/srv/irclogs.ubuntu.com/2012/07/13/#ubuntu-server.txt

=== pdtpatrick_ is now known as pdtpatrick
=== Nolar_ is now known as Nolar
=== cpg is now known as cpg|away
=== Ursinha` is now known as Ursinha
=== Ursinha is now known as Guest53271
hallynsmoser: and so your container would use both your new bridge and virbr0?  it's actually a tinge more complicated than what i was suggesting imo but should work :)  just make sure that libvirt doesn't start a dnsasq for the new network02:14
=== n0ts_off is now known as n0ts
=== Guest92579 is now known as cpg
overriderHello - i am running a 10.04 Server, and i would like to upgrade a few packages on it to newer version despite them not being in the archives. Example is id like to use rsnapshot1.3.1 but only have 1.3.0 on my system. Another is i kind of need wkhtmltopdf 0.9.9 but only have 0.9.0 on my system. Is the only obvious answer to upgrade to 12.04 ?03:25
arusselHi, I'm having a lot of difficulties install postgres on the latest LTS. I've done it on 2 other servers, I can't manage to know what is going wrong with this one.03:35
arusselI'm doing: 'sudo apt-get install postgresql'03:36
arusselbut no files are installed in '/etc/postgresql', so the init script doesn't find any 'version' and doesn't start anything03:37
arusselI've got no error either when installing or '/etc/inid.d/postgresql start'03:37
arusselwhat package is supposed to create the directory '/etc/postgresql' ?03:41
ScottKarussel: I think you want /etc/postgresql-common and it's postgresql-9.1.03:55
arusselScottK: thanks. I've tried a billion times with all the packages, I switch from aptitude to apt-get and got it installed now.03:57
arusselnot sure this is the root cause03:57
arooni-mobilehey folks;  running ubuntu 10.04 LTS;  i want to get php-fpm + nginx.  i already have nginx setup... but trying to figure out how to get php-fpm.  ideas?04:23
=== 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
S0undwaveHow does one go about manually configuring DHCP for Ubuntu Server? I'm new.05:53
ScottKThey didn't last long.06:00
=== n0ts is now known as n0ts_off
=== n0ts_off is now known as n0ts
xfxchildemorning07:19
=== zyga-afk is now known as zyga
=== smb` is now known as smb
=== ubuntu is now known as Nathan_S
=== exocaesar is now known as Guest71714
=== Guest71714 is now known as exocaesar
=== n0ts is now known as n0ts_off
=== n0ts_off is now known as n0ts
=== ubuntu is now known as Guest86821
=== Guest86821 is now known as Nathan_S
=== ninjak_ is now known as ninjak
=== fenris is now known as Guest73296
=== cpg is now known as cpg|away
=== n0ts is now known as n0ts_off
reisiwhat script is ran after kernel image installs to update grub? i'd like to add grub installing on all my hard drives (which are part of md raid) instead of first one10:29
=== Taftse is now known as Taftse|NL
RoyKhi all. setting up a couple of machines as a small kvm cluster with iscsi/gfs2 storage11:24
* glance would sugest using lvm and drbd instead11:25
RoyKhuh - why??11:26
reisiffs this grub/grub-pc thing for 10.04 is unbearable, apparently back in the day no one did booting from md+lvm11:26
RoyKthey're both connected to a SAN, shared storage11:26
* RoyK glances at glance 11:26
RoyKand I'm using lvm on that11:27
RoyKand gfs2 on top11:27
glancei would atleast drop gfs211:27
RoyKand then use what?11:27
glanceplain lv's11:27
glanceone less layer -> less problems and more preformance11:28
RoyKand how would you plan for the two machines to share that?11:28
=== al-maisan is now known as almaisan-away
glancedidn't you say that you used shared storage? =)11:28
=== almaisan-away is now known as al-maisan
RoyKglance: yes, shared block device, meaning: NEED FOR FILESYSTEM SUPPORTING SHARED BLOCK DEVICE11:29
* glance just shakes his head11:29
=== al-maisan is now known as almaisan-away
glancedon't use any filesystem on your host machines. use lv's directly.11:30
RoyKglance: if you have a good alternative to that, feel free to comment11:30
glancei just did.11:30
RoyKglance: oh - no filesystem - just dd your VMs to the LV?11:30
glancejepp11:30
glanceor rather qemu-img convert -O host_device11:31
RoyKhm...11:31
RoyKthen - how can I make sure those VMs aren't started on two nodes at the same time?11:31
glanceimplement that in control logic instead of filesystem locking.11:32
glancekvm doesn't lock the file anyhow.11:32
RoyKsanlock works well for that11:32
glanceand what prevents you from starting that vm two times on the same machine then? =)11:33
RoyKanyway - first problem is: iscsiadm -m node --login: works well, but after a reboot, the device is lost and I need to --login manually again - any idea when?11:33
RoyKglance: sanity and stuff like virsh/virt-manager11:34
glanceso, use your sanity to not start the vm's on multiple machines =)11:35
glanceor use ex. ganeti11:35
RoyKany idea about that iscsi problem?11:36
RoyKand about sanity and multiple machines - I'm not going to run this all by myself, and I want it foolproof11:37
RoyKany sane ones around?11:40
=== n0ts_off is now known as n0ts
glanceRoyK: you can use clvm if you like locking-by-shared-storage =)11:57
RoyKsorry, but what's wronga about gfs2?11:58
RoyKredhat uses that in production in its clusters11:58
RoyKalso, do you have iscsi experience? my LUNs fail to attach after a reboot11:59
_rubenRoyK: which target software you use?11:59
RoyKiscsiadm -m node shows the lun, but I have to manually login to it to see it11:59
_rubenor wait, you're just the initiator12:00
RoyKdunno - some SAN12:00
RoyKyep12:00
_rubenno experience with that12:00
=== n0ts is now known as n0ts_off
=== zyga is now known as zyga-food
=== n0ts_off is now known as n0ts
=== zyga-food is now known as zyga
=== almaisan-away is now known as al-maisan
shortdudeHi, does anyone have any experience with rsyslog?13:40
mdeslaurlynxman: are you the puppet expert?13:44
glanceshortdude: ask your question13:46
mdeslaurlynxman: puppet in quantal is likely to be busted, as default ruby is now 1.9....I've got a merge ready from debian, but it doesn't solve the problem13:46
jcastrosmoser: jamespage roaksoax Ursinha m_3 G+ in about 10 minutes?13:49
shortdudei have been running rsyslog to gather remote syslogs from other ubuntu servers and today it decided to start crapping out.  Most of the local syslog files are under 1M but today its over 600M w/ a repeated message from 2 hours ago.  It is complaining about the line "$RuleSet remote" and the 2 places i bind it to ports.  The rsyslog process is also hammering /dev/log.13:50
glancelog loops?13:51
glancewe just use $AllowedSender13:52
shortdudesorta, it continuosly logged a series of errors about the 3 lines 2 hours ago.  no other logs are duplicated over and over13:52
glancelet servers have if $fromhost-ip != '127.0.0.1' then ~13:54
glancebefore *.* @13:54
glance...13:54
glanceso that it doesn't send away any logs but its own.13:54
shortdudeif it helps, my config is at http://pastebin.com/mejj7v5r13:56
glanceis the config-file the same at cerberus13:58
glance?13:58
shortdudewhoops forgot to take that name out, cerberus is the name of the server13:59
shortdudei have restarted the rsyslog process but after 15 seconds it started hammering /dev/log and the cpu wait is 25-75% according to top.  I have not rebooted it to see what happens since i dont want to go rebotting things unless i know it will do some good.14:03
koolhead11hi al14:05
glanceshortdude: http://pastebin.com/D77nrj3x14:07
shortdude@glance... Thanks, another admin and i are working to chance our config to remove the ruleset similar to yours.  The other admin here with me said he found something about rsyslog 4.2 not working well with rulesets14:18
glanceshortdude: that config is runing in a 4.2 rsyslogd on lucid, and is working just fine.14:20
=== n0ts is now known as n0ts_off
glancein two years it have had maybee two unexpected crashes.14:20
glanceand its about 1k hosts logging to that machine14:21
shortdudesounds good, i like to hear that.  we are running on lucid also14:25
=== mrmist is now known as groupcat
=== al-maisan is now known as almaisan-away
philipballew__What would I use to see all ip address in use on my lan?15:51
philipballew__nmap for sure, but what is the command if anyone knows off hand15:51
ikoniaphilipballew__: the switch/router should show that from an arp15:52
philipballew__ikonia, well my router kinda sucks, but Im ssh's in from far away15:53
philipballew__Ah, ikonia I see what you mean!15:54
philipballew__thanks!15:54
=== jimbaker` is now known as jimbaker
koolhead11hi philipballew_16:02
shortdude@glance... you still on?16:10
=== ubuntu is now known as Nathan_S
hallynzul: bad news.  0.9.13 libvirt doesn't seem to work for creating a new qemu based vm through virt-manager (at least remotely)16:17
zulgah?16:17
glanceshortdude: sortof16:24
shortdude@glance... thought i would let you know that we figured out what was going on and we did have a logging loop on the log server16:36
shortdudewe took off a line and it is working now16:36
shortdudethanks for you help16:36
=== dendro-afk is now known as dendrobates
|2ump|2oastHello - I'm building a SOHO samba fileserver using ubuntu server 12.04 and I have been looking for bare metal backup / restore solution.  I've looked at mondo/mindi and tested it (both on 32bit and 64bit editions of ubuntu server) but there were numerous issues and I was unable to restore.  Does anyone here have a bare metal backup / restore solution that you use or recommend?17:14
koolhead11adam_g: hey there18:24
adam_gkoolhead11: heya18:28
koolhead11adam_g: will you be at OSCON>18:29
koolhead11adam_g: cool stuff. Folsom on Q and precise18:29
adam_gkoolhead11: of course ill be there, i can practically see the convention center from my window right now. :) will you?18:31
koolhead11adam_g: yes. am in portland as well18:31
adam_gkoolhead11: cool!18:31
koolhead11adam_g: you from portland itself?18:32
adam_gyup18:35
kirklandSpamapS: howdy!18:59
kirklandSpamapS: could I get you to promote something from -proposed to -updates?18:59
kirklandSpamapS: http://pad.lv/u/ecryptfs-utils18:59
kirklandSpamapS: there's a proposed package that's been sitting there for a long time18:59
SpamapSkirkland: I don't see any verification on the bugs19:02
SpamapSkirkland: bug 884407 seems to be verified19:04
uvirtbot`Launchpad bug 884407 in ecryptfs-utils "mount.ecryptfs_private is broken on arm" [Low,Fix committed] https://launchpad.net/bugs/88440719:04
=== cpg|away is now known as cpg
SpamapSkirkland: bug 576133 needs verification still tho19:04
uvirtbot`Launchpad bug 576133 in ecryptfs-utils "Shouldn't allow autologin for ecryptfs users" [Low,Fix committed] https://launchpad.net/bugs/57613319:04
kirklandSpamapS: lemme get that for you19:05
kirklandSpamapS: ugh;  downloading 11.10 iso;  will be a few minutes19:07
roaksoaxSpamapS: howdy! If you have some time, could you please take care of bug #102401019:28
uvirtbot`Launchpad bug 1024010 in maas-enlist "[SRU] After Commission Action 2 no longer exists" [High,Fix released] https://launchpad.net/bugs/102401019:28
kirklandSpamapS: done!  https://bugs.launchpad.net/ubuntu/+source/ecryptfs-utils/+bug/576133/comments/1019:29
uvirtbot`Launchpad bug 576133 in ecryptfs-utils "Shouldn't allow autologin for ecryptfs users" [Low,Fix committed]19:29
SpamapSroaksoax: if its in the precise-proposed queue, it will get looked at soon..19:58
roaksoaxSpamapS: awesome, thanks19:58
SpamapSroaksoax: we have a crazy backlog.. several days behind.. if its time sensitive though, I'm happy to queue jump it19:58
roaksoaxSpamapS: it would be nice to have asap as breaks maas enlistment into precise maas version20:02
roaksoaxinto quantal maas verrsion*20:02
hallynmdeslaur: hey - with virt-manager, connecting remotely, do you usually connect to root?  is connecting to qemu:///host/system (rather than session) the intent?20:07
mdeslaurI never use session, and I connect to my user on the remote machine20:07
mdeslaurhallyn: you use session?20:08
hallynmdeslaur: well system wasnt' working in one setup, so i had quickly hacked connect.py to use session...20:09
hallynbut after i did that, i noticed it created ~/VirtualMachines and used that in place of /var/lib/libvirt/images/,20:09
hallynand was wondering whether i'm just messing with things, or if there's a real problem :)20:09
hallynok, let me try clearing everything and using /system20:09
mdeslauryeah, as it doesn't have sufficient rights when you use session20:09
mdeslaurI suspect a bunch of stuff will fail if you use session20:10
mdeslauryou should put your user in the libvirt group on the remote machine, and use system with your user20:10
hallynok - i was in the libvirt group, but i think my problem was due to 0.9.13 in quantal (which we'll look at monday)20:11
hallynthanks (retrying)20:11
mdeslaurI've only used remote connections for testing virt-manager though, so I'm not exactly an expert of using it that way20:11
mdeslaurhallyn: let me know if you need any help20:11
hallynmdeslaur: i thought i'd tried this before, but i think ihadn't deleted my ~/VirtualMachines.  resetting everything worked.  thanks!20:14
mdeslaurhallyn: cool!20:14
hallynnow to figure out whether grub1 failing was fixed by switching from precise->quantal, or amd->intel20:16
stgraberhallyn: are you aware of an easy way to know whether an ethX entry is veth?20:19
stgraberhallyn: I need to make NM deal with veth for when it's running in a container. It currently ignores any virtual interface which makes it quite broken in there.20:20
hallynstgraber: hm, not really.  you can at least tell whether it's virtual by looking under /sys/devices/virtual/net20:35
stgraberhallyn: yeah, the problem is that NM ignores anything that's virtual, so I'm trying to make it a bit more clever, ideally without allowing all virtual devices in the process :)20:36
hallynstgraber: looking at the kernel code, ethtool may know how to get the information.20:37
hallynstgraber: 'sudo ethtool -i veth' shows'driver:veth'20:38
hallynstgraber: http://paste.ubuntu.com/1090532/  is the code which detects that.  ioctls will get you what you need20:41
stgraberhallyn: thanks! I'll add that to NM's detection code then20:41
hallyn0x00000003 is the ioctl #.  not sure yet if you can do ita gainst any socket or what20:42
hallynstgraber: yeah looks like just socket(AF_INET, SOCK_DGRAM, 0);20:43
=== dendrobates is now known as dendro-afk
hallynstgraber: surely you can make it prettier, but this ripped off hacked bit from ethtool works for me http://people.canonical.com/~serge/detectveth20:56
hallyn(shows what you need)  'sudo detectveth veth1' worked for me20:56
stgraberhallyn: cool. There's already a function in NM doing a similar ioctl, will change it a bit to also retrieve the driver and then add a check for veth20:58
stgraberthanks20:58
hallynstgraber: excellent :)  have a good weekend21:04
stgraberyou too21:04
zastaphoops.. I made a software RAID on another machine, moved the harddisks, now I can't even DBAN them, and most rescue disks fail to boot because something is confused21:16
zastaphtried many mdadm commands, fail, remove, stop.. --zero-superblock and even dd if=/dev/zero21:17
zastaphthe latter faults the OS during the process, segfaults21:17
zastaphwhat boot iso can do this gracefully ?21:18
zastaphguess I need dmraid21:34
hallynDaviey: are you still around?21:40
=== dendro-afk is now known as dendrobates
hallynDaviey: well, nm.  was wondering if your amd x130e was available and pre-loaaded with quantal server :)21:41
hallyni'll just sacrifice my install.  bbl21:41
=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
=== bazhang_ is now known as bazhang
=== dendrobates is now known as dendro-afk
=== Lcawte|Away is now known as Lcawte
=== dendro-afk is now known as dendrobates
=== cpg is now known as cpg|away
=== cpg|away is now known as cpg
=== dendrobates is now known as dendro-afk
=== cpg is now known as cpg|away
=== Lcawte is now known as Lcawte|Away

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!