=== n0ts_off is now known as n0ts === dendrobates is now known as dendro-afk [00:18] anyone up at this hour? [00:18] trying to set up pptpd server [00:25] anyone able to assist me? [01:25] stgraber: ok - i'm out this week so i'm unlikely to be pushing anything. but will look for it when i do. === Aaton is now known as Aaton_off === n0ts is now known as n0ts_off === n0ts_off is now known as n0ts === n0ts is now known as n0ts_off === n0ts_off is now known as n0ts === n0ts is now known as n0ts_off === n0ts_off is now known as n0ts === n0ts is now known as n0ts_off [04:21] i mount the 12.04 iso to /media/cdrom then run apt-cdrom ,nothing happened .just say "Repeat this process for the rest of the CDs in your set" [04:21] the sources.list is nothing added [04:27] New bug: #1004494 in bacula (main) ""Job not run" message after any attempt to run a job using bconsole" [Medium,Expired] https://launchpad.net/bugs/1004494 [04:47] How can I verify that no files are missing from installed packages? [04:53] undecim_ maybe apt-get check would do [04:54] mvp: I don't see anything in the man page that does what I'm looking for === n0ts_off is now known as n0ts === n0ts is now known as n0ts_off === n0ts_off is now known as n0ts [06:13] hello, I can not restart the network interfaces! i have assigned multiple IP addresses http://paste.ubuntu.com/1120895/ === n0ts is now known as n0ts_off === smb` is now known as smb [07:11] morning [07:16] I wish there had been a way to do caching with SSD on linux, like there is on ZFS... === obelus|2 is now known as obelus [08:03] When I run crontab -e, it opens it with nano, how do I change it to vim? [08:03] found it… ~/.selected_editor :) [08:06] Pupeno_W, i was not aware of this file, usually i just set the EDITOR env variable [08:06] melmoth: seems to be new, on ubuntu 12.04, when I run crontab -e, it asks me which editor I want, and then it configures the editor env variable in .selected_editor. [08:12] Pupeno_W: thats how it works i suppose :P [08:12] melmoth: hello btw :) [08:12] hola koolhead11 [08:13] melmoth: reading up on quantum [08:14] good luck [08:14] i played with it a bit, but not long enough to understand === nibalizer is now known as n1balizer === n1balizer is now known as niba1izer === niba1izer is now known as nibalizer [08:16] Pupeno_W: no, that is 'sensible-editor' [08:21] melmoth: and how far you reached? [08:21] Daviey: hello sir [08:21] not far :) Just reading some doc, trying to play with it, realising i had no idea what i was doing [08:21] and then i had to go back to something else. [08:21] melmoth: http://networkstatic.net/2012/05/openstack-essex-and-quantum-installation-using-openvswitch-from-scratch/#comments [08:21] this might help [08:22] not on my radar anymore. [08:22] hi, having serious issues with raid performance on 12.04 [08:22] melmoth: what are you on these days :P [08:23] as yesterday, i cannot make a single host openstack install work. [08:23] s/single host/single_host mode/ [08:23] this seems to be related to a kernel bug? quite amazed it's not solved yet... anyways, max_hw_sectors_kb is set to 127 which is, to say the least, a pretty inefficient number [08:23] melmoth: lol. what was the issue. are yu on folsom [08:24] on all the disks below it, it's 512 (a nice binary multiple) === n0ts_off is now known as n0ts [08:24] hey koolhead11 [08:24] am on essex. I cannot reach floating ip for vms that are hosted on another node that the one running nova-network. [08:24] it's a 7 disk raid-5, so I assume 6*512 (the parity disk shouldn't be counted I presume) would be the optimal value [08:25] in any case, 127 is not a binary multiple and will thus always suck [08:25] iptables rules seems ok, but what i found funny is, the compute node receive the tcp connextion, but the source address is the kvm bridge, not the actualo box sending it. [08:25] melmoth: you need to run nova-network and nova-api on same machine [08:25] i have have others in distributed setup [08:26] already the case [08:29] melmoth: It should be too easy :P [08:29] morning o/ [08:30] hola lynxman [08:30] * koolhead11 wonders if its really lynxman [08:30] melmoth: bonjour :) [08:30] koolhead11: it's-a me [08:30] koolhead11, well, it s not, and i have no idea why the connexion appears to be coming from the kvm bridge ip and not the node ip itself. [08:30] wich i think, is the reason why the connexion cannot happen, as the packets comes back to another place it has been originated from. [08:31] <_ruben> freakynl: doesn't seem to be 12.04 specific, i see similar numbers on for instance mdadm volumes on 10.04 [08:32] melmoth: i smell some screwed network config [08:32] lynxman: how is life man? seems like puppet is keeping you too busy [08:32] it s all just a kvm network. I try with both isolated and nated one. [08:33] i try with setting the compute nodes bridge in /Etc/network/interfaces, and letting nova deals with it. [08:33] sames beaviour [08:33] _ruben: I've seen reports on it happening since 2.6.38 iirc [08:33] <_ruben> freakynl: 10.04 is 2.6.32 [08:33] _ruben: the shitty thing is, I can't set it higher than 127kb due to the restriction in max_hw_sectors_kb === n0ts is now known as n0ts_off [08:33] _ruben: ouch... it's fixed in 3.3 from what I saw :) [08:34] <_ruben> freakynl: in that case i'd suggest you'd try to confirm that by testing with a mainline kernel build from the kernel team ppa [08:37] I'm having authentication problems with samba server, the server works when I make the file owned by one user, but I want to be able to use the write list and read list syntax, however when I do that with the directory permissions at 0775, I get access denied, even though the user is in the write list. [08:38] <_ruben> obelus: i'm guesing the user will still need proper rights on dir/file level .. unless you do stuff like force user/group in samba [08:40] Well, the idea is that I don't want every user on the server to have write permissions, I want a few to have write, and the others to have read... if I force user will it allow me to just use write list/read list? [08:43] _ruben: I can do that, but this is a production backup/iscsi server. Filing a bug now, will wait a couple of days to see if they respond [08:44] _ruben: I've just set force user = nobody, force group = nogroup, and write list is working now. [08:44] Thanks :) [08:47] _ruben: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1031260 [08:47] Launchpad bug 1031260 in linux "RAID write performance is horrible. max_sectors_kb is set to the odd value of 127" [Undecided,New] [08:49] _ruben: I think you can click a link there that you're affected too, might speed up things :) [08:49] koolhead11: yeah, quite busy lately :) === ninjak_ is now known as ninjak [09:06] <_ruben> freakynl: how about setting it to 64, as that's a "nice" number? might be worth a try [09:08] lynxman: hehe. where are you these days? [09:08] as in London or somewhere else [09:09] <_ruben> freakynl: how you liking lio btw? curently a scst user myself [09:09] <_ruben> stuff like vaai might make it rather interesting to make the switch [09:10] koolhead11: London still [09:10] lynxman: cool. enjoying olympics? :) [09:11] koolhead11: well, they've been busier in another side of the city so far :) === Lcawte|Away is now known as Lcawte [09:14] :) === n0ts_off is now known as n0ts === n0ts is now known as n0ts_off === mcclurmc_away is now known as mcclurmc === n0ts_off is now known as n0ts === n0ts is now known as n0ts_off === n0ts_off is now known as n0ts === n0ts is now known as n0ts_off [09:40] _ruben: I did that, but since that's still way less than a stripe it doesn't increase performance at all (ideally, you'd write entire stripes, now it writes partially on 1 disk, reads the reads the rest and writes parity for N data disks every friggin' time. In our case that means it's load * 7 :/ [09:41] _ruben: liking it so far, better than iet anyways. However, whilst I saw VAAI on the roadmap (free) quite some time ago it has vanished (can't find the page any more anyways and it was targeted for 3.3 or 3.4 back then... those are out...) [09:41] seems like they've reserved it for their paid version (it's beyond me why a party that wants to make profit of their modules got included in the kernel over scst...) [09:42] don't like the config tool tho', it didn't want to let me export the MD device. Had to hack rtslib to add the major numbers of the md devices to their list [09:42] I rather have tools giving me the ability to fuck up, than that it prevents me from doing things that should be possible. But I guess that's what you get when you allow a company that wants to sell their modules into the kernel [09:44] but I'm not the first that accuses it from being overzealous in protecting the admin (I don't think end-users will export volumes...). But alas, it's just the config tool. It seems to run way more stable than IET and it has lvl 2 error correction [09:44] iirc, iet has none :D [10:02] Daviey: any idea whom to poke to get help in testing LXC with essex === mcclurmc is now known as mcclurmc_away [10:04] koolhead11: what issue are you seeing? [10:06] Daviey: no issue as of now http://wiki.openstack.org/LXC is a doc [10:06] and we have this too http://wiki.openstack.org/LXC [10:06] so it appears bit confusing to me [10:06] http://docs.openstack.org/essex/openstack-compute/admin/content/lxc.html [10:06] i meant [10:08] koolhead11: probably best wait for zul [10:09] Daviey: cool. i want to test KVM LXC Qemu everything on my single setup :P [10:12] koolhead11: super! [10:22] guys where is the binary for "source" located? [10:23] "which source" does give me no result [10:27] adac: 'source' command is a shell builtin [10:27] so, dash, bash, etc. [10:28] DavidLevin_, hmm I wanted to run it within an init.d script. Is this even possible? [10:28] Sorry Daviey [10:29] adac: yeah [10:29] adac: people tend to use "." [10:30] Daviey, hmm but how? when I do "source PATH" then i do get: "source: not found" [10:31] adac: So, ". ~/.bashrc" ~= "source ~/.bashrc" [10:33] Daviey, "." seems to work [10:34] adac: source is a bashism, "." is portable [10:34] Daviey, well now another python binary should be used since the source has been set. but still the old one is running [10:34] Daviey, oh I see [10:35] * Daviey gets back to stuff [10:36] Daviey, thank you for your help! [10:49] Hi there, got massive problems :s ive got xen vm's and have ubuntu server 12.04 lts installed one one of them, just done sudo apt-get update then upgrade, took a few mins(ok) then asked for a reboot(i did) it rebooted, i can ping it, but I cant ssh to it, I can't get to the websites on it either. Any ideas please? === matsubara is now known as matsubara-lunch [10:51] Daviey: seems like some shit has happned. it has to do with networking. will try it again :) [10:57] AdvoWork: do you have access to dom0? [10:57] if so, connect to the console and see what's going on :) === cpg is now known as cpg|away [11:02] mattt, i can do xm console myvm but its stopped displaying anything: http://pastebin.com/rnSUQtSV [11:02] it doesn't generate anything past that last line [11:07] AdvoWork: were you ever able to console on that instance? [11:07] (and by console, i mean a login prompt) [11:10] mattt, i only ever ssh'd into it really [11:10] the thing is, its came back up now, i dunno if one of our ext support guys has done something or just it took time :S [11:10] i doubt time to be fair === al-maisan is now known as almaisan-away [11:28] AdvoWork: odd, well sorry i couldn't be more helpful :) === mcclurmc_away is now known as mcclurmc [11:39] hello [11:42] hey guys one of my friends has around 200 pc's and he would like to deploy ubuntu on those machines in one go, they are all connected in a lan, and have identical configuration, any suggestions will be helpful, thank you. [11:44] anyone ? [11:45] <_ruben> freakynl: the config tool (targetcli i think?) seems a tad nasty indeed. guess i'll be sticking with scst for now then. iet i didn't like much either [11:45] mi3: did you do any research before asking the same? [11:46] yeah, unsuccessfully. [11:46] <_ruben> freakynl: and indeed, vaai is enterprise only, bah [11:47] issa: http://serverfault.com/questions/192719/best-way-to-deploy-ubuntu-onto-100s-of-laptops [11:47] issa: http://fai-project.org/ [11:48] issa: piece of advice, dont go for mint, use ubuntu instead, they give you 5 year LTS. [11:48] :D [11:48] koolhead11: issa needs help :D [11:48] mi3: thanks for the advice [11:48] issa: no problem. [11:53] <_ruben> freakynl: also sucks that the enterprise edition doesn't have any pricing listed on the website. they're probably ashamed of how high the prices are :p [12:01] good morning === matsubara-lunch is now known as matsubara [12:14] hi all. how do i get the non-headless version of openjdk-7-jre installed on a server? [12:15] this is a 12.04 server spawned off the standard image available on ec2 [12:18] _ruben: the config tool is odd compared to other stuff yea, but I haven't seen the big picture. The enterprise version is rtsadmin (go figure :)) and is supposedly able to manage a lot of machines from 1 console/server [12:18] <_ruben> freakynl: ah [12:19] it's also immature (targetcli that is, haven't seen rtsadmin). As in if I export a device through fileio with buffering enabled, the targetcli will just show an exported device, it doesn't say whether buffering is enabled or not [12:19] <_ruben> freakynl: i like scst's approach: simple config with with a perl based cli tool that can read/write/verify the config file [12:19] <_ruben> hehe [12:20] yea I much rather had scst in the mainline kernel than lio... it replaces a lot more of the scsi stuff and enables virtual tape libraries and other stuff. Overkill for most but very nice to build your own storage solutions [12:36] list === railsraider_ is now known as railsraider === ninjak__ is now known as ninjak === dendro-afk is now known as dendrobates === ninjak_ is now known as ninjak === dendrobates is now known as dendro-afk [13:36] so I haven't had a change to look at ubuntu cloud in the past couple years (last time I looked it routed all traffic through 1 server..) is it reasonably performant now? we're in a crisis and I can either build my own or (preferably) throw a reasonable upstream-supported solution together [13:37] for starters it'll just be vms (with lots of heavy network IO but very little disk) hosted on an existing iscsi cluster. eventually it might get more cpu/memory intensive with new vms [13:41] jamespage: whats the url for the bug tracking that you showed me yesterday [13:43] Just booted my machine up (ubuntu server 12.04) and it says: pyGRUB version 0.6 an gives 2 options to select Ubuntu 10.04 LTS and Ubuntu 10.04 LTS (Single-User) whats going on with that, why isn't it 12.04? [13:47] advowork, that is a xen thing [13:48] and that would be cause the xen loader is looking at grub 1 config, instead of the grub2 config 12.04 installs [13:48] most likely [13:50] patdk-lap, how would i swap where it slooking? [13:51] that is easy, but then it wouldn't work [13:51] grub2 uses a totally different config file format [13:51] hopefully you can locate a pygrub upgrade [13:51] oh, what can I do then? this was a totally fresh install of Ubuntu, and then an upgrade [13:51] ubuntu isn't the issue [13:52] it's what is outside of it, that is [13:52] your xen host [13:52] so whats up with the xen host, its a probelm for this specific VM, or all of them? [13:52] it will be a problem for all that don't use grub1 [13:52] how can i tell if a VM is using grub1? [13:59] christ. 10 node minimum for cloud? any way to get started with 2 nodes? (I've got 6 more in the building but they aren't released from inventory yet. and thats still only 8.) or is there a better channel for this? [14:00] Disconnect: you can start with one if you want. [14:01] jamespage: Disconnect ^^^ is deploying own cloud, anyone can help out? [14:07] xnox: thanks :) i'm looknig at https://help.ubuntu.com/community/UbuntuCloudInfrastructure which seems to say there is a live image for single node, but I can't find info on whether that is expandable to a "real" cloud as I get the other servers. [14:07] * Disconnect has 1 node now, 2 nodes once I have a vm migrated to it (its on the other node right now) and 6 downstairs waiting for inventory tags [14:08] damn. "Cannot join #ubuntu-cloud (Channel is invite only)." [14:08] Disconnect: you do want "Ubuntu Cloud Infrastructure with MAAS and Juju" [14:09] minimum 10 nodes, as in it doesn't make sense to do that unless you will be adding more nodes now or later. [14:09] and I want to apologize straight off, I'm sure 99% of what I will ask is online somewhere, but I've got 100 scientists sitting waiting for a solution (they don't care what) so I want to knock out the biggest show-stopper questions now before I start the deeper evaluation/installation/etc. [14:09] public cloud?! [14:09] ah ok. I think in theory we'll hit 8-10 total. not quite identical but they are all vm capable, etc. [14:10] xnox: private. @ a .gov. just had a really really terrible storage system eat itself again, this time beyond repair [14:10] MAAS is basically pxe boot + magic sauce. and then you will get a $ juju add-node to add those to the existing cloud i believe [14:10] jdstrand: ping your nova bug is on quantal right? [14:11] zul: all the stuff I filed within the last few days is quantal, yes [14:11] the short term is to get openfiler running on a redundant cloud (using shared storage, no drbd mess) using those servers. the long term is that we are mandated to go to vms and such, and ubuntu priv cloud seems like the direction to take things.. [14:11] jdstrand: ok cool [14:11] jdstrand: im just going through them and fixing them up now, django-compressor should be in the archive real soon and there will be a MIR coming for it [14:12] New bug: #1031359 in cinder (universe) "[MIR] cinder" [Undecided,New] https://launchpad.net/bugs/1031359 [14:12] the good news is they bought beefy storage servers just before I got here. they didn't get them set up (fortunately, since I saw how they did the last set. hacked up rhel/centos blend with heartbeat and nfs, fencing each other every day or 2) [14:12] zul: thanks-- I have a todo to attend to django-compressor [14:12] jdstrand: cool beans [14:13] does openstack do any sort of automatic failover? (rebooting the vms is ok, so long as its automatic) [14:14] crap. this might be a moot point. http://wiki.openstack.org/IscsiBlueprint says iscsi for backend storage isn't supported [14:15] Disconnect: no you will have to setup the failover stuff yourself [14:15] Disconnect: openstack + libvirt -> iscsi is fine === n0ts_off is now known as n0ts [14:16] ok. so I can stick the vms on iscsi. thats good. and i can set up heartbeat if I can shove the vm configs into an iscsi drive. looks like its time to start building. === niemeyer_ is now known as niemeyer === n0ts is now known as n0ts_off [14:41] hey guys, ive installed the latest version of ubuntu server, i have a problem with my dns server, adding it to /etc/resolv.conf doesnt work, it says at the top that "anything i write here manually will be overwritten" .. [14:41] so where do i add the dns server then? [14:43] <_ruben> Righton: /etc/network/interfaces .. dns-nameserver a.b.c.d [14:46] _ruben, is that new in 12.04? [14:46] i just tried that anyway and i still cannot resolve any hostnames [14:47] adding it there, will just make it populate /etc/resolv.conf I think [14:47] <_ruben> you'll need to down/up the respective interface for it to populate /etc/resolv.conf [14:48] <_ruben> or you could update the latter manualy for the time beign [14:48] i did try out both, restarted the network interface and i still cannot ping/resolve any hostnames for some reason [14:49] Righton, did it populate resolv.conf with your DNS servers? [14:50] _ruben, is it dns-namserver or dns-nameservers [14:51] Righton, seems it should be dns-nameservers a.b.c.d according to http://manpages.ubuntu.com/manpages/lucid/man8/resolvconf.8.html [14:52] <_ruben> Righton: what does `dig www.google.com` show? [14:52] right, now t does! [14:52] it is dns-nameserverS with an S at the end [14:52] thansk guys =) [14:52] yw [14:53] Where can I rant about how horrible some ubuntu packages are in a constructive manner? [14:53] just one question though, is this new in 12.04? [14:53] cuz i dont remember doing it that way with older versions [14:53] Righton, by default, yes [14:53] i see [14:54] jMCg, is it something the ubuntu-server team can fix? [14:54] tgm4883: I was just wondering, aloud, what the appropriate channel would be. It's about tomcat7 and tomcat7 not being installable side-by-side. [14:55] tomcat7 and tomcat7? [14:55] isn't that the same thing? [14:55] jMCg: launchpad bugs [14:55] tgm4883: typo. 6 and 7. [14:56] jMCg, I agree with Disconnect file a bug on LP [14:56] <_ruben> jMCg: #ubuntu-server / #ubuntu-motu / #ubuntu-devel / etc depending on the nature of the package, see the maintainer directive in the source packages [14:56] *nod* [14:56] <_ruben> and ofcourse bug filing indeed :) [14:56] after they reject the bug, then come here and start yelling :) [14:56] Yeah, but I can't *rant* in bug report, that'd be unproductive and very unconductive to the bug's solution :D [14:57] you can after its definitively rejected [14:57] there is nothing good about ubuntu's tomcat management, really. [14:58] * Disconnect doesn't miss it. Last job was 300 vms, ubuntu-server running tomcat. [14:58] Disconnect: how did you manage it? [14:58] lots of kvm+libvirt, lots of puppet, lots of cussing. and the devs just bundled basically everything into the deployment. [14:59] the only thing worse was the ruby "gems"-that-aren't-gems [14:59] ? [14:59] I think your devs need some serious talking to. [15:00] no. the way ubuntu handles ruby libraries is beyond brain damaged. [15:00] Oh, yeah. That's true. [15:00] in perl, for example, a packaged cpan module can be installed from the package and used as if cpan had installed it. in ubuntu, a packaged gem cannot. [15:01] damnit. think I broke my bonding config. bbiab. [15:01] I install rvm and go with that, because, no.… No Because, it's just ARGH [15:01] Disconnect: might also wanna look into fpm [15:02] jMCg: rbenv is so much better though! [15:02] rbenv? Better than rvm? Haven't seen that yet, or needed to look into it because rvm just works so well for me. [15:03] jMCg: rvm polutes shell environment. Just type $ env and see how much stuff you have there [15:03] and rvm is shell functions which break certain things. [15:03] rbenv is lightlight and is simply $PATH and that's it. [15:03] xnox: like sudo, which is why you need rvmsudo? [15:04] exactly. [15:04] with rvenv sudo just works the way it always does. [15:04] rvm has negative shell start up performance [15:04] rvenv is faster =) [15:05] negative? [15:05] has a negative impact on shell start-up, but I don't think there is such a thing as negative time. At least I haven't seen it proven yet. [15:06] jMCg: https://github.com/sstephenson/rbenv [15:06] jMCg: as in it negatively affects the start up time, i.e. huge increase (3-4 seconds) [15:07] i'm not a fan of ruby in general ("lets throw away decades of knowledge about how to manage systems, environments, etc and put the developers in charge of reinventing it!" .. they're literally making all the same mistakes over again. it'd be fun to map perl milestones (cpan, packages, etc) to ruby and see how soon it won't totally suck [15:08] :-\ [15:11] One of these days gems will stop sucking, I'm sure of that. [15:13] its like slackware+cpan all over again. (or solaris+cpan) [15:14] Solaris is dead (to me) I hope the Illumos folks are doing that one better. [15:14] Disconnect: never used slackware/cpan [15:15] On Solaris (10) I wasn't able to use it, because all of our servers where behind firewalls with no access to the internet, no compilers or libraries installed, etc… Can I get this CPAN package installed - NO! [15:15] basically the old days of "download compile install" where upgrading was best done by starting with a new box (since there is almost no installation tracking). cpan made it a little better, and *nix eventually got useful packages beyond core system stuff, but.. [15:16] * Disconnect had almost 300 identical servers (well, sets of servers. obviously memcache vms aren't identical to tomcat..) by avoiding crap like that [15:16] This reminds me: https://twitter.com/hirojin/status/230118763444465664 [15:17] I "updated" my FreeBSD VM yesternight. [15:18] yah. exactly.. at least ports do installation tracking, of sorts. [15:18] of sorts. [15:18] there used to be a tool to let you install apps to /usr/local/appname/ (/bin /etc /sbin etc) that managed symlinks. thats how bad it was. and ruby still is.. [15:18] Disconnect: GNU stow [15:19] thats the one! [15:19] Ah, right! I remember that one, I didn't use it when I built my LFS, back in the day. [15:19] I used NetBSD's pkgsrc [15:27] https://bugs.launchpad.net/ubuntu/+source/tomcat7/+bug/1031395 << should I open the same thing in tomcat6? Or should I have opened this bug in libtomcat7-java in the first place? [15:27] Launchpad bug 1031395 in tomcat7 "tomcat6 and tomcat7 are not installable side-by-side" [Undecided,New] [15:36] New bug: #1031395 in tomcat7 (main) "tomcat6 and tomcat7 are not installable side-by-side" [Undecided,New] https://launchpad.net/bugs/1031395 [15:40] hey all, looking for some advice for mysql-cluster and how to know how many nodes i need to set up? anyone have a good tutorial or advice? === dendro-afk is now known as dendrobates [15:42] Anyone here able to help with trying to re-route incoming requests from a specific IP range to port 443 to go to port 22 instead? I'm using ufw at the moment. [15:43] I am looking for some direction on setting up sendmail to send mail out from my 11.04 server. I have 5 total, one of which already can send email. [15:46] r3dLunchb0x: why sendmail? [15:47] jMCg: it is what i'm used to, but if postfix is just as easily setup that would work too. I am running into disk space issues and need to set up alerting... [15:48] And instead of sending snmp traps, you're putting *more* stuff on your machine, stuff that keeps queues, and produces big logs. [15:50] jMCg: I have nagios installed, but just the basics. still need to send alerts. [15:52] I'm confused, doesn't nagios notify a central server of such things, which can then do the alerting for you, or am I crack? [15:54] jMCg: I believe you're right. I'll check into getting nagios alerting turned on. The nagios master is running on the server that has mail capabilities. [15:56] next I need some help with settings for openfile handlers on the servers. any one got a good guide on this? I know how to set it up on redhat enterprise...ubuntu server is waaaaay different lol [15:56] Did oem-config-server ever make it out of the dream state? (https://wiki.ubuntu.com/OemConfigServer) [15:57] r3dLunchb0x: what *is* openfile handlers? [15:58] it is a limit/setting for the amount of open files a process or shell or tcp connection is allowed to open. Also there is a base setting for when the OS starts. if it is too low, after running say a big java program it'll "crash" with Broken Pipe errors. increase the amount of open file handlers and things go more smoothly. [16:00] right now I have about 5 HP proliant servers running ubuntu server 11.04 with open file handlers set to 1024, for about 30 developers whose shells get stopped by this very error. === Ursinha` is now known as Ursinha [16:03] r3dLunchb0x: it would appear you're looking for /etc/pamd.limits.conf -- that's same on RHEL, innit? [16:06] anyone have any experience getting bridged networking to work with kvm/libvirt? [16:08] ssmy: most people in #virt (on oftc) do. [16:11] jMCg: thanks, I'll try there === babyface_ is now known as baby_face [16:13] jMCg: not really looking to "limit" the devs, more like open it up. lol === baby_face is now known as babyface_ [16:14] r3dLunchb0x: same/same, really. [16:15] jMCg: now to just track opening that up is actually proving my point that it will go smoother. ;-) [16:33] is https://wiki.ubuntu.com/ServerTeam/MAAS/AvahiBoot expected to actually work? it insists on a working ubuntu mirror [16:35] adam_g: http://pastebin.ubuntu.com/1121750/ [16:35] adam_g: the nova-compute stuff is the correct one right? === railsraider_ is now known as railsraider [16:37] damn. booting the ub cd (even after its listed as 'commissioning') results in immediate shutdown. what a mess. [16:39] Disconnect: what hardware? [16:40] virtual. trying to get private cloud stuff working. the maas server is working and shows the 2 nodes as commissioning, but no matter waht i try (booting the usb image, booting the server cd, etc) it fails in some or another spectacular way [16:41] the avahi usb image fails because its trying to use precise and - for whatever reason - can't find it on archive.ubuntu.com (I checked, its there..) and the server cd works up to where it contacts the maas server, then it shuts down. [16:42] hmm. if anyone is doing wiki editing on the maas pages, it'd sure be nice if those screenshots were zoomable. at least far enough to read what the node status should be (commissioning, done, on fire, etc) [16:47] hmm. i think i see whats going on. the next step is to leave them off and get juju going. [16:50] nope. same problems with the usb key [16:51] oooh. so close. the commandline assumes a dhcp server. ok i think i can fix this. yay. ish. [16:52] bug 1028453 [16:52] Launchpad bug 1028453 in ubuntu-meta "Quantal Ubuntu Server minimal install oversized" [High,Confirmed] https://launchpad.net/bugs/1028453 [16:52] smoser, Daviey: ^^ gonna fix this? [16:54] jamespage: good question [16:54] jamespage: and a good place to discuss it :) [16:54] Daviey, shall I make the case for dropping then? [16:54] jamespage: I'm not sure it /can/ be fixed, it's just kernel bloat now? [16:55] Daviey, I think that is the key difference [16:55] and now its just the module set [16:55] the -virtual kernel still gets used in cloud images.... [16:55] so we should not drop it but I think we should consider dropping the minimal virtual install from the ISO [16:56] jamespage: I'd support that [16:57] I suspect it's not used, and if i am honest.. i'd rather not encourage it.. cloud images should be used instead [16:58] i need a walk through for a ubuntu server install any takers? [16:58] Daviey, agreed [17:00] Anyone knows about a DC++ hub in the Ubuntu repositories? === mcclurmc is now known as mcclurmc_away [17:06] i need a walk through for a ubuntu server install any takers? [17:07] !guide [17:07] The Ubuntu server guide may be found at http://help.ubuntu.com/12.04/serverguide/C/ [17:08] thank you much [17:09] halvors: I don't think there are any DC++ hubs in the repos anymore. I used to run opendchub, but it's gone in 12.04 as well. So I just don't bother anymore. [17:10] jMCg: just a thanks for the #virt recommendation. They were able to help me solve my problem [17:18] ssmy: I know, and they're mostly nice too! [17:22] i typed init 6, init 0, shutdown, and even pressed the power button. I get messages saying the system is going down for reboot or shutdown, but nothing happens.... how can I tell it I want to reboot it without hitting the reset button? [17:22] ninjai: why is it hanging? [17:22] no idea [17:22] it acts as if nothing has happened [17:23] i can still use the machine as normal.. [17:24] ninjai: dmesg? /var/log/syslog? Why is it not shutting down? Something must be preventing it from doing that. [17:24] ssmy: Is there any unoffcial opendchub ppas? [17:25] dmesg http://pastebin.com/qAjYrXNe [17:26] syslog: http://pastebin.com/BzkUq7sk [17:28] I think I have to press the button :( [17:32] reboot and its happy. Weird. [17:37] can someone tell me how to remove LVM. [17:38] i messed up on a raid install [17:41] SpamapS: alive? [17:42] or anyone else on the server team or anyone who can answer an SRU related thing (regarding php5) [17:43] can someone tell me how to remove LVM? [17:45] LordOfTime: sure whats up? [17:46] SpamapS: can we get https://bugs.launchpad.net/ubuntu/+source/php5/+bug/1014044 SRU'd for Precise? this is a significant regression, and causes issues with nginx and other web servers with php5-fpm [17:46] Launchpad bug 1014044 in php5 "PHP5-FPM not reporting errors to web server (nginx)" [Medium,Fix released] [17:46] a fix was included in 5.4.x in Debian [17:46] which was synced to Quantal [17:46] but not in Precise [17:46] there *is* a 5.3.x patch [17:46] on the PHP main site [17:46] i can't dget the package just yet, to test if the patch applies [17:47] and it'd be ubuntu-only changes since it was fixed upstream [17:47] LordOfTime: it would go more smoothly if you attached a debdiff to that bug.. ;) [17:47] SpamapS: indeed,but as I said i can't dget from here [17:47] LordOfTime: and if you added the necessary info as per https://wiki.ubuntu.com/StableReleaseUpdates [17:47] can we at least get the SRU process started, pending the debdiff upload? [17:48] i'll do that shortly, i'll just need someone who can nominate against Precise (apparently bugcontrol can't set that, and lp times out when i try to) [17:48] LordOfTime: I've accepted it for precise.. just neesd a Test Case / Regression Potential added to the top of the description [17:49] LordOfTime: if you could attach the patch to the bug that would be great as then somebody could at least package up the fix for precise [17:51] SpamapS: i plan to do that anyways [17:51] one thing at a time, at work i've only got iexplore [17:51] which is a piece of ********** [17:51] it lags like crazy :/ [17:53] SpamapS: i fully anticipate generating the deb diffs myself anyways :P [17:53] * LordOfTime is trying to get his name on more debdiffs xD [17:54] can someone please tell me how to remove LVM? [17:54] i need help urgently [17:55] LordOfTime: good plan. :) rememebr tos ubscribe sponsors [17:55] vychune: remove it in what way? [17:55] SpamapS: not the first SRU i've done :P [17:56] that, and i've had to deal with security bugs, they're even more difficultish xD [17:56] i need to partition the drive i installed it without doing RAID first [17:56] i cant do raid with it it on there === Aaton_off is now known as Aaton [17:57] SpamapS: the only reason i am getting this SRU'd is because (1) its a significant functionality failure, and (2) even though its not a bug against nginx, its a highly blocking bug for running PHP programs. And apparently, the verison in Oneiric is unaffected, so it is a very specific regression. [17:57] any help will be appreciated SpamapS [17:58] LordOfTime: your time is very much appreciated. If you can get it done this week it will be included in 12.04.1 [17:58] SpamapS: its marked as Triaged, but can I assign it to me and set it as In Progress? I fully plan to have this fixed by 00:00 UTC-4 [17:58] LordOfTime: yes do that [17:59] done, i'll have this fixed by the end of the day [17:59] assumign nothing horrible, such as Comcast going "ZOMGWTHBOOM" again [17:59] , happens [17:59] BAH, evil keyboard is evil. [18:01] SpamapS: when i generate the debdiff, the changelog entry... should it point to precise-updates, or precise-proposed? [18:01] * LordOfTime ran into this issue previously [18:01] * LordOfTime plans on testing within a PPA or a local repository separately, targetted to 'precise' for testing purposes only [18:02] (I ran into this with a security bug, they wanted me to target it to a specific repo with the debdiff) [18:02] damnit. 3 hours arguing with squid-deb-proxy. [18:03] well, with what turned out to be squid-deb-proxy anyway [18:03] SpamapS: any suggestions? [18:04] LordOfTime: precise-proposed [18:04] vychune: I am not sure I understand what you're saying [18:04] vychune: you have a drive, not your root partition, which has RAID+LVM ? [18:05] i have a raid 1 with two drive and both drives got used for LVM somehow === bitmonk_ is now known as bitmonk [18:06] vychune: probably old partition tables [18:06] vychune: can you lose all of the data? [18:06] yes there is no data [18:08] gudie me sensei lol [18:09] guide* [18:09] vychune: so I would make sure its all unmounted.. [18:09] vychune: and stop the RAID1 [18:09] vychune: then 'pvs' to see the "physical volumes" that lvm thinks it has [18:09] how do i stop it? [18:09] vychune: then for each of those "pvremove /dev/..." [18:10] SpamapS: when's 12.04.1 releasing? [18:10] vychune: I forget, I don't have any mdadm boxes to look [18:10] so i know the absolute latest i need to get the debdiff in by [18:10] LordOfTime: August twenty-something [18:10] LordOfTime: August 9 you need to have it in proposed so it can make it to -updates [18:10] oh so i have to get this in by Friday for consideration/inclusion [18:10] i'll have it done by sooner! [18:10] wait im on the cd system in recue mode [18:11] rescue [18:11] LordOfTime: 9 is the absolute drop dead time.. get it in way sooner [18:11] Did oem-config-server ever make it out of the dream state? (https://wiki.ubuntu.com/OemConfigServer) [18:12] vychune: anyway, pvremove all those physical volumes. Then 'fdisk /dev/...' and make sure the partition table is clear or has only one thing (RAID1 as the whole disk) [18:13] SpamapS: you mean like at the earliest 5 hours from now? :P [18:13] pfft, this is an easy thing to apply [18:13] afaict at least [18:13] (lets hope its not like one of the patches for an nginx cve a while back, that actually had to be redone manually) [18:13] s/redone/recreated/ [18:14] (I should know, i was the one who created the debdiffs) [18:14] SpamapS: "Device contain neither a vaild DOS partions table, nor Sun......." [18:16] s/partion/partition [18:16] SpamapS: while i'm digging in the package anyways, want me to gen the debdiff for this too? https://bugs.launchpad.net/ubuntu/precise/+source/php5/+bug/1006738 [18:16] Launchpad bug 1006738 in php5 "php5-fpm segfaults with error 4 in libc-2.15.so" [High,Triaged] [18:17] or should i wait until the other debdiff i'm working on is accepted [18:17] vychune: ok. Then you should just be able to re-create the RAID1 and lvm should not return. [18:19] ok trying it now [18:20] thanks in advance [18:23] SpamapS: WORKED [18:23] thank you kindly [18:24] anyone know the console login of one of the maas nodes? [18:24] vychune: np [18:24] Disconnect: "one of the maas nodes" ? [18:25] the same mess I've been arguing with all day: https://help.ubuntu.com/community/UbuntuCloudInfrastructure [18:26] it "conveniently" cobblers it. which would be good, if it didn't require assigning it to some random ip in the dhcp pool. i think i may have to start over with the nodes now that i've got other issues straightened out, but i'm not sure and i was hoping to get access to one of tehse things to figure it out === cpg|away is now known as cpg [18:35] SpamapS: repoke. doyou want me to generate the debdiff for LP Bug 1006738 as well, separately? [18:35] Launchpad bug 1006738 in php5 "php5-fpm segfaults with error 4 in libc-2.15.so" [High,Triaged] https://launchpad.net/bugs/1006738 [18:35] * LordOfTime never got an answer due to lagspike [18:36] New bug: #1005298 in maas "warning: kernel option length exceeds 255 during maas-import-isos (dup-of: 1003460)" [Low,Confirmed] https://launchpad.net/bugs/1005298 [18:37] LordOfTime: afk for a bit. It would be good to group those together .. I think [18:45] SpamapS: i'll first make sure they apply, then i'll combine :P [18:48] SpamapS: the server is seeing the two drives as one device. why? [18:48] i thought RAID was two devs? [18:50] SpamapS: i'm tempted to not fix the bug for Quantal if its already in upstream. I did assign myself for the Precise SRU though. just would like you to confirm this was applied upstream (LP Bug 1006738) [18:50] Launchpad bug 1006738 in php5 "php5-fpm segfaults with error 4 in libc-2.15.so" [High,Triaged] https://launchpad.net/bugs/1006738 [18:50] might want to sync that in if its already fixed upstream [18:51] LordOfTime: do you know what i can do on this problem? === azop is now known as azopfreenode [18:51] vychune: unfortunately not, if i did i'd have given help. [18:52] of course lol i'm just flustered [18:52] i'm here right now to stab SpamapS, because of bugs related to php5, and getting them updated and fixed in precise by the Stable Release Updates process [18:52] understandable [18:54] being a student in this field is hard enough lol [18:54] real world hits a knocks the wind out of my sails and everything goes nuts [18:54] vychune: should be one device, /dev/mdX [18:54] vychune: that is a RAID device [18:54] s/a/and [18:55] oh [18:55] vychune: hey, being a student in IT Security is worse :P you have to sift through $arbitrarysize of data and identify whether someone's leaking personally identifiable info and stuff [18:55] just as part of a *test* [18:55] :P [18:55] jesus [18:56] that, and a bunch of other things [18:56] * LordOfTime returns to sifting through bugs. [18:57] i see lol [19:09] SpamapS: i dont know what im doing wrongf' [19:10] wrong [19:24] ok. i clearly should have done more reading. someone sanity-check me - if i get ubuntu private cloud (maas+juju) working, I can deploy charms. charms are effectively ubuntu apps, and there are no charms for (for example) deploying a centos VM.. is that right? === cpg is now known as cpg|away [19:28] disconnect, true dat [19:29] ok. so its basically useless to me. damn :( can anyone recommend a vm-based cloud system for ubuntu? [19:29] * Disconnect isn't going down the manual-virsh road again [19:29] * bellsouth * epic4 * epic5 * ircii [19:29] ukpl;@werld [19:29] * epic4 * epic5 * ircii ! [19:29] * bellsouth , [19:30] disconnect, i'm thinking openstack? if you want a platform for provisioning vms [19:30] on 1:TEXT:*:#: { commands } on 1:TEXT:!trigger *:#: { commands } [19:30] ;D [19:30] i thought ubuntu virtual could was based on openstack [19:30] * bellsouth on *:DEOP:#:{ ;check if we are being deopped if ($opnick == $me) { ;requesting op msg BOTNICK op $chan ;kick the person that deopped you kick $chan $nick Don't deop me! } }! [19:31] http://kthx.net/ftb/bbCodeHelp.php [19:31] 4ÐÐ 4 70p1( [70p1( n4m3=70P1(_N4M3] 4ÐÐ 4 ƒr4m3 [ƒr4m3 n4m3=ƒR4M3_N4M3] 4ÐÐ 4 £1n|{ [£1n|{=£1N|{]73X7[/£1n|{] 0r [£1n|{]73X7[/£1n|{] 174£1( 73x7 [1]73X7[/1] µnÐ3r£1n3 73x7 [µ]73X7[/µ] B0£Ð 73x7 [b]73X7[/b] 1n£1n3 (0mm4nÐ [(]73X7[/(] m1R( (0Ð3 [(0Ð3](0Ð3_H3R3[/(0Ð3] === Aaton is now known as Aaton_off [19:31] * bellsouth ],uk[ [19:31] ukpe43lo9c! [19:31] * bellsouth ],leolkic9![ [19:32] b3||$0uth [19:32] !ops [19:32] Help! Channel emergency! soren, lamont, mathiaz, Pici, Daviey, Tm_T or pmatulis [19:32] D&d G@m3 = |!nk [19:32] O_o [19:32] bastid [19:32] Aw, too late. [19:32] close though! [19:34] :(( [19:38] gotta love that ops command [19:39] * genii-around makes sure IdleOne gets a reward cookie [19:39] * IdleOne eats cookie promptly [19:39] genii-around is known for taking back cookies [19:39] a server hosting a website needs DNS and mail, right? [19:40] vychune: meh, not necessarily [19:40] You need to be able to resolved DNS, not serve it. [19:40] resolve [19:40] ohhhhhhh [19:42] so the mail can go somewhere else? [19:45] Generally. [19:47] well... why does a server hosting a website even need to resolve dns? the only thing a web server has to do is translate an http request into a document and serve it... [19:49] Good point. [19:50] There may be off site resources it needs to resolve perhaps, but I guess it's not inevitable. [19:51] so the domain host will do the DNS thing? [19:51] and mail? [19:52] Could be on any server. [19:52] yeah... most domain reg services come with basic dns management (a, cname, mx), but not actual mail service as in receiving and storing mail [19:52] though many offer that as an addon [19:53] at least i assume the reg is who you're talking about when you say "domain host": [19:53] right [19:54] registar lol [19:55] My linode has been behaving very strangely. I've installed gitlab recently and mysql is occasionally crashing and I'm finding nothing but empty logs. A moment ago I could not access the server by any means, simply timing out. I issued a reboot from the linode dashboard and it took it about 20 minutes to shut it down. I tried connecting via the ajax ssh console and was given an out of memory error. Any ideas on what's going on? === cpg|away is now known as cpg [20:04] How do I run KVM VMs headless? [20:04] wizonesolutions: check out libvirt/virsh [20:04] when I run them with curses they still occupy a terminal [20:04] with the kvm command [20:04] nathwill: looks like there is a charm for openstack. so lets see if i can put some of today's adventure to good use [20:05] Disconnect: k. I was following the tips at https://help.ubuntu.com/community/UEC/Images#Ubuntu_Cloud_Guest_images_on_Local_Hypervisor_Natty_onward to get started, but it doesn't start them headless. [20:05] any links that are good to start with? [20:05] 12.04 [20:06] https://help.ubuntu.com/12.04/serverguide/libvirt.html [20:06] and I still have to figure out how to get the guests Internet access...getting my head around how it all works, I guess (used to VirtualBox) [20:07] and virt-manager is a decent GUI front-end (lower on that page) [20:07] Disconnect: In that case I've been told have to use X over SSH forwading yeah? Is it worth the trouble? [20:11] not necessary [20:12] Disconnect: which is not necessary? using virt-manager or forwarding X? [20:12] forwarding x. === cpg is now known as cpg|away [20:13] this ssh thing is pissing me off. does anyone know what the default password for a maas-installed node is? 'juju status' is telling me invalid ssh key, adn I can't get logged in to see why [20:14] wizonesolutions: forwarding is not necessary. If you need to use the gui, it sets up a vnc connection accessible via virt-manager [20:15] Disconnect: Is it passw0rd like the Ubuntu cloud images? [20:15] damnit. that'd do it. authorized_keys are empty. [20:15] ssmy: ah, ok. thanks. [20:15] wizonesolutions: hmm. might be. i just bounced it into single user [20:16] Disconnect: that works too, heh. [20:24] is there any good reason to work with ubuntu private cloud (maas+juju) if I'm actually just in need of openstack? does it provide any sort of gain or should Ij ust blow it away and fire up openstack? [20:24] MOTD, how to change it and personalize it per user who logs in? === Guest41633 is now known as TheLordOfTime [20:32] SpamapS: and again i poke thee. === Aaton_off is now known as Aaton [20:56] TheLordOfTime: Just back from lunch. Whatsup ? [20:57] r3dLunchb0x: per-user stuff should probably go in /etc/profile.d [20:58] SpamapS: cool. will check it out. [21:01] the only thing in that directory is bash_completion.sh [21:01] can I just add a new script and it get run everytime a user logs in? [21:06] r3dLunchb0x: yes [21:08] SpamapS: coolio, thanks appreciate it [21:16] SpamapS: do you want me to merge the two SRU-worthy items into the same debdiff, and upload for both bugs (LP Bug 1014044 and LP Bug 1006738)? [21:16] Launchpad bug 1014044 in php5 "PHP5-FPM not reporting errors to web server (nginx)" [Medium,In progress] https://launchpad.net/bugs/1014044 [21:16] Launchpad bug 1006738 in php5 "php5-fpm segfaults with error 4 in libc-2.15.so" [High,In progress] https://launchpad.net/bugs/1006738 [21:16] SpamapS: also, can you confirm if the segfault bug was fixed upstream? [21:17] so I installed with a 64 bit iso, but im running a x686 install when I check uname. Any ideas how that could have happened? [21:17] and if so, has it been put in Debian, and then synced to Ubuntu Quantal? [21:18] TheLordOfTime: the segfault bug is in 5.4.5.. [21:18] TheLordOfTime: which is in the Debian PHP team's git repo [21:19] TheLordOfTime: I'll make sure to merge it as soon as it comes out in quantal [21:19] TheLordOfTime: re the debdiff, yes, a single debdiff that we can just review and upload to precise-proposed would be ideal. [21:19] SpamapS: a debdiff that includes both fixes in Precise? [21:19] TheLordOfTime: Oh and to be clear, the segfault is *fixed* in 5.4.5 [21:19] because the patches apply cleanly-ish [21:19] (the cleanly-ish refers to the NEWS portion of the error reporting fix patch not applying) [21:19] TheLordOfTime: Yes. Use quilt refresh on them before creating the debdiff. [21:20] TheLordOfTime: right, the NEWS bit can likely be hand merged or even dropped. [21:22] mhm, when doing quilt import /path/to/patch for that upstream patch, and then doing quilt push -a, i had to do quilt push -a -f to force the one with the NEWS issue, and then refreshed that patch to remove the NEWS part [21:22] i'm doing the debuild -S now for build testing in a PPA [21:29] TheLordOfTime: sounds good [21:31] bleh, i do not like translations. there's a package or two i actually need to create a translation template for, and i'm so new to that, i have no idea how to proceed... [21:33] ... whoops [21:33] wrong GPG key [22:05] SpamapS: geez, PHP takes ages to build. [22:06] TheLordOfTime: yes, the tests are LONG [22:07] well, at least it started building about when i uploaded to my php-sru-tests ppa [22:07] well.... [22:07] after i fixed the PGP key thing [22:09] SpamapS: i will be ***GLAD*** once this reporting bug is fixed [22:10] * TheLordOfTime has been inundated with emails about this being a bug [22:10] then i can say "STOP EMAILING ME, A FIX IS IN -PROPOSED" === bitmonk_ is now known as bitmonk [22:16] SpamapS: i take it the time for tests is > 30 minutes? [22:20] TheLordOfTime: on a good fast machine yeah [22:21] 45m on mine here I think [22:21] does that include the PPA builders? [22:21] PPA builders will do it in 1 hour usually [22:21] been a while since I waited on them tho [22:21] heh [22:21] * SpamapS just uses sbuild.. nearly the same thing [22:21] well pbuilder and pbuilder-dist are broken, and sbuild doesnt run right here [22:28] SpamapS: for the segfault bug in PHP, is that SRU worthy in Precise if it doesnt yet exist in Quantal? [22:28] granted i'm still working on the build testing first, but... :P [22:30] TheLordOfTime: yes its ok, you just have to explain that it is upstream already [22:30] ah okay. i didnt set up the SRU template on the segfault bug yet. and i took the extra time to DEP3 tag the patches [22:30] TheLordOfTime: the point of making sure it is fix released in dev is just to make sure people focus efforts on the dev release first and don't just scratch their stable release itches ;) [22:30] indeed. [22:30] TheLordOfTime: working fine here (pbuilder/-dist) [22:31] ScottK: its been narrowed down to this system itselkf [22:31] apparently there's some weird bug where pbuilder doesnt get the right files or something [22:31] and its specific to this system [22:31] Did you try recreating the pbuilder chroot from scratch? [22:31] *shrugs* [22:31] mhm [22:32] the chroot doesn't build [22:32] it errors out [22:32] its not worth debugging now though [22:32] * TheLordOfTime was going to look into getting and configuring sbuild in future [22:32] then work got in the way :P === dendrobates is now known as dendro-afk === cpg|away is now known as cpg === Lcawte is now known as Lcawte|Away [23:14] SpamapS: the amd64 builders are faster than the i386 o.O [23:17] does virt-install run under a different user? [23:17] I'm running it as root but I'm getting permission errors [23:18] says that it can't access the .img file, which is under /root. and obviously root itself has permissions to that. [23:18] for the file I specified as disk --path= [23:18] I added root to the kvm and libvirtd groups for good measure [23:25] SpamapS: well, amd64 works, so... [23:25] time to debdiff! [23:25] TheLordOfTime: woot [23:25] (i386 will probably build as well, given prior track history with SRU and backport and security tests) [23:26] SpamapS: i won't be shot for uploading the debdiff to two separate bugs will i? [23:28] oh, uncommenting the user/group to run libvirtd as in /etc/libvirtd/qemu.conf and restarting libvirtd-bin actually fixed it. [23:29] SpamapS: lets hope that after dinner, the thing will be finished building on i386. [23:29] once that happens i'll push the debdiff up to LP [23:30] SpamapS: you may as well just ping me when php5 bugs get triaged or fixed upstream, i'm very likely to be able to just do the debdiffs xD === n0ts_off is now known as n0ts [23:35] does virsh -c qemu:///system suspend survive a restart? [23:35] or do I have to save/restore for that? [23:35] and is there a way to tell the hypervisor to suspend or save state of open VMs when the system is about to restart and then restore them once it is back up? [23:36] like to script that in [23:41] what are the correct last 4 digits of md5sum ubuntu-12.04-server-amd64.iso ? [23:41] TheLordOfTime: I suggest only putting the debdiff on one of the bugs, and referencing it from the other. === n0ts is now known as n0ts_off [23:52] i think i found a problem [23:53] http://bpaste.net/show/o87CdsGW75uVBXCmTRQg/ [23:53] I go to the 64 bit url, but it gives me the 386 version. [23:54] who's to blame? wget? [23:55] osmosis: the & must be escaped [23:55] osmosis: notice [2]+ Done bits=64 [23:55] osmosis: you started a background process that ran 'bits=64' [23:55] osmosis: try wrapping the url in ' [23:56] ok..now Ill go reinstall my servers too