=== freeflying_away is now known as freeflying === peter is now known as Guest71690 [01:42] since #ubuntu has decided to troll instead of help, I"ve got a zombie sh process that I'd like to terminate... how do I do that === zastern_ is now known as zastern [02:00] TheLordOfTime: the zombie's parent needs to wait on it, or the parent needs to die and init will wait on it. But really there's no need. A zombie process is dead as far as everything is concerned. It only remains so that the parent can observe its demise without a race. [02:01] rbasak, so there's nothing bad about letting a defunct sh process sit there... [02:01] (I've reported to upstream of the parent program that the system isn't grabbing the information about the process's death) [02:02] TheLordOfTime: not really. Except that its parent should wait on it. Depending on what the parent does, there's a risk that it'll fill up the process table with zombie processes. [02:03] TheLordOfTime: OTOH if the parent will eventually wait on it (eg. when it needs to start another child), then there's not really a problem, apart from slight sysadmin annoyance. [02:03] Of course it could indicate a different bug in the parent. [02:04] i doubt the parent process is going to wait on it [02:04] ZNC isn't nice that way when it spawns sh [02:04] :/ [02:04] unfortunately the system's got upwards of 75 users on it... [02:04] so i can't exactly restart the parent process [02:10] Sometimes a parent wants to background child processes, and isn't set up to be able to wait on a child. Eg. Python's subprocess module, which IIRC can only wait on backgrounded processes when a new process is started because of the way it's arranged. [02:11] yeah, well, i'm going to just chmod 000 the module that spawns those `sh` processes anyways.. [02:11] that way i don't have it sitting there lagging out the parent process [02:11] (which is kinda what's happening) [02:11] i've got to install the kernel updates anyways, and maintenance reboot... [02:12] that should not only terminate the zombies with cybernetic oblivion... it should also fix everything. [03:14] ello === freeflying is now known as freeflying_away [04:18] getting this with Xubuntu alternate CD 12.04lts [04:18] http://imgur.com/19Gkzfr [04:18] can anyone help [04:20] trying to build a LTSP environment === _thumper_ is now known as thumper === DavidDuffey is now known as dduffey === Smark[Gone] is now known as Smark === Tzunamii_ is now known as Tzunamii === freeflying_away is now known as freeflying [05:38] Hi, I'm trying to connect to a VPN using nm-applet, but everytime I try to connect it says "The VPN connection 'vpn name' failed because the connection attempt timed out". An hour ago I had no problems, but now I can't connect to any of my servers === Sling_ is now known as Sling === PaulePan1er is now known as PaulePanter === smb` is now known as smb === vila is now known as vila-afk-biab [09:28] jamespage: https://code.launchpad.net/~soren/python-jenkins/add-crumb/+merge/188155 [09:28] soren, ooo [09:28] * jamespage looks === freeflying is now known as freeflying_away [09:28] jamespage: I have a hunch that might disappear in the noise of the scores and scores of mp's I'm sure you have to look at each day. [09:29] soren, :-) [09:31] soren, OK - I'm mid deployment of an openstack havana test environment right now [09:31] I'll take a look PM today [09:34] jamespage: Cool. [09:58] hey guys I have migrated 2 websites of mine to a new server and im getting an error saying the following Your PHP installation appears to be missing the MySQL extension which is required by WordPress. I have installed the following php extension php5-mysql [09:58] im a bit stumped as to what else i would be missing do i need the native msql driver? === vila-afk-biab is now known as vila [10:08] hallyn: still in europe? [10:10] cause if you are, https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1233075 may entertain you. [10:10] Launchpad bug 1233075 in linux "resizefs failure with raring" [Undecided,New] [10:13] lifeless, It would help a lot if you had also the top of the stack trace which includes the exact kernel version [10:14] smb: I would love to have it [10:14] smb: I'm obtaining this from an ilo textcons, which is 80x25 only [10:15] smb: as it happens in early userspace I can't even attempt a login - there is no getty yet [10:15] nor even the fallback logic glue for when root fails to mount [10:15] lifeless, Ok, you probably could manually mount the cloud image you use and check there. [10:16] https://lkml.org/lkml/2013/3/19/554 [10:16] looks very similar [10:17] Yeah at least flexbg stuff as well and probably something occuring in upstream stable too. [10:17] smb: I'm not sure what you mean by manually mount the cloud image and check.. [10:17] Thats why I would be interested to know which kernel version of us this happens on and whether maybe there is already something int the pipes [10:18] smb: oh, you want the kernel version? [10:18] yes [10:18] linux-image-generic, latest as of a few hours ago [10:18] when I built the image; apt-get updated in there [10:18] ok [10:18] I'm just building a quantal image now to see if it happens there [10:19] I had a bunch of machines deploying perfectly with quantal 4 months back or so [10:19] from memory [10:19] but we've twiddled our fs creation code since then too... [10:20] Ah, hm, so it may or may not be present in original raring or have been caused by a stable change. we would not know [10:21] yup [10:22] or it might be present in Q as well and depend on the ext4 fs we're creating. [10:22] That I'll know in a little bit === shauno_ is now known as shauno [11:23] smb: quantal worked fine. [11:25] lifeless, ok, so somewhere between that and now... would you have the option of trying an older raring kernel in the image? [11:26] 3.8.0-19.29 would have been on the release and 3.8.0-31.46 is in theory the current one [11:29] smb: notnow, but toss it in the bug and I'll see about us doing some track-down for you [11:30] lifeless, ok will do as well as probably asking for latest saucy (to check whether it might have been fixed but just not backported) [12:09] ok, if im doing a hostname schema like http://foo.bar.domain.com/, should i make my /etc/hostname entries foo.bar? [12:09] oops, li meant like foo.bar.domain.com [12:10] jamespage: lovely the neutron tests are timing ut [12:14] MACscr: no, the host is "foo" [12:15] zul, \o/ [12:16] mardraum: should the domain in /etc/resolv.conf be just domain.com? seems to be .net.domain.com because my routers hostname is fw1.net.domain.com [12:17] if you mean the "search" parameter, it can be anything you like. but it is most useful probably in you sub-domain. It's also entirely optional and won't break things to not have it. [12:18] unless you rely on referring to local machines by host only, which I would not advise you do usually [12:22] mardraum: these hostnames will be used only internally, so just trying to make sure i am doing it right. guess i have no option but to manually create a dnsmasq entry for each server so they can communicate with each other by their fqdn [12:23] usually, you would just setup your dns zone propery [12:23] properly [12:23] you shouldn't be manually adding things on servers, if that's what you mean? [12:24] im talking about dnsmasq on the pfsense system (aka, the router) [12:25] these systems will only have internal ip addresses [12:25] I have no idea what you are talking about then [12:25] internal? as in rfc1918? that does not preclude you getting DNS sorted out in any way. [12:26] you can even do it with the same DNS server if you wanted via zones [12:26] personally I like to keep them separate [12:29] well for internal dns, i am thinking a simple dns forwarder (dnsmasq) on the gateway is the best route to go. Which means that all the servers have 192.168.0.1 as their dns entry, then it queries the dnsmasq entries before it queries the public dns servers [12:52] I am installing ubuntu-server 13.04 in expert mode. Any real difference between choosing the kernel linux-generic and linux-server? the main purpose of the machine is BOINC (basically, there will be a GUI and several math threads at 100% cpu running all the time). [12:54] I ask this because I already googled, but I found out that "Since 12.04, there is no difference between the Desktop linux-generic and Server linux-server kernels; they have been merged.". I am confused. [13:21] z302, why ask if you know the answer? [13:25] hi! I'm setting up a Jenkins server. I'd like to bind it to port 80. Is it a good idea? === grn_ is now known as grn [13:27] grn: why would it be good/bad ? [13:28] ikonia: is exposing an application server directly w/o a reverse proxy a good idea? [13:29] jamespage/roaksoax/adam_g: https://code.launchpad.net/~zulcss/ceilometer/havana-alarms-ftbfs/+merge/188331 [13:29] grn: it can be good/bad, what are you worried about ? [13:30] ikonia: I have no particular worries, just asking for general advice. The instance is going to be accessible from the Internet if that matters. [13:31] grn: there can be pros/cons [13:31] it's up to you to work out if they met your needs [13:32] patdk-lap, I asked because the installer still prompted me [13:47] adam_g, hey - I updated the swift-storage redux branch to bind to 0.0.0.0 [13:47] if was trying to bind to 'private-address' [13:48] not work so well on maas [14:16] How does the MOTD system work on 12.04? I have a bunch of systems (originally booted off the cloud image) that are not updating the MOTD. [14:45] adam_g, how do you feel about having an 'external-network' configuration item on the nova-cloud-controller for neutron? [14:45] before it was configured when the neutron networks where configured [14:45] but that's been dropped [14:45] but it sits in /etc/nova/nova.conf so really need to be injected somehow [14:46] has anyone seen this: [14:46] https://bugs.launchpad.net/neutron/+bug/1210121 [14:46] Launchpad bug 1210121 in neutron "Havana staging:ERROR neutron.common.legacy [-] Skipping unknown group key: firewall_driver" [Low,In progress] [14:47] Hello, recently I purged an important package (don't ask why). It removed many dependencies (A LOT). I connected to the internet through ethernet because it purged network-manager. Now I installed network-manager again, but when I try to install the important package it doesn't install any of the dependencies I deleted. Is there a way to reinstall them without manually doing it by copying the purged packages one by one? [14:47] justanotheruser: you can extract removed packages from the log and then reinstall them [14:48] justanotheruser: /var/log/apt/history.log [14:53] ivoks: thanks [14:55] Hi. Running Ubuntu 12.04 on a server rented at a hosting provider (Hetzner), I cannot use USB or other ports like graphics. Do you folks disable such things for fewer points of failure or just leave it? [14:58] PaulePanter: i have ubuntu servers in hetzner [14:58] PaulePanter: it runs normal ubuntu kernel, therefore nothing is disabled [14:59] PaulePanter: and i can see usb ports on my servers [14:59] zul, how much packaging related change do we have across the openstack packages this cycle? it would be good to get that uploaded asap [15:00] so as we move towards release its just upstream only changes [15:00] jamespage: like patch changes? [15:00] zul, I was mainly thinking about dependencies [15:01] jamespage: well all of the dependencies are up to date [15:01] zul: btw I'm just uploading mariadb to debian :-) [15:01] jamespage: \o/ [15:01] will we sync it? [15:01] zul, in the packages in the archive? or just in the branches? [15:01] ivoks, depends how quickly it gets through the NEW queue [15:01] things like oslo.config, etc we are up to date [15:01] ivoks, but TBH is not mega critical for this release [15:01] Another option would be to build a Linux kernel, with not-needed parts disabled. This would even boot faster and the image from disk is smaller. [15:01] more for next [15:02] zul, you did a load of updates to control files to add versioned deps right? is that in archive yet? [15:03] ivoks: Yes. But as you are not in their datacenter you are not able to use the USB ports. So it would be better to disable that. [15:03] smoser, utlemming: in between precise and quantal, ubuntu was removed from the admin group in our cloud images. Was this intentional, and if can you point me to the rationale, please? It stops ubuntu from being added to the libvirtd group by the libvirt-bin postinst. [15:03] PaulePanter: i wouldn't be so sure about that... [15:03] jamespage: not yet it will be when we upload next, but the dependencies changes that i did to the control file we have the latest [15:03] PaulePanter: when you ask for their intervention, they might be plugging in usb keyboard [15:04] rbasak: the admin group was dropped at precise, if I recall correctly [15:04] PaulePanter: anyway, most of ubuntu kernel is modular; you can just re-create your initrd [15:04] utlemming: it was replaced by the sudo group, but the admin group still exists. [15:04] zul, OK - I think I understand [15:05] jamespage: we should be good :) [15:05] (in our cloud images) [15:05] rbasak, between precise and quantal the admin group ceased to be recommended i think. [15:05] replaced by the 'sudo' group. [15:05] to which the user should be a member [15:05] smoser: right, but we used to add ubuntu to the admin group, and now we don't add ubuntu to the sudo group. [15:05] rbasak: https://bugs.launchpad.net/ubuntu/+source/policykit-1/+bug/893842 [15:05] Launchpad bug 893842 in policykit-1 "Move "admin" group to "sudo"" [Critical,Fix released] [15:05] PaulePanter: kernel is less than 5MB, but initramfs can go up to 20MB [15:06] rbasak: checking on the latest cloud image, we do add ubuntu to the sudo group [15:06] rbasak: that is for Quantal [15:06] utlemming: I just did: wget http://cloud-images.ubuntu.com/quantal/current/quantal-server-cloudimg-amd64-root.tar.gz; tar xzf quantal-server-cloudimg-amd64-root.tar.gz etc/group; egrep admin\|sudo etc/group [15:06] sudo:x:27: [15:06] admin:x:111: [15:07] rbasak: right [15:07] rbasak: on quantal the user is created at first boot [15:07] rbasak: the ubuntu user does not exist in the tarball [15:07] adam_g, I just spotted the changes to the cloud-archive reports [15:07] nice [15:07] utlemming: OK, I'll check by starting instances. But in my test saucy image, ubuntu does not seem to get added to sudo. [15:08] (upon login) [15:08] rbasak: yeah, so that is definately a bug [15:09] rbasak: care to fill one out for it? [15:09] In cloud-init then I guess, if it is supposed to happen on first boot? I'll investigate - thanks. [15:09] Will do. [15:09] Thanks for your help. [15:09] rbasak: I found the problem [15:10] rbasak: I'll get smoser a MP on it [15:11] ivoks: Yes. That initramfs is another thing to optimize. Probably not needed at all on a server, which is not encrypted and where you have static components and now what you run. [15:11] utlemming: thanks! Bug 1228228. [15:11] Launchpad bug 1228228 in uvtool "ubuntu is not added to sudo group" [High,New] https://launchpad.net/bugs/1228228 [15:12] PaulePanter: there's nothing to optimize, you can decide what to put in it [15:12] rbasak, /etc/cloud/cloud.cfg lists the groups that the default user will be added to [15:12] and there is no 'sudo' there. [15:12] does libvirt copy users in the sudo group to the libvirt group ? [15:12] PaulePanter: /etc/initramfs-tools/initramfs.conf [15:12] smoser: yeah, mp coming your way for that [15:12] smoser: yes - it copies both admin and sudo members [15:13] rbasak, your tool should address the case where the user is not in the sudo group. [15:13] ie, you should have a sane path for "you do not have libvirt group access" [15:13] smoser: you mean a test and a sensible error message, or something more? [15:13] smoser: shouldn't the default user be in the sudo group? [15:13] PaulePanter: distribution can not ship different kernels for different machines; that's why there's generic kernel + initramfs [15:14] PaulePanter: if you want a kernel that's specific for your hardware, you have to do it yourself [15:14] rbasak, i don think you shoudl do anything more than a sane error message: [15:14] ack. [15:14] could not connect to qemu:///system . You need to be lin the libvirt group. sudo adduser $USERNAME libvirtd [15:14] somethign to that affect [15:15] also, kind of sucks but then after that you hvae to log out and back in or use 'sg' to get into that group. [15:15] utlemming, its really a "meh" whether or not they should be in sudo group. [15:15] does 'd-i' add the first user to sudo ? [15:15] if so, then i'd agree to following that ocnvention [15:16] smoser: checking.... [15:19] smoser: It does. [15:19] utlemming: ^ [15:19] http://bazaar.launchpad.net/~ubuntu-installer/user-setup/master/view/head:/user-setup-apply#L164 [15:19] soren: ack, I thought it did...just didn't have the code reference [15:20] ivoks: Yes. [15:20] soren: thank you kindly [15:21] ivoks: So blacklisting the modules is the only thing to do, when you want to stay with the default Linux image. [15:21] PaulePanter: 'default linux image' doesn't exist [15:21] PaulePanter: ubuntu kernel without modules is almost non-functional [15:24] ivoks: »default Linux image« meant the image shipped by Ubuntu. [15:24] PaulePanter: kernel modules are part of the ubuntu operating system [15:25] it's like linux kernel works [15:25] if you disable modules on a server, you lose: kvm, iscsi, parport, network card drivers, etc, etc... [15:26] i think even sata is lost then [15:26] ivoks: Surely I would only blacklist the modules, which I do not need. [15:26] if you've never done that... i would strongly advise not to do it on production server :) [15:27] that's a famous sentence all of us said once [15:27] and only for 1-2 seconds faster boot [15:29] What would prevent motd from updating… but then start working after I ran run-parts on it manually? === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [16:15] jamespage, you dont have anything for cloud-archive that checks ubuntu -> staging, do you ? [16:16] smoser: http://reqorts.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/ [16:16] we should update that to compare ubuntu->staging for cloud-tools [16:16] is there code that generated that ? [16:16] lp:ubuntu-reports I think - adam_g would know definatively [16:17] k. yeah, i'd like to have 'ubuntu' 'next' 'staging' 'proposed' 'updates' coluns [16:17] and really... [16:17] i just want a tool like cloud_do_work that does it. [16:21] smoser: just testing the cloud-tools pocket uvtool for the first time. [16:21] smoser: uvtool-libvirt : Depends: python-simplestreams but it is not going to be installed [16:21] smoser: python-simplestreams : Depends: python-glanceclient but it is not installable Depends: python-swiftclient but it is not installable [16:21] smoser: we need to drop those down to Recommends I think? [16:22] rbasak, well dropping to recommends isn't a fix. [16:22] is it ? [16:22] smoser: that functionality is optional, though, isn't it? [16:23] well, if you consider code that just doesn't have to run "optional". [16:23] i think maybe just do python-simplestreams-openstack package . [16:23] that was my intent, but i forgot to do that. [16:24] That seems reasonable [16:24] rbasak, of course i dont know how we're ogin gto handle that... [16:24] can you think of a way ? [16:24] Ah I get your use case now. You want to be able to create a package that depends on the openstack functionality provided by python-simplestreams. [16:25] hm.. [16:26] if we just put the openstack stuff into a binary package, that binary package will not be installalbe on precise. but, oh well. [16:27] That's fine by the cloud-tools pocket's defintion through, right? [16:28] smoser: about how to do it. Must we separate the code? What if python-simplestreams-openstack were a simple metapackage with the appropriate Depends? [16:28] smoser: packages that require simplestreams openstack packages can depend on it. Packages that don't need it depend on python-simplestreams. [16:29] And the code can move in the future if necessary. [16:34] yeah, its fine for cloud-tools pocket. [16:35] rbasak, well, i already cut the files out of the package for python3 [16:35] (as none of that stuff will run, so i dont ship the files that wouldnt run) [16:36] smoser: I filed bug 1233269. [16:36] Launchpad bug 1233269 in cloud-archive "python-simplestreams in cloud-tools pocket is broken" [Undecided,New] https://launchpad.net/bugs/1233269 [16:37] smoser: I'm not familiar enough with Python packaging to understand the interaction between Python module packages and files and stuff :-/ [16:38] rbasak, http://paste.ubuntu.com/6176232/ [16:39] well, we'll see. i'll try tof ix. [16:40] rbasak, i think i like your idea of the metapackage though. [16:41] roaksoax: around? [16:41] zul, he's out till manana [16:41] smoser: arrgh [16:41] smoser: can you have a quick look https://code.launchpad.net/~zulcss/ceilometer/havana-alarms-ftbfs === jdstrand_ is now known as jdstrand === psivaa is now known as psivaa-afk [18:12] hi [18:12] kernel 3.11 is only for ubuntu saucy right ? [19:41] Hi guys. I'm trying to install alsa and the snd kernel module on ubuntu but running into some trouble with getting that module. [19:42] I tried this guide: https://wiki.ubuntu.com/Audio/InstallingLinuxAlsaDriverModules . But on the install step, the paackage isnt found. [19:57] zul, https://code.launchpad.net/~zulcss/ceilometer/havana-alarms-ftbfs/+merge/188331 [19:58] th eonly thing that looks curious is the oslo.config dependency version change [19:58] is/was there a reason for that? [20:04] zul, ^^ [20:10] any ideas why motd would fail to update on login? as far as I can tell, pam_motd is properly configured in both sshd and login pam files [20:10] smoser: yeah it requires a newer version of oslo.config [20:13] actually, it looks like it's updating it… but displaying the previous version [20:14] zul, approved. [20:15] sorry that took so long. [20:15] smoser: thanks no worries === gary_poster is now known as gary_poster|away === gary_poster|away is now known as gary_poster [22:24] LTSP Question [22:24] Xubuntu 13.04 Host OS and LTSP server, 1 wifi card connected to Internet. Virtual Box installed with Thin Client Guest OS. Is it possible to have Thin Client boot-up with this LTSP server plz? [22:25] just a qq. Someone install apache mpm-event with phpfpm on my server the other day; just tested it a few minsa ago and realised that php pages aren't being intepreted [22:26] all googling points to the php5 module not being loaded, but the guy says that's not needed. Is this true?