=== Ursinha is now known as Ursinha-afk [00:24] i'm running tail -f 'file.log' | grep --line-buffer 'query' > output.txt on an nginx log file that gets written about 3.4 mil lines a day, would that consume any memory? i'm looking at top right now and it doesn't look like it's consuming memory as far as i can tell [00:25] dougb: I'd expect both to consume less than one megabyte of RSS [00:26] ok great, thank you === gary_poster is now known as gary_poster|away === freeflying_away is now known as freeflying === Ursinha-afk is now known as Ursinha === freeflying is now known as freeflying_away === freeflying_away is now known as freeflying [04:48] Out of curiosity, has anyone here used BloomIP as a hosting provider? [04:49] jamespage: hello there === thumper is now known as thumper-afk [07:27] good morning everyone. I've installed some updates on one of our servers this morning and two things did happen after a reboot: naming of my second NIC did change from eth1 to eth2 (did have to delete /etc/udev/rules.d/70-persistent-net.rules) and I did get a message for EDAC at boot (first time ever) that /usr/sbin/edac-ctl was missing, installed edac-utils, but there is still a message at boot: [07:27] Not enabling Memory Error Detection and Correction since EDAC_DRIVER is not set [07:28] edac-ctl --status states that the drivers are loaded and I can find the driver (i7core_edac) per lsmod. Does that mean that I can ignore this message because the drivers were loaded anyway? [08:50] Hi there. Can somebody please clarify what will happen in my case if I run 'sudo do-release-upgrade'? http://paste.debian.net/46339/ Apparently my ubuntu server is up to date === freeflying is now known as freeflying_away === thumper-afk is now known as thumper === bitnumus_ is now known as Guest76595 [09:52] hi, I can't ssh from ubuntu server to ubuntu desktop, all 12.04 === Guest76595 is now known as bitnumus [10:01] quietone: does the ubuntu desktop has the SSH server (openssh-server) installed? IIRC the desktop has only the client installed by default [10:03] that's correct, unless you chose it during installation [10:05] or install it later, obviously === gary_poster|away is now known as gary_poster [11:06] adam_g, a load of the charms have grown exec.d support - but I don't think we have that implemented [11:06] I think there is a helper for that [12:10] Looking for help and info on Landscape service [12:11] HI I would like to download a binary ubuntu package and its dependencies to my server, but not installed, how could I do that? thx [12:11] apt-get whatever ? [12:12] Hey jacalvo [12:21] rostam: "sudo apt-get -d install $package" (but it only works for packages/dependencies which you doesn't have already installed) [12:22] geser: thank you so much [12:59] anyone have thoughts on this ? [12:59] https://wiki.ubuntu.com/ServerTeam/CloudArchive [12:59] that (and cloud-tools archive) are more complex than ideal to enable. [13:00] any thoughts on a way to make that easier ? [13:07] smoser, add something to python-software-properties - thats alwaysbeen on the plan [13:07] but we #failed at that [13:08] smoser, that incorrect support diagram is back again [13:08] folsom, grizzly and havana are not supported until the end of 12.04 support [13:09] hi, i did apt-get upgrade [13:10] and udev did not upgrade properly, because the post install script hangs without error message [13:10] i tried setting the package to hold [13:10] hello i want that my ubuntu machine is able to send out email via an external smtp server how should this be done ? [13:10] but it does not work, because the package is partially installed [13:11] how do i fix that? [13:11] ws2k33, apt-get install postfix, select to use a smarthost, put your external SMTP server as the smarthost [13:12] jamespage, nice. [13:12] jamespage, so should we do the python-software-properties thing? [13:12] smoser, something like 'add-apt-repository cloud:grizzly' [13:12] the wierd thing is we'd add it in saucy [13:12] but it sin't really saucy related. [13:12] smoser, yeah - we would need to backport it to precise [13:12] can someone tell me how i could fix the partially installed udev package? [13:12] and it would be *wrong* on saucy, jamespage [13:13] smoser, yes - but that could be included in the code [13:13] we'd be adding broken / incorrect code :) [13:13] that just seems weird [13:13] smoser, check for 12.04 and warning message/error out [13:13] qman thanks you [13:14] jamespage, maybe i will do that today. [13:16] smoser, add-apt-repository cloud:tools ? [13:16] ... [13:17] maybe cloud:openstack-grizzly and cloud:tools ? [13:19] smoser, I'd like to keep it consistent with the charm openstack-origin stuff [13:19] smoser, which is currently cloud:precise-grizzly [13:19] but the precise- is not relevant (that should be fixed during the redux) [13:20] hm.. [13:20] i'm just trying to think of logic for that. to "know" without a hard coded list if the release is good. [13:20] so we dont have to SRU for 'j' name. [13:21] maybe i just try to allow 'cloud:j*' and try. [13:21] :) [13:27] smoser, while you are here [13:28] smoser, I'm seeing update-apt-xapian-index running on first boot of cloud-images [13:28] smoser, which is chewing a fair bit of memory and causing some juju/charm related issue [13:28] jamespage, https://bugs.launchpad.net/ubuntu/+bug/1227425 [13:29] Launchpad bug 1227425 in ubuntu "Cloud images do not need apt-xapian-index" [Undecided,New] [13:29] i dont kno when that got into the images. [13:29] smoser, maybe sometime recently as this was not problem until sometime in the last 3 weeks [13:31] jamespage, its been in images since 12.04 [13:32] smoser,trying to figure out why is just started to cause problems [13:32] so something else must have changed that made it decide to rupdate itself. [13:32] smoser, bootstrapping on a m1.tiny now fails [13:32] as i said in that bug, i'm not opposed to dropping tasksel [13:32] so does deploying mysql on a m1.small [13:32] with default settings... [13:35] jamespage, dropping tasksel would drop aptitude [13:35] * jamespage shrugs [13:35] whats aptitude? :-) [13:36] and aptitude is what Recommends apt-xapian-index. [13:36] i agree with your feelings on aptitude [13:36] but i often see people document [13:36] aptitude safe-upgrade [13:36] or the like [13:36] Hi, I've got Ubuntu 12.04 with Nginx. However, css isn't loaded for some reason. Any idea why? [13:36] so i'm somewhat weary of dropping that from a cloud iamge. [13:39] In other words, the html of my site loads, but not the css. And I've tested the site on a development server, with no problems whatsoever... [13:40] smoser, yeah - I'm wary as well tbh [13:40] Dry_Lips: in my experience, you'll have to ignore the cookie header: http://paste.ubuntu.com/6158741/ [13:43] pesky, I don't use cookies, I've only got .js = adobe edge webfonts... What file was that paste from, btw? [13:44] Dry_Lips: from one of our own servers, we have a CDN [13:45] Dry_Lips: are you really sure you don't use cookies (checked response header)? [13:46] pesky, this is a simple "page under construction" site... There is one java script in order to load the Adobe font, but that's it... [13:46] Dry_Lips, a simple google for "nginx css" reveals others having issues defining the mime types [13:46] yeah, I saw that too qman___ [13:47] But I checked the mime.types, and I didn't see any suspicious [13:48] I don't use nginx myself, it just seems to fit, and keep in mind that text/css is a different mimetype from text/html [14:13] jamespage: https://code.launchpad.net/~zulcss/nova/lxc/+merge/187796 [14:13] another question: when i log in to ssh, it prompts me immediately for my username [14:13] but after pressing enter [14:13] it takes more than 10 seconds [14:14] to ask for my password [14:14] after that it is fast again [14:14] what does take so long? [14:14] jamespage, if you want to dig on that bug, that'd be good. i'm really very ok with not having that silly index thing in the images. [14:32] zotta, the openssh client doesn't prompt for a username; putty prompts for a username before it tries to connect [14:33] the delay is the encryption handshake taking place, the username being submitted, and waiting on the authentication challenge [14:34] it can happen if you have a poor network connection or if the server's CPU is slow or overloaded, or if something is wrong with your authentication system server-side [14:39] zotta: DNS timeout, while looking up the ptr record of the ip you are connecting from? [14:53] jamespage/adam_g: https://code.launchpad.net/~zulcss/python-keystoneclient/bash-completetion/+merge/187806 [15:50] smoser: I have a cc_write_files issue [15:50] wonder if something has been borked there recently [15:50] try to create a file, and set owner to ubuntu:ubuntu [15:50] and the cloudinit fails with: [15:51] OSError: Unknown user or group: 'getpwnam(): name not found: ubuntu' [15:51] mgz: which version of ubuntu? [15:52] I note that cc_write_files happens ridiculously early (before key generation even), and before cc_users_groups, but thought ubuntu should exist in the image anyway [15:52] oh. bugger. [15:52] utlemming: tried both raring, image from this month, and saucy beta1 [15:52] mgz, it makes sense. [15:52] there is no 'ubuntu:ubuntu' at that point [15:52] should cc_write_files just be bumped a bit later, to after cc_users_groups? [15:52] write_files runs very early. before the user is created. [15:53] well if it runs later, it can affect less. [15:53] smoser: I swear this used to work :) [15:53] it works on 12.04 [15:53] because ubuntu is builtin to the image [15:53] shall I bug a file? [15:53] i cant imagine how i'd fix it. [15:54] split cc_write_files in two, defer any with owner till after user creation? [15:54] adds a nice annyoing level of complexity [15:54] its not that easy [15:54] write_files runs as part of the init portion [15:54] while create users runs a cloud-config module [15:55] they both run as cloud-init-modules [15:55] mgz, you can actually chnage the order of those yourself . [15:55] oh you're right... [15:55] or just add another 'write_files' afterwards. [15:55] :) [15:55] i dont have a good solutoin. [15:56] you have a sane use case. [15:58] mgz, i'd consider adding a 'write_user_files' module. [15:58] that would seem reasonable [15:59] alright. well file a bug. [15:59] and submit a patch if you'd liek [15:59] filing [15:59] should be pretty simple. [16:04] smoser: filed bug 1231541 [16:04] Launchpad bug 1231541 in cloud-init "Using owner with write_files in cloud config breaks" [Undecided,New] https://launchpad.net/bugs/1231541 [16:36] Hi I am trying to download a binary package, xorg, with its dependencies, the command "apt-get -d install xorg" looks to my local system and since the package is already installed it does nothing. Is there a way I can work around this please? thx [16:44] zul, has your lxc driver been pitched upstream yet? [16:44] adam_g: not yet its way past feature freeze im going to pitch it for icehouse [16:53] rostam, apt-get install --reinstall --download-only xog [16:53] smoser, thank you so much. [16:56] smoser, how could i issue the command that it install the packages to the current directory? thx [16:58] I Need to Setup a Chat server in Ubuntu server , im planning to Buy a VPS For that , Which Chat Software i can Choose and It want to be a Open Source .... thanking u all [16:58] "install into current directory" ? [16:58] what ever .... [16:59] for viewer it want to be like a messenger [16:59] same as yahoo [17:00] smoser, the apt-get command install the packages to the /var/cache/apt/archive I want them to install in different directory. That what I meant by current. [17:04] rostam: That default value for that is Dir::Cache="var/cache/apt/" if you want to change it, you need to add to the apt-get command an option like: -o Dir::Cache="/full/path/to/where/you/want" [17:08] genii, I think I made an error, it gives me error: E Unable to locate package install. Here is what I issued: [17:08] apt-get install -o Dir::Cache="/home/Work/CarboniteDepo" install --download ffmpeg [17:09] this seems to work [17:09] http://paste.ubuntu.com/6159522/ [17:15] smoser, yes it worked thank you and genii thank you. [17:15] rostam, note, that you dont even have to be root there. [17:15] which is nice. [17:16] smoser, yes and thx again [17:18] Are there any known issues with running squid 3.3 on ubuntu 12.04 server lts? The distro wont pull anything over 3.1 through apt for me. [17:20] bugzc: you could try compiling the saucy package on 12.04 LTS, or requesting a backport, but you'd have to be the one to keep up with security updates for squid if you do it yourself, and ask for new backports when there are security issues.. [17:21] bugzc: or, you could download and compile the source and skip the packaging system, but that's harder to manage long-term. I'd probably try compiling the saucy package on 12.04 LTS first.. [17:23] Thanks guys. To be honest I just need it stable. It sounds like the version from apt is just that. [17:24] Im using it in a transparent/inline cache proxy. ~1.5TB cache [17:29] zounds, that's a lot of cache :) hehe [17:38] Hi, what TTL value do you guys recommend using? [17:39] that is down to your preference and needs/desires [17:39] it's unique to that person and their setup [17:41] I've never fiddled with it and never seemed to suffer for it :) [17:41] Okay, if I'm in the process of setting up a site, a short TTL would be good? [17:41] sarnold: Yeah, it's for windows/av/etc updates for a small network [17:42] Dry_Lips: as I said, it depends on a a few things, [17:42] bugzc: nice! I bet that helped immensely. I tried doing that once, cheaply, and promptly found out that different hostnames in the update pool really did lead to multiple copies of data... for as tiny as my use was, it wasn't worth it. [17:43] it's not a "setting up new site = low ttl" model, you need to think about what you are doing and the systems you are interfacing with [17:43] It's a small site with static html [17:44] Just in the process of setting it up for the first time [17:44] what has that got to do with it ? [17:44] Dry_Lips: but where are your users? what will their internet links look like? how many IX centers will their packets need to traverse? [17:44] the content has nothing to do with it [17:44] right... [17:44] Dry_Lips: I strongly recommend changing nothing. [17:45] Okay, the default TTL of my registrars is 172816... [17:45] roaksoax: on bug 1231578, stgraber just fixed it i believe [17:45] Launchpad bug 1231578 in dnsmasq "dnsmasq binds bootp on all interfaces preventing other DHCP servers from starting" [Undecided,New] https://launchpad.net/bugs/1231578 [17:45] ah, good, I'll add the bug number to the upload [17:45] oh, DNS TTL. I thought you meant IP TTL. stupid name collisions. :) [17:45] Yep, DNS TTL [17:45] lol [17:46] stgraber: :) [17:46] Dry_Lips: so, uh, yeah, 3600 or something might be fine to start with :) hehe, sheesh, sorry for being completely wrong. heh. [17:46] roaksoax, hallyn: uploaded [17:47] sarnold, no problem, mate! and thanks for the suggestion :) [17:47] thanks [17:50] zul, whast the deal with the keystoneclient patch that hacks around httpretty? [17:51] adam_g: its not in main and didnt have a change to do a MIR, dont think its in the archive either [17:51] sarnold: I've had a proof of concept type setup for a couple years now with a smaller drive running off an ancient p4 ibm box and even that helped quite a bit. but a lot of the stuff i have to cache is for software where you cant configure a proxy manually or its otherwise impractical. a transparent one is the way to go. I hope it works out :) The connection here is very slow.. [17:52] zul, are you planning a new upstream release before saucy? [17:52] (of ksc) [17:52] yeah [17:52] zul, httpretty is used all over the test suite now [17:52] adam_g: yeah i know :( [17:52] not sure how limited its use was when you added the patch [17:52] it was pretty limited [17:53] meh ill look into it [17:53] zul, is there any reason why we need new clients at this point? [17:53] unless they do a version bumb betwen now and release no [17:53] bump even [17:54] ya [17:54] wonder if they are frozen yet [17:54] clients are never frozen [17:54] they arent [17:54] but the requirements of the projects will be at some point [18:09] zul, http://paste.ubuntu.com/6159741/ [18:10] all look to be satisfied currently. if we can get some kind of confirmation that those wont be bumping before havana, we should stop our saucy client builds and consider the current clients we have in saucy to be the versions we ship. [18:10] jamespage, thoughts? [18:11] +1 from me [18:17] zul, that said, it looks like the requirements repo's global-requirements.txt lists python-ceilometerclient>=1.0.5 [18:18] (we have 1.0.3 currently) [18:18] ill do it today [18:18] VersionConflict: (oslo.config 1.2.0a3 (/usr/lib/python2.7/dist-packages), Requirement.parse('oslo.config>=1.2.0')) [18:18] also ^ [18:18] oslo.config is pending beta freeze [18:18] (did that yesterday) [18:18] oh, awesome [18:22] adam_g: I am looking at this bug: https://bugs.launchpad.net/ubuntu/+source/python-quantumclient/+bug/1170849 [18:22] Launchpad bug 1170849 in python-quantumclient "Cliff 1.3.2 is not available " [Medium,Invalid] [18:24] adam_g: how can I get python-cliff and finally nova and quantum clients installed for grizzly on precise? [18:24] I'm slightly confused on the resolution to this [18:24] kurt_, that bug shouldn't be valid. you're hitting it? [18:24] yeah, I can't install cliff [18:25] kurt_, from which grizzly repository? [18:25] I've tried precise-proposed, stable, etc [18:26] kurt_, so python-novaclient and python-quantumclient are uninstallable on precise from the cloud archive grizzly-proposed/staging/updates pockets? [18:26] correct :) [18:26] I haven't tried all pockets [18:27] testing [18:27] kurt_, do you have a system up right now that is affected? [18:27] yup [18:28] kurt_, which versions of the client packages are you trying to pull? [18:29] just one for example is precise-proposed python-novaclient [18:29] http://pastebin.ubuntu.com/6159835/ [18:30] kurt_, http://paste.ubuntu.com/6159834/ [18:31] what's your set up? what am I missing for precise? [18:32] adam_g, +1 [18:32] kurt_, nothing special, just "deb http://ubuntu-cloud.archive.canonical.com/ubuntu precise-updates/grizzly main" [18:32] you are grabbing from precise-updates... [18:33] kurt_, apt-cache policy python-novaclient ? [18:34] http://pastebin.ubuntu.com/6159856/ [18:34] its cacheing from main? [18:35] adam_g: how can I fix that? [18:36] kurt_, i dont know what you've got configured in your sources. you should be able to add a pocket of the cloud archive (paste that deb line into /etc/apt/sources.list), apt-get update, then 'apt-get install python-novaclient' [18:36] you should'nt specify '-t precise-proposed' [18:36] all the dependencies required by anyhint offered in the grizzly cloud archive should be resolvable either in the main ubuntu archive for precise, or the cloud archive pocket [18:36] the only thing I added for various reasons was "deb http://archive.ubuntu.com/ubuntu/ precise-proposed restricted main multiverse universe" [18:38] kurt_, i'd remove the precise-proposed entry unless you're actually testing something from precise-proposed, and add a source entry for the cloud archive [18:38] I am. A lot of the juju stuff I need is there [18:39] ok [18:39] maybe temporarily change it? [18:39] or is there another way to do this? [18:39] no [18:39] setup access to the grizzly cloud archive, and install the clients from there [18:39] any hints on that? a web page RTFM referral is fine - just need a direction to be pointed in [18:40] you mean like good ol' wget? [18:40] kurt_, https://wiki.ubuntu.com/ServerTeam/CloudArchive#How_to_Enable_and_Use [18:40] thnx [18:41] zul, jamespage so it looks like there is a requirements freeze in effect upstream. if we can get ceilometer updated to match what is in global-requirements.txt, lets pull the client builds and pick them up when saucy is out [18:42] ceiloemeterclient right? [18:42] we can focus on getting and keeping the remaining projects building . seem to be some issues to be resolved [18:42] zul, oh yea. [18:43] adam_g: is order critical in the sources.list? and will this affect any of my existing precise-proposed stuff? [18:43] kurt_, ordering doesn't matter. you can even put the line by itself in a file /etc/sources.list.d/ for easy deleting. [18:43] kk [18:44] kurt_, if this is your desktop machine, and all you are pulling is those two clients, you might want to consider removing the entry after you have installed them. there are lots of other packages in the repository that may inadvertently upgrade unrelated stuff [18:45] adam_g: did you have a look through the nova-lxc stuff? [18:45] adam_g: good idea [18:49] zul, briefly. commented on lack of tests and general pep8 stuff. we need test coverage [18:49] adam_g: ack [18:49] zul, and i'd also like to see this deployed and tested with the rest of our CI [18:53] grrr [19:04] adam_g: https://code.launchpad.net/~zulcss/python-ceilometerclient/1.5.0/+merge/187879 [19:14] adam_g: Yargh! LOL http://paste.ubuntu.com/6159997/ Something broken with quantum client [19:16] adam_g: apparently running in to this now: https://bugs.launchpad.net/ubuntu/+source/quantum/+bug/1211764 [19:16] Launchpad bug 1211764 in quantum "Grizzly's python-quantumclient wrong dependencies" [Undecided,Confirmed] [19:17] kurt_, 'sudo apt-get install python-pyparsing' [19:17] kurt_, as a workaround [19:18] kurt_, the package lacks the correct versioned dependency for the cloud archive [19:18] jamespage: ok, i'll try that === mrmist is now known as mist [20:21] adam_g: can you file a FFE for ceilometerclient....im out [20:33] * hallyn heading out for a bit - be around later and tonight [20:40] jamespage, smoser http://people.canonical.com/~agandelman/ca/folsom/policykit-1-0.104-2ubuntu1.1~cloud0/ and http://people.canonical.com/~agandelman/ca/grizzly/policykit-1-0.105-1ubuntu1.1~cloud0/ are new packages needed to fix FTBFS of libvirt security fix in folsom + grizzly [20:42] oh. joy. [20:46] adam_g, is there a bug for that ? [20:46] smoser, for the CA FTBFS? no [20:47] sure. i guess. did libvirt gain a build-depends ? [20:47] what was the security bug? [20:50] smoser, no bug, CVE-2013-4311. part of libvirt security update was adding a version requirement on policykit in respective distro to pull in a ensure it pulls in a security fix there [20:51] adam_g: ** RESERVED ** This candidate has been reserved by an organization or individual that will use it when announcing a new security problem. When the candidate has been publicized, the details for this candidate will be provided. (http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2013-4311) [20:51] k [20:53] smoser, per jamespage last comment in https://bugs.launchpad.net/ubuntu/+bug/1227425 [20:53] Launchpad bug 1227425 in ubuntu "Cloud images do not need apt-xapian-index" [High,New] [20:53] will the SRU to apt-xapian-index resolve the issue, or is the the SRU that is causing the issue? [20:54] arosales, you need a new cloud image to reoslve it. [20:54] the sru caused it. === Ursinha is now known as Ursinha-afk [20:55] by causing 'apt-get update && apt-get dist-upgrade' to install a new package [20:55] well, newerver version of that package [20:55] and installation of that packkage then does the silly stuff [20:56] smoser, ok. So this should resolve it self in subsequent updates. --thanks. [20:56] when a new released cloud iamge is made the problem will go away. [20:56] untill the next time there is an update to that package. [20:57] ugh [20:57] smoser, your thoughts on just removing that package? [20:57] I guess it only affects smaller instances, but is it worth the cost? === Ursinha-afk is now known as Ursinha [20:59] arosales, the package is pointless and annoying to us. [20:59] but it is a 'Recommends' of aptitude, which is probably a commonly used package by real users. [21:00] ah [21:01] smoser, so your suggestion is to keep it as is [21:02] no. i hate that package. [21:02] and i dont actually know why or if aptitude needs apt-xapian-index to run. [21:03] i'd like to fix it by dropping the package from the cloud iamges, but doing that means dropping the 'Recommends' in aptitude to a Suggests. [22:31] it seems the esacpe characters in the /etc/issue file are not being recognized. at ssh login they just show up as \n \l. any ideas why? i cant see that the issue file needs anything more than the '\' to control to codes. [22:40] xander452: I'm not sure they're meant to be; that's why you'd usually use /etc/issue.net for such applications (those sequences are usually parsed by getty, which ssh doesn't use) [23:39] shauno: thanks for the reply. i made sure the BANNER line sshd points to issue.net. the issue.net file was empty so i added \n \l. there was no change when i ssh to that ip. [23:46] the lts of 12.04 is until 2017? [23:49] is no one here? [23:50] 384 people in the room and no chat? [23:53] 5 years, so yes [23:54] thanks janc [23:54] janc is that full support or does it taper down slowly? [23:55] like in 2016 limited support and 2017 no support [23:57] it's updates for security and (major) bugs until April 2017 for packages in main and (to some degree) for packages in universe too [23:57] thanks JanC [23:57] :D [23:57] I built a couple of small servers with ubuntu 12.04 [23:58] I was just wondering what the deal was with it [23:58] nothing business critical [23:58] :)