[00:51] jjohansen: ok, giving it a try now [01:02] jjohansen: still no luck [01:27] stgraber: what was the error you got from apparmor that kernel has a tweak to say where it failed the mount [01:27] sorry, /me is in and out today so haven't really gotten a chance to work with it yet [01:30] [ 227.801504] type=1400 audit(1331427700.899:53): apparmor="DENIED" operation="mount" info="failed flags match" error=-13 parent=3224 profile="/usr/bin/lxc-start" name="/usr/lib/lxc/root/" pid=3245 comm="lxc-start" srcname="/home/stgraber/data/vm/lxc/containers/qatracker01/rootfs/" flags="rw, rbind" [01:31] the profile is still with only "mount", so no restriction on source, location or flags [01:34] stgraber: okay it will be a bit before I can get to it, but I will have another kernel up later tonight. [01:34] jjohansen: cool, thanks. I'm trying setting the flags now in the profile, see if I can get it to work with that [01:37] jjohansen: even with "mount options=(rw,rbind) /** -> /usr/lib/lxc/root/," it doesn't seem to match... === sixstringsg|away is now known as sixstringsg [02:38] Hello, I really need some help. I got Ubuntu Server to install just fine. But I'm trying to get a GUI installed but its not working [02:38] I would have gone with just Ubuntu Desktop but I can't get that to install. While the server edition installed just fine === Darkwing is now known as DWonderly [02:55] how does keeping systems updated fits into ubuntu orchestra? as I understand we can deploy new software using juju, but can we choose which updates to install, see which updates are available etc. without paying for Landscape? [03:04] running a virtual kvm of ubuntu 10.04 server…. when running rsyslog, I am getting "Could no open output file '/dev/xconsole' [try http://www.rsyslog.com/e/2039 ]" in the logs... [03:05] apparently the /dev/xconsole device doesnt exist… can I point this to another location? IE, a file in /var/logs maybe and avoid the above error? === sixstringsg_ is now known as sixstringsg === sixstringsg is now known as sixstringsg|away === Savvis_ is now known as Savvis === koolhead17|away is now known as koolhead17 [07:46] https://launchpad.net/~brianmercer/+archive/php << anyone if there is an official lucid package of php5-fpm? === alaing is now known as funkymonk [10:31] anyone have experience setting up SIP VOIP on ubuntu and android? === milkshake_ is now known as Guest28996 [12:16] hi, where's the server guide for 12.04 at now? [12:17] i want to read it in its current form, whatever that is [12:17] mhh... ldap is more difficult than it seems :) [12:17] MatBoy: i'm working on ldap right now [12:17] MatBoy: what are you trying to do? [12:18] jacobw: ah nice... I'm totally new to it... I'm a decent admin but always tried to avoid ldap :) [12:19] jacobw: just setting up... I have installed ldap and phpldapadmin [12:19] MatBoy: ok, does olcRootDN and cn=config mean anything to you right now? [12:20] jacobw: little bit... I have setup my own part alreayd that I need for my doamin [12:20] domain [12:20] cool :) [12:21] be aware that all configuration in openldap > 2.4 is done by modifying the cn=config DIT which contains the directives previously in /etc/ldap/slapd.conf [12:22] has anyone tried http://salt.readthedocs.org/ [12:22] i've seen it mentioned before [12:22] jacobw: thanks... I already thought something like it... but I need users in it now :) [12:23] it seems to perform the same function as puppet, which makes me want to know why i should consider it over puppet [12:26] MatBoy: that's quite a simple use case, remember than in rfc2307bis which the schema that ships with openldap that posixAccounts and posixGroups are both auxillary objectclasses, you need a structural objectclass, like inetorgperson, to define each user [12:27] MatBoy: i've been struggling with openldap oddities for a while now :p [12:38] New bug: #952199 in apache2 (main) "oneiric LAMP server: Apache consumes 100% of cpu in QEMU" [Undecided,New] https://launchpad.net/bugs/952199 [12:40] ubuntu server, i have a " 06:00.0 RAID bus controller [0104]: Marvell Technology Group Ltd. 88SE6145 SATA II PCI-E controller [11ab:6145] (rev a1) ", and ubuntu doesn't see the 4 2tb drives i have connected as proved by "sudo fdisk -l " (the four drives are seen by the controller on POST bootup) [12:40] version 12.04 [12:50] jacobw: yeah me too... normally I use MySQL for a lot of stuff [12:59] New bug: #950589 in mysql-5.1 (universe) "unmet dependencies on installing mysql-server-5.1" [Undecided,New] https://launchpad.net/bugs/950589 [14:17] http://shrubbery.mynetgear.net/c/display/W/Marvell+SATA+RAID+Controller+-+Linux+Not+Happy [14:23] please help me kill this pata_marvel driver [14:25] http://dpaste.com/714958/ [14:26] how can I find out a list of all the automated processes running on the server? [14:26] I'm getting a weird error that seems to crop up every now and again and suspect it may be due to some scheduled process … [14:27] but the crontabs are empty [14:27] imanc, even the root crontab is empty ? [14:28] imanc, whats the error ? [14:28] hmm - ther eis no home/root is the root crontab in etc or something? [14:28] exim is bailing with a memory alloc error for like a few seconds. It happened at 5am this morning when there's no server activity [14:28] 2012-03-11 05:42:52 daemon: accept process fork failed: Cannot allocate memory [14:29] on top of that, there's a weird python error - the site can't find a specific modue. It happesn only for about a minute every once in awhile, and seems to bring down the site. But again, there's no reason for it - and if it was occuring, no reason that it'd suddenly self rectify. [14:30] something freaky is going on [14:30] imanc, check out /etc/cron.d/php5 [14:31] do you have any entries there ? [14:32] actually … there's a few entries in cron.daily including an entry for sendmail, which I uninstalled [14:32] probably comment it out [14:32] when you apt-get remove , probably it did not 'clean' it [14:33] yeh [14:33] hmm [14:33] there's also apache2, apt.dpkg-dist [14:34] yes there's also the php5 script on cron.d but nothing on the server is running php, AFAIK [14:36] is there something in it [14:36] gotta check [14:38] yes, the php5 stuff is just clearing any expired sessions [14:38] can't see how it'd be affected django [14:38] or exim [14:39] apache2 script is just clearing the cache too [14:41] there seems to be sysklogd and logrotate, which both claim to rotate the logs [14:57] imanc: any disks near full? Also do you have swap configured? [14:57] hmm checking disc usage now [14:58] there's only one fiesystem .. /dev/vzfs and it's at 16% usage, so I assume that means no swapping has been configured [14:58] as there isn't a swap partition [14:59] mem usage is at 485 mb of a total 1265mb [15:00] The server hosts a website and app used by architects, and I'm assuming they upload huge files - so that may account for some of the issues. But they're definitely not doing that at 5am on a Sunday [15:14] imanc: I'd recommend hooking up detailed logging of system status. collectd is useful for that [15:15] imanc: why not at 5am on a sunday? Architects are weird. ;) [15:15] ha ha [15:18] awesome, I'll check that out SpamapS [15:18] imanc: also to check if there is swap, 'free -m' would be a good command.. (-m means it shows all numbers in Megabytes) [15:18] It just dawned on me tho, the server is in US time, and we're uk based. so I think 5am is more likely 9am here, and it's entirely feasible they were doing some work on the server. [15:19] SpamapS: yeh free -m shows 0 for cache [15:20] imanc: cache != swap [15:20] Mem: 1751 487 1263 0 0 0 [15:20] -/+ buffers/cache: 487 1263 [15:20] Swap: 0 0 0 [15:20] imanc: can you install the 'pastebinit' package and run 'free -m | pastebinit' ? That will give you a URL you can paste here [15:20] imanc: ahh while I typed you just pasted it. 3 lines is ok I guess. ;) [15:21] imanc: 0 cache is... something is wrong! [15:21] cache is not zero, but swap is .. hmm [15:22] wow, pastebininit is cool [15:22] imanc: shared, buffers, cache is all 0 there [15:22] maybe he never uses the disk :) [15:22] http://pastebin.com/055yHWqV [15:23] imanc: is this some kind of shared host? [15:24] nope - it's a dedicated or cloud server [15:24] running ubuntu [15:24] I've inherited the project and am seeing strangeness [15:25] I assume the guy who was managing it before me configured ubuntu [15:25] so its a VM? [15:26] imanc: I've never seen a box have 0 cache and buffers [15:26] oh, probably a hmm, what do they call that thing [15:26] SpamapS: what does it mean? :) [15:27] the pre-lxc thing [15:27] jacobw: I assume so. The guy who handed over the project is a graphic designer and not overly clued up on tech stuff. I just got "here's the ssh details, it's with media temple" === alaing_ is now known as alaing [15:27] hmm [15:27] where 'media temple' is? [15:28] suspect it's their 'dedicated virtual' package [15:28] but there is no control panel [15:29] Patrickdk: OpenVZ [15:29] my cat just can't get enough of my mouse on my screen, following it like it was a laser pointer [15:29] ya, openvz thing [15:29] so you think it's the virutalisation system that is reporting 0 for cache, buffers, etc.? [15:30] looks like it is openvz .. lots of reports in the googles about openvz and mediatemple [15:30] aha [15:30] so perhaps openvz does not report cache/buffer utilization to the containers [15:30] * jacobw thinks so [15:30] which kind of sucks.. but makes sense [15:31] imanc: so.. you're on a shared host [15:31] imanc: and likely overcommitted [15:31] interesting [15:31] imanc: so this is probably *not* your "server"'s fault [15:32] imanc: call MediaTemple. Explain. *COMPLAIN* [15:32] It would explain the weird, random problems [15:32] yes it would [15:32] including this error, which makes no sense: Try using django.db.backends.XXX, where XXX is one of: [15:32]    'dummy', 'mysql', 'oracle', 'postgresql', 'postgresql_psycopg2', 'sqlite3' [15:32] its my main reason for thinking the container idea is a fail for anything but testing [15:32] For that to happen, a python module would have to be removed [15:33] imanc: or memory allocation failures to cause your app to perform stupidly. ;) [15:33] I don't know how well python handles ENOMEM [15:33] SpamapS: is that the same for cloud based solutions? I've never had any issues with Linode, but anytime I've used VPSs in teh past, I've always had trouble. And typically the hosting companies support dept. deny that the server is overtaxed [15:34] I thought linode was all vps also [15:35] Nooo [15:35] Linode is VMs [15:35] plus, the definition of overtaxed, is not always the same [15:35] far less opportunity to overcommit [15:36] ahh linode are Xen VPS, apparently [15:36] imanc: If a hosting company treated me like that, I'd take my business elsewhere. [15:36] SpamapS: yes, well apparently they have been pretty crappy in general [15:36] heh, I hate people used these new amazon t1-micro thing [15:36] then asking why they don't work [15:37] but my client didn't see a clear reason to trasfer to linode. I think now we have one. [15:37] I went to a party they threw in Portland where they paid for about 200 peoples' bar tabs all night. They can afford to not f*** you over. [15:37] I see them contantly going with 0 cpu availablity for >min at a time [15:37] SpamapS: media temple? [15:37] whoa [15:37] Patrickdk: oh yeah, t1.micro is a total fail for anything but the tiniest load. [15:37] * SpamapS waves his hands in front of his own wordpress site running on t1.micro .. "NOTHING TO SEE HERE" [15:38] ya, nothing like 2min page load times :) [15:38] Patrickdk: yeah, you're at the mercy of all the m1.small's on the same hardware. ;) [15:38] Patrickdk: aggressive caching is where its at. :) [15:38] this was a pure static site :) [15:39] Patrickdk: if I haven't updated anything on the site, it pretty much just serves everything out of the ondisk cache. [15:39] Patrickdk: that has never happened to me.. but.. I think I get 100 page views per day. ;) === Lcawte|Away is now known as Lcawte [15:45] a [15:45] Thanks for your help guys! [16:26] New bug: #952340 in autofs5 (main) "autofs5 fails to install: post-installation script returned error exit status 1" [Undecided,Confirmed] https://launchpad.net/bugs/952340 === hazmat` is now known as hazmat [16:54] hey guys hi all.. I have an issue, I've lost a db schema of an ubuntu server [16:54] The only solution I think that i can perform is to try and restore the deleted data files [16:54] a db schema? [16:54] does anybody think that will help? [16:54] yes [16:54] what is that? [16:54] sorry a mysql database [16:54] then you have to restore from backup, yes... [16:55] (unless the database resided on a snapshot-capable filesystem and you took snapshots regularly) [16:55] RoyK: well thats the issue, I have a month old backup [16:55] RoyK: nope [16:55] what happened to the db? [16:55] disk crash? [16:55] I'm between a rock and a very hard place right now [16:56] RoyK: no someone messed up using the workbench [16:56] oh... [16:56] DELETE FROM blah; ? [16:57] no "DROP schema" [16:57] :S [16:58] heh - so "drop database" (schema is a synonym for database) [16:58] meaning, basically, you're in deep brownies [16:58] yeah [16:59] I have a month old backup [16:59] but thats just bad business [16:59] I'm afraid that's all you have, then [16:59] I'm trying to user photorec [16:59] to restore .frm files [16:59] unless you want to attempt to do low level recovery [16:59] and if so, unplug the server asap [16:59] yes thats what i'm thinking of [17:00] if you want to do low level recovery, you need a full copy of the disk as it is [17:00] but the problem is that the recovery tools in ubuntu are queit complicated for me [17:00] and I'm not that experienced [17:00] I somewhat doubt you'll get much out of it unless you're really good, though [17:00] yes i understand that, but it won't hurt to try [17:01] then poweroff the machine, boot on something like a live usb stick, connect a largeish disk and dd off the entire partition or disk with that filesystem [17:01] photorec tells me no hard disk [17:01] photorec sucks [17:01] it probably only knows fat32 [17:01] and you probably use ext3 or ext5 [17:01] and skalpel and magic rescue requires my to make header and footer in the extension files [17:01] erm, ext4 [17:01] hmm, ext5 [17:02] one sec, I'll tell you [17:02] recovery from ext[234] is generally very hard, MUCH harder than with fat32 [17:02] would you know of a tool that would help? [17:03] not really [17:03] even if there was an undelete tool, it wouldn't help much, since the database is removed from inside mysql [17:04] yes, but there's a place where the files reside.. thats where the data is kept in binary format [17:04] if i get to them, i can savor some importand data from my VM [17:04] than update the db with the difference [17:05] those last couple steps seems of a less challenge to me than recoviring lost files from ubuntu server [17:06] problem is, with fat32, only the index is removed when a file is deleted. with ext[234], the filesystem is cleaned for references for the file [17:06] meaning the file is GONE [17:06] ah,, )(**&&( [17:07] indeed [17:07] that's why it's crucial to keep a backup... [17:07] Code_Factory: what was in that database? [17:08] LOL. I'm on the mysql channel hearing the same words [17:08] :) [17:08] RoyK: its a database for a social website ('self -driven content') [17:09] so it loox very poor now when users are sent back to a month old system, some users will need to resign up === koolhead17|away is now known as koolhead17 [17:09] its really a very bad situation we have jere [17:09] here [17:13] daily backups :) [17:14] yep [17:14] unless you shut off the server pretty much immediately after it happened, there's almost zero chance of recovery [17:14] consider it lesson learned, backups are important [17:20] RoyK: thank you I gtg now and jump in my problems [17:22] one more question if possible: how to find out what filesystem i'm running? [17:23] lots of ways, mount, cat /etc/fstab, sudo fdisk -l [17:23] cat /proc/partitions [17:23] df -h [17:24] qman__: df and fdisk gave me info except the type of filesystem [17:25] qman__: actually none of them did [17:25] mount does for sure [17:25] df doesn't, that's my bad, fdisk only shows 'linux' which could be any linux type [17:25] qman__: yes thank you [17:26] qman__: I"m on ext3 do you know of any file recovery tools? [17:26] nope, like I said [17:26] unless you shut off the server almost immediately after it happened, it's gone [17:26] qman__: thank you === alaing_ is now known as alaing === alaing is now known as funkymonk [20:46] scuse me. what's the command (i need to read and study) to control services such as sshd and a new service i've just installed (named murmur) a link to a manual about the same would be appreciated. thanks in advance [20:47] Gallomimia: try with sudo service ssh status (or start or stop) [20:50] service. okay. i want to read manuals and pages first [20:50] thanks guntbert === Leseb_ is now known as Leseb [21:35] running a virtual kvm of ubuntu 10.04 server…. when running rsyslog, I am getting "Could no open output file '/dev/xconsole' [try http://www.rsyslog.com/e/2039 ]" in the logs... [21:35] apparently the /dev/xconsole device doesnt exist… can I point this to another location? IE, a file in /var/logs maybe and avoid the above error? [22:09] scubes13, you can configure rsyslog to not log to console, and log to files only === Lcawte is now known as Lcawte|Away [23:35] Good news everybody! After a few hours of testing, I've managed to add dbconfig-common support to glance (mysql ATM). [23:38] pabelanger: I *think* that is cool [23:39] pabelanger: tho dbconfig-common can sometimes be a double-edged sword.. since you *have* to have a running db server to complete installation [23:39] SpamapS, indeed! Automated database configuration [23:39] yar [23:40] reminds me that I need to write the dbconfig-common charm helper so juju charms can make use of it more easily [23:40] Will likely have to talk more about that, but for now the default database is sqlite (boo). Perhaps we can see about changing it to mysql or pgsql [23:40] SpamapS, Ya, I have some existing puppet manifests that will use dbconfig-common, the main reason for making the changes [23:43] pabelanger: I'd think sqlite being default is a good way to go. Casual users will appreciate it.. larger systems require testing/planning/etc and so people will have to pre-seed values anyway. [23:48] SpamapS, Ya, I'm just going to leave it as the default.