qman__ | is there a PPA or some other low-hassle method to get kernel 3.2 on 10.04? | 00:49 |
---|---|---|
qman__ | I need it for driver reasons | 00:49 |
three18ti | any way I can move /lib to its own partition post-install? | 02:13 |
three18ti | /tmp /usr /boot /var and /home are all on their own partitions.. | 02:15 |
three18ti | I just forgot /var | 02:16 |
three18ti | derr.. I mean /lib | 02:16 |
tdotr6 | hey guys | 03:46 |
tdotr6 | im trying to setup multi nic cards on my box | 03:46 |
tdotr6 | and my issue is , I cant get the default route to stay added | 03:46 |
tdotr6 | I understand you can not have 2 default routes on the same gateway | 03:46 |
tdotr6 | sorry subnet * | 03:46 |
tdotr6 | and to me It doesnt make sense , you cant have 2 default gateways, But When the 1 NIC fails, how can I have the second nic take over like it should.. | 03:47 |
tdotr6 | ? | 03:54 |
tdotr6 | anyone? | 03:54 |
tdotr6 | home? | 03:54 |
qman__ | you have to do one of a number of things to make it work | 03:55 |
qman__ | you either have to set metrics so one route is always preferred | 03:55 |
qman__ | or configure the kernel to recognize which interface traffic is coming from, and send it back out that way | 03:56 |
qman__ | or configure load balancing | 03:56 |
qman__ | I don't have time to go into the details of how to do those things, but hopefully that will set you in the right direction | 03:56 |
qman__ | without doing one of the above, having more than one default gateway will not work correctly | 03:58 |
qman__ | tdotr6, ^ | 03:58 |
mistigi | Hello, I am having problem with updates: linux-image-server depends on linux-image-3.2.0-37-generic; however: Package linux-image-3.2.0-37-generic is not installed. Would appreciate if anybody could help, here is my session dump: http://pastebin.com/dFpvBVFb | 04:02 |
ScottK | mistigi: Do sudo apt-get dist-upgrade | 04:07 |
mistigi | k, let me try that | 04:08 |
mistigi | same issue: root@system:/boot# sudo apt-get dist-upgrade | 04:09 |
mistigi | Reading package lists... Done | 04:09 |
mistigi | Building dependency tree | 04:09 |
mistigi | Reading state information... Done | 04:09 |
mistigi | You might want to run 'apt-get -f install' to correct these. | 04:09 |
mistigi | The following packages have unmet dependencies: | 04:09 |
mistigi | linux-image-server : Depends: linux-image-3.2.0-37-generic but it is not installed | 04:09 |
mistigi | linux-server : Depends: linux-headers-server (= 3.2.0.37.44) but 3.2.0.39.47 is installed | 04:09 |
mistigi | E: Unmet dependencies. Try using -f. | 04:09 |
mistigi | oh, let me try that with -f, one moment | 04:10 |
mistigi | its pretty much the same, it forced some packages to upgrade and kernel too, from .38 to .39, after reboot kernel is on .39. But still any apt operation complains " linux-server depends on linux-image-server (= 3.2.0.37.44); however: Package linux-image-server is not configured yet." | 04:22 |
mistigi | "linux-image-server depends on linux-image-3.2.0-37-generic; however: Package linux-image-3.2.0-37-generic is not installed." | 04:23 |
ScottK | The correct version is published: linux-image-server | 3.2.0.39.47 | precise-updates | amd64, i386 | 04:28 |
ScottK | Next I'd try, sudo apt-get -f install linux-image-server | 04:30 |
mistigi | k, one sec | 04:30 |
mistigi | sudo apt-get -f install linux-image-server ... "The following packages have unmet dependencies: linux-server : Depends: linux-image-server (= 3.2.0.37.44) but 3.2.0.39.47 is to be installed Depends: linux-headers-server (= 3.2.0.37.44) but 3.2.0.39.47 is to be installed" | 04:32 |
ScottK | Try sudo apt-get -f install linux-image-server linux-headers-server | 04:33 |
ScottK | Actually, run sudo apt-get update first. | 04:33 |
mistigi | k | 04:33 |
ScottK | I just installed this kernel off the US mirror and it's there. | 04:33 |
ScottK | (at least for i386) | 04:33 |
mistigi | still same error, like apt is stuck on .37 though I am running .39 "The following packages have unmet dependencies: linux-server : Depends: linux-image-server (= 3.2.0.37.44) but 3.2.0.39.47 is to be installed Depends: linux-headers-server (= 3.2.0.37.44) but 3.2.0.39.47 is to be installed" | 04:36 |
mistigi | would it make sense to copy /etc/apt/sources.list from another machine running .39 ? | 04:37 |
ScottK | No. It does appear that you've got the right entries. | 04:38 |
ScottK | mistigi: Try this and pastebin what you get: grep linux-server /var/lib/apt/lists/* | 04:43 |
mistigi | k | 04:43 |
mistigi | http://pastebin.com/PiQhfwut | 04:44 |
ScottK | Try taking the i386 entries out of your sources.list, apt-get update, and try again. | 04:47 |
mistigi | k | 04:47 |
ScottK | You can see you've got the correct version in your Packages files, so I am wondering if things are somehow confused due to multi-arch. | 04:49 |
ScottK | (it's a guess) | 04:49 |
mistigi | hmm, all 386 are commented out already in sources.list | 04:49 |
ScottK | They were being checked in your first pastebin. | 04:50 |
ScottK | Check isn sources.list.d too. | 04:50 |
mistigi | k | 04:50 |
ScottK | By comparison to your output, here's what mine (for an i386 server) shows: http://paste.ubuntu.com/5627289/ | 04:52 |
mistigi | I see what you mean about 386 being check out, but I do not see them anywhere under /etc/apt | 04:54 |
ScottK | OK. | 04:55 |
ScottK | Here's what I'd do then ... | 04:56 |
ScottK | wget https://launchpad.net/~canonical-kernel-team/+archive/ppa/+build/4332112/+files/linux-server_3.2.0.39.47_amd64.deb | 04:56 |
ScottK | sudo dpkg -i linux-server_3.2.0.39.47_amd64.deb | 04:56 |
ScottK | sudo apt-get -f install | 04:56 |
mistigi | k, let me try that | 04:56 |
ScottK | It's kind of doing it the hard way, but I don't see why you aren't gettting the new one. | 04:57 |
mistigi | same error: http://pastebin.com/7n2btqpL | 04:59 |
ScottK | Progress. Not quite the same error. | 05:00 |
ScottK | Do the same thing with https://launchpad.net/~canonical-kernel-team/+archive/ppa/+build/4332112/+files/linux-image-server_3.2.0.39.47_amd64.deb | 05:01 |
mistigi | right, that did not display any errors, one sec | 05:04 |
mistigi | wow, it looks like it worked :-) Really appreciate your help :-) | 05:05 |
=== mikehale__ is now known as mikehale | ||
ScottK | Except I was giving you the wrong version. Crap. | 05:08 |
ScottK | Somehow I ended up on the kernel team PPA. | 05:08 |
mistigi | its a VM, i have a snapshot I can rollback to | 05:09 |
ScottK | Acutally, it's fine. | 05:09 |
mistigi | k | 05:09 |
ScottK | I forgot the kernel updates are all built in a special PPA for testing. | 05:09 |
ScottK | So those are actually the right ones. | 05:09 |
mistigi | sweet, thanks for your time, really apreciate it :-) Saves me a lot of effort so I do not have to rebuild this box from scratch | 05:10 |
Keith_Pawson | What is the best option for quickly building new test servers for either home setup or a test env, have been looking at Cobbler or considering fog for images? | 05:25 |
Keith_Pawson | I would also like to build both Ubuntu servers and CentOS | 05:25 |
Sargun | How has your guys' luck with Ubuntu MaaS been? | 06:50 |
ttx | jamespage: swift milestone-proposed branch cut | 07:59 |
jamespage | ttx, marvellous - thanks! | 08:03 |
ttx | jamespage: in the process of tagging 1.8.0-rc1 | 08:03 |
sonne | hey there | 09:09 |
=== security is now known as fire | ||
=== smb` is now known as smb | ||
phretor | I followed ubuntu's practice of putting the images under /var/lib/libvirt/images/<vm name>/ubuntu-kvm/*.qcow2 - is there any suggested practice about snapshots (i.e., where to place them on the host filesystem)? | 09:40 |
* RoyK wonders what sort of genius that came up with moving to a new kernel release at 12.04.2 | 09:48 | |
ogra_ | the policy of supporting new HW with newer point releases | 09:53 |
ogra_ | (complain in #ubuntu-kernel instead of -server though) | 09:54 |
=== phretor_ is now known as phretor | ||
alex88 | koolhead17: o/ | 10:12 |
koolhead17 | alex88: my friend. howdy/ | 10:13 |
alex88 | just configured some new servers for mysql and storage, finally with hw raid, it flies :) you? | 10:14 |
alex88 | btw, why I need to install a package to set as manually installed? | 10:19 |
=== island219_ is now known as island219 | ||
alex88 | is there a way to set a package as installed from source in apt-get? | 10:33 |
=== phretor_ is now known as phretor | ||
mardraum | hi, anyone here using qemu-kvm with libvirt on raring who is still able to live migrate with virsh? I can't do so after upgrading two hosts for testing, trying to work out if it is a bug or me | 10:34 |
ikonia | alex88: no | 10:38 |
alex88 | ikonia: just found https://help.ubuntu.com/community/CheckInstall | 10:38 |
ikonia | alex88: I very strongly advise against that | 10:40 |
ikonia | but of course it's up to you | 10:40 |
alex88 | ikonia: usually generated problems? | 10:40 |
ikonia | no, I just feel the whole process is bad | 10:40 |
ikonia | but it's a personal opinion | 10:40 |
ikonia | it's a sloppy package basically | 10:40 |
Daviey | smb: Do you have that xen dsc/debdiff? | 10:41 |
Daviey | mardraum: Hmm, yeah it should still work.. What are you seeing? | 10:42 |
smb | Daviey, if that means the one between 4.2.0 and 4.2.1... maybe not yet but simple to create | 10:42 |
Daviey | smb: raring current, to what you are proposing :) | 10:42 |
Daviey | alex88: what source is it? | 10:43 |
smb | Daviey, its now where the src pkg is | 10:45 |
Daviey | smb: thaks | 10:45 |
mardraum | Daviey: I get "error: operation failed: migration job: unexpectedly failed" from something like "virsh migrate --live --verbose somevm qemu+ssh://somehost/system" - nfs shared storage, openvswitch networking | 10:46 |
mardraum | dns is still ok, same as it was before upgrade, forward and reverse ok for both hosts | 10:46 |
Daviey | mardraum: Ah, i haven't played with shared storage migration as much.. In a few hours, hallyn will be around.. he will be a much better person to help work this out. | 10:48 |
mardraum | Daviey: cheers | 10:48 |
alex88 | Daviey: vips (package is libvips-tools) | 10:51 |
alex88 | I should do the same also for libtiff since I need 4.0.3 on ubuntu 12.04 | 11:06 |
alex88 | and that version is just on ubuntu 12.10 | 11:06 |
alex88 | maybe I'll just install its dev | 11:06 |
alex88 | maybe I'll just install its deb | 11:06 |
Daviey | alex88: if that doesn't work (quite likely), grabbing the source package from a newer release.. and rebuilding it for 12.04 is better | 11:07 |
Daviey | alex88: out of interest, what is it you need a later version for? | 11:07 |
alex88 | Daviey: bigtiff support | 11:07 |
alex88 | files more then 4gb | 11:07 |
Daviey | ah | 11:07 |
alex88 | btw, I need the dev package | 11:08 |
alex88 | Daviey: http://packages.ubuntu.com/quantal/libtiff5-dev this is what I need to install, can you advice me how to do that? | 11:09 |
Jeeves_ | 'apt-get install libtiff5-dev' | 11:09 |
alex88 | Jeeves_: I'm not on 12.10 | 11:10 |
alex88 | I'm on 12.04 | 11:10 |
Daviey | alex88: I'm sorry, but this is going to take longer than I have time to spare right now. | 11:11 |
alex88 | Daviey: oh sure, no worries | 11:11 |
Jeeves_ | w | 11:21 |
Jeeves_ | oops | 11:21 |
alex88 | the problem is that installing from source installs both libtiff4 and libtiff4-dev pacakges | 11:29 |
alex88 | I mean, the same contents of both | 11:29 |
patdk-lap | exactly what do you mean, installing from source? | 11:32 |
alex88 | patdk-lap: downloading source package from authors site, ./configure, make, make install | 11:33 |
patdk-lap | well, that wasn't the right way to do it | 11:33 |
alex88 | mmhh no? | 11:33 |
patdk-lap | not if you ever planned to uninstall it or needed to fix your system | 11:34 |
alex88 | patdk-lap: well, I'm using checkinstall | 11:34 |
alex88 | instead of make isntall | 11:34 |
alex88 | *install | 11:34 |
patdk-lap | you should have downloaded the source from 12.10 | 11:36 |
patdk-lap | done a dpkg-buildpackage | 11:36 |
patdk-lap | then installed the .deb | 11:36 |
alex88 | patdk-lap: from where? http://packages.ubuntu.com/quantal/libtiff5-dev -> [tiff_4.0.2-1ubuntu2.1.debian.tar.gz] ? | 11:37 |
patdk-lap | yes. atleast 3 files are needed | 11:37 |
patdk-lap | the debian.tar.gz, the source(.tar.gz?) and the .dsc | 11:37 |
alex88 | ok so all o fhtem | 11:38 |
alex88 | *of them | 11:38 |
patdk-lap | but if you did a make *install* | 11:38 |
patdk-lap | not sure what it did to your system | 11:38 |
patdk-lap | and how hard it will be to put it back to normal | 11:38 |
alex88 | patdk-lap: I didn't, just checkinstall, then dpkg -r | 11:39 |
alex88 | and it's now clean | 11:39 |
alex88 | since no traces of libtiff in /usr/lib | 11:39 |
patdk-lap | you shouldn't ever :) | 11:39 |
alex88 | patdk-lap: btw, it's still clean now, also, what should I do with those 3 files? | 11:40 |
patdk-lap | even when it is installed, there will be no traces of it there (except maybe if you use i386?) | 11:40 |
patdk-lap | dpkg-source -x | 11:40 |
alex88 | patdk-lap: it's not installed, no files, no packages installed | 11:40 |
patdk-lap | find /usr -name 'libtiff*' | 11:41 |
alex88 | just in /usr/local/src that's where I'm working on | 11:41 |
patdk-lap | ok | 11:41 |
alex88 | btw, did dpkg-source -x tiff_4.0.2-1ubuntu2.1.dsc | 11:42 |
alex88 | extracted, unpacked, applied patches and so on | 11:42 |
patdk-lap | go into that folder | 11:42 |
alex88 | done | 11:42 |
patdk-lap | apt-get build-dep libtiff4 | 11:42 |
alex88 | patdk-lap: version 4.0.2 from quantal is libtiff5, should I apt-get build-dep libtiff5? | 11:43 |
patdk-lap | is there a libtiff5 in your os version? | 11:43 |
alex88 | libtiff4 is version 3.9.2 from precise, libtiff5 is this one 4.0.2 from quantal | 11:43 |
alex88 | nope | 11:43 |
patdk-lap | then it won't work | 11:43 |
patdk-lap | use libtiff4 | 11:43 |
alex88 | as build-dep right? | 11:43 |
patdk-lap | yes | 11:43 |
patdk-lap | your just installing what it needs to build | 11:44 |
patdk-lap | the differences between libtiff4 and libtiff5 are normally nothing | 11:44 |
alex88 | btw | 11:44 |
alex88 | Picking 'tiff' as source package instead of 'libtiff4' | 11:44 |
alex88 | E: Unable to find a source package for tiff | 11:44 |
alex88 | btw, ./configure already works | 11:44 |
alex88 | since I've already installed packages | 11:44 |
patdk-lap | who said to use ./configure? | 11:44 |
alex88 | (i think it works, it does for the 4.0.3 I manually downloaded) | 11:44 |
patdk-lap | now it's probably messed up and need to delete that folder and do dpkg-source again | 11:44 |
patdk-lap | ok | 11:45 |
alex88 | nope, haven't run it here | 11:45 |
alex88 | just said that it worked for the other source I downloaded by hand | 11:45 |
patdk-lap | dpkg-buildpackage -us -uc | 11:45 |
alex88 | k, trying | 11:45 |
alex88 | dpkg-checkbuilddeps: Unmet build dependencies: cdbs (>= 0.4.106~) libxmu-dev libglu1-mesa-dev freeglut3-dev libxi-dev libjbig-dev | 11:45 |
alex88 | ok, trying to install those | 11:45 |
alex88 | damn, libjbig-dev is not there on 12.04, just libjbig2dec0-dev | 11:46 |
patdk-lap | same thing | 11:46 |
patdk-lap | or maybe | 11:47 |
patdk-lap | just give it a try | 11:47 |
alex88 | try? I get http://pastie.org/pastes/6625376/text | 11:48 |
alex88 | btw, I've installed all the other packages | 11:48 |
patdk-lap | heh? | 11:48 |
alex88 | patdk-lap: I get that running dpkg-buildpackage -us -uc | 11:49 |
alex88 | it says that libjbig-dev is missing | 11:49 |
alex88 | maybe I need to do the same thing for libjbig-dev? | 11:51 |
alex88 | build from source? | 11:51 |
alex88 | as it just depends on libjbig0 which just needs libc6 so not that many dependencies to go trough | 11:52 |
patdk-lap | ya, looks like you need that | 11:53 |
patdk-lap | the one in precise is 0.11 vs 2.0 | 11:53 |
patdk-lap | though it seems to compile with it, it's probably going have bugs | 11:54 |
alex88 | what? libc6? | 11:54 |
patdk-lap | and you will need to edit cdbs to say 0.4.100 | 11:54 |
patdk-lap | libjbig-dev | 11:54 |
patdk-lap | edit debian/control | 11:55 |
alex88 | ok, let me try to build libjbig first | 11:55 |
alex88 | I've already downloaded source | 11:55 |
alex88 | doing dpkg-source -x jbigkit_2.0-2ubuntu1.dsc | 11:55 |
alex88 | now dpkg-buildpackage -us -uc | 11:55 |
patdk-lap | ya | 11:56 |
alex88 | good, I've the debs (party) | 11:56 |
alex88 | now install jbigkit debs | 11:56 |
patdk-lap | yes | 11:56 |
NaGeL_Work | I have a question about munin can it be used that the munin and munin-node is installed on the same server? and what is munin-node anyway? | 11:56 |
patdk-lap | dpkg -i *.deb | 11:56 |
alex88 | done, now back to tiff | 11:56 |
alex88 | ok, just dpkg-checkbuilddeps: Unmet build dependencies: cdbs (>= 0.4.106~) | 11:57 |
patdk-lap | I said you needed to edit that | 11:57 |
alex88 | sure, going to do that now (was just to do step by step) | 11:57 |
alex88 | k, I'll set 0.4.100~ as version | 11:58 |
alex88 | since precise has that | 11:58 |
alex88 | ok building | 11:58 |
alex88 | there we go.. I've the deb :) | 11:59 |
alex88 | all debs installed, great! | 11:59 |
alex88 | thanks patdk-lap! | 12:00 |
alex88 | thanks alot! | 12:00 |
patdk-lap | now to see if things will, just use it | 12:00 |
patdk-lap | or if they need to be recompiled based on that new version | 12:00 |
alex88 | patdk-lap: well, vips now says tiff is installed correctly | 12:00 |
alex88 | the problem is that libmagick++-dev installs libtiff4.. | 12:01 |
alex88 | I think it will replace mine | 12:01 |
alex88 | isn't possible to say that libtiff5 is the same and just "use that"? | 12:01 |
patdk-lap | oh, probably uninstall the deb, make another change and rebuild | 12:01 |
alex88 | say that it provides libtiff4? | 12:01 |
alex88 | btw, quantal has both packages, so maybe they can work together? | 12:02 |
alex88 | btw, sorry patdk-lap, have to go to lunch, I'll talk to you later if you still have time.. thanks alot anyway for now!! | 12:03 |
patdk-lap | they can *work together*, but everything in precise will default to the old package :) | 12:04 |
patdk-lap | what you can do is edit the control file again, and change tiff5 to tiff4 | 12:05 |
patdk-lap | that should do it | 12:05 |
patdk-lap | forget how to say, also provides | 12:05 |
=== cronus__ is now known as cronus | ||
zul | jamespage/yolanda: https://code.launchpad.net/~zulcss/python-keystoneclient/0.2.3/+merge/154076 | 13:14 |
yolanda | ok | 13:15 |
zul | Daviey: ping...can you review python-ceilometerclient its still stuck in binary-new | 13:23 |
Daviey | zul: src new? | 13:24 |
jamespage | zul, swift rc'ed btw | 13:24 |
zul | Daviey: yeah | 13:24 |
zul | jamespage: ill get to it next | 13:25 |
alex88 | patdk-lap: "forget how to say, also provides"? | 13:27 |
zul | jamespage: tarball was cut? | 13:27 |
jamespage | zul, no idea - I switch to the milestone-proposed branch for ci | 13:27 |
jamespage | zul, I think 1.8.0-rc1 was being prepared | 13:27 |
zul | jamespage: oh ok | 13:27 |
zul | yeah they actually havent caught a tarball for it yet though :) | 13:28 |
=== wedgwood_away is now known as wedgwood | ||
=== luminous_ is now known as luminous | ||
zul | jamespage: http://people.canonical.com/~chucks/ca/ | 13:38 |
mardraum | jamespage: I think you helped me once before with a libvirt regression, do you have any vm hosts running raring you can try some live migrations? | 13:41 |
mardraum | mine are failing since I started testing beyond 12.10 | 13:42 |
jamespage | mardraum, give me 20 mins and I can switch my testing from precise->raring for grizzly | 13:43 |
mardraum | thanks. if you scrollback you'll see what I am getting using virsh migrate | 13:43 |
jamespage | zul, typo in changelog btw | 13:44 |
jamespage | zul, 'releaes' | 13:44 |
zul | jamespage: frigging | 13:44 |
jamespage | mardraum, I don't have a rig with shared storage; but I should be able to test it with live-block migration I think | 13:48 |
mardraum | ok, thanks. if that works for you I'll try that too | 13:49 |
hallyn | mardraum: I don't have a rig with shared storage either, but usually reproduce with some cloud instances with nfs when I need to. But for something like this with so many variables, a launchpad bug where we can track info is usually best | 14:07 |
hallyn | mardraum: do you mind opening one, with xml files for your vms and storage, and your exact commands? | 14:07 |
mardraum | sure. also I just finished testing using --copy-storage-all after creating the img on the receiving end and got the same error. | 14:08 |
mardraum | heh, the attempt also makes the vm mount ro, sigh | 14:14 |
mardraum | and a bsd vm panic as a filesystem disappears | 14:14 |
mardraum | I'm going to rule out openvswitch too first | 14:15 |
mardraum | I'll go back to the old bridging and see if that works | 14:15 |
=== phretor_ is now known as phretor | ||
f4c3p4lm | can anyone help me with postfix+courier-imap ? i just need a correct mysql database structure | 14:31 |
f4c3p4lm | :< | 14:36 |
histo | f4c3p4lm: Are you following a guide in particular? flurdy.com has one | 14:37 |
alex88 | patdk-lap: btw, got everything to work.. it seems..thanks man! | 14:37 |
f4c3p4lm | histo, i followed a guide cant recall which one | 14:38 |
f4c3p4lm | i just need the database | 14:38 |
f4c3p4lm | all the rest is OK | 14:38 |
f4c3p4lm | my db is incorrect | 14:38 |
f4c3p4lm | i know that there are missing tables | 14:38 |
f4c3p4lm | dont know the names and fields | 14:39 |
histo | f4c3p4lm: i would look to flurdy.com for examples. How are we going to guess what sort of tables your configs use etc... | 14:39 |
f4c3p4lm | ill look for the guide but | 14:39 |
histo | f4c3p4lm: each guide is different they could have used any format/table name they wanted | 14:39 |
f4c3p4lm | isnt there a generic one? | 14:40 |
f4c3p4lm | http://flurdy.com/docs/postfix/edition8.html | 14:41 |
f4c3p4lm | would this be the one? | 14:41 |
histo | f4c3p4lm: http://flurdy.com/docs/postfix/index.html | 14:44 |
histo | f4c3p4lm: specifically http://flurdy.com/docs/postfix/index.html#config-simple-database section | 14:45 |
f4c3p4lm | hum | 14:45 |
f4c3p4lm | thanks | 14:45 |
f4c3p4lm | my database looks exactly like that one :D | 14:46 |
eagles0513875 | hey guys i have dvecot setup using sieve which conf file do i need to look at to ensure its enabled | 14:46 |
eagles0513875 | !sieve | 14:46 |
=== dduffey_afk is now known as dduffey | ||
jamespage | zul, http://people.canonical.com/~jamespage/ca-updates/ | 15:27 |
jamespage | quantum and cinder | 15:27 |
zul | jamespage: +1 | 15:29 |
histo | I'm setting up a mail server for a small client. SHould I use virtual mailbox maps? Or just linux accounts with nologin | 15:30 |
histo | So far I've installed the mail-stack-delivery package and my self generated certs | 15:31 |
histo | Just looking for someone's opinions | 15:32 |
vila | hallyn: ping, I got a weird error while trying to create a kvm since I upgraded to raring, can you help me refining the diagnosis ? virt-install (which was working fine with quantal and even on raring on my laptop) is now failing with: error: ERROR An error occurred, but the cause is unknown | 15:54 |
hallyn | vila: do you mind opening a bug? add teh full command line you're using, and /var/log/libvirt/libvirtd.log and /var/lib/libvirt/qemu/<vm>.log if it exists. | 15:55 |
hallyn | i don't actually use virt-install, so not sure offhand if it'd be a bug in that, or libvirt, or qemu | 15:56 |
vila | hallyn: ~/.virtinst/virt-install.log is slightly more helpful: http://paste.ubuntu.com/5628501/ | 15:56 |
hallyn | oh, ok, usually it isn't :) checking | 15:56 |
hallyn | oh you're using virtualbox | 15:56 |
vila | hehe, at least there is a stacktrace | 15:56 |
vila | hallyn: no, not anymore, but it's still installed | 15:56 |
vila | hallyn: but I too found it weird that vbox appeared there | 15:58 |
hallyn | vila: Tue, 19 Mar 2013 16:48:26 virt-install 15392] DEBUG (cli:333) Received libvirt URI vbox:///system | 15:58 |
hallyn | vila: waht command line are you using for virt-install? | 15:58 |
vila | hallyn: it's in the paste, second line | 15:59 |
hallyn | vila: oh, or you might look through /etc/libvirt/libvirtd.conf for any defaults set to vbox | 15:59 |
vila | hallyn: only comments there (so only default values ?) and I didn't touch it. The only thing I can remember changing are: /etc/libvirt/storage/default.xml to use /caviar3/libvirt/images and using br0 (though I can't remember whre exactly), but I made these changes while running precise I think | 16:02 |
hallyn | vila: allr ight, please open a bug aginst lbivirt. lbivirt is telling virtinst that vbox is the default uri. i dont' know why offhand | 16:04 |
* hallyn must move to server meeting | 16:04 | |
vila | hallyn: ack | 16:04 |
hallyn | thx | 16:04 |
vila | hallyn: https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1157272 | 16:12 |
uvirtbot | Launchpad bug 1157272 in libvirt "fail to create vm since upgrade to raring" [Undecided,New] | 16:12 |
=== tsaavik_ is now known as tsaavik | ||
vila | hallyn: uninstalling vbox worked around the issue, comment added to the bug | 16:23 |
hallyn | vila: thanks much | 16:41 |
hallyn | note, i don't t hink you canrun kvm and vbox at the same time | 16:42 |
vila | hallyn: my pleasure. It may still be worth fixing this bug, my team has use cases where kvm is not appropriate yet | 16:42 |
vila | hallyn: yup, fully understood | 16:42 |
vila | hallyn: but in the past, I could at least run one or the other | 16:43 |
vila | hallyn: which was crucial to allow me to migrate to kvm... | 16:43 |
hallyn | vila: but if you install vbox, but don't have any vobx instances running, | 16:43 |
hallyn | oh wait, lemme go check the code... | 16:43 |
vila | hallyn: that was the case, I tried uninstalling vbox by supersitition and because you mentioned you didn't expect to see vbox there | 16:44 |
vila | hallyn: but I haven't started a vbox vm for.. months and certainly not since I rebooted ;) | 16:44 |
sarnold | I've had success in the past unloading the vbox module, modprobing the kvm module, and continuing on... | 16:44 |
hallyn | vila: oh, sorry, i thought you said somewhere you had two vbox instances up and running, | 16:46 |
vila | sarnold: would have been delicate in my case, most of them were started by jenkins | 16:46 |
hallyn | sarnold: you're *supposed* to no longer need to do that | 16:46 |
hallyn | sarnold: but my guess is libvirt simply checks for vbox before kvm | 16:46 |
hallyn | so that vbox becomes teh default if availab.e zul^ do you know if that's true? | 16:46 |
vila | hallyn: no, two vbox related processes (the services.. can't remember the names) | 16:46 |
sarnold | hallyn: no kidding! :) There's no end to the wonders in the world... :) | 16:46 |
sarnold | vila: heh, delicate indeed :) it's not the same as just driving your laptop.. heh. | 16:47 |
hallyn | vila: oh, got it, thx. never used vbox, didn't realize it left taskshaning :) | 16:47 |
vila | sarnold: hehe | 16:47 |
vila | hallyn: yeah, something may have changed, I thought those services were only started when the first vbox vm was... | 16:48 |
vila | hallyn: but now that you mention it, that may have been libvirt starting them... | 16:48 |
vila | hallyn: like: libvirt: hmm, what do we have here ? Do we have vbox ? vbox: Me ? Let me start ! | 16:49 |
vila | hallyn: AFAIR, those processes were supposed to shut down themselves when inactive and uninstall complained that they were still running and failed, I had to kill-9 them | 16:50 |
hallyn | vila: right. if you specify '-c qemu:///' it'll use kvm, but if you don't it uses a default | 16:50 |
hallyn | we would want the default to be kvm though | 16:50 |
hallyn | heh, perahps a vbox bug is in order | 16:51 |
vila | as in --connect qemu://system ? I tried that but got another obscure error | 16:51 |
vila | oh wait 3 / not 2 ? | 16:51 |
hallyn | yeah | 16:52 |
hallyn | it's 'qemu://host/{system,session}' | 16:52 |
* vila reinstalls vbox to try ;) | 16:53 | |
hallyn | vila: can you check /etc/libvirt/libvirt.conf for a uri_default setting? | 16:54 |
hallyn | and env | grep URI | 16:54 |
hallyn | libvirt honors those if set | 16:54 |
vila | hallyn: I did, nothing there, ha, well, commented out that is | 16:54 |
vila | #uri_default = "qemu:///system" | 16:54 |
vila | urgh, 3 / there, I hate my tyops :-( | 16:55 |
hallyn | vila: yeah, that's commented out in the distro file. not sure if we want to uncomment that | 16:55 |
vila | hallyn: virt-install -c qemu:///system will take precedence right ? | 16:56 |
vila | hallyn: I have a script that encapsulates the calls to virt-install so I can fix it there | 16:57 |
vila | hallyn: and yeah, given the comment in libvirt.conf, it's a fallback | 16:57 |
vila | # This can be used to prevent probing of the hypervisor | 16:57 |
vila | # driver when no URI is supplied by the application. | 16:57 |
TristamWrk | I've got a Dell Power Edge R515 with an Opteron 4334 6 core processor with 12.04 server installed on it. I'm only seeing 1 core. Anybody have any ideas what I should look at? Uname -a to follow: | 17:02 |
TristamWrk | Linux XXXXXXX 3.5.0-26-generic #42~precise1-Ubuntu SMP Mon Mar 11 22:17:58 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux | 17:03 |
mistigi | Not sure how you "seeing 1 core", but you can check: cat /proc/cpuinfo or start top and press "1" | 17:05 |
TristamWrk | cpuinfo is only showing 1 core. | 17:05 |
TristamWrk | as is htop | 17:05 |
sarnold | TristamWrk: is the system bios configured to only run one core? | 17:06 |
TristamWrk | sarnold, it better not be or I'm going to smack Dell. I will have to run downstairs and check. | 17:06 |
TristamWrk | I will have to do so after this person calls me back. | 17:08 |
mistigi | How about: cat /var/log/dmesg | grep SMP also: lshw -class cpu | 17:13 |
=== unreal is now known as CaptainBirdseye | ||
hallyn | plars: hey, what is the protocol for filing a bug against packages in linaro ppas? | 17:20 |
hallyn | just email whoever is in the changelog, i assume? :) | 17:21 |
plars | hallyn: what's the project? | 17:21 |
hallyn | plars: qemu-linaro | 17:22 |
plars | hallyn: https://bugs.launchpad.net/qemu-linaro | 17:23 |
hallyn | plars: that pkg doesn't exist in raring though... only in ppa... stillu se it? | 17:23 |
hallyn | ok, i'll mark it affecting that, thanks :) | 17:24 |
plars | hallyn: probably - it wouldn't hurt to email the package maintainer also though | 17:24 |
plars | hallyn: if nothing else, ping fabo on #linaro - I'm sure he can get you to the right place | 17:26 |
hallyn | plars: ok, thanks | 17:26 |
TristamWrk | alright, I checked the bios and all the cores are enabled. | 17:34 |
mistigi | Tris: can you check your dmesg log like I posted above | 17:35 |
TristamWrk | mistigi, I didn't scroll up when I got back to my desk..sorry | 17:42 |
sarnold | TristamWrk: < mistigi> How about: cat /var/log/dmesg | grep SMP also: lshw -class cpu | 17:42 |
TristamWrk | Thanks | 17:43 |
TristamWrk | Weird... | 17:43 |
TristamWrk | [ 0.000000] SMP: Allowing 6 CPUs, 0 hotplug CPUs | 17:43 |
TristamWrk | [ 0.018996] SMP disabled | 17:43 |
mistigi | are you passing strange parameters to kernel in grub ? | 17:43 |
TristamWrk | mistigi, I didn't change grub from what the installer setup | 17:43 |
mistigi | ok, but SMP should not be disabled afaik, one sec | 17:43 |
sarnold | TristamWrk: what does /proc/cmdline look like? | 17:44 |
TristamWrk | Grub has this: GRUB_CMDLINE_LINUX="acpi=off noapic nolapic nodmraid" | 17:44 |
TristamWrk | BOOT_IMAGE=/vmlinuz-3.5.0-26-generic root=/dev/mapper/fs1-root ro acpi=off noapic nolapic nodmraid | 17:44 |
mistigi | yeah, remove acpi=off and reboot | 17:44 |
sarnold | hunh. I think ACPI is required to do SMP. | 17:45 |
TristamWrk | I didn't even think about that. I had to turn stuff off to get the install CD to boot. Didn't even think about that carrying through. | 17:45 |
TristamWrk | Still no joy | 17:49 |
TristamWrk | Still has "noapic nolapic nodmraid" on the kernel line. | 17:49 |
mistigi | yeah, they may be problem too | 17:50 |
mistigi | remove them and give it a try, especially nolapic | 17:50 |
TristamWrk | Yeah, just found a line near the "SMP Disabled" line that mentioned lapoc | 17:52 |
TristamWrk | lapic | 17:52 |
histo | TristamWrk: after you remove the stuff you have to update-grub | 17:53 |
histo | TristamWrk: or add things | 17:53 |
sarnold | I'd just fiddle with the command line in grub at boot time until you get sometihng that works reliably | 17:53 |
TristamWrk | histo, I did, thanks | 17:54 |
TristamWrk | Yay! 6 cores! | 17:54 |
TristamWrk | Thanks folks | 17:54 |
sarnold | mistigi: nice work :) | 17:54 |
mistigi | gratz :-) | 17:54 |
TristamWrk | Sorry I didn't look close enough at dmesg and pestered y'all | 17:54 |
TristamWrk | NOW rsync is maxing out my 100Mbit link...haha | 17:55 |
sarnold | haha :) | 17:55 |
stgraber | sarnold: hey, so /etc/dnsmasq.d-available/lxc is a symlink on your system? | 18:17 |
sarnold | stgraber: yes lrwxrwxrwx 1 root root 28 Feb 28 16:19 /etc/dnsmasq.d-available/lxc -> /etc/dnsmasq.d-available/lxc | 18:17 |
stgraber | sarnold: wow, that's clearly very wrong... have you been only using the packages from the archive or were you on some of the PPA builds for a while? | 18:18 |
sarnold | stgraber: I had the juju ppa enabled previously, but no lxc-specific ppas | 18:18 |
stgraber | sarnold: it sounds like the migration script we pushed a couple of releases ago, moving /etc/dnsmasq.d/lxc to /etc/dnsmasq.d-available/lxc backfired for you and created that mess | 18:18 |
stgraber | sarnold: did you have lxc on that machine pre-raring? | 18:19 |
stgraber | sarnold: btw, if you just want to fix it, remove /etc/dnsmasq.d-available/lxc and make it a regular file containing http://paste.ubuntu.com/5628886/ | 18:20 |
sarnold | stgraber: yes, precise, quantal, and now raring.. | 18:20 |
sarnold | stgraber: oh, hey, dpkg left behind the lxc.dpkg-new with those contents :) | 18:21 |
stgraber | ok, I tested the new package on precise and didn't see that problem, so I'll just run a quantal => raring upgrade test to make sure that's not the trigger | 18:21 |
stgraber | I've re-checked the logic in the current maintainer scripts and can't find anything wrong, but it's possible that an earlier upload triggered this | 18:21 |
sarnold | hehe :/ | 18:22 |
zul | stgraber: ping | 19:10 |
stgraber | zul: pong | 19:13 |
hallyn | stgraber: i asked zul to ping you on a name for the libvirt liblxc driver :) | 19:13 |
hallyn | he's going with liblxc i think | 19:13 |
hallyn | i would've probably done lxc2 or lxcext, but liblxc actually makes sense i think | 19:14 |
stgraber | liblxc sounds good, because, well, that's what it's ;) | 19:14 |
stgraber | yeah, lxc2 may sound like some kind of new release of libvirt-lxc which it clearly wouldn't be | 19:14 |
Catbuntu | Hi | 19:21 |
=== highvolt1ge is now known as highvoltage | ||
mikehale | smoser: Hi, I was wondering if you had any thoughts on an implementation strategy for https://bugs.launchpad.net/cloud-init/+bug/1153626/. | 19:52 |
uvirtbot | Launchpad bug 1153626 in cloud-init "Multiple Interfaces and IPs not detected in AWS VPC" [Medium,Triaged] | 19:52 |
mikehale | I had some thoughts, but I'm not sure if I'm headed in the right direction or not. | 19:52 |
smoser | mikehale, oh. hey, sorry for not having seen that. :-( | 19:52 |
smoser | (please feel free to ping me like this) | 19:53 |
mikehale | cool, yeah @flacoste said I should :) | 19:53 |
smoser | i'm not actually sure what i think about updating /etc/network/interfaces. | 19:55 |
smoser | that does make sense, as that is where people would expect to go making changes. | 19:55 |
smoser | but then you have to deal with removing those entries on interface removal. | 19:55 |
smoser | (and also, cloud-init can be a bit of a PITA here... | 19:56 |
smoser | it blocks boot waiting for any interface listed in /etc/network/interfaces to come up) | 19:56 |
mikehale | does cloud-init handle the initial creation of /etc/network/interfaces? | 19:56 |
smoser | no. | 19:57 |
smoser | there is a hard coded minimal one in the images. | 19:57 |
mikehale | ah | 19:57 |
smoser | actually, adding entries to /etc/network/interfaces would be OK as long as they're set to 'manual' | 19:58 |
smoser | i think i like that. | 19:58 |
smoser | on interface add, have a rule that fires. | 19:58 |
smoser | that rule then tries to figure out if its in ec2, and if it is, gets the data necessary, modifies /etc/network/interfaces and ifup | 19:59 |
mikehale | cool | 19:59 |
smoser | we can even make the code quickly exit doing nothing if a /etc/cloud/do-nothing-your-not-in-ec2 file exists or something | 20:00 |
smoser | and cloud-init would on first boot set that setting. | 20:00 |
mikehale | makes sense | 20:02 |
smoser | mikehale, so for 'manual' all you do is put whatever ocnfig you *would* put there (see man interfaces) except for rather than 'auto eth0' you put 'manual eth0' | 20:02 |
smoser | and that will be perfectly fine to be manual, as it actually be automatic on plug/remove of the interface. | 20:03 |
mikehale | oh I see that is auto/manual are classes and /etc/init/network-interface.conf only attempts to ifup interfaces defined as auto | 20:06 |
lg188 | Good day, I was plugging in my Hard drive (with usb) in my server and it doesn't mount. | 20:31 |
lg188 | this is my dmesg output | 20:31 |
lg188 | http://paste.ubuntu.com/5629262/ | 20:32 |
lg188 | What am I doing wrong? | 20:32 |
RoyK | lg188: an ubuntu server install won't auto-mount an usb disk | 20:49 |
RoyK | s/an usb/a usb/ | 20:49 |
RoyK | :P | 20:49 |
RoyK | lg188: it says sda for the usb device - what does 'cat /proc/partitions' tell you? | 20:50 |
shauno | it does enumerate sda1 on line 17 too | 20:52 |
lg188 | RoyK: Ah it's alreight I had to manually mount it because there was something wrong with being shutdown with windows | 20:53 |
lg188 | And the server uses scisi as root dir if that matters | 20:53 |
RoyK | scisi? what is that? scsi? iscsi? | 20:54 |
lg188 | scsi* | 20:54 |
RoyK | well, if it works, \o/ | 20:54 |
lg188 | Yeah, Thanks for the effort though :) | 20:55 |
ttx | jamespage: Glance milestone-proposed branch was cut | 21:57 |
ttx | jamespage: same for heat | 22:08 |
=== keithzg_ is now known as keithzg | ||
mimizone | imi2000 | 22:48 |
mimizone | Hi in there, general question on the upgrades for 12.04. with the LTS, do I get all the updates that are included in 12.10, 13.x etc… until 2017? are they backported quickly? my main interest is the upgrades for juju, maas and openstack packages. | 22:51 |
sarnold | mimizone: there is a Stable Release Process that is used to backport specific fixes when desired. (There is also the "hardware enablement stack" which will backport the kernel, X, and maybe related packages, but it's a much smaller subset of the distribution than everything...) | 22:54 |
sarnold | mimizone: you may wish to use the juju, maas, and openstack ppas -- at least, I'm pretty sure juju's ppa is intended to also work on 12.04 LTS | 22:54 |
sarnold | (it is my assumption that the other projects have PPAs..) | 22:54 |
mimizone | thanks sarnold. you mean I should use the ppas instead of the ubuntu apt packages right? ppas being the most up to date packages from the developers typically? | 23:00 |
sarnold | mimizone: yes; if you want to follow what is _new_ in the packages, you'd use their PPAs. | 23:01 |
mimizone | sarnold: great. thanks | 23:01 |
island219 | Finally got Ubuntu Server to install on my machine, although I'm still getting "Input Out Of Range" errors on my monitor. Any known fixes? I'm runnin 12.10. | 23:21 |
island219 | Anyone? | 23:34 |
jackads | hello guys, someone can help me in a DNS/mail server configuration? I get a VPS in digitalocean, ubuntu server 12.04 64bits... I installed iRedMail, but I think I had configured wrong the dns or maybe some other configuration, because I can't send or receive e-mail, always returns. Someone would help me? | 23:34 |
=== wedgwood is now known as wedgwood_away | ||
sarnold | island219: back in the day, that meant you set your X modelines to a resolution that the monitor couldn't support | 23:43 |
sarnold | island219: these days of course the edid from the monitor is supposed to prevent out-of-range inputs, but perhaps your monitor doesn't report them properly | 23:43 |
island219 | Hmmm… I reckon it's the graphics in the server. | 23:58 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!