[00:01] bdmurray: indeed, I need to go fix the script on limequat so that it runs again. I'll do that tomorrow [00:07] stgraber: well rolling out the new version would be ideal [00:09] bdmurray: yeah, I'll run it from the db account while waiting for the next code update on the server [02:15] bdmurray: hmm, did you manage to run the current python script from the branch on your machine? [02:16] bdmurray: I'm trying to run it on limequat and I'm getting quite a few errors mostly around the raw SQL that was between simples " instead of """ and then later on in the code (still fixing the issues as they appear) [02:17] stgraber: How could I test it? [02:17] bdmurray: I think most of these errors would have appeared even without a working database [02:18] stgraber: Okay, I can work on it more tomorrow then [02:18] bdmurray: I think I'm almost done with the fixes, last one is: [02:18] >>> from launchpadlib import LPNET_SERVICE_ROOT [02:18] ImportError: cannot import name LPNET_SERVICE_ROOT [02:18] hmm, >>> from launchpadlib.launchpad import LPNET_SERVICE_ROOT [02:19] actually but same result [02:19] ok, found it (I think), testing [02:20] maybe launchpadlib.uris [02:21] dinner here bbiab [02:22] yeah, .uris did the trick [02:48] bdmurray: https://bugs.launchpad.net/ubuntu/+source/update-manager/+bug/914794 [02:48] Launchpad bug 914794 in update-manager (Ubuntu) "Gdk-CRITICAL errors output when updating from command line (affects: 1) (heat: 6)" [Medium,Confirmed] [02:48] bdmurray: I tested the code on this one [02:55] bdmurray: pushed the changes to the branch and re-enabled the cron on limequat using the new code [02:57] bdmurray: looking at that script again, I really need to find some time to clean it up ... it should be working for now but I see hundreds of corner cases where a part or another of that script is going to fail [02:58] bdmurray: anyway, thanks for your changes, you should see bug comments appear in the next few hours for existing bugs [03:53] bdmurray: looking at my e-mail notifications, I think having the script set the subject is a bit annoying (I'm used to see the bug's title instead), maybe we should change that? [03:53] anyway, enough poking at the ISO tracker for today :) [04:23] stgraber: you are used to seeing the bug's title because almost everybody uses the web interface to comment and that just happens to use the bug's title. Having a standard subject would allow people a way to filter away those messages. Honestly, I'm indifferent though === bladernr_ is now known as bladernr_afk [09:17] jibel, gema good morning! could one of you link me to the upgrade testing blueprint if/when you get a chance? [09:19] rickspencer3, good morning. https://blueprints.launchpad.net/ubuntu/+spec/foundations-p-lts-upgrades [09:20] mvo, do you know how to configure squid-deb-proxy to not cache *tar.gz and *.tar.gz.gpg ? [09:20] thanks jibel [09:20] mvo, there are test failures because it fails to validate the upgrader tarball [09:24] jibel: hm, odd, let me look [09:26] mvo, for example https://jenkins.qa.ubuntu.com/view/Precise%20Upgrade%20Testing%20Dashboard/job/precise-upgrade-i386/PROFILE=universe-i386,label=aldebaran-upgrade/lastSuccessfulBuild/artifact/universe-i386/bootstrap.log [09:26] "Authenticating the upgrade failed. There may be a problem with the network or with the server." [09:26] I removed the proxy and it pass [09:28] jibel: could you please try adding "refresh_pattern tar.gz$ 0 0% 0" and the same for tar.gz.gpg$ ? [09:29] mvo, k [09:29] mvo, I'll let the current run finish just to make sure there's nothing broken and will retry with this settign [09:30] thanks [09:31] mvo, ah there's this setting already "refresh_pattern tar.gz$ 129600 100% 129600" I think that's the culprit [09:32] squid must be downloading the gpg key but uses the tarball from the cache [09:35] yeah [09:36] jibel: I wonder if we can add a " dist-upgrader-all/current/.*tar.gz$" pattern with 0 0% 0 to prevent that apt-get source stuff is uncached [09:40] mvo, I excluded all tar.gz from the cache. the main concern is caching deb files, fetching a couple of tar files won't hurt. [09:41] * mvo nods [09:41] ah more stable builds this morning thanks to stgraber === _salem is now known as salem_ === bladernr_afk is now known as bladernr_ [16:13] stgraber: Do you think we should change the subject of the ISO tracker comment? [16:16] whoa, topic updated [16:19] bdmurray: I think it'd be less confusing for people reading their bug mail and seeing 50 or so e-mails with "ISO tracker notification" as the subject [16:20] bdmurray: I usually do an early filtering of my bug inbox and having all of them with the same title makes that difficult as I need to read them to find the actual bug title [16:20] stgraber: the subject also contains the bug number but okay [16:22] bdmurray: right, though I'm not that good at remembering bug numbers :) I think the sender is already a clear enough indication that it's an ISO tracker notification. [16:27] stgraber: okay, pushed [16:28] stgraber: I noticed that get_list in the xml rpc interface returns some linaro bugs too [16:28] stgraber: is there some way to filter other than milestone? [16:31] bdmurray: oh, sounds like a bug. I'll have a look. It should filter based on what URL you're using (iso.qa.ubuntu.com shouldn't return bugs for iso.qa.linaro.org or whatever their URL is nowadays) [16:33] stgraber: it looks like $site isn't used in the query [16:33] bdmurray: yeah, I'll fix xmlrpc/qatracker.xmlrpc.bugs.php [16:33] stgraber: cool, thanks [16:35] bdmurray: do you have an example? so I can check that my fix works [16:35] stgraber: bug 660604 [16:35] Launchpad bug 660604 in linux-linaro "Linaro images should utilize EDID (affects: 2) (dups: 1) (heat: 5)" [Low,Triaged] https://launchpad.net/bugs/660604 [16:38] bdmurray: ok, fixed in the branch. I'll update the lp-integration.py script on limequat, fix a few more bugs and ask IS to update the PHP code too [16:40] stgraber: Do you know how lp-integration stopped working? [16:41] bdmurray: no, the cronjob was still correct when I looked yesterday. Might have been an LP authentication issue that broke it at some point, I did a clean authentication yesterday to get a new .cred file [16:41] stgraber: oh, it was probably edge vs production [16:42] that sounds likely indeed [16:51] stgraber: thanks for your help [16:51] np [17:05] stgraber: okay, this is interesting. bug 517992 got tagged by the tracker but it seems that is not the reporter's bug [17:05] Launchpad bug 517992 in debian-installer (Ubuntu) ""No kernel modules were found." message from Lucid powerpc alternate cd. (affects: 2) (heat: 12)" [Undecided,Fix released] https://launchpad.net/bugs/517992 [17:05] as they were testing xubuntu i386 [17:05] with precise [17:07] bdmurray: yeah, seems like gridcube picked an old bug instead of filing a new one for precise [17:07] so removing the tag from the old bug seems right but what about contacting this person? [17:07] the "no kernel" error is fairly standard when d-i and the seeded kernel are out of sync [17:08] IIRC gridcube is Xubuntu's QA guy and is fairly often on IRC [17:09] bdmurray: are you an admin on iso.qa.ubuntu.com? if so, you probably should unlink these reports to the bug before untaging it or it'll get tagged again [17:09] oh yeah, you should be as you are in ~ubuntu-qa-website-devel [17:09] ah but I thought it was a good idea to prevent modifying the history, so no you can't change them... [17:10] bdmurray: do you know if there's a way to know if a bug was tagged with a given tag in the past? so we can avoid re-tagging a bug that was manually untagged? [17:10] stgraber: yeah there is a way but you have to parse the activity log :-( [17:11] stgraber: I'll whip something up [17:13] k. Out to grab something for lunch, will be back in a bit [17:20] mvo, I spend few hours today on bug 917173 . I found other bugs but was unable to reproduce it. [17:20] Launchpad bug 917173 in update-manager (Ubuntu Precise) (and 1 other project) "lucid -> precise upgrade failed: Resolver failed to calculate the upgrade - dpkg-dev held back (affects: 1) (heat: 6)" [High,Confirmed] https://launchpad.net/bugs/917173 [17:20] mvo, I think it is related to multiarch but cannot find what's wrong [17:21] mvo, at the top there should be a something like 'installing libc6 as depends of multiarch' but there isn't [17:22] for some reason installation of libc6 is deferred and breaks the upgrade of libc6-i686 [17:22] " libc6-i686:i386 PreDepends on libc6 [ i386 ] < 2.11.1-0ubuntu7.8 -> 2.13-24ubuntu2 > ( libs ) (= 2.11.1-0ubuntu7.8) can't be satisfied!" [17:22] stgraber: actually it'd be rather involved to find out if the tag was removed - plus it could have been accidental. I'd rather make the script respect the bot-stop-nagging tag [17:23] bdmurray: hmm, indeed, removing iso-testing + adding bot-stop-nagging sounds good and is easy to check [17:23] stgraber: right, I'll make that change [17:31] GridCube: hey we were just talking about bug 517992 which you reported on the iso tracker. It'd be better if you were to report a new bug. [17:31] Launchpad bug 517992 in debian-installer (Ubuntu) ""No kernel modules were found." message from Lucid powerpc alternate cd. (affects: 2) (heat: 12)" [Undecided,Fix released] https://launchpad.net/bugs/517992 [17:34] oh, okay, i reported it back in the day because it seemed just like a standard mistake (like it was) so i did not see a need for a new report, i mean it was just a silly mistake after all [17:35] because the tracker wont allow me to put it as failed without a bug [17:36] the message no kernel modules found can appear for multiple reasons though and that was powerpc specific [17:36] that's why a new bug would be best [17:36] stgraber: bot-stop-nagging change pushed === yofel_ is now known as yofel [17:41] ok :) will do next time i find it again [17:52] bdmurray: branch updated on limequat === salem_ is now known as _salem [20:20] jibel: oh, didn't you write in the bugreport that you can reproduce it every day? [20:32] mvo, I did. There's a vm ready with the bug. we can look at it tomorrow if you wish, I'm off for today. [20:33] jibel: ok [20:33] * mvo waves goodnight [20:33] mvo, good night o/ === bladernr_ is now known as bladernr_afk [22:12] yup, that works :) [22:13] works [22:14] good show [22:50] balloons, I read the mail [22:50] hello [22:51] yea, my mini novel? [22:51] did you read mine? Don't need an answer [22:51] yep ;D [22:51] I did.. since your here we can discuss a bit [22:51] Do you have more information about the autotesting sofware? [22:51] software* [22:51] oki [22:52] First of all, I want to say that my primary goals are: [22:52] 1) To be made in a home network [22:52] 2) To be the most easy to configure [22:53] 3) to speed up the installation of the server an packages as posible [22:53] those are the primary goals [22:53] I meant the installation of the OS [22:53] and packages [22:55] jibel, what means o/ ? I thought it was hello, but I've seen it in too many contexts ... [22:56] txomon|home, ok, I got it [22:56] and then [22:56] 4) to automatically execute tests? [22:56] yes [22:56] that was implied ;) [22:57] but I didn't get there yet [22:57] Making the installation of fresh systems in 4 minutes/ 5 minutes was the main goal [22:57] gotcha [22:57] (at the end, the most part of the time is spent in the installation) [22:57] did you look at the tools page on the wiki [22:58] let me see if I can find it [22:58] i'm not sure pts will do what you want, but I would try it if for nothing else than an excuse to do so [22:58] also it may be worthwhile to have a conversation with some of the orchestra folks to see what it might offer. I know you tried it and find it too much, but you might just be able to use cobbler [22:59] balloons, I really tried hard with cobbler [22:59] but it was very... complicated to me [22:59] * balloons has no experience with cobbler sadly [23:00] I know that with a little help I would have had it running, but as I didn't get any support, I started on what I knew [23:00] i am most interested in your project however. if you can get a nice solution put together, I'm sure there are others who would want to replicate it :-) [23:00] what cobbler gives, is really a lot more that what I need [23:01] that is true [23:01] which has a real inpact in the points 1) and 2) [23:01] i should re-read your mail again [23:02] my idea was to install *just* the necessary. I am not very confident in apache installs [23:02] don't worry, I am here now ;) [23:02] my thoughts are now along the lines of having a local mirror and pxe booting the boxes [23:02] of course, then what.. heh [23:03] cobbler I am sure it is a very good solution to deployment in complicated frameworks... but my solution intends to be the minimal [23:03] mine tooo [23:03] too* [23:04] I started developing them as part of a _rescue_ system for a office I am the admin from (don't know if this is said in a good way) [23:04] it's interesting because of the physical nature of things [23:05] in the cloud, this wouldn't be such an issue [23:05] :-) [23:05] but I realized that it would be a benefit for any application in which you needed to install many systems, regularly [23:06] agree with that "cloud" thing [23:07] but, the cloud means somewhere, and there, would be nearly for sure more usefull orchestra [23:07] I am trying to create a pxe image that gives you the ability to install _any_ system, debian, fedora, ubuntu (or at least, in the same distro, different options) [23:08] in that way, you wouldn't need to create different profiles [23:09] (I am thinking on cobbler) [23:09] what I can sure you, is that using apt-cacher-ng + pxe is a *very* *high* boost in testing [23:10] at least, in installation in real hardware [23:10] if you are using VM, just apt-cacher-ng would be perfect [23:11] its a shame that normal CDs don't allow you to specify mirror (<-- haven't seen it, thought, I am going to check that now) [23:11] ahh yes [23:12] i know the debian netboot cds do [23:12] i was going to make a one cd to rule them all cd.. aka, netboot cd that would grab the latest ubuntu version and install it [23:12] maybe give a choice for development or stable or lts [23:13] netboot gives the option to mirror [23:13] dont know about standard CDs [23:14] I have chosen apt-cacher-ng because: 1) doesn't need configuring 2) doesn't install anything weird, more than itself [23:14] squid-proxy was a little more complicated to me [23:15] I read about apt-cacher-ng and I thought that would be a nice solution [23:16] for example, my solution, as I have a NAS in the home network, is to mount a NAS folder in /var/cache/apt-cacher-ng, and in that way, I can save space in my laptop, and If really needed grab packages installing the apt-cacher-ng in another computer [23:17] but that is for me [23:20] ohh.. having a nas does make that cool [23:21] it's always bad when your home server doubles as something else [23:21] eek [23:21] but that is not an usual home network [23:21] that is why I didn't include it in my solution [23:22] but the idea is the same [23:22] yes, I don't have a nas [23:22] you can make a hdd be mounted in that location throw fstab etc. [23:22] yep, same story [23:22] there are lots of solutions [23:22] or using the computers storage, normal disks [23:23] at the end, you are to use that space when installing ;D [23:23] so I think I will re-read your post and try and dump some thoughts via IRC or mail tomorrow.. For now, the stomach is growling! [23:23] haha [23:24] I am happy to see someone thought about the idea of having various distros/versions available in the same cD [23:24] CD * [23:24] i can say if you want I can help connect you with some folks who do know about cobbler.. the other stuff I'll have to think on [23:24] oki [23:24] ohh yea.. i tried to make a starcraft live cd once too [23:24] hahah [23:24] it was a cool concept [23:24] boot live ubuntu, preinstalled wine and sc [23:25] well, the idea was not that difficult in theory [23:25] I had to add a pair of things more and modify the configuration [23:25] but... where to touch is the difficult thing ;D [23:25] anyways, food calls. have a wonderful evening.. just know your not alone on the radical ideas front [23:25] oki [23:26] so have a full happy meal :) [23:26] filling*