[00:00] What the flying.. [00:01] For some reason, when my server reboots.. the RAID5 array I've build as /dev/md0 appears as /dev/md127 and its always stuck in a "PENDING" state, then it moves into "RESYNC" [00:15] yeah, I've had that before [00:15] it was caused by a screwed up disk and some uuid confusion [00:21] Hmm.. [00:21] I've given them the disk by its ID though [00:21] Which.. shouldnt change [00:28] your array is probably fine, for me it was just a matter of correcting the problems and getting it to assemble again [00:28] I did it by removing the troublesome disk and manually assembling, then letting it fail over to another [00:28] then fixing the messed up disk === cpg is now known as cpg|away [02:46] New bug: #1095180 in etckeeper (main) "incorrect bash-completion" [Undecided,New] https://launchpad.net/bugs/1095180 [02:56] New bug: #1095181 in etckeeper (main) "terrible breakage on git rebase" [Undecided,New] https://launchpad.net/bugs/1095181 === cpg|away is now known as cpg === Graet is now known as Guest55006 === Guest55006 is now known as Graet === fhd_ is now known as fhd [06:11] hi people! [06:12] somebody knows about linux drivers installation? [06:12] helpme please [06:17] ~ask | neologico [06:17] !ask | neologico [06:17] neologico: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) See also !patience [06:18] ok understand [06:24] I can copy a compiled driver: wl.ko in one machine to another and install it?(wl.ko is the driver compiled for wireless card broadcom 43xx) [06:24] ? [06:27] neologico: for the same kernel version only [06:28] neologico: best to use dkms to trigger a a compile and install on the other machine a swell [06:35] lifeless:dkms in any application installable ? [06:36] neologico: its a standard package, if thats what you mena [06:36] *mean* [06:38] yes understand; even though my language is spanish:) [06:41] I can not even compile the driver after installing backtrack with updated repositories [06:44] well; thanks lifeless in another time i continue this, is too late here.Thanks!! [06:47] bye everybody happy 2013 [07:47] hiya all, planning on using U12.04LTS server on a few HP-NAS boxes. Out of curiosity, is there much trouble in getting it all to fit within 4GB? ( rsync, ssh, and samba are the only services being offered ) [08:55] inflex: will fit just fine [08:57] tx. Planning on using a Sandisk Cruze Fit USB 4GB as the OS/boot. [08:57] (Amazing how tiny those things are) === cpg is now known as cpg|away === Graet is now known as Graet_AFK_sick === yofel_ is now known as yofel [13:16] Hello there! How can a machine be resolving names if it's resolv.conf is blank? [13:17] I do have a dns server running, but does it affect my external name resolution? === binni is now known as Remion [14:22] LuizAngioletti: DNS can be passed in via DHCP iirc and Network config handles DNS from outisde of resolv.conf [15:03] ahs3: hey, new netcf release is out :) prevents segfault in use by threaded app iiuc. i assume now is still a bad time for debian upload/ [15:09] zul: good morning! could I pursuade you to sync netcf from debian-experimental into raring? [15:10] hallyn: i could be persuased [15:23] zul: looking at bug 591489. i don't get it. vnc doesn't fwd audio, so why would this bug matter? (hoping you understand :) [15:23] Launchpad bug 591489 in libvirt "No sound in virt-manager (QEMU_AUDIO_DRV set to none by libvirtd)" [Medium,Confirmed] https://launchpad.net/bugs/591489 [15:24] hallyn: hrm. good question; i'll check on the freeze, but yeah, i suspect it may be "hurry up and wait" a bit [15:26] hallyn: cant tell you but its in lucid have you tried something a bit newer? [15:27] hallyn: ill put the 1.0.1 up for you somewhere this week btw (with the lxc attach stuff backported) [15:38] zul: i havne't tried any, no :) [15:40] zul: IIUC, they're saying they want to start a VM with virt-manager, and have the qemu guest use the host's sound card even though they're using vnc [15:40] so if they were doing vnc remotely, the server would be playing sound [15:40] right it wouldnt [15:40] i see yeah that's what they want (comment #8) [15:41] I installed 12.04 server before the holidays.. everything was working fine, I have a static IP and downloaded software and updates.. today I am not able to get to any sites I tried apt-get update and it fails with "unknown host", same error if I try to ping anything. I have looked at /etc/resolvconf/resolv.conf.d/original and the correct DNS servers are in there.. any ideas where to look next? [15:42] zul: tempted to set it to wontfix rather than ignoring it... [15:42] but i guess i won't [15:42] i would [15:42] but i would also get them to test something newer [15:43] since they are like an lts behind [15:46] zul: the proposed fix reuires running the vm as the logged-in user (to connect to the pulseaudio session) so i doubt i'ts "fixed" [15:46] (where "fixed" == broken imo, but...) [15:47] so i don't really want to ask - seems like a waste of time [15:47] it does....besides its not a fix that we would SRU imho either === JanC_ is now known as JanC [16:34] hey guys on precise server I just ran sudo apt-get update and got this error any ideas? W: GPG error: http://gb.archive.ubuntu.com precise Release: The following signatures were invalid: BADSIG 40976EAF437D05B5 Ubuntu Archive Automatic Signing Key [16:36] davmor2, you need to import a public key. [16:36] wich key.. i m not sure [16:37] melmoth: it's the archive repo key and seems to be in place [16:38] http://en.newinstance.it/2009/06/22/the-following-signatures-were-invalid-badsig-40976eaf437d05b5-ubuntu-archive-automatic-signing-key/ [16:40] melmoth: many thanks [16:48] whats the 6euro a month vps provider's website again? [16:51] sw0rdfish: you asked yesterday, and I even did some googling for you [16:52] I notice that the default conf for rsyslog specifies $FileGroup adm, but then $PrivDropToGroup syslog. The latter appears to negate the former - I'm seeing new files created as syslog:syslog. [16:52] errr [16:52] sorry RoyK..... plz highlight me next time, and I found what you said, thanks man [16:53] damn man how did you that... that is a nice search technique ;) [16:53] sw0rdfish: I guess I to prefix with your nick... [16:53] * RoyK knows search engines :) [16:54] hehe [17:02] zul: not sure you care to follow my process on this :) but in case you do, i've pushed my changes to the debian qemu tree to github.com/hallyn/qemu. i'm going to now push that (based on 1.2 from end of october, i.e. what is in ppa:serge-hallyn/crossc) to raring, then i will merge in the new 1.3.0 qemu from debian's git tree (which hasn't been tested) [17:02] though i see it's in debian-experimental [17:02] hallyn: okie dokie [17:02] so, seems impossible that *something* would not go wrong with the 1.2.0 qemu push, [17:02] which will remove qemu-kvm and qemu-linaro source packages [17:03] oh wait. what am i thinking [17:03] clearly i won't have rights for this [17:03] hallyn: famous last words... [17:03] RoyK: i'm not happy about it :) but it's gotta be done... [17:04] hi guys. [17:04] so, I have Ubuntu Quantal authing to AD for logins and it seems to be working well. I created /home/DOMAIN for domain accounts to have their home directories created, but instead of creating home directories it's just complaining they don't have one and putting them in /. How do I get it to create their homedir on login if it doesn't exist? [17:04] zul: so actually i think i'll have to (a) ask someone (you?) to push the source from my ppa, and (b) ask for the new src pkg to be added to ubuntu-server group [17:04] hallyn: most likely :) [17:04] hallyn: or anyone else on the team [17:06] zul: you'd prefer someone else? [17:06] hallyn: nah but if you can spread the love that would be cool [17:09] zul: since i expect problems with this one, i'd (sorry :) rather have you do this, and ask roaksoax to do the netcf sync [17:11] hallyn: sure [17:11] zul: thanks! so after you push, i'll ask on #ubuntu-devel to have qemu added to server set [17:11] roaksoax: ping (when you're around) :) [17:13] hallyn: here [17:15] roaksoax: could you sync netcf from debian-experimental into raring? [17:15] hallyn: sure :) [17:17] roaksoax: thanks. [17:17] (honestly i need to package 0.2.3 too, but i prefer to go through debian first for all netcf now) [17:17] (and that potentially awaits freeze) [17:20] w00t! my american alienvps is doing ok for paltalk usage! [17:20] I'm soo happy [17:20] in that case I don't need a european one :D [17:23] hallyn: yeah I agree, but if we are close to freeze and it is not in debian yet, then I would suggest to package it [17:23] roaksoax: we're not close to freeze just yet right? [17:24] hallyn: nope [17:24] we still have time :) [17:24] do you remember off hadn when ff is? [17:25] hallyn: March 7th is FF [17:25] hallyn: https://wiki.ubuntu.com/RaringRingtail/ReleaseSchedule [17:26] thx just foudn it :) [17:26] ok so i'll put this down in my feb tickler file to package it for sure [17:28] hallyn: yeah, another thing you could do to help Debian, is package it yourself, ping the maintainer to try to get the new prepared in (specially of there are packaging changes) [17:28] roaksoax: no no, i'm the maintainer :) [17:28] roaksoax: i just want to wait until i'm sure wheezy is not in freeze [17:28] hallyn: yeah just noticed :) [17:29] roaksoax: thanks again. (going to step away now for a few mins) [17:29] hallyn: syncpackage: Request succeeded; you should get an e-mail once it is processed. [19:09] I'm trying to install quantal under xen(suse) . Last time I followed these directions for natty ( http://www.mmacleod.ca/blog/2011/05/ubuntu-natty-narwhal-and-xen/) and it worked fine. Now I cannot seem to get it to go with the latest version. After install, I get "boot loader didn't return any data" [20:12] adam_g: Hey, are you able to fix up the CA report to present for grizzly aswell? [20:14] Daviey: yeah, *should* be able to just add a new set of cronjobs to pull grizzly/raring instead of folsom/quantal. ill confirm in a bit and let you know what to add [20:19] "Running /etc/init.d/networking restart is deprecated because it may not enable again some interfaces" What should I do then? [20:20] restart networking? [20:22] RoyK, when I rand networking restart. it gave that warning [20:22] so if one way is deprecated, another should be used instead, right? [20:22] anyways, I am running into other issues :-\ [20:23] did you try that command I just gave you? [20:24] oh, that is a command XD [20:24] nope [20:24] doesn't appear to be an upstart job [20:24] http://fpaste.org/M2vw/ I added eth1:1 to my /etc/network/interfaces and then did /etc/init.d/networking restart [20:25] root@bcache:~# restart networking [20:25] networking start/running [20:26] that's on quantal, though [20:26] I am on precise === Graet_AFK_sick is now known as Graet [20:26] works for me too [20:26] dunno - I have no VM on precise, and wouldn't like to test on a box in roduction ;) [20:27] *production* [20:27] RoyK, this is odd [20:27] RoyK, http://fpaste.org/SGCd/ [20:39] ;quit [20:41] Daviey: yeah, the currentl script in lp:ubuntu-reports should work the same way. './gather-versions.py grizzly && ./ca-versions.py -r grizzly' should dump a ca_versions_grizzly.html [20:43] adam_g: Ah, i thought you might try and put it on the same output file.. but yeah, ok, that makes sense [20:43] Thanks [20:47] Daviey: oh, ya i could do that i suppose. [20:47] * hallyn looks for utlemming [20:48] oh well [20:55] adam_g: hmm, unless you think it is easy.. don't stress about it.. two pages is fine. === guampa_ is now known as guampa [21:12] so, I used to be able to install 32bit packages like so: apt-get install libx11-6:i386 in server 12.04, no longer, it seems for 12.10 [21:12] are i386 packages still avilable? [21:13] yes [21:14] is package:i386 no longer the way to do it? [21:20] last i checked that's still how to do it [21:21] unless there's a bug in apt/apt-get which i missed during my rounds of watching the bug announce channel [21:21] sudo apt-get install libx11-6:i386 .... E: Unable to locate package libx11-6 [21:24] * TheLordOfTime pulls up a quantal VM [21:25] (I am using the canonical EC2 image) [21:25] and which mirror is it pulling from :P [21:26] us-west-2 === giovani_ is now known as giovani [21:26] bleh my VM segv'd. [21:26] REINSTALLAITON TIME [21:26] hehe [21:28] might have figured it out [21:29] yep [21:29] dpkg --add-architecture i386 [21:29] apparently that used to be default in 12.04 amd64 server [21:29] and it no longer is [21:32] woo success === erichammond1 is now known as erichammond === medberry is now known as med_ [22:35] What is the best bare-metal backup strategy for an ubuntu server? [22:36] veeam? acronis? backupexec? [22:40] checking them out now [22:40] both acronis and backupexec seems to do what I want [22:40] or did you mean free backup software? [22:41] It doesnt have to be free if it works [22:41] I played with backupexec, but never put it into production [22:41] I mainly use veeam and acronis [22:41] only two free ones I know of, and I haven't digged deep into either [22:42] bacula, and hmm, someting that starts with an a [22:42] Seem like you have to boot from a CD [22:42] you always have to [22:42] to do a baremetal restore [22:42] coworkers use rsnapshot, suggested looking into e.g. horcrux for scripting around rsnapshot [22:42] but that isn't a back-to-working sort of tool, more of a "keep my data backed up" sort of tool === cpg|away is now known as cpg [22:43] ya, I do multible types of backups [22:43] including rsync, backuppc, and baremetal [22:43] all depending on how paranoid, and whattype of recovery I want [22:43] I looked into clonezilla [22:43] that also required a boot [22:44] also, nice, to know if some backup method fails, the others will likely still be working [22:44] acronis installs onto ubuntu, it shouldn't need reboots at all to backup [22:45] Hmm interesting [22:45] other thing you could do [22:45] use lvm [22:45] create snapshot [22:45] backup snapshot [22:45] release snapshot [22:45] but now your basically building your own backup solution [22:45] all depends what you want [22:46] I think I use lvm now.. =P I am not that pro with this [22:46] I'm all virtualized, so I just use veeam to backup everything [22:46] Lvm really makes the snapshot consistent? I have a database running as well [22:46] hallyn: ping [22:46] it makes a *consistant* disk snapshot [22:47] okey [22:47] flusing your applications to disk is a different story, and all backup solutions will have that issue [22:47] that is why microsoft created the vss stuff, so databases and things just plug their own vss driver in, and when someone requests that, they all flush [22:48] though really, I would not depend on that for a database backup [22:48] while it is nice, I would defently also backup the db's seperately, using just a db specific thing [22:48] adam_g: what's up? [22:48] it's more likely for your db to have issues, or get corrupted from user error or something [22:48] and easier to restore that way also [22:48] Ok, so the db is most prone to these kind of problems [22:49] well, the more you use sometihng, the more prone it is to issues [22:49] the rest of the os would backup fine [22:49] patdk-lap: too true. [22:49] okey [22:49] likely, doing weekly backups baremetal, and daily backups of the db would be good [22:50] but all depends on what it is you need, and what goals you want [22:50] hallyn: slowly making my way thru comments in bug #1057024 but, testing some openstack updates in quantal-proposed and running into a permissions issue on /dev/kvm with the qemu-kvm version thats in quantal-proposed. what is responsible for setting group ownership of /dev/kvm with the newer package? [22:50] Launchpad bug 1057024 in qemu-kvm "kvm kernel module always loaded, without setting /dev/kvm permissions" [High,Fix committed] https://launchpad.net/bugs/1057024 [22:50] Goal #1: Not having to reboot the server once a week and carry a monitor into the basement :P [22:50] like my db servers, nothing changes on them, so just doing database backups is enough, as long as I have one baremetal, or I could just easily rebuild it quickly enough [22:51] maybe there is a way to automate a boot-backup solution? [22:51] mkander, likely, doing a baremetal backup once every 3 months might be enough [22:51] and backing up /home /var/lib /etc daily [22:52] well, if websites and mail on it, /var/www and /var/mail also maybe [22:52] zul: if you're pushing a libvirt soon-ish, there's a trivial buglet in bug 1095140 [22:52] Launchpad bug 1095140 in libvirt "documentation flaw "libvirt", rather than "libvirtd" group" [High,Triaged] https://launchpad.net/bugs/1095140 [22:52] Yeah agree, but I want everything automated [22:52] maybe I could get clonezilla to do it automated [22:52] well, you could automate all that easily, except for the baremetal backups [22:52] adam_g: /dev/kvm perms are supposed to be fixed by the end of qemu-kvm.postinst [22:53] by the call to udevadm [22:53] adam_g: i thought i'd verified that one already fwiw [22:54] hallyn: hmm. im installing nova-compute, which pulls in qemu-kvm and i end up with root:root. ill keep poking [22:54] Seems like Clonezilla is an open source program that creates complete disk backups. Looks like some people have managed to automate the process as well. [22:54] adam_g: hm [22:55] adam_g: oh wait! yes there is (we believe) a bug in udev [22:55] lemme see where i filed it [22:56] adam_g: https://bugs.launchpad.net/ubuntu/+source/udev/+bug/1092715 lemme ping slangasek in -devel [22:56] Launchpad bug 1092715 in udev "udevadm trigger --action=change not working since quantal?" [Undecided,New] [22:57] hallyn: yippie, thanks! [22:59] hallyn: so is it unlikely the qemu-kvm update will be moving to -updates anytime soon? [22:59] not sure [23:00] in precise it should be :) [23:00] maybe i should spend tomorrow looking into the udev bug [23:01] hallyn: im concerned with quantal ATM. testing these nova SRUs assumes qemu-kvm works as expected. :\ i suppose i can work around the issue by manually setting permissions on /dev/kvm during deployment but i'd consider it a regression in the qemu-kvm in quantal-proposed, no? [23:02] adam_g: no, bc as slangasek points out that never worked right in quantal, so it's not a regression [23:03] adam_g: symptom is slightly different of course - before it would mark it the right group, but it wouldn't give group rw perms [23:06] adam_g: i'm sneaking away - ttyl [23:06] hallyn: well [23:07] hallyn: 1.2.0+noroms-0ubuntu2 gives me correct crw-rw---- 1 root kvm 10, 232 Jan 2 17:08 /dev/kvm [23:08] adam_g: getfacl /dev/kvm [23:09] (gotta run, bbl) [23:09] cya [23:32] how do I setup a proper bidirectional replication FROM a windows DNS server TO a bind server? [23:33] bidirectional? I thought the whole point of dns replication was master / slave [23:34] sarnold, ideally, however, the BIND server will need to do updates to the Windows enviroment. [23:35] sarnold, Idealy, the BIND server will be the slave [23:36] jeeves_moss: slave-only ought to be easy, configure the axfr access controls on your windows server to let your bind zone transfer, and set up the bind system to do zone transfers.. [23:36] jeeves_moss: .. but getting that bind to update the windows dns, well, that just doesn't seem 'normal' to me. [23:38] sarnold, right now, I need to setup the BIND server to be a slave, and pull all newly created domains, etc FROM the windows servers. there is a technet howto posted by microsoft that's a step by step, but I can't find it