[01:16] I guess that's what you get for idling... ;) [01:16] @comment 43860 that was a test guys! [01:16] The operation succeeded. [01:25] that was uncalled for and mean. I am reporting you for op abuse [01:26] :P [01:26] awwww [02:00] this is the most unhelpful channel [02:00] but FYI, I fixed my problem [02:01] it ended up being a VM + Linux Kernel issue [02:01] http://ubuntuguide.net/virtualbox-4-on-ext4-error-host-io-cache-will-now-be-enabled [02:09] lol @ opabuse :P [04:05] Hey guys, I need some help with software raid [04:05] Is this the right place to ask? [04:07] I created an md1 (radi1) with sdb1 and the 2nd drive missing. I then ran mkfs on md1, mounted it, copied some files, and then umounted, rebooted. Now i cant mount md1. Says there is no filesystem [04:07] be right back [04:08] Right, back. [04:18] Is there a better channel for mdadm questions? [04:56] hi [04:57] could anyone kindly help me download this file please:http://www.filesonic.com/file/2311849181/0470540834.pdf === jj-afk is now known as jjohansen [05:24] Is there a way to get I/O stats for tmpfses? === Ursinha is now known as Ursinha-afk [06:54] hey does anyone here test security? === smb` is now known as smb === Ursinha-afk is now known as Ursinha [08:00] morning all [08:05] jamespage: morning! [08:07] morning [08:09] Has anyone encounterd nfs issues where it times out most of the time but does eventualy connect? ubuntu 10.10, nfs v3 on both ends [08:10] daxroc: at boot? [08:10] twb: no, manual mounts. autofs fails too with timeouts [08:11] haven't seen that, no [08:11] I know the boot process as at lucid has cyclic dependencies (i.e. never works) for some NFS use cases [08:12] It seems to work fine on fresh installs, but dist upgrades seem to have problems [08:12] Hm, the main difference I know of is that hardy had server-side groups off, and lucid had them on [08:13] Which only matters if your NFS server can't resolve uids/gids on its own exports [08:14] daxroc: I would be doing some tcpdumping and rpcinfo -p'ing and such, see if you can spot a bork-bork [08:14] lynxman, morning [08:38] twp only thing that stands out in the tcpdump is .rpassword incorrect for the two hosts [08:43] to install the man pages, is it sudo apt-get install man or manpages or both, as i see both exist? [08:43] AdvoWork: the man pages are installed by default [08:44] AdvoWork: you shouldn't need to install them [08:45] ikonia, hmm, ive just used debian and xen to install ubuntu 10.04 server,and its sort of installed a bare bones system, ive had to install like vim, ssh etc, and it doesnt appear to have man either [08:46] AdvoWork: what install media did you use [08:51] ikonia, on the debian machine, i specified the link for lucid # mirror_lucid = http://archive.ubuntu.com/ubuntu [08:51] and installed via xen that way [08:51] it seems to have installed, but as i say, barebones [08:51] AdvoWork: no, I asked what ubuntu install media you used [08:53] well, i specified that path, xen downloaded and installed it, is that what you mean? [08:53] so you didn't install from an iso for example ? [08:55] no [08:55] first time ive ever made an ubuntu VM, and it seemed to go ok, but im wondering if it didnt now [08:56] I'd suggest using one of the iso images so that it goes through the proper install process, gets the right packages, configures them etc etc [09:15] ikonia, not that easy with xen [09:16] AdvoWork: really, I've never had an issue with it [09:16] how would I specify it then? instead of a mirror, in the /etc/xen-tools/xen-tools.conf? this is only what ive been told mind you [09:17] let me see if I can grab a read through the docs, I've not used xen for ages since kvm came usable [09:18] --mirror=/file/system/location/file.iso [09:23] hi all [09:25] ikonia, ok cool, and one further thing, would it be this file? http://releases.ubuntu.com/lucid/ubuntu-10.04.3-server-amd64.iso [09:27] AdvoWork: that looks like the current 10.4 64bit iso [09:27] cool [09:27] hallyn_: yesterdays libvirt update looks to have fixed the spawning multiple machines with lxc error in nvoa [09:28] I've hit something racey on termination though - looking now [09:32] ikonia, and is that also still using debootstrap? [09:34] AdvoWork: I don't believe so [09:35] i can't seem to find an example or anything for using an iso instead, i know i can specify the iso file as you say, but it still asks in my xen-tools.conf dist = whatever which refers to a mirror, so i'd comment out the mirrors, but dont know about that bit, nor the debootstrap bit [09:35] AdvoWork: dies = lucid [09:35] not a mirror [09:40] dies? sorry i dont get you [09:41] AdvoWork: sorry, typo [09:41] distro=lucid [09:42] ahh, so really i'd comment out the mirrors, specify the new mirror for the iso, leave dist=lucid as it is? [09:43] thats how I see it [09:44] ok cool, and would i also comment out the debootstrap bit or not? [09:49] AdvoWork: have a quick read onthe docs, that's all I'm doing [09:49] I've not done this for years so just referencing the docs [09:50] if ive firstly done xen-create-image --hostname=whatever and then xm create /etc/xen/cfg/whatever.cfg and now i want to change it, ive done xm destroy whatever, how do i undo the first bits? ie how do i properly delete the whole vm? i keeep reading steps that only seem to do the same as destory [09:54] ah that does work [09:54] just missleading [09:57] RoAkSoAx: hey [10:00] koolhead17: pretty sure he's sleeping :) [10:04] lynxman: hey. :) [10:35] jamespage: yay (on the first) [10:35] jamespage: by chance, are you getting msgs in syslog from ext4 on shutdown? [10:35] hallyn_, I don't think so [10:36] bug 868206 is the termination issue [10:36] * hallyn_ goes to look, and wonders where the bot went [10:37] jamespage: don't lie to me :) [10:37] ah - you mean on instance shutdown [10:37] they yes I do [10:38] it looked like the devices where getting unmounted or something before the lxc instances where actually terminated [10:39] jamespage: can you try the trivial patch from bug 861656 ? [10:39] yes - just re-imaging now [10:39] though really you're right, and we might ought to do it differently [10:40] i dont' know if we add a wait on the container init, if we cna prevent the umount from failing in the first place [10:50] ikonia, ok, i downloaded the iso, set the mirror to the iso, commented out the other mirrors, created the vm, but still, man command is not found.. [10:50] AdvoWork: did it go through an install process ? [10:50] AdvoWork: eg: ask you to create a username and password, partition the disk etc etc [10:50] ncurses based install routine [10:52] ikonia, no, it basically only asked me for the root password, so nothing else :S [10:53] so i dont know what im doing wrong? [10:53] then it's not using an ubuntu server iso [10:53] it's not even installing ubuntu [10:53] ubuntu doesn't have a root password set [10:56] no i mean, its asking me to set the root password [10:56] and to confirm it [10:56] thats the only thing it asked [11:02] just ssh'd into the machine, and cat /etc/issue shows Ubuntu 10.04 LTS \n \l so its installed ubuntu, just not the right one? [11:02] AdvoWork: I don't know where it's getting the install from [11:03] nor me :/ [11:03] whats wrong with using this base version, is it just not worth it? [11:03] well the fact that you don't know anything about it is what's wrong with it [11:08] hallyn_, no IO errors - but 4/6 did not complete termination according to nova [11:08] although no lxc instances where found on the hypervisor [11:08] running the terminate again did clear them [11:08] so I would say its better [11:09] hallyn_, then I looked at the console [11:09] lots of ext4 errors still :-( [11:09] no opps tho [11:11] ikonia, well i know its a base/bare system of ubuntu server, it appears to be ok from what ive seen so far, but i dont know whats not in it that should be [11:11] AdvoWork: well your missing packages, you've not got the sudo security model setup, so that's 2 core things which I wouldn't trust [11:13] how do you know ive not got the sudo security model setup? only wondering [11:13] can i not just do an update/upgrade? [11:14] AdvoWork: because you set a root password [11:17] dont you normally when you do a new server installation? its been a while since i did one [11:18] no [11:18] Ubuntu doesn't have root password [11:18] no matter what kind of installation you are doing [11:18] there should be no root password [11:21] hmm,in which case, im stuck then [11:26] jamespage: eh that's just how ext4 rolls :) [11:26] jamespage: can you pb some examples? [11:26] yeah - sure [11:27] ikonia, any further suggestions then? [11:29] ikonia, also just done uname -a which shows: Linux test-system 2.6.32-34-server #77-Ubuntu SMP Tue Sep 13 20:54:38 UTC 2011 x86_64 GNU/Linux does that help at all? [11:59] jamespage: still not seeing any here though fwiw [12:01] hallyn_, http://paste.ubuntu.com/702714/ [12:02] sorry - got distracted trying out juju local lxc provider [12:05] jamespage: how many were you shutting down? [12:05] 6 [12:05] ok [12:51] * HarryPanda is looking for a simple package/upgrade manager, like Spacewalk, but for ubuntu [12:52] * koolhead17 points HarryPanda to landscape [12:52] I'm trying the demo at the moment, quite slick - but I need *hands on* [12:52] *Landscape [12:53] HarryPanda: expalin handson [12:53] *hands on" [12:53] source code, that I can tinker with [12:54] HarryPanda: is Spacewalk free product? [12:54] yes [12:54] or a Red Hat enterprise product? [12:54] it was a redhat enterprise product - their 'Satellite' network, and was only open-sourced reasonably recently [12:56] * HarryPanda will look at 'landscape-common' package and see if anything interesting can be used [13:38] hallyn_, any idea why plymount --ping might hang? having trouble with juju lxc containers not starting up properly [14:06] jamespage: no, i don't even know what that is. [14:06] * hallyn_ checks the source [14:09] jamespage: I assume the PLY_BOOT_PROTOCOL_SOCKET_PATH socket is not created. Wonder waht creates that [14:10] (abstrace unix socket \0/ply-boot-protocol) [14:52] does ubuntu run on i386 machines? [14:52] jdstrand: quick question about the bash auto completions for virsh - is it better to have libvirt-bin drop those in, or to add it to the bash-completions package? [14:56] aliverius: Yep [14:56] ty [14:57] aliverius: We have a couple old 486DX machines here using 6.06 [14:58] hallyn_: while the bash-completion package does have a lot, in this case it feels more right to have whatever package ships the virsh command (libvirt-bin) ship it, especially if you expect it to change and are going to be maintaining it [14:58] hallyn_: so, for Ubuntu, ship in libvirt-bin. when pushing that to Debian, the maintainer may decide bash-completion is better [14:59] hallyn_: it would also make the most sense if upstream libvirt picked it up to ship it in libvirt-bin [14:59] hallyn_: (that way we can always get their updates to it easily) [15:00] jdstrand: yup, i'll definately send it to the list. THough I want to fix up the container name completion first [15:00] thanks [15:00] cool. thanks for picking that up :) [15:04] (right now i'm trying to build netcf :) [15:15] jamespage, Daviey or zul: around? I'm quite concerned about bug 868349. I can probably fix it quick but need help reproducing it - it doesn't in my dev environment running out of the source tree, and I don't know how to set up a working environment straight from the packages. [15:15] rbasak, I am [15:15] and where is the bot? [15:16] also, I don't know what to do with the bug metadata to indicate that I think it needs looking at immediately [15:16] (apart from Importance) [15:17] rbasak, working environment straight from packages - https://wiki.ubuntu.com/ServerTeam/Oneiric/OpenStackTestPlan [15:17] its all in one [15:17] ah yes, of course [15:18] rbasak, lemme tear down my test rig and I'll have a go as well [15:19] rbasak, I've added the server-o-rs tag to the bug report - that will get it on Ursinha and Davieys reports and visible [15:20] yes sir [15:25] hi guys! which could be the reason for overwriting my /etc/apache2/ports.conf after every system restart (I'm using ubuntu 10.04)? [15:25] with the default ports.conf [15:28] jamespage: OK I've reproduced it [15:29] rbasak: sounds like something we really need to fix then [15:30] Can you marked it as confirmed and 'High' for importance [15:32] f18: Do you have a separate partition/mount for /etc ? It may be reading whats underneath before mounting happens [15:34] genii-around: no I don't, but the strange thing is that as far as I saw only to files are affected: the ports.conf and the sites-available/default [15:34] *two [15:35] jamespage: OK I've got a one line fix. The problem was a double close. What I don't understand is why I can't reproduce it in a test, but the fix is trivial. I'll keep trying the test for a bit. [15:56] Hi! [15:56] Has anyone here figured out a way to allow people to use vnc towards kvm-vm's via some kind of a proxy? [15:57] Via a webserver? [15:57] With password-authentication? [16:19] jamespage: I've attached a patch to the bug: https://bugs.launchpad.net/ubuntu/+source/nova/+bug/868349/comments/4 - are you OK to update and upload from there, or would you like me to update the bzr branch and do a merge request? Only I know that time is critical and I don't want to make a mistake there. [16:24] rbasak: I'd prefer todo this through a merge request; I'm happy to upload but I would like smoser or zul to review [16:24] jamespage: OK, I'll prepare one now [16:25] rbasak, are you happy with what you need todo? i.e. branches etc.. [16:25] btw - I just reproduced so pretty easy to test the fix [16:26] jamespage: yes - i'm branching lp:~ubuntu-server-dev/nova/diablo, right? [16:26] yep - thats the one [16:29] jamespage: shall I update backport-libvirt-console-pipe.patch or add a new patch? [16:30] rbasak, I would update that patch; patch upon patch is just going to get complicated [16:35] jamespage: trouble is that I now have a diff of a diff to look at which is quite confusing! quilt refresh has walked all over the patch as well. Is this OK - is this the normal way of doing it? [16:36] hmm - not sure quilt should have changed to much in the patch [16:36] I think zul already refreshed then against the diable codebase [16:36] diablo [16:36] can you push the branch and let me have a look? [16:37] I did: quilt push backport-libvirt-console-pipe.patch; quilt fold < /tmp/868349.patch; quilt refresh; quilt pop -a [16:37] OK will do === gustav- is now known as beerbro [16:41] jamespage: what should I be pushing to? I tried ~racb/nova/diablo/868349 but got permission denied [16:41] rbasak, I think thats OK [16:42] (the use of quilt) [16:42] try ~racb/nova/868349 [16:42] you can specify which branch to target when you generate the MP [16:43] at which point use ~ubuntu-server-dev/nova/diablo [16:44] Yep found it. https://code.launchpad.net/~racb/nova/868349/+merge/78284 - I hope that's right! [16:47] zul, smoser: any chance one of you can find the time to review ^^ [16:48] or even Daviey :-) [16:56] rbasak, to complete the round trip the review.openstack.org change should be updated as well [16:56] jamespage: will do [16:56] that way the patch headers match up nicely - will keep reviewers more happy [17:08] rbasak, I'm eod - hopefully Daviey, zul or smoser will pickup and review [17:08] and we can upload am tomorrow [17:08] jamespage: OK, see you tomorrow [17:09] ttfn [17:13] Any idea on how I can figure out why Samba won't let me save a file as a guest (even though the guest account has R/W access to the shares). I can write a file as the gues account with no issues via ssh [17:28] I can write a file via notepad, but not illustrator or photoshop oddly enough. This is driving me mad hehe [18:06] Hi everyone, I was looking for some help with powernap on ubuntu server. [18:07] I have Apach2 running on the machine but its connecting via TCP6 which I am guessing the TCP monitor isnt catching [18:08] I thought that I would have it run an IOMonitor for apache but that isnt working either. [18:37] !google | test [18:37] test: While Google is useful for helpers, many newer users don't have the google-fu yet. Please don't tell people to "google it" when they ask a question. [18:37] Results for | test on Google: [18:37] Results for | test on Google: [18:37] -- [18:37] -- [18:37] Results for | test on Google: [18:37] -- [18:37] g0t_ g0t__ and g0t___ could you please disable that script in all Ubuntu channels. [18:43] Tm_T: g0t____ has rejoined [18:45] every keyboard action has a delay and possibility to be lost, nice [19:09] close [19:13] close [19:13] close [19:15] could i create a raid 6 array, then mirror that array to another HD in raid 1 and just swap out the single HD on the other side of the raid 1 as my take home backup? [19:16] i know it's not going to let me restore files or anything like that, but as a quick way to coordinate an off site backup? [19:16] dkn: I'd suggest having the two meta devices, then using a technology such as lvm to keep them in sync, assuming the sizes of the arrays are compatible [19:16] hmmm [19:16] what do you mean by two meta devices? [19:17] dkn: 1 raid6 meta device and 1 raid 1 meta device [19:18] (or the meta device and single hard disk) [19:19] ya, i didn't know lvm let you manage sync, i just though it was a layer you could move things around on [19:19] you can "mirror" with lvm, so if you use it that way, it's a good otion [19:20] would the lvm recognize the performance of the raid 6 over a single disk and prefer it? [19:20] dkn: your mirror the data within lvm it will use both [19:21] but, for example, if you raid1 a SSD, and a HD together, it's going to prefer the SSD for reads, and the HD will be the bottleneck for writes [19:21] will it do the same with lvm mirror? [19:21] it won't do that at all [19:21] if you raid 1 a device it will access the meta device not the individual hard disk [19:21] raid or lvm?' [19:22] either [19:22] really? [19:22] it will request info from the meta device, and the OS will offer the data from what it best can at that moment in time [19:23] but i thought the meta device requested the info, and whatever disk provided it first it would ship it off, so the SSD would supply the data faster and that data would get sent as soon as it's available [19:24] you may get it for a read, but it still has to confirm the second disk is in a valid state [19:24] mmm true... [19:24] eg: it needs to know both are in sync before offering the data or how does it know which one is valid [19:24] bah... [19:25] stupid bottlenecks [19:25] it makes sense when you say it that way though [19:25] I don't believe it actually checks the data on both, mearly checks the device is marked as sync [19:25] I'm not %100 been a while since I looked really under the hood at it [19:26] ya.. [19:26] but that's still a read [19:27] does the lvm mirror support hot swapping? [19:28] depends on the hardware [19:28] ok, so i'm good them [19:30] it's to bad mirror or raid 1 doesn't really come with any performance advantage... [19:30] why would it ? you're writing to 2 disks and constantly keeping them in sync [19:31] cause i can either get a purchase order approved for the redundancy level we need, or the capacity we need, but not both.. [19:32] thanks for the help === IdleOne is now known as PreciseOne [19:56] "Precise Pangolin" === PreciseOne is now known as pangolin === pangolin is now known as PreciseOne === PreciseOne is now known as pangolin [20:22] Anyone know a way to skip dpkg-reconfigure when installing a package that requires configuration; or some way to specify configuration values from a script? [20:28] pardon me [20:28] how do I specify what mirror does the first stage of pxe/netinstall gets used? [20:29] right now it always uses some archive.ubuntu.com mirror [20:32] * patdk-wk finds it easier to just tell it to use a web proxy server [20:32] hmm [20:32] not a bad thought [20:32] as I'm running apt-cache-ng locally, it helps out :) [20:32] plus it autoconfigures apt to use that, on the installed system [20:33] it seems liek archive.ubuntu.com is hardcoded in a .c file [20:33] na, probably in the seed file [20:33] nope [20:33] set everything in preseed we could [20:33] odd [20:33] reading C code now [20:39] patdk-wk: this is stage 1 ... where it gets the intaller gets retrieved [20:45] patdk-wk: does your proxy server intercept traffic to archiv.ubuntu.com [20:47] RoAkSoAx: ping [20:47] kirkland: pong [20:48] kirkland: howdy man [20:48] RoAkSoAx: hey man [20:48] RoAkSoAx: i'm curious about ubuntu-natty-x86_64.seed in the orchestra source [20:49] RoAkSoAx: seems kinda outta place [20:49] kirkland: yeah I just saw it today too [20:49] kirkland: I'll get rid of it [20:49] kirkland: it's a leftover from preivous orhcestra [20:49] RoAkSoAx: okay, hold on [20:49] RoAkSoAx: i'll kill it [20:49] kirkland: ok [20:49] RoAkSoAx: i'm fixing a couple of bugs i hit with SpamapS last night [20:50] kirkland: the reconfigure one and the partition one? [20:50] RoAkSoAx: yup [20:50] RobinBAwesome, no, no need to [20:50] RoAkSoAx: reconfigure is committed [20:50] kirkland: ok cool [20:50] when you start the installer, it asks for a proxy server to use [20:51] RoAkSoAx: where's the seed that orchestra uses for non-juju installs? [20:51] kirkland: that's supposed to be the one which I was planning to rename and cleanup [20:52] RoAkSoAx: the one that was named: provisioning-server/var/lib/orchestra/kickstarts/ubuntu-natty-x86_64.seed ? [20:52] patdk-wk: do you know if the proxy server be set in the kernel params? [20:53] kirkland: yeah, that's the only seed file that was there before I added the juju one [20:53] RoAkSoAx: okay [20:53] kirkland: but technically, the juju one will work [20:53] RoAkSoAx: there used to be one called ubuntu-orchestra-client.preseed [20:54] kirkland: uhmmm i might have got rid oof it by mistake [20:54] kirkland: but the juju one, the only difference is the cloud-init stuff for juju, which if it is non-existant, shouldn't affect [20:55] kirkland: http://bazaar.launchpad.net/~orchestra/orchestra/trunk/revision/239 [20:56] RoAkSoAx: okay, cool [20:56] kirkland: yeah the ubuntu-orchestra-client.seed was an old preseed that should probably be dropped and just copy the juju one and remove the juju sepcific stuff [20:57] kirkland: my parking is expiring in 10 mins (im at my former school). I'll take care of it [20:57] as soon as I get home [20:57] RoAkSoAx: agreed [20:57] RoAkSoAx: i just called it "orchestra.preseed" === kentb is now known as kentb-out [23:13] anyone known a descent web management tool voor samba file server? [23:35] Hi [23:36] is there a glassfish version 3 package available for 8.04 ? [23:43] kaushal: is 8.04 anymore supported? [23:43] yes it is [23:43] I mean server [23:45] till 2013.04 :) [23:46] or is it 2013.10? [23:48] Patrickdk: yes [23:48] Any clue to my answer ? [23:48] i mean to my question [23:49] you don't know how to use apt? [23:50] http://pastebin.ubuntu.com/703055/ [23:50] i dont see version 3 available [23:50] but as it's in universe, it's not supported [23:52] Patrickdk: are there deb packages available for 8.04 ? [23:52] I mean pertaining to version 3 [23:52] how should I know? [23:53] there are no official, or any in the unoffical ubuntu builds [23:53] you would have to find someone that did it themselfs [23:53] I would suggest google [23:53] Patrickdk: i did it already