda> They actually make better pets than dogs. Much more independent.
[10:31:17] *** vivid has quit IRC
[10:31:20] <EugeneKay> Unconventional? Sure. But so was the caveman who kept wolves.
[10:32:07] <CodeWar> a permit might be required ?
[10:32:09] <CodeWar> is this california?
[10:32:32] <EugeneKay> You need to have a permit to piss in your own yard in CA
[10:32:35] <EugeneKay> Or did they ban that, too?
[10:32:36] <palowoda> Yep right in the SF bay area.
[10:33:49] <palowoda> In fact when I have a BBQ with a bunch of friends they come around and join the party.  The friends are getting use to it. :)
[10:34:39] *** sysspoof_ has quit IRC
[10:34:59] <palowoda> The cats got ripped up a couple of times though.
[10:36:14] <palowoda> Here is stupid: http://fiver.net/misc/two_plus_stupid.jpg
[10:36:41] *** ewdafa has joined #opensolaris
[10:36:49] *** ss11 has joined #opensolaris
[10:37:58] <palowoda> And the local gang: http://fiver.net/misc/raccoons/
[10:38:46] <CodeWar> do you do a count every night to make sure they didnt scale over the wall and into the streets ..
[10:39:40] <palowoda> I just get pissed when they eat all my grapes. I'm not worried about the streets.
[10:39:55] *** ss1 has quit IRC
[10:40:01] *** jaywalk has joined #opensolaris
[10:40:21] *** sysspoof_ has joined #opensolaris
[10:42:27] *** LouisJB_ has joined #opensolaris
[10:44:22] *** Fallen_Demon has joined #opensolaris
[10:45:34] *** jgracin has joined #opensolaris
[10:47:15] *** ss11 has left #opensolaris
[10:49:55] *** CodeWar has quit IRC
[10:50:37] *** Vagrant has joined #opensolaris
[10:51:00] *** Nardul has quit IRC
[10:54:29] *** lateau has joined #opensolaris
[10:54:47] *** Nardul has joined #opensolaris
[10:55:49] *** micols has quit IRC
[10:57:35] *** micols has joined #opensolaris
[10:59:39] *** timsf has joined #opensolaris
[11:00:32] <AukeF> Hi! I have 2 J4500 storage arrays hooked up with a SAS connector to an opensolaris box. The 96 disks are detected without problems (numbered c2t0d0 through c2t95d0). Is there a tool to create a mapping of disks to enclosures (and preferably also slot-within-enclosure)?
[11:04:29] <madwizard> Coffee
[11:04:50] *** EugeneKay is now known as EugeneKaway
[11:06:20] <Stric> AukeF: put traffic on a disk, see which slot led twinkles, put a sticker on it with the number..? :)
[11:07:04] *** Vagrant has quit IRC
[11:07:57] <tsoome> gief back good old parallel scsi jbods with target numbers!
[11:07:59] <tsoome> :D
[11:08:26] <houst0n> AukeF: Do you have CAM setup?
[11:08:39] <AukeF> I should have that, yes
[11:08:59] <houst0n> You're really going to want to use that w/ 96 disks :P
[11:09:20] <houst0n> You can do all the usual stuff w/ it, locate disks, turn on leds etc
[11:09:26] <AukeF> Or I can always install it if needed. But I was hoping for a simple 'cXtY to 'enclosureXslotY' listing cmdline tool ;)
[11:09:40] <houst0n> There is a CLI for cam also
[11:09:58] <AukeF> Ah, I just spotted the webinterface, I'll give it a whirl, thanks for the pointer.
[11:10:15] *** merzo has quit IRC
[11:10:16] <houst0n> http://dlc.sun.com/pdf/820-4192-12/820-4192-12.pdf
[11:10:29] <houst0n> The webui is nice though
[11:11:00] <palowoda> It's just not part of Opensolairs right?
[11:11:08] <houst0n> Nope
[11:11:19] <houst0n> http://www.oracle.com/us/products/servers-storage/storage/disk-storage/031603.htm
[11:11:58] *** merzo has joined #opensolaris
[11:13:32] <houst0n> I'd like a j4500 to play with. We have a few 4200s - work very well
[11:13:53] <palowoda> Hey we are all Oracle sluts. :)
[11:14:28] <houst0n> I'll reserve comment there
[11:14:44] <houst0n> Not in a good mood w/ oracle atm
[11:14:54] <palowoda> I was wondering if I could say that in public.
[11:15:10] <houst0n> Well, they yanked our primary mirror yesterday
[11:15:49] *** wdp has joined #opensolaris
[11:19:09] <Dagobert> What's a server? Storage without disks ;-)
[11:20:51] <bonsaikitten> Dagobert: an interesting perspective ... but what do you need storage for?
[11:21:25] <palowoda> Yah everything is stored on the cloud now days.
[11:21:33] <houst0n> No it's not :P
[11:22:16] <palowoda> Oh that's right what ever happened to the Sun Cloud stuff?
[11:22:17] <houst0n> I hate this cloud bullshit
[11:22:29] <houst0n> palowoda: More dead than opensolaris I believe
[11:22:34] <houst0n> So, very dead
[11:22:44] *** p3rror has quit IRC
[11:22:51] <palowoda> Larry ran his boat over it?
[11:23:09] *** twisti has quit IRC
[11:23:12] <bonsaikitten> cloud just means "not your hardware" ;)
[11:23:22] <bonsaikitten> a rather scary idea for paranoid people like me
[11:23:25] <houst0n> I think oracle don't give a fuck about offering hosted oracle
[11:23:32] <houst0n> or hosted w/e
[11:23:59] <palowoda> Ah I see just buy the support and shut up.
[11:24:15] <houst0n> Anyone found any decent solaris VPS hosting btw?
[11:24:28] *** het has quit IRC
[11:24:38] <houst0n> ... besides joyent
[11:24:52] *** twisti has joined #opensolaris
[11:25:07] <palowoda> joyent is kind of pissed why would Oracle come up with another solution?
[11:25:25] <houst0n> ? what ?
[11:26:17] <palowoda> Why would Oracle believe in marketing a virtual environment to resale on the net?
[11:26:25] *** xtrondo has joined #opensolaris
[11:26:41] <houst0n> I imagine the only reason for a company like oracle to do so is for appearances
[11:26:59] <houst0n> They'll certainly not make any real money from it
[11:27:09] <palowoda> They seem to be failing that from a marketing prospective.
[11:27:21] <houst0n> hence why so many are jumping ship to redhat
[11:27:22] <houst0n> heh
[11:27:25] *** gebi has joined #opensolaris
[11:27:47] <palowoda> Oracle repackages Redhat too you know.
[11:28:00] <bonsaikitten> so does CentOS
[11:28:12] <bonsaikitten> (which amusingly has the same amount of letters as redhat so patching is easier)
[11:28:17] <houst0n> Well, would you choose a rebuild called 'oracle unbreakable linux' over rhel ?
[11:28:37] <palowoda> But Oracle's advantage is indemnification.
[11:28:43] <houst0n> We use CentOS mainly, where we don't need support
[11:28:51] <houst0n> 90% solaris, tho
[11:29:13] <houst0n> Did centos release 5.5 yet?
[11:29:18] <houst0n> oracle beat em to it, it seems
[11:31:17] <palowoda> Really Oracle will have their legal department protect you if you buy the ultimate premium support for Unbreakable Linux from lawsuits.  It has to be an advantage.
[11:32:07] <palowoda> It's got to be worth at least 1K per hardware unit.
[11:32:15] <houst0n> Suprisingly, I'm not very scared of a lawsuit resulting in my choice of server OS
[11:32:36] <palowoda> Got to sell something these days.
[11:33:40] <Wes-> Don't worry, sales will pick up once Oracle starts making bright red servers
[11:33:43] <bonsaikitten> houst0n: as far as I can tell centos is a bit slow at the moment, they take an unexpectedly long time for updates
[11:34:04] <houst0n> bonsaikitten: Well, last update I heard was a few days ago, apparently on 2 or 3 pkgs left which throw errors
[11:34:12] <houst0n> then of course it needs to go through RC and so on
[11:34:19] *** LenZ has joined #opensolaris
[11:34:21] <houst0n> They do very well, I think
[11:34:50] <houst0n> I'm not fussed about how close between rhel releases they are, though
[11:34:52] <palowoda> But if you buy Oracle Linux some of the profits are reinvested in Opensolaris.
[11:34:56] <houst0n> I wouldn't adopt 5.5 straightaway
[11:35:14] <houst0n> That, I really don't care about
[11:35:23] <bonsaikitten> I am unwilling to use the mix of bad package manager and bad init scripts
[11:35:34] <bonsaikitten> since I end up fixing stuff anyway :)
[11:35:34] <houst0n> They can invest the money in solaris 10 instead, or firing the retards who thought IPS was a good idea
[11:35:57] *** p3rror has joined #opensolaris
[11:36:07] <RoyK> I don't get it - I have a supermicro box with osol svn134, and osol only sees 2GB of the 8GB installed
[11:36:27] <RoyK> I've tried to remove some, replace them, it still only sees 2
[11:36:38] <palowoda> Oracle Enterprise Linux is helping support the development of IPS. Can't all be that bad.
[11:36:51] <RoyK> not 4, not 8, just two - BIOS sees the memory correctly
[11:37:01] <bonsaikitten> palowoda: doesn't change the fact that it's fundamentally screwed up
[11:37:02] <houst0n> RoyK: Hmm, I'm on a SM X7DA8 atm, sees all mem file
[11:37:10] <houst0n> Kind of weird
[11:37:25] <palowoda> RoyK: Almost sounds like your booting a 32bit kernel.
[11:37:30] *** swift has joined #opensolaris
[11:37:38] <RoyK> this is a X7SBU
[11:37:39] <houst0n> RoyK: isalist ?
[11:37:52] <RoyK> sec - rebooting
[11:38:09] <palowoda> Actuall 32bit should see more than 2G.
[11:39:35] <RoyK> and this should be 64bit
[11:39:48] <RoyK> the CPU is definitely 64bit
[11:40:01] <Stric> RoyK: what does the kernel say when booting? (I've asked this before)
[11:40:13] <Stric> add -v   and/or check kernel log
[11:40:14] <CosmicDJ> someone asked for isalist, too
[11:40:14] <RoyK> Stric: I'll check
[11:40:31] <tsoome> bios update?:P
[11:40:42] <Stric> and does any other OS see them?
[11:40:57] <Stric> could be screwed up ACPI/DMI tables or something
[11:41:13] * RoyK waits for the PoS to finish booting
[11:41:27] <palowoda> if it's the ACPI tables good luck.
[11:41:38] <CosmicDJ> piece of shit?
[11:42:13] <palowoda> Oracle has no idea what ACPI tables should mean beyond what they sell.
[11:42:52] <RoyK> isalist says 'amd64 pentium_pro+mmx pentium_pro pentium+mmx pentium i486 i386 i86'
[11:42:55] *** perlmongo has joined #opensolaris
[11:42:57] <palowoda> It's about open standards not that they comply.
[11:43:24] <palowoda> Pentium Pro, is that a joke?
[11:43:36] <RoyK> not my joke
[11:43:45] <palowoda> Not my hardware either.
[11:43:47] <RoyK> /var/adm/messages says 'Apr 28 11:46:00 mime unix: [ID 168242 kern.info] mem = 2096128K (0x7ff00000)'
[11:44:03] <palowoda> isalist
[11:44:18] <RoyK> [11:48]  <RoyK> isalist says 'amd64 pentium_pro+mmx pentium_pro pentium+mmx pentium i486 i386 i86'
[11:44:44] <RoyK> prtdiag shows four memory modules
[11:44:58] <RoyK> they're all 2 gigs (although prtdiag doesn't tell)
[11:45:18] <houst0n> RoyK: under prtdiag, i see 'Bank1\nBank1\nBank2\nBank2' etc
[11:45:24] <palowoda> smbios work?
[11:45:26] <houst0n> DIMM1A DIMM1B etc
[11:45:31] <houst0n> Do you have the mem in the right place?
[11:49:36] *** micols has quit IRC
[11:50:05] *** micols has joined #opensolaris
[11:50:32] <madwizard> Coffee
[11:51:18] <RoyK> houst0n: http://pastebin.com/bjWs4uQe
[11:52:15] *** het has joined #opensolaris
[11:53:52] *** |AbsyntH| has quit IRC
[11:54:40] <CIA-33> Casper H.S. Dik <Casper.Dik at Sun dot COM>: PSARC 2009/377 In-kernel pfexec implementation., PSARC 2009/378 Basic File Privileges, PSARC 2010/072 RBAC update: user attrs from profiles, 4912090 pfzsh(1) should exist, 4912093 pfbash(1) should exist, 4912096 pftcsh(1) should exist, 6440298 Expand the basic privilege set in order to restrict file access, 6859862 Move pfexec into the kernel, 6919171 cred_t sidesteps kmem_debug; we need to be able to detect bad hold/
[11:55:32] <palowoda> RoyK: Not that it's related to the memory but the CPU id of 00000000000000000000000000000000 CPU 1 looks funky.
[11:56:04] <Andys^> move pfexec into kernel ... ?!
[11:56:25] *** ozux has joined #opensolaris
[11:56:55] <RoyK> palowoda: erm - yes
[11:58:29] <Andys^> RoyK: I presume you have enabled EM64T in the bios?
[11:58:40] *** kwinz3 has quit IRC
[11:59:14] <Andys^> you may have to try fiddling with some boot settings of opensolaris
[11:59:15] <lewellyn> yay more pf*sh
[11:59:20] <Andys^> like apic, ioapic, etc
[12:00:42] *** lucalenardi has joined #opensolaris
[12:01:05] <Andys^> i'm pretty sure the X48 chipset can handle four 2gb sticks in just about any configuration
[12:01:11] *** het has quit IRC
[12:02:29] *** lucalenardi has quit IRC
[12:02:50] *** lucalenardi has joined #opensolaris
[12:04:50] *** tosh has joined #opensolaris
[12:05:07] <houst0n> RoyK: Do you only have one CPU in there?
[12:05:34] <houst0n> I guess if yes you need to put all the memory on one side..
[12:06:52] <Andys^> its a single CPU board
[12:07:00] <Andys^> previous generation without directly attached memory
[12:09:40] <palowoda> In other words an architecture Oracle and Intel do not agree on?
[12:09:51] *** het has joined #opensolaris
[12:10:22] *** hohum_ has joined #opensolaris
[12:10:25] *** Nardul has quit IRC
[12:11:12] *** jwit_ has joined #opensolaris
[12:11:13] * lewellyn doesn't agree on palowoda
[12:11:35] <palowoda> I'm hurt.
[12:12:08] <lewellyn> and i'm feeling friendly right now! ;)
[12:12:28] *** taemun_ has joined #opensolaris
[12:12:41] <lewellyn> and i'm out of pistachios. so time to wash up the dinner pans and find the warm body in my bed
[12:12:45] <lewellyn> &
[12:12:48] <palowoda> I should get a free dinner when your nasty than. :)
[12:13:01] <lewellyn> so should i.
[12:13:15] *** encode_ has joined #opensolaris
[12:13:27] *** crazed_ has joined #opensolaris
[12:13:34] *** sand_man has quit IRC
[12:13:56] *** kitche_ has joined #opensolaris
[12:14:11] *** dama0_ has joined #opensolaris
[12:14:45] *** totzillo has joined #opensolaris
[12:15:04] *** sand_man has joined #opensolaris
[12:15:17] *** soro has quit IRC
[12:15:45] *** houst0n1 has joined #opensolaris
[12:15:50] *** Edgeman2 has joined #opensolaris
[12:16:01] *** sergiusens has quit IRC
[12:16:37] *** Tygrys^ has joined #opensolaris
[12:16:49] *** sickness has quit IRC
[12:17:08] *** RoyK^ has joined #opensolaris
[12:17:46] *** Edgeman has quit IRC
[12:17:57] *** Edgeman2 is now known as Edgeman
[12:18:25] *** gosx_ has joined #opensolaris
[12:18:29] *** houst0n has quit IRC
[12:18:29] *** RoyK has quit IRC
[12:18:29] *** nickw has quit IRC
[12:18:29] *** leito has quit IRC
[12:18:29] *** omry|work has quit IRC
[12:18:29] *** Hedonista has quit IRC
[12:18:29] *** gosx has quit IRC
[12:18:29] *** Tiger^ has quit IRC
[12:18:29] *** taemun has quit IRC
[12:18:29] *** crazed has quit IRC
[12:18:29] *** kitche has quit IRC
[12:18:30] *** ottom has quit IRC
[12:18:30] *** hohum has quit IRC
[12:18:30] *** kjetilho has quit IRC
[12:18:30] *** dama0 has quit IRC
[12:18:30] *** kokotier has quit IRC
[12:18:30] *** jwit has quit IRC
[12:18:30] *** paul has quit IRC
[12:18:30] *** encode has quit IRC
[12:18:30] *** sickness has joined #opensolaris
[12:18:30] *** taemun_ is now known as taemun
[12:22:27] *** kjetilho has joined #opensolaris
[12:22:28] *** leito has joined #opensolaris
[12:23:09] *** omry|work has joined #opensolaris
[12:23:55] *** calumb has joined #opensolaris
[12:24:05] *** calumb has joined #opensolaris
[12:25:03] *** dalibor has joined #opensolaris
[12:25:39] *** nickw has joined #opensolaris
[12:32:31] *** hugo has joined #opensolaris
[12:33:47] *** popster has joined #opensolaris
[12:36:04] <popster> Can anyone help with resolving the following error? ld: fatal: library -lmcpp: not found
[12:36:30] <sickness> I'm back
[12:36:55] *** LouisJB_ has quit IRC
[12:37:18] *** LouisJB_ has joined #opensolaris
[12:37:57] <Berny> popster, install libmcpp?
[12:41:05] *** timeless_mbp has quit IRC
[12:41:05] <popster> yeah
[12:42:35] <popster> Berny: I am not sure the proper way to go about that?
[12:46:09] *** calumb has quit IRC
[12:46:51] <madwizard> Coffee
[12:46:57] <madwizard> Have a coffee
[12:48:15] *** calum has joined #opensolaris
[12:48:38] <popster> Need to make a fresh cup of tea :/
[12:50:16] <Berny> popster, i don't think it is in the repos (maybe source juicer has it), so i'd say go build ;-)
[12:51:17] *** Dagobert_ has joined #opensolaris
[12:51:19] <popster> Only src I have found is linked with a debian package but don't know if thats what I want.
[12:52:37] *** zenroom has quit IRC
[12:52:59] *** Phenox_ has quit IRC
[12:54:05] *** Phenox_ has joined #opensolaris
[12:54:32] *** Mousk4r has joined #opensolaris
[12:54:45] *** Dagobert has quit IRC
[12:54:46] *** Dagobert_ is now known as Dagobert
[12:55:07] *** LenZ has quit IRC
[12:56:25] <popster> This is what I have found and I see in the .tar.gz a debian folder with the libmcpp but when I build from source I dont see it making a shared library.  Is there a way to correctly tell it to make one? http://packages.debian.org/lenny/libmcpp-dev
[12:57:50] *** twilling has joined #opensolaris
[12:59:43] *** merzo has quit IRC
[13:01:32] *** merzo has joined #opensolaris
[13:04:08] *** tosh has quit IRC
[13:06:07] *** LaidBack_01 has joined #opensolaris
[13:06:24] *** zenroom has joined #opensolaris
[13:07:07] *** lateau has quit IRC
[13:07:20] <popster> Sorry I am tired, found the original source but I am still stalled in making the lib.
[13:08:25] *** calum has quit IRC
[13:08:27] *** tru_tru has quit IRC
[13:08:48] *** LenZ has joined #opensolaris
[13:09:28] *** tosh has joined #opensolaris
[13:12:44] *** kwinz3 has joined #opensolaris
[13:17:51] *** tru_tru has joined #opensolaris
[13:19:18] *** timeless_mbp has joined #opensolaris
[13:19:18] *** timeless_mbp has joined #opensolaris
[13:20:23] *** houst0n1 is now known as houst0n
[13:20:29] *** houst0n has joined #opensolaris
[13:24:17] *** perlmongo has quit IRC
[17:22:15] *** echelog has joined #opensolaris
[17:22:25] <lucalenardi> Does anybody know ho to change the driver for an nForce ethernet chipset from nge to nfo? This site (http://tinyurl.com/2v5j7ck) shows how to do the trick, but I still get an error anytime I try to load the driver: "Warning: Driver (nfo) successfully added to system but failed to attach"
[17:23:09] <houst0n> Tonnerre: Cacti is just for graphing/stats
[17:23:20] <houst0n> Plays well with switches/snmp
[17:24:01] *** popster has left #opensolaris
[17:24:19] *** timsf has joined #opensolaris
[17:24:36] *** kinkie has quit IRC
[17:25:15] *** perlmongo has joined #opensolaris
[17:27:20] *** timeless_mbp has quit IRC
[17:29:31] *** ingenthr has joined #opensolaris
[17:30:53] *** kwinz3 has joined #opensolaris
[17:33:25] *** dijenerate has joined #opensolaris
[17:36:00] *** piwi_ has joined #opensolaris
[17:38:22] *** olivier_54 has quit IRC
[17:39:01] *** Rach has joined #opensolaris
[17:39:45] *** randw has quit IRC
[17:39:49] *** randw has joined #opensolaris
[17:39:49] *** kwinz3 has quit IRC
[17:42:10] *** popster has joined #opensolaris
[17:45:14] <Rach> Is it possible to change the response to a fault in the Fault Manager ?
[17:45:30] <tsoome> ?
[17:47:07] *** raichoo has joined #opensolaris
[17:47:16] *** calSHOP is now known as calumb
[17:47:17] <Rach> I want to change the "fault.fs.zfs.io_failure_wait" fault.
[17:47:22] *** datadigger has quit IRC
[17:48:19] *** vivid has joined #opensolaris
[17:48:32] <tsoome> you wanna clear it?
[17:49:22] <Rach> No, just change the action from its current nothing, to rebuild on a hot spare if available.
[17:49:24] *** RoyK has joined #opensolaris
[17:49:48] <Rach> Current response is: "No automated responce will be taken."
[17:49:58] <tsoome> if you have hotspare defined, it will be used automatically
[17:50:12] *** Dagobert has quit IRC
[17:50:38] <Rach> Only in some situations.
[17:50:47] *** datadigger has joined #opensolaris
[17:50:58] <tsoome> ?
[17:51:28] <Rach> The defined hot spare is not always used when a drive fails.
[17:51:39] <tsoome> like when?
[17:52:00] <Rach> And for the FMA fault class "fault.fs.zfs.io_failure_wait" no action is taken.
[17:53:02] <tsoome> i just dont get it - if you have physical io error, why spare is not used?
[17:54:10] *** sndcrb has joined #opensolaris
[17:54:19] *** ozux has joined #opensolaris
[17:54:41] <Rach> Its do to the simple fact that a hot spare will only be used automaticly if the fault class trickered in the fault manager is setup to do something.
[17:54:55] <CIA-33> Anurag S. Maskey <Anurag.Maskey at Oracle dot COM>: 6944720 upgrade with ephemeral location causes network/location to go into maintenance on boot
[17:55:23] <RoyK> how can I check the disk type/model attached?
[17:55:41] <Rach> type format
[17:55:43] <Stric> iostat -En
[17:55:54] <RoyK> Stric: thanks
[18:00:47] *** piwi_ has quit IRC
[18:04:28] *** houst0n has quit IRC
[18:04:31] <Shoggoth> I've got a solaris box running as a paravirt VM and I've just hot added a disk to it... how do I get solaris to recognise the disk?
[18:05:07] *** popster has quit IRC
[18:05:37] <Stric> devfsadm -v
[18:05:58] *** popster has joined #opensolaris
[18:06:04] <thepickle> hey all.. just starting to test/learn openhacluster, and running into a problem.  I have two nodes, "cluster1" and "cluster2 and a quorum server "quorum1"
[18:06:25] <thepickle> when i reboot one of the nodes, the other node goes down... message is "WARNING: CMM: Reading reservation keys from quorum device quorum1 failed with error 2"
[18:06:28] <Rach> tsoome: look at this bug report: http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6905317
[18:07:05] <thepickle> can't figure out what's wrong..
[18:08:06] <Shoggoth> Stric: tried that and it didn't seem to work
[18:08:36] <Stric> Shoggoth: Then I guess the disk wasn't added to the VM as it should
[18:08:53] <tsoome> hm, interesting, 5months open bug.
[18:09:07] <Rach> Yeah.
[18:09:16] <Rach> But it seems to be an old bug.
[18:09:17] <Shoggoth> hmmm.... block-attach didn't give any errors and it's return value was 0
[18:09:30] <RoyK> http://pastebin.com/6JgB9Kvd are these error numbers since last reboot?
[18:10:12] <Rach> Problem is the fault manager isn't "programmed" to take action in all the different ways that a disk can fail, but only some of them.
[18:10:32] <Rach> And for those ways that aren't programmed, it does nothing.
[18:12:14] <RoyK> wierd - last one of my drives died, the spare kiked in
[18:12:33] <RoyK> kicked, even
[18:12:46] *** nage_ has joined #opensolaris
[18:12:48] <tsoome> well, go src.opensolaris.org and check it out
[18:13:03] *** xtrondo has quit IRC
[18:13:05] <Rach> RoyK: its all to do about what fault class is triggered in the Fault Manager.
[18:13:29] *** ozux has quit IRC
[18:13:34] <Rach> Some fault classes don't have any default auto response defined.
[18:13:49] <RoyK> is this possible to control/redefine?
[18:14:02] <Rach> That was my initial question :-)
[18:15:06] <Rach> If you can't setup your own auto responses, then the Fault Manager is not very useful I think.
[18:16:05] <RoyK> perhaps some mdb hack
[18:16:13] <tsoome> never researched it.
[18:16:28] <Rach> mdb ?
[18:16:40] <tsoome> no, fma
[18:16:53] <RoyK> got a hint on how to change zfs_scrub_limit with that, so that the system performs better during scrub - echo zfs_scrub_limit/W0t2 | mdb -kw
[18:17:31] *** fOB has joined #opensolaris
[18:17:42] <Rach> What does that scrub limit do ?
[18:17:45] <RoyK> that lowers max I/Os queued per vdev from 10 to 2
[18:17:59] <Rach> Ahh..
[18:18:12] <Rach> But won't the scrub take longer then ?
[18:19:00] <RoyK> I don't think so - it didn't change much on my test system, but that was mostly idle
[18:19:23] <RoyK> I guess the scrub will take longer on a busy system
[18:19:31] <Rach> Ok
[18:19:50] <RoyK> but then, some might want a slower scrub rather than a slower system
[18:20:20] *** quenelle has joined #opensolaris
[18:20:48] *** quenelle has left #opensolaris
[18:21:02] *** quenelle has joined #opensolaris
[18:21:18] *** quenelle has left #opensolaris
[18:22:09] *** KermitTheFragger has quit IRC
[18:23:15] *** balrog has quit IRC
[18:24:05] *** derchris has quit IRC
[18:24:11] *** derchris has joined #opensolaris
[18:24:19] *** xtrondo has joined #opensolaris
[18:24:22] *** balrog has joined #opensolaris
[18:25:30] *** kwinz3 has joined #opensolaris
[18:26:12] *** Erwann has quit IRC
[18:28:16] *** houst0n has joined #opensolaris
[18:28:27] <houst0n> Boink. Where the fuck is gnome-keyring? ... I.e .... what  pkg?
[18:28:29] <houst0n> lewellyn: ping.
[18:30:02] <alanc> if it's in a pkg, then pkg search should tell you
[18:30:14] <richb> $ pkg search gnome-keyring
[18:30:14] <richb> INDEX      ACTION VALUE                                PACKAGE
[18:30:15] <richb> basename   dir    usr/lib/gnome-keyring                pkg:/library/gnome/gnome-libs at 0 dot 5.11-0.137
[18:30:15] <richb> basename   dir    usr/share/gtk-doc/html/gnome-keyring pkg:/library/gnome/gnome-libs at 0 dot 5.11-0.137
[18:30:15] <RoyK> I have this 27TiB server with 3x7 2TB drives in 3xRAIDz2 and we're running low on disk space, so I just got an offer for a new disk shelf for 12 more 2TB drives. The box is mostly used for archiving satellite data, so speed isn't too much of a concern so long as the data is available. Would using these new 12 drives in a single raidz2 be appropriate for this?
[18:30:23] <alanc> but didn't they do something silly like rename it to seahorse?
[18:30:29] <richb> ignore the 137 bit.
[18:30:49] <richb> alanc; looks like the binary isn't there.
[18:30:58] *** icarus902 has joined #opensolaris
[18:31:04] <alanc> oh, seahorse is the gui frontend: http://live.gnome.org/Seahorse
[18:31:43] <houst0n> alanc: We're using osol on sunray, gnome-keyring-daemon is hanging on logout
[18:31:53] <houst0n> same with iiim-daemon
[18:32:04] <houst0n> Can we remove/disable this shit? gconf-editor or something?
[18:32:12] <alanc> no clue
[18:32:27] *** n1c0b has joined #opensolaris
[18:32:30] <alanc> I use OpenSolaris on Sun Ray as my primary desktop, but never logout...
[18:32:54] <Stric> RoyK: what you will lack is random iops..
[18:32:58] <houst0n> Well, poses us a bit of a problem.. We logout a fair bit as we use the same boxes for sgd
[18:32:59] <alanc> iiim would be input methods, so disabling them may annoy any users who ever enter text in a non-English language
[18:33:04] <houst0n> Unless you know a clever way to resume? :P
[18:33:12] <houst0n> We only have 10 guys here, so I think we're ok
[18:33:18] <lblume> IS there a way to know if a PCI-X slot is 66 or 133 MHz?
[18:33:22] <rkeene> The SunRay is too slow to use as a primary desktop... scrolling down a web page gives me a headache due to lag
[18:33:37] <Stric> lblume: rtfmanual?
[18:33:41] <tsoome> houst0n: what build? if there is newer build, it may be fixed?
[18:33:45] <houst0n> 134
[18:34:04] <houst0n> And I don't like using -dev on our workstations..
[18:34:53] <alanc> rkeene: sounds like your Sun Ray server or network is too slow - I see no lag on scrolling web pages, but then in the office I'm using high speed ethernet to a lightly loaded X4150 with dual 3ghz Xeons
[18:34:55] <lblume> Stric: I wish, but I don't have the paper version, and the online version seems rather lacking for details :-/
[18:34:56] *** houst0n has quit IRC
[18:35:00] *** Laidback01 has quit IRC
[18:35:23] <alanc> I don't normally notice much lag when I login to the same session over DSL+VPN from home though
[18:35:32] <rkeene> alanc, My SunRay server is on an M9000 and I'm connected to it on the same VLAN at GigE on a Cisco 6509
[18:35:59] <alanc> dunno then
[18:36:04] *** sipior has quit IRC
[18:36:13] *** Laidback01 has joined #opensolaris
[18:36:28] <rkeene> It's slow whether I use the Desktop Access Client on a 3GHz workstation or the thin client (SunRay 1)
[18:36:48] <rkeene> Aside from moving to 10GigE, there's nothing I can do to make it any faster.
[18:36:55] <Stric> lblume: tg3 driver in linux can report the speed.. ;)
[18:37:07] *** Laidback01 has quit IRC
[18:37:14] <RoyK> Stric: it's just that 6 drives in a raidz2 seems wasting space
[18:37:28] *** Laidback01 has joined #opensolaris
[18:37:34] <lblume> I'm rather far from the box to boot it on something different :-)
[18:37:41] <Stric> RoyK: space, performance, cost.. pick two..
[18:37:51] *** hsp has quit IRC
[18:38:49] *** sushiandbeer has joined #opensolaris
[18:39:02] <Laidback01> RoyK, http://www.stringliterals.com/?p=77
[18:39:19] *** sushiandbeer has left #opensolaris
[18:39:23] <RoyK> thanks
[18:39:26] *** sushiandbeer has joined #opensolaris
[18:39:31] <RoyK> I guess I'll ask the guys paying for it
[18:39:33] <Laidback01> he's got an interesting discussion there of various raidz technolgies dealing with large drives like you are talking about.
[18:39:44] *** sushiandbeer has quit IRC
[18:39:59] *** sushiandbeer has joined #opensolaris
[18:40:58] *** Odin- has joined #opensolaris
[18:42:06] <RoyK> Laidback01: reading :)
[18:42:20] *** lateau has joined #opensolaris
[18:43:09] *** ozux has joined #opensolaris
[18:43:42] *** hsp has joined #opensolaris
[18:44:17] *** lblume has quit IRC
[18:45:59] <stallion_work> alanc: any idea where I can find templates for submitting an ARC case?
[18:46:19] *** jnss has quit IRC
[18:48:05] *** Mdx4 has joined #opensolaris
[18:49:17] *** Mech0z has quit IRC
[18:50:11] *** Odin- has quit IRC
[18:50:20] <RoyK> Laidback01: supermicro now have 45 drive enclosures in 4U :)
[18:50:49] *** |AbsyntH| has quit IRC
[18:51:16] <Laidback01> bet those enclosures are more than $300 too.
[18:51:44] <alanc> stallion_work: http://hub.opensolaris.org/bin/view/Community+Group+arc/onepager - though that seems couple versions behind the internal copy
[18:51:49] <RoyK> http://www.supermicro.com/products/chassis/4U/?chs=847
[18:52:24] <alanc> for a fasttrack though, the only section that matters is "4. Technical Description" - all others are omitted or automatically generated for fasttracks, and the sponsor only has to provide the section 4 text
[18:52:47] <stallion_work> alanc: gotcha. I'm not sure if this will get fasttracked, but it should be an automatic case.
[18:53:06] <alanc> automatic approval is same as fasttrack in terms of paperwork/template
[18:53:24] <stallion_work> Gotcha
[18:53:36] <stallion_work> @(#)sac_nextcase is the one I'm seeing in a similar case Garrett did for elxl
[18:54:19] <RoyK> Laidback01: about NOK 20k from the Norwegian importer, it seems, so about $3k - about the same as the 9U 50-drive enclosure the article linked to
[18:54:49] <CIA-33> Rich Burridge <rich.burridge at sun dot com>: 6354836 tar attempts to unlink a temporary file which it doesn't create
[18:56:55] *** ziyourenxiang has quit IRC
[18:57:45] *** vraa has joined #opensolaris
[18:59:11] *** kwinz3 has quit IRC
[18:59:45] *** dalibor has quit IRC
[19:00:00] *** swift has quit IRC
[19:01:35] *** EugeneKaway is now known as EugeneKay
[19:03:25] <vraa> i screwed up with phpmyadmin on webstack, i ran it before i copied over the manifests for http apache, mysql, and memcache
[19:03:41] <vraa> so now when itry running it again, it says "it appears phpmyadmin has already been configured earlier"
[19:03:41] <vraa> '
[19:05:01] *** melbogia has quit IRC
[19:07:15] *** dunc has joined #opensolaris
[19:07:26] *** wrapster has joined #opensolaris
[19:08:46] <wrapster> im trying to get referrals working on my b134 machines.
[19:08:59] <wrapster> i end up with a silly error like this.. http://pastie.org/939449
[19:09:02] *** comay has joined #opensolaris
[19:09:06] <wrapster> what am i doing wrong?
[19:09:10] *** ChanServ sets mode: +o comay
[19:10:34] *** jthunder_ has joined #opensolaris
[19:11:10] *** Odin- has joined #opensolaris
[19:13:29] *** lattera has joined #opensolaris
[19:14:28] *** TomJ has quit IRC
[19:14:35] *** Odin- has quit IRC
[19:17:50] *** Mick27 has joined #opensolaris
[19:18:11] *** Beket has joined #opensolaris
[19:21:46] *** melbogia has joined #opensolaris
[19:24:26] *** Transformer has joined #opensolaris
[19:24:46] *** Odin- has joined #opensolaris
[19:26:59] *** Transformer has quit IRC
[19:27:49] *** timeless_mbp has joined #opensolaris
[19:27:49] *** timeless_mbp has joined #opensolaris
[19:27:57] *** ravenslay3r has joined #opensolaris
[19:28:20] *** ravenslay3r has left #opensolaris
[19:29:11] *** n1c0b has left #opensolaris
[19:34:23] <popster> Does anyone know where I can find libiconv?
[19:34:51] *** thepickle has quit IRC
[19:35:42] <RoyK> popster: sunfreeware.com might be a start
[19:35:56] <RoyK> I thought it was in osol already, but doesn't look like it
[19:35:57] <datadigger> popster: I think it is integrated in the C library
[19:36:41] *** Mousk4rak1 has joined #opensolaris
[19:38:51] <popster> there is an iconv() in libc but it doesn't seem to satify this software
[19:39:03] *** Mousk4r1 has quit IRC
[19:39:32] <cfs> perhaps it wants gnu iconv
[19:39:39] <cfs> which is SUNWgnu-libiconv
[19:41:06] <tsoome> there was only 1 thing gnu libiconv does what solaris wont
[19:41:07] *** achester has joined #opensolaris
[19:41:27] <tsoome> it cann accept empty tring as to or from argument.
[19:41:39] <tsoome> string*
[19:43:26] <cfs> well that is the general rule for most, if not all, of libc on solaris
[19:43:35] <popster> tsoome: I am trying to build ICE-3.4.0, any suggestions on what course I should take?
[19:44:08] <cfs> popster, I'm guessing configure is barfing, what does config.log say is wrong with libc's iconv
[19:44:24] <alanc> if it needs gnu libiconv, then download it from ftp.gnu.org and build it
[19:44:30] <popster> cfs: configure succeeds, make fails
[19:45:32] <cfs> er, ok, but _how_ does it fail?
[19:46:09] <popster> cfs: ld: fatal: library -liconv: not found
[19:46:12] <cfs> alanc, or as I suggested install the package from the repo.
[19:46:26] <cfs> popster, fix the Makefile, remove -liconv
[19:46:28] *** _setuid_H has joined #opensolaris
[19:46:31] <tsoome> you can fix configure or go for easy way
[19:46:35] *** balrog has quit IRC
[19:46:42] <popster> alanc: I have SUNWgnu-libiconv installed and up to date from the repo.
[19:47:11] <popster> tsoome: Okay, I will try that, just was worried it might break things.
[19:47:32] <tsoome> i personally see no point poisoning system with foreign bits, esp gnu bits, but sometimes its not worth the effort
[19:47:33] *** hugo has quit IRC
[19:48:13] *** calumb has quit IRC
[19:48:13] <alanc> I didn't think there was a current version of gnu-libiconv in the repo
[19:48:36] *** balrog has joined #opensolaris
[19:48:41] <tsoome> iconv is iconv really. depending on time frame, gnu iconv can have more code pages defined.
[19:49:10] <alanc> pkg on 134 reports:
[19:49:12] <alanc>           Name: SUNWgnu-libiconv
[19:49:16] <alanc>          State: Not installed (Obsolete)
[19:49:19] <alanc>           Size: 0.00 B
[19:49:54] <tsoome> and if sun makes iconv("name", ""); to be accepted, there is no need for gnu iconv:P
[19:50:04] <tsoome> err, oracle:P
[19:50:06] <popster> Maybe its not installed then, I thought I had, but by now this zone is so f'd its not even funny.
[19:50:10] <cfs> tsoome, not the solaris way
[19:50:22] <cfs> Solaris is not HP/UX!
[19:50:26] <cfs> :P{
[19:50:31] <cfs> s/{//
[19:50:33] <tsoome> ?
[19:50:38] <alanc> I thought the i18n folks were working on that, as part of the GNU compatibility project
[19:51:11] <tsoome> would be nice really
[19:51:14] <popster> Careful, I hear this channel is indexed by google. Rather I found questions I had asked previously while searching.
[19:51:30] <tsoome> :D
[19:51:40] <tsoome> why you need to be careful?:D
[19:51:53] <alanc> ah - http://arc.opensolaris.org/caselog/PSARC/2009/561/mail was what I was thinking of - "Pass-through iconv code conversion"
[19:52:46] <alanc> which should already be in google's index 8-)
[19:54:21] *** jthunder_ has quit IRC
[19:54:36] <tsoome> well, i find it disturbing to see OS flooded with several alternate instances of software, each adding little or virtually no additional bonus. thats why i try to stick as much on os  parts as possible.
[19:54:43] <tsoome> or as reasonable.
[19:54:56] <cfs> tsoome, I agree.
[19:55:05] *** dalibor has joined #opensolaris
[19:55:57] <popster> tsoome: How do I remove this from the Makefile?
[19:56:19] <tsoome> since you are asking that, you probably shouldnt:D
[19:56:44] <popster> tsoome: Well I thought I knew what I was looking for, but evidently not.
[19:56:46] <tsoome> it depends on how iconv functions are used.
[19:57:27] <tsoome> if incompatible functionality is used, it may be not worth troube.
[19:57:31] <tsoome> trouble*
[19:57:42] <tsoome> but that needs code review
[19:58:25] *** Laidback01 has quit IRC
[19:58:45] <cfs> popster: find . -name Makefile | xargs grep '-liconv'
[19:58:55] *** derchris has quit IRC
[19:58:55] <cfs> search and destroy from there on in
[19:59:00] *** derchris has joined #opensolaris
[19:59:56] <tsoome> check configure first, if iconv is optional
[19:59:56] *** dalibor has quit IRC
[20:00:03] *** dalibor has joined #opensolaris
[20:05:54] *** derchris has quit IRC
[20:08:17] *** Calibellus has quit IRC
[20:08:20] *** aruntomar has quit IRC
[20:11:51] *** madwizard has quit IRC
[20:14:26] *** ravenslay3r has joined #opensolaris
[20:14:27] *** nage_ has quit IRC
[20:15:29] *** ravenslay3r has left #opensolaris
[20:16:42] *** MuffiXXL_ has joined #opensolaris
[20:17:23] <PrestonConnors> how do i enable an ethernet interface to be "active on boot" in opensolaris?
[20:18:59] *** MuffiXXL has quit IRC
[20:19:00] *** MuffiXXL_ is now known as MuffiXXL
[20:19:07] *** lattera has quit IRC
[20:19:57] <RoyK> PrestonConnors: it's quite easy in the gui
[20:20:12] <PrestonConnors> RoyK: I was looking for the non-gui method.
[20:20:33] <RoyK> PrestonConnors: a /etc/hostname.devname is created
[20:20:49] <PrestonConnors> Yes, /etc/hostname.e1000g0 exists.
[20:20:54] <RoyK> create a new one and that will be enabled after restarting the network
[20:21:46] *** twilling has quit IRC
[20:22:07] *** lattera has joined #opensolaris
[20:22:27] <RoyK> PrestonConnors: should work then, just update /etc/netmasks
[20:22:35] <PrestonConnors> RoyK: Okay, thanks!
[20:26:00] <mui>  /w 121
[20:26:02] <mui> rr
[20:26:02] *** Laidback01 has joined #opensolaris
[20:29:15] *** dganitch has joined #opensolaris
[20:29:40] <dganitch> question about SAMBA AD intergration in opensolaris
[20:30:57] <RoyK> http://www.youtube.com/watch?v=SEVU-YLpM8A lol
[20:31:15] *** ozux has quit IRC
[20:32:13] *** MuffiXXL has quit IRC
[20:35:25] *** joaocep has joined #opensolaris
[20:35:38] *** ultrasparc-viii has joined #opensolaris
[20:36:00] <cfs> RoyK this assumes you turned off nwam, yes?
[20:37:41] *** madwizard has joined #opensolaris
[20:38:26] *** Erwann has joined #opensolaris
[20:38:41] <RoyK> nwam?
[20:38:59] <madwizard> Coffee
[20:39:16] <datadigger> madwizard: I'll join in ;)
[20:39:49] <cfs> oh , what's the name of the bot in here.
[20:40:17] <cfs> http://hub.opensolaris.org/bin/view/Project+nwam/
[20:40:52] <madwizard> smrt: explain: nwam
[20:40:52] <smrt> Um. I seem to not know anything about explain: nwam...
[20:40:58] <madwizard> smrt: explain nwam
[20:40:58] <smrt>  No really, NWAM is quite evil. I promise. Currently Phase Zero for 2010.03, Phase One was integrated to b135 (after 2010.03).  See http://hub.opensolaris.org/bin/view/Project+nwam/Phase1Spec  See also: static ip
[20:41:27] <cfs> smrt explain static ip
[20:41:35] <cfs> smrt: explain: static ip
[20:41:36] <smrt> Um. I seem to not know anything about explain: static ip...
[20:41:49] <cfs> smrt needs some re-edu-ma-ca-shun
[20:42:13] <RoyK> smrt: explain sanity
[20:42:14] <smrt> Um. I seem to not know anything about sanity...
[20:43:14] *** dalibor has quit IRC
[20:43:22] *** icarus902 is now known as stevemaresca
[20:43:44] *** Calibellus has joined #opensolaris
[20:43:55] *** joaocep has left #opensolaris
[20:44:37] *** stevemaresca is now known as icarus902
[20:44:48] *** fOB has quit IRC
[20:45:41] *** timeless_mbp has quit IRC
[20:46:49] *** OclkdMan has joined #opensolaris
[20:49:04] *** joaocep has joined #opensolaris
[20:50:44] *** Transformer has joined #opensolaris
[20:50:58] *** kwinz3 has joined #opensolaris
[20:51:15] *** mavhc has quit IRC
[20:51:59] *** Transformer has quit IRC
[20:52:55] <Wes--> smrt: explain women
[20:52:56] <smrt> Um. I seem to not know anything about women...
[20:53:09] <Wes--> Join the club! lol
[20:53:34] *** kwinz3 has quit IRC
[20:53:38] *** kwinz3 has joined #opensolaris
[20:54:01] *** MuffiXXL has joined #opensolaris
[20:54:06] *** xtrondo has quit IRC
[20:55:07] <CIA-33> Mark J. Nelson <Mark.J.Nelson at Oracle dot COM>: 6936772 nightly.sh inclusion of tools proto in allprotos causes issues for source builds
[20:55:36] <joaocep> hello all
[20:55:49] <joaocep> i dont need routing proto gre in ipfilter
[20:56:06] <joaocep> using opensolaris
[20:56:09] <joaocep> and ipfilter
[20:57:48] *** InTheWings has joined #opensolaris
[20:57:56] *** hsp has quit IRC
[20:58:01] *** jpd1234 has joined #opensolaris
[20:58:30] *** OclkdMan has quit IRC
[20:59:56] *** kwinz3 has quit IRC
[21:02:04] *** pksu has quit IRC
[21:03:13] *** timeless_mbp has joined #opensolaris
[21:03:14] *** timeless_mbp has joined #opensolaris
[21:05:10] *** melbogia has quit IRC
[21:07:07] *** webmink_ has joined #opensolaris
[21:13:54] *** EugeneKay is now known as EugeneKaway
[21:14:08] *** EugeneKaway is now known as EugeneKay
[21:15:03] *** joaocep has quit IRC
[21:16:18] *** timsf has quit IRC
[21:17:12] *** mph has joined #opensolaris
[21:18:16] *** ingenthr1 has joined #opensolaris
[21:18:16] *** ingenthr has quit IRC
[21:18:26] *** icarus902 has quit IRC
[21:21:51] *** jthunder has joined #opensolaris
[21:37:48] *** mavhc has joined #opensolaris
[21:39:21] *** dunc has quit IRC
[21:42:57] <raphi> 9 .-, lohu9tawr
[21:44:28] <trygvis> nice password :)
[21:44:47] *** balrog has quit IRC
[21:45:58] <lewellyn> i thought it was his username
[21:46:24] *** balrog has joined #opensolaris
[21:48:25] *** vraa has quit IRC
[21:50:29] *** robo_ has joined #opensolaris
[21:50:33] *** Rach has left #opensolaris
[21:54:26] *** baza11215 has quit IRC
[21:57:11] *** EugeneKay is now known as EugeneKaway
[21:57:22] <eklof> Is it expected behaviour that scrubbing a dedup-volume will take quite a bit longer than without dedupe?
[21:57:56] *** EugeneKaway is now known as EugeneKay
[21:58:47] *** icarus902 has joined #opensolaris
[21:59:58] *** ingenthr1 is now known as ingenthr
[22:00:59] *** jenky has quit IRC
[22:02:31] *** sysspoof_ has joined #opensolaris
[22:03:02] *** fleers has quit IRC
[22:03:07] *** fleers has joined #opensolaris
[22:03:48] <lewellyn> eklof: yes
[22:04:05] <lewellyn> i expect everything to do with dedupe to take longer ;)
[22:04:20] *** timeless_mbp has quit IRC
[22:04:59] *** CosmicDJ has joined #opensolaris
[22:05:04] *** TomJ has joined #opensolaris
[22:05:42] <eklof> That's what i thought as well. However, 24 hrs for 1TB :(
[22:05:54] *** Erwann has quit IRC
[22:07:19] *** gebi has quit IRC
[22:10:09] *** jthunder has quit IRC
[22:10:35] <eviljames> eklof: That seems slightly excessive.
[22:10:51] <eviljames> to me, anyhow.  but then again, I'm just a guy on irc with absolutely no evidence to back it up.
[22:10:52] <eklof> it's also compressed
[22:11:02] *** raichoo has quit IRC
[22:11:06] <eklof> gzip-9 :)
[22:11:07] *** sysspoof_ has quit IRC
[22:11:18] <eklof> no just kidding, gzip-6 i think it is
[22:11:39] <eviljames> I think that's the default, no?
[22:11:41] *** nitrile has quit IRC
[22:11:42] <eklof> or whatever the default is.
[22:11:44] <eklof> yes
[22:12:05] <eklof> no wait, that pool is uncompressed.
[22:12:11] <mbreitba> question - I know this isn't a nexenta channel, but since it's based on OpenSolaris, I figured I'd ask
[22:12:12] *** nitrile has joined #opensolaris
[22:12:20] <mbreitba> what's the best way to get IPoIB working?
[22:12:37] <eklof> what's that?
[22:12:39] <eviljames> eklof: Which hash are you using on dedup?
[22:12:42] * lewellyn sees IB and runs
[22:12:46] <eklof> eviljames: sha256!
[22:12:49] <mbreitba> It detects my adapter, and I've got it configured with an IP address, but it's unreachable from other hosts
[22:12:51] *** melbogia has joined #opensolaris
[22:12:53] <mbreitba> IP over InfiniBand
[22:12:57] <eklof> no verify
[22:13:27] *** piwi has joined #opensolaris
[22:13:42] <eviljames> I know it is processor dependent but what is the time it takes to hash 1TB of data in sha256 (approximately)?  Is it linear?
[22:14:19] <bonsaikitten> eviljames: openssl has a benchmark function ... should give you an idea
[22:15:14] <bonsaikitten> I get 81230.74kB/s on a 2.8Ghz amd64
[22:15:25] <lewellyn> bonsaikitten: he'd need to benchmark it then get back to eklof to let him know if his scrub took too long ;)
[22:15:28] <TommyTheKid> mbreitba: I assume you have TWO hosts on IPoIB and you are attempting to ping between them?
[22:15:33] <mbreitba> Correct
[22:15:43] <mbreitba> actually, I have three hosts
[22:15:47] <eklof> Actually, i have new info. it took 38hrs for backup 1.07T   548G   322G  /backup
[22:15:51] <eklof> :)
[22:15:54] <mbreitba> two windows clients, and a Nexenta client
[22:15:56] <TommyTheKid> did you sacrifice a goat and spill hte blood in a circle around your three servers?
[22:16:03] <mbreitba> windows clients can ping each other
[22:16:11] <eviljames> bonsaikitten: just ran on openbsd machine - I get around half that on a P4
[22:16:19] <mbreitba> Nexenta can ping itself, but neither of the other clients
[22:16:23] <bonsaikitten> eviljames: yeah, P4 is pretty sad
[22:16:27] <mbreitba> the other clients cannot ping the nexenta box
[22:16:40] <stallion_work> why even bother with IPoIB?
[22:16:46] <eviljames> bonsaikitten: *shrug* works fine with openbsd, as a home server.
[22:16:48] <mbreitba> it's stupid fast?
[22:17:03] <bonsaikitten> eviljames: I prefer pcengines alix boards :)
[22:17:07] <mbreitba> fastest datacenter interconnect you can get
[22:17:20] <TommyTheKid> true true
[22:17:27] <TommyTheKid> what protocol are you using over it?
[22:17:39] <mbreitba> iSCSI and NFS
[22:17:42] <stallion_work> mbreitba: Infiniband is "stupid fast", IPoIB is like wrapping a PDU in bacon and shoving over IB
[22:17:42] <lewellyn> hrm. lblume isn't around. i'm curious whether this email i got is legit and he can probably find out more readily than i can ;)
[22:17:46] <TommyTheKid> NFS should use RDMA
[22:17:56] <lewellyn> i'm afraid to open the attachments :)
[22:17:57] <TommyTheKid> I am not sure about iSCSI
[22:18:12] <lewellyn> mmmbacon
[22:18:17] <balrog> lewellyn: what are the attachments?
[22:18:24] <mbreitba> they can both use RDMA, and will bypass the IP stack if I can get that configured properly, but I want to at least try to get the IPoIB working
[22:18:27] <TommyTheKid> i want bacon
[22:18:36] <mbreitba> plus, the RDMA is even less well documented than the IPoIB
[22:18:37] <stallion_work> mbreitba: be a man and use RDMA
[22:18:39] <lewellyn> balrog: a .doc and a .pdf
[22:18:44] <lewellyn> i've apparently been invited to beijing in july
[22:18:44] <balrog> hmm.
[22:18:51] <balrog> I see
[22:18:52] <eviljames> bonsaikitten: I have found my next purchase.  Hats off to you! :D
[22:18:55] <balrog> why are you afraid then? :)
[22:19:05] <mbreitba> not that IB is well documented at all
[22:19:10] <wrapster> is there a way i can figure out what internationalization support is provided by default in osol and what can be downloaded later?
[22:19:11] <rkeene> You won't see the advantages of infiniband over 10GigE if you're using only IP over it.
[22:19:11] <bonsaikitten> eviljames: really nice hardware, I like them :)
[22:19:20] <stallion_work> mbreitba: its documented well enough.
[22:19:25] <lewellyn> because i don't have any ties to the sector, and it was sent with a gmail.com return address over 163's mail servers
[22:20:01] <mbreitba> the other reason to use IB is that our bladecenter supports it, and the total cost was less than what 10GE would have costed to impliment
[22:20:01] <balrog> ahh.
[22:20:08] <balrog> open it in google docs :)
[22:20:08] <TommyTheKid> I would focus on RDMA, although we had talked about using IBoIP because its fanstastically cheaper than 10GIG (I think)
[22:20:37] <mbreitba> well, I want to get the simpler protocols working first, then move on to the RDMA
[22:20:42] <lewellyn> i'd still have to download the attachment ;)
[22:20:55] <lewellyn> huh. bing says it's likely legit
[22:20:57] <mbreitba> that, and I don't know that the windows boxes that we're currently using work that well with NFSoRDMA
[22:21:16] <balrog> I don't see how downloading the attachment could cause issues, unless you open it
[22:21:17] <mbreitba> here's the situation - we've got a large infrastructure running Windows/Hyper-V
[22:21:36] <mbreitba> Everything is on iSCSI right now, formatted NTFS
[22:21:51] <mbreitba> we want to move to either Xen or VMWare (haven't decided which yet)
[22:22:01] <TommyTheKid> mbreitba: I would check the simple stuff first, dladm show-phys, make sure you have link, ifconfig -a .. make sure its UP, make sure it has a proper IP/netmask/broadcast, make sure there is no firewall (ipfilter?)
[22:22:04] <mbreitba> want to utilize storage on NFS volumes for data portability
[22:22:20] <TommyTheKid> unfortunately not sure how any of that maps into nexenta :)
[22:22:37] *** sysspoof_ has joined #opensolaris
[22:22:52] <mbreitba> Everything that I've checked looks fine
[22:23:06] <mbreitba> all of the IB queries show that the interfaces are all connected and up
[22:23:20] <lewellyn> phone
[22:23:21] <mbreitba> and the nexenta box is able to ping itself on the IB interface
[22:23:23] <TommyTheKid> make sure the d0ze side has no firewalls as well
[22:23:37] <mbreitba> no firewalls on the Win side
[22:23:40] <TommyTheKid> ofcourse if they can ping eachother...
[22:23:59] <mbreitba> I'm about 20k deep into this project, and it's really starting to piss me off
[22:24:11] <stallion_work> mbreitba: well... I wouldnt have high hopes for using just NFS
[22:24:17] <TommyTheKid> so is your interface ibd0?
[22:24:27] <mbreitba> ibd0 and ibd1 on the nexenta box
[22:24:30] <stallion_work> you are better off with a NAS head which supports CIFS/NFS access to the same filesystems
[22:24:55] <mbreitba> well, the nexenta box would be the NAS head end - I can provide CIFS storage out of it also
[22:25:28] <mbreitba> also - more info about the build here : http://www.zfsbuild.com/ - details about the hardware and intentions
[22:26:21] <TommyTheKid> ok yuck, our IB router is running S10 :(
[22:26:55] *** Mousk4rak1 has quit IRC
[22:27:00] <TommyTheKid> IB router -> system with 10G and IB connecitivy, routing between them
[22:27:14] <mbreitba> I've already tried using bare OpenSolaris - I actually got the IBoIP working ok on it using 2009.06, but it would frequently just drop the connection
[22:27:31] *** sventhomsen has joined #opensolaris
[22:27:31] <mbreitba> iSCSI over the IPoIB link would reset every 35 seconds (literally, like clockwork)
[22:27:47] <mbreitba> NFS wouldn't mount, but I'm wondering if that was because it was only offering it over RDMA
[22:27:57] *** pjfloyd has joined #opensolaris
[22:27:59] <mbreitba> IPoIB wouldn't work at all in 134a
[22:28:10] *** mroconnor has quit IRC
[22:29:36] *** sventhomsen has left #opensolaris
[22:30:29] <TommyTheKid> yea, so I wish I had some great knowledge to impart, I know we are messing with IB in our lab, for the same reason you are, but I don't know where we are with it.. I was just asked for a subnet to route to :)
[22:30:41] <TommyTheKid> a subnet and a route to it on serverX that is
[22:31:21] <TommyTheKid> and the reason I grumbled about S10 was nmap isnt there ;)
[22:31:47] <mbreitba> well, maybe it's time to jump back into OpenSolaris and hack on it from that end.....the Nexenta stuff is a little more black-box'ish and that seems to be complicating things
[22:31:48] <lewellyn> TommyTheKid: er... yes it is
[22:31:51] <lewellyn> install the companion cd :P
[22:31:54] <TommyTheKid> bah
[22:31:57] <rkeene> Atleast you aren't trying to do IP-over-Fibre-Channel in Solaris 10.  It's completely broken and Sun/Oracle refuses to fix it.
[22:32:00] *** phil_ has quit IRC
[22:32:09] <TommyTheKid> I am not going to go installing software on their router
[22:32:20] <TommyTheKid> hehe
[22:32:25] <TommyTheKid> IBoFC, yikes
[22:32:27] <lewellyn> your statement is still false ;)
[22:32:32] <rkeene> IBoFC ?
[22:32:36] <TommyTheKid> typo
[22:32:41] <rkeene> Oh
[22:32:50] <rkeene> fcip(7D) is completely broken
[22:33:13] <mbreitba> Nexenta actually told me that the IB card that we had wasn't on the OpenSolaris HCL......which was totally wrong, in any event though, I'm hoping to get something running
[22:33:35] <TommyTheKid> lewellyn: its not there by default, companion CD (/opt/sfw I assume?) is generally hopelessly out of date, so we stopped installing it
[22:33:51] <mbreitba> how is the CIFS support on OpenSolaris - is it going to work for my Windows hosts until I can get migrated to Xen/VMWare and use NFS?
[22:34:08] <TommyTheKid> smrt: explain cifs
[22:34:08] <smrt>  File sharing protocol commonly used on Windows.   OpenSolaris includes two cifs servers, one builtin to the kernel, and the samba userspace daemon.   (See also: cifs server, samba)
[22:34:09] <lewellyn> TommyTheKid: it may be out of date, but it is part of a "complete" install. hell, most of the os is hopelessly out of date ;)
[22:34:16] *** LenZ has quit IRC
[22:34:20] <TommyTheKid> thats true lewellyn  :)
[22:34:40] <TommyTheKid> I used to install it once, centrally and mount it NFS.. but I don't know how long its been since I had that :)
[22:34:42] <mbreitba> and I'm guessing that if I use CIFS, I can export that via NFS later and Xen/VMWare will still be able to read it?
[22:34:55] *** Metamucil has joined #opensolaris
[22:35:09] <TommyTheKid> sharesmb=yes?
[22:35:18] <TommyTheKid> smrt: explain cifs server
[22:35:19] <smrt>  OpenSolaris has a new CIFS (Windows file sharing in SMB) server in the kernel. Good resources for it include http://dlc.sun.com/osol/docs/content/SSMBAG/troubleshooting.html and http://wiki.genunix.org/wiki/index.php/CIFS_Service_Troubleshooting . The CIFS provides better performance and easier administration for CIFS than Samba but Samba provides other SMB services (See also: samba)
[22:35:26] <lewellyn> mbreitba: if you use the kernel cifs, you can manage nfs and cifs with the same interface
[22:36:30] <mbreitba> alrighty, I think my answer is to turn back to OpenSolaris then and hack some more on it
[22:37:11] <mbreitba> also, just out of curiosity, as I didn't find this in my digging - if the NFS server detects RDMA capable hardware, does it only offer NFS over RDMA, or does it offer it up over IP at the same time?
[22:38:21] <TommyTheKid> I *think* if both ends are on IB, it skips over the IP stack (uses RDMA)
[22:38:51] <TommyTheKid> but maybe thats because it doesn't offer NFS over IP on IB interfaces,,, I couldn't tell you there ;)
[22:39:12] *** pjfloyd has left #opensolaris
[22:39:31] <mbreitba> alrighty - thanks for the info - back to hacking on OpenSolaris - maybe I just need a bigger axe ;)
[22:39:32] *** PrestonConnors has left #opensolaris
[22:39:48] <TommyTheKid> mbreitba: good luck :)
[22:41:21] <TommyTheKid> i dont honestly know why they used S10, probably to keep from scaring the OPS folks :p
[22:41:51] <TommyTheKid> ... apparently either the routing is messed up or they are firewalling, either way, I am not mucking with their config :)
[22:42:22] <lewellyn> hrm. has anyone packaged tortoisehg for opensolaris yet? it seems overly logical to have ;)
[22:42:30] <TommyTheKid> lewellyn: quick question... if I build a package with SFE, can I like scp it to another system and install it?
[22:42:50] <TommyTheKid> not that I hafe SFE working yet ;)
[22:43:09] *** LadySerena has joined #opensolaris
[22:43:47] <alanc> lewellyn: GUI's are illogical though
[22:44:00] * LadySerena is a happy kitty
[22:44:12] <Tonnerre> LadySerena, good!
[22:44:28] <TommyTheKid> i suppose thats better than beeing a sad salamander ?
[22:44:38] <LadySerena> At school, I got my pin and senior scrubs today.  :D
[22:44:49] <lewellyn> TommyTheKid: if you build an svr4 package.
[22:45:02] <lewellyn> alanc: they're still faster than opengrok at getting history ;)
[22:45:21] <TommyTheKid> lewellyn: otherwise, I can just point at the pkg server that will somehow supposedly be running there on that machine and install it over the network?
[22:45:35] <lewellyn> and if you're poking around in nautilus, it's fater than opening a terminal and navigating
[22:45:45] <lewellyn> TommyTheKid: in theory. i've not built ips yet ;)
[22:46:07] <TommyTheKid> this could be interseting :)
[22:46:11] <TommyTheKid> interesting even
[22:48:05] <lewellyn> remember last night how i was asking about pkgtoolrc? ;)
[22:48:56] <TommyTheKid> hmm, I was bashing my head on the bootstrap script, so I didnt see it :)
[22:49:16] <lewellyn> i still don't know why you were editing it.
[22:49:21] <lewellyn> just run the damn thing :P
[22:49:28] <TommyTheKid> it doesnt work
[22:49:38] <lewellyn> smrt: explain doesn't work
[22:49:39] <smrt>  Then make it file for unemployment benefits.
[22:49:41] <lewellyn> heh
[22:49:45] <TommyTheKid> hah
[22:49:59] <TommyTheKid> I have an NFS home directory and it doesn't cooperate
[22:50:30] <TommyTheKid> so I tried to edit the USERHOMEDIR= setting to point at /ws/tm.... and that sorta works, except that it still somehow makes some stuff appear in ~/
[22:51:20] <lewellyn> yeah. i keep meaning to patch it to deal with nfs homedirs.
[22:51:50] <lewellyn> but tomww is quiet and i don't know how out-of-date the current download is vs his local copy
[22:52:24] <TommyTheKid> :)
[22:52:35] <TommyTheKid> if he doesn't commit his copy up, then its not there :P
[22:52:50] *** timeless_mbp has joined #opensolaris
[22:52:51] *** timeless_mbp has joined #opensolaris
[22:53:39] <lewellyn> patches are less useful when they're against virtually every line and you're trying to commit them against something fairly different ;)
[22:54:01] <TommyTheKid> agreed
[22:55:49] *** CompanionCube has quit IRC
[22:56:54] <lewellyn> that was a needlessly long address
[22:58:21] <TommyTheKid> see one of the last couple comments in myemail... it looks like someone was attempting to move all the "tunables" up above the "code" parts and missed some stuff and/or the logic in the "code" parts needs changed, as ... "# setting XXX can override the above methods" doesn't make sense :)
[22:59:33] <lewellyn> imo, there shouldn't be tunables :P
[22:59:39] <lewellyn> i think i've told him that a few times though ;)
[22:59:55] <TommyTheKid> what if I don't want spew in my home directory ?
[22:59:56] *** afrasi has quit IRC
[23:00:20] <lewellyn> i also kinda want to take another stab at fixing his os detection
[23:00:26] <lewellyn> TommyTheKid: then do it all by hand ;)
[23:00:31] <TommyTheKid> heh
[23:00:38] <lewellyn> imo, the bootstrap script exists to get a consistent environment
[23:01:20] <TommyTheKid> sure, but the consistency can have a different "base" directory to store stuff in, I basically wanted it to be in a local directory instead of $HOME
[23:01:24] *** CompanionCube has joined #opensolaris
[23:01:47] <lewellyn> mv spec-* /where/ever
[23:01:49] <lewellyn> :)
[23:02:06] <TommyTheKid> heh
[23:02:37] <lewellyn> also, i find it works best if you run it for a special build-only user
[23:03:01] <lewellyn> then you can source the environment in .profile
[23:03:03] <TommyTheKid> see that is against policy, we aren't supposed to have accounts that aren't associated with an "employee" :)
[23:03:26] <TommyTheKid> I can see how that would be beneficial though
[23:03:35] <TommyTheKid> simpler at least
[23:03:45] <lewellyn> you can't have zones?
[23:03:56] *** ivo_ has joined #opensolaris
[23:04:37] <TommyTheKid> the only way I can think of making it even remotely work is calling it a "role" that you assume to build stuff :-/
[23:05:00] <TommyTheKid> it would effectively be like an "applicaiton user" then? right?
[23:05:14] <TommyTheKid> this is "company policy" stuff, sorry
[23:05:49] <lewellyn> i find it odd that there's no provision to allow employees clean build accounts
[23:06:10] <TommyTheKid> SOX complaince crap
[23:06:10] <cfs> RBACs ftw?
[23:06:16] *** sushiandbeer has quit IRC
[23:06:46] *** jnss has joined #opensolaris
[23:06:58] <lewellyn> TommyTheKid: a role may work
[23:07:08] <lewellyn> it won't now
[23:07:20] <lewellyn> but with what i had in mind to clean up the script anyhow, it should.
[23:07:28] *** bhall has joined #opensolaris
[23:07:41] <cfs> 14:08 < TommyTheKid> see that is against policy, we aren't supposed to have accounts that aren't associated with an "employee"
[23:07:42] *** boyd has joined #opensolaris
[23:07:44] <TommyTheKid> the USESU apparently fails to set properly, cause it makes me su to myself :D
[23:07:48] <cfs> ^ if that is the case, can you have root?
[23:07:50] *** ingenthr1 has joined #opensolaris
[23:07:56] <lewellyn> TommyTheKid: i've never used that
[23:08:01] <TommyTheKid> root is a role, thats where I got the idea of a role for build
[23:08:32] <TommyTheKid> apps are supposed to use role accounts as well, but some fail miserably with that, and are thus given a *NP* local user account
[23:08:37] <lewellyn> i really want to go through and do another hatchetjob on the bootstrap script :)
[23:08:53] <TommyTheKid> the main thing is not having to manage WHO is logged in as that user and what they have access to and who they are and what are they doing
[23:08:57] <TommyTheKid> (I think?)
[23:09:21] <TommyTheKid> its really quite messy ;)
[23:09:33] <TommyTheKid> i like the idea in principal tho
[23:12:04] *** radsy has joined #opensolaris
[23:12:08] *** ingenthr has quit IRC
[23:15:48] *** ingenthr1 is now known as ingenthr
[23:15:53] *** projectmoon has quit IRC
[23:16:51] *** projectmoon has joined #opensolaris
[23:16:58] *** projectmoon has quit IRC
[23:18:51] *** jollyd has joined #opensolaris
[23:18:57] *** jollyd has left #opensolaris
[23:21:30] *** hunter has joined #opensolaris
[23:24:43] *** robo_ has quit IRC
[23:26:03] *** psychicist has quit IRC
[23:26:54] *** totzillo has quit IRC
[23:27:25] *** Hedonista has joined #opensolaris
[23:27:44] *** tomaw has quit IRC
[23:28:04] *** wdp_aao has joined #opensolaris
[23:31:47] *** Amse has joined #opensolaris
[23:31:51] *** tomaw has joined #opensolaris
[23:33:41] *** RoyK has left #opensolaris
[23:35:39] <gosx_> is there an easy way to see how many interrupts/sec a given driver is processing?
[23:35:43] *** gosx_ is now known as gosx
[23:36:17] <gosx> my load is ~1.4 with no zpool traffic, and i think it's just getting flooded by open explorer windows (that's what snoop seems to think)
[23:36:39] <gosx> 20% kernel, 80% idle, 0% iowait
[23:37:03] <gosx> and for any experts, would enabling TCP/checksum offload help?
[23:37:31] * duckinator looks in, has no clue what gosx is talking about, and goes back to playing with the bootloader he's coding ;P
[23:38:28] *** bklang has quit IRC
[23:39:01] <TommyTheKid> gosx: maybe something with dtrace, but I am probably not a good person to ask :)
[23:39:24] <bdha> intrstat?
[23:40:41] *** merzo has joined #opensolaris
[23:42:54] <stallion_work> gosx: lockstat -s 255 -I sleep 10
[23:43:22] <lewellyn> stallion_work: is that your answer to everything? :)
[23:43:37] <stallion_work> lewellyn: bah, use -s 50
[23:43:48] <lewellyn> :)
[23:43:50] <stallion_work> lewellyn: when the kernel runs amok, there isn't much you can do to narrow it down :P
[23:43:56] <lewellyn> yeah.
[23:44:00] <stallion_work> gosx: rather, issue: lockstat -s 50 -I sleep 10
[23:45:14] *** Thrae has quit IRC
[23:47:34] *** sysspoof_ has quit IRC
[23:49:19] *** nanase has joined #opensolaris
[23:49:51] <jbit> why does hte pkg source contain lots of bobcat references? :)
[23:50:15] <lewellyn> jbit: there should be a url mentioned ;)
[23:50:24] *** macros73 has quit IRC
[23:50:26] <jbit> lewellyn: there is, but there are lots more too
[23:50:30] <jbit> http://src.opensolaris.org/source/xref/pkg/2009.06/src/tests/api/t_history.py#323
[23:50:36] <jbit> bobcat-alive-plus-poison-in-box
[23:50:41] <lewellyn> oh yes.
[23:50:43] *** niq has joined #opensolaris
[23:50:44] *** niq has joined #opensolaris
[23:50:46] <lewellyn> schrodinger's bobcat ;)
[23:51:03] <lewellyn> awesome pun imo
[23:51:08] <jbit> a friend is debuggin pkg issues
[23:51:13] <jbit> and was like... wtfh
[23:51:16] *** EugeneKay is now known as EugeneKaway
[23:51:26] <lewellyn> the bobcat's been declawed some
[23:51:35] *** EugeneKaway is now known as EugeneKay
[23:52:34] <gosx> thx stallion_work: the top consumer (other than i86_mwait) was bnx_intr_1lvl
[23:53:36] <gosx> followed by various smb_ functions
[23:53:45] <jbit> (it is the same friend who made  http://temp.jpgclan.com/plan.html , since he's seeing it alot)
[23:54:52] <lewellyn> jbit: doesn't work in w3m but i get the idea
[23:54:55] <CIA-33> Mark J. Nelson <Mark.J.Nelson at Oracle dot COM>: 6936772 nightly.sh inclusion of tools proto in allprotos causes issues for source builds (fix quotes)
[23:56:12] <jbit> i want a creating plan tshirt :P
[23:56:26] *** synegy34 has quit IRC
[23:56:27] <lewellyn> ok. i'm about to try installing firefox 3.6.3 from the .pkg... using the old package names, what do i need to remove first? SUNWfirefox, SUNWfirefox-devel, and SUNWfirefoxl10n-*?
[23:57:05] <lewellyn> or should i remove SUNWfirefox-bookmark too?
[23:57:26] <jbit> does pkg remove dependant packages?
[23:57:34] * lewellyn is using pkgrm
[23:57:34] <jbit> because a bunch of stuff depeneds on the firefox package
[23:57:37] <jbit> ah okay
[23:57:58] <lewellyn> not ballsy enough to attempt this on indiana (yet) ;)
[23:58:24] <lewellyn> supposedly 3.6.3 will fix a javascript issue i was banging my head against earlier today
[23:59:06] <lewellyn> well, not necessarily 3.6.3, but $newerThanInstalled
[23:59:08] *** sndcrb has quit IRC
[23:59:13] <jbit> tempted to try to get chromium compiled on solaris at some point
[23:59:24] <jbit> there'S no real reason why it shouldn't work without the sandboxing
[23:59:28] <lewellyn> there's a few people working on that, i thought
[23:59:35] <jbit> ah okay
[23:59:41] * lewellyn may be wrong

   April 28, 2010  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >