=== Chrisfu- is now known as Chrisfu [00:28] rbasak: fyi, done refactoring up to remote, so only submit and tag to finish [00:28] rbasak: these are going to be a pain to review [00:44] jamespage: beisner: pike-proposed is ready to promote [02:52] coreycb jamespage - pike-proposed promoted now. [05:41] nacc: well, all those upload tags are from before I was allowed to push anything [05:42] nacc: maybe I experimented with them back then and the recent git ubuntu submit on a totally different branch pushed them? [05:42] I'd not expect that, but that would be the first explanation that comes to my mind [05:43] nacc: I certainly have not set upload tags in that repo this week, and we have one in the paelzer repo that I might have pushed in the past before it was clear how to use them, but three are in pkg/... === pavlushka is now known as Guest19993 === Guest19993 is now known as pavlushka [06:59] good morning btw [07:38] Good morning [08:15] I'm working on a preseed file for pxe uefi install... I'm wondering where to look for documentation on what tasksel tasksel/first multiselect ubuntu-server would correspond to in terms of a normal install. Could someone point me in the right direction? [08:17] powersj: nacc: rharper: the sync only clears images that are not reuqested (as you assumed) and "not used" [08:17] so anything that has a guest defined will not be cleared [08:17] IIRC [10:19] jamespage: seeing that you have added python3-ceph pkgs in artful, is there any chance to get these for xenial/jewel, too? my target is OpenStack py3 development, so having something within UCA only would be a valid option [10:20] not jewel, but luminous is in the pike uca with py3 packages I think [10:20] frickler: ^^ [10:43] jamespage: ah, perfect, I was only looking at ocata uca up to now, but that should be fine, thx [11:05] jamespage: what would you think about adding a python3-ceph metapackage, similiar to python-ceph? [11:17] frickler: we could do that [11:17] frickler: most of what I added was sync'ed from the upstream ceph packaging, so if we can get it into github.com/ceph/ceph +1 as it makes my life a whole lot easier :-) [11:25] jamespage: understood, I'll take a look === kyle1 is now known as kyle__ [14:25] rbasak, or cyphermox , anyone interested in reviewing [14:25] https://code.launchpad.net/~smoser/ubuntu/+source/plymouth/+git/plymouth/+merge/330286 [14:25] simple change and i feel fine to upload but figured i'd ask for a review [14:33] Anyone run into issues upgrading to the latest msyql lately? [14:33] http://paste.ubuntu.com/25478468/ [14:34] smoser: +1 [14:37] rbasak, is there a easy-ish way to run 'submittodebian' ? from git ubuntu tree ? [14:37] with a commit-ish [14:43] smoser: not that I know about, sorry. But "git send-email" should be available to you now. [14:49] cyphermox rbasak nacc, i'm applying for 'sru developer' position next week, and I'd like to ask if any of you have time and are willing, you have all sponsored sru uploads for me in the past, could you add an endorsement to my application page? https://wiki.ubuntu.com/ddstreet/UbuntuSRUDeveloperApplication [14:59] lordievader: you know how to add a module to initrd in ubuntu server iso? [15:05] ddstreet: will do it this week! [15:05] naac thanks! [15:05] nacc thanks! [15:18] madLyfe: add it to the config. It is too long ago that I made Ubuntu initrd's to remember where exactly. [15:19] cpaelzer: oh it's fine (upload tags), just surprising [15:32] rbasak: given scheduling, etc. how do you want to handle review of the refactor? would it make sense to push to edge/beta/candidate and just test heavily as the review process by our squad? [15:34] nacc: good question. We should a agree a way to avoid blocking you while I'm out. [15:35] nacc: maybe ask ahasenack to review? [15:35] rbasak: yeah, good idea [15:35] ahasenack: sorry for that :) [15:36] rbasak: i'm also trying to make sure pylint3 doesn't indicate any fatal errors (ignoring style generally) for now [15:38] That's a good idea. [15:39] I had it in mind that we should eventually introduce something like that and then perhaps start ratcheting it up. [15:40] rbasak: yep, i'm cleaning up the smaller modules as much as I can in the process (indentation, etc.), as it also does make it easier to see some of the variable names [15:40] rbasak: but importer.py, e.g., is quite large :) [15:40] we also are able to do some consolidation now that we can share code easier (e.g., submit and lint can use the same logic for figuring out the target-branch) [15:41] which i think will resolve cpaelzer's issue with having to specify --target-branch at all for `git ubuntu submit` [15:42] yeah would be nice [15:42] I mean "issue" is already a too hard word for it [16:07] is there a way to dist-upgrade 14.04 to 16.04 without problems? yesterday i tried upgrading and it ended in a new install. initscripts was broken and most of the binaries in bin und sbin were missing [16:08] i couldnt even initiate a reboot anymore [16:08] init was gone [16:08] done many 14.04-16.04 dist-upgrades with no issue [16:08] are you using do-release-upgrade? [16:08] ybaumy: you don't dist-upgrade anymore, use do-release-upgrade [16:08] joelio: it was a mess and it was my first time [16:09] nacc: thats what i did. i just used the term dist-upgrade [16:09] sounds like you broke it then, as missing bin and sbin stuff sounds a bit off [16:09] or there were non-standard things added to 14.04 that br0k the upgrade? [16:09] joelio: i broke it? well i dont have logs but how could i have done that if i used that do-release-upgrade [16:10] joelio: well on that server there were plesk repos enabled [16:10] generally it should disable the repos, the problem is if you've got packages that are non-standard that interact in strange ways with upstream packages [16:10] that problem expresses itself even more when doing a dist-upgrade [16:11] if it's a vanilla install, then generally (well 100% of the time imhe) then it just works [16:12] well newly install was straight forward and in the end maybe i have a clean install and its ok for me. but if im in the situation again i would like to know how to progress from start [16:12] I'd basically no dist-upgrade if you're doing stuff non-standard [16:13] but use config management etc. to rebuild the box on a new target distro [16:13] difficult to say what broke if there are no logs mind [16:13] perhaps could have been easy fix :) [16:14] well next time i will ask here but it was this box im on and im using for irc [16:14] though [16:16] its painful to new install. its like sles11 to sles12 on power from big endian to little [16:17] Actually, as someone who runs SLES and RHEL on power....it isnt that bad at all [16:17] i dont know how many system you have running on sles but new install 1000 of them its a pain is the ass [16:18] its als SAP instances [16:19] surely you dont do this manually [16:21] no of course not [16:21] but it would be easier to uprade in place IMO [16:24] on AIX things were easy. add a disk make the upgrade .. reboot .. finished [16:31] have a better setup in place and it's painless to new install [16:32] this sounds like aprocess problem, not a technical one (beyond the do-release-upgrade failing) [16:32] but if you dont' have logs for d-r-u failing (they are stored in /var/log, iirc), then it's not possible to really debug/help [16:32] nacc: will have the next time. just wanted to ask if there are general problems [16:33] ybaumy: there are not known general problems between 14.04 -> 16.04, I think many have done it [16:34] nacc: is it possible to do something like alternate disk install on linux? [16:36] ybaumy: you mean install to a different disk? I'm not familiar with the term (it's been a while since i used RHEL/SLES and i've never used AIX) [16:37] nacc: you add a disk make a copy of your system to this disk. make the upgrade on the new disk and then boot from the new one [16:37] ybaumy: i mean, you can just do that in linux [16:37] ybaumy: i don't think there is a special tool in the installer to do it (but i might be wrong) [16:38] nacc: well i just wanted to ask. thats one the cool AIX features [16:39] ybaumy: probably means its patented [16:39] nacc: probably even if ibm has contributed alot to the OSS community [16:41] nacc: i had a system migrated from 4.3.3 to 5.1 to 5.2 to 5.3 to 6.1 but then got decommissioned. but the pain i had was minimal [16:42] ybaumy: and i know people who have done upgrades of servers over many ubuntu. But also, in the new world, you just reinstall. Separate your data, use backups, etc. [16:42] ybaumy: use containers, or VMs, there are lots of other choices than big metal [16:43] nacc: im still learning. im into linux for 2 years and im doing openstack and and docker and shit. but this is all like a whole new world to me [16:44] its like everyday i learn 20 things new [16:44] IT was so easy :D [16:44] 10 years ago [17:01] cpaelzer: ever seen this before https://paste.ubuntu.com/25479276/ [17:01] that's on the power system trying to launch a kvm guest [17:04] ah this looks like LP: #1709784 [17:04] Launchpad bug 1709784 in linux (Ubuntu Xenial) "KVM on 16.04.3 throws an error" [Critical,Fix committed] https://launchpad.net/bugs/1709784 [18:35] ybaumy: ever thought of using lxc or some virt to handle your plesk stuff, if you abstract it away then it's a bit easier to manager. Plus then you can do test upgrades, snapshotting, rollback etc etc etc [18:36] although given it's plesk, probably want to have some deep seated system stuff [18:36] (which could still work!) [18:37] what I'm getting at, is if you abstract your systems a bit into containers or virt (where you can) then they're a bit easier to deal with imho [18:38] or use lvm/zfs/btrfs snapshotting on the os level itself etc [18:38] loads of ways to skin it :) [18:39] ahh, I see nacc mentioned that too, good stuff :) [18:40] nacc: ever have issues with not finding the git-ubuntu binary? [18:42] also who should I tell git-ubuntu is running the tests? e.g. git config --global gitubuntu.lpuser [19:28] powersj: from the snap? [19:29] powersj: given that we are doing a no-push import and no other writes, i think you can use usd-importer-bot [19:29] nacc: https://paste.ubuntu.com/25479991/ [19:29] powersj: that's ... weird [19:29] yeah... [19:30] SSH'ing directly in and it worked as expected [19:30] powersj: sshing, `which` finds it? [19:30] powersj: oh you might need to do a login shell [19:30] ah [19:31] nacc: ok here are the results so far: https://paste.ubuntu.com/25479994/ [19:31] there is one error that pops up [19:31] powersj: looking [19:31] powersj: this is with master? [19:31] yes [19:35] powersj: debugging [19:40] nacc: thx, here is the full run as it just finished https://paste.ubuntu.com/25480030/ took just under an hour [20:32] nacc, https://code.launchpad.net/~smoser/ubuntu/+source/open-iscsi/+git/open-iscsi/+merge/330315 [20:33] that'd be nice if you looked. there is a mess of stuff to fix with netplan transition, but that is one of the things. [20:33] i have to run [20:35] smoser: will look [20:35] powersj: reading that too [20:35] powersj: ok, on my bastion (running the current snap), it didn't fail to reimport ipsec-tools [20:37] powersj: will fix the popd message [20:37] powersj: fix pushed to master for that [20:39] nacc: th [20:39] thx [20:42] powersj: i'm not sure on the reimport failure still [20:42] nacc: ok I'm planning on re-running shortly and will see if it reproduces. Should I expect a run to take ~hour? [20:42] powersj: yeah, the from-scratch import takes a while [20:50] Mmm nic bonding [21:39] how do I tell if a package is seeded in the install I'm running, or if it's something that I installed via request, or dependency of a request I made? [21:40] dpb1: apt-mark for the latter bits, i think [21:40] a) seeded b) installed by me direct or through dep [21:40] ok [21:40] * dpb1 runs [21:44] its too bad the installer doesnt offer bonding out of the gate [22:11] powersj: i wonder if we can also add `pylint3 -E gitubuntu --ignored-modules=pygit2` to our jenkins job? [22:11] powersj: it will probably fail on master for now, but is clean on my refactored branch [22:12] nacc: sure can [22:12] can you add a card for that too? [22:12] powersj: ack [22:13] powersj: done [22:13] nacc: thx [22:26] is puppetserver 2.8 puppet 4 or 5? [22:29] hey hey [22:29] sarnold: whatsu [22:30] I was talking to some people - hiring process for some project, and I felt uneasy about hiring a girl, dunno some kind of fear maybe [22:31] not cool. [22:34] well I met 1 guy from Israel 20 plus years C++ [22:34] he said there nearly no one hires girls [22:34] as they cause troubles [22:35] :) [22:35] hehehe: please stop talking like that [22:35] why [22:35] I simply want to ask fellow people opinion [22:35] not acceptable behavior on this channel [22:35] shutting subject up simply reinforce the impression [22:35] that girls are best not hired [22:36] just look at your behaviour [22:36] ty for the feedback dpb1 [22:40] hehehe, here is a opinion, you are a sexist idiot [22:40] there [22:40] Ussat: ok and for statistics you are male? [22:41] and yes girls are way more likely to cause issues at work [22:41] statistically [22:41] however I was thinking maybe .... [22:41] can we get a op to regulate here [22:41] Ussat: regulate what? [22:41] was just looking at how to do that [22:42] pm hehehe to complain [22:42] use pastebin :) [22:42] Dude you have been banned in a few other channels I am in for similar stuff [22:42] Ussat: 0 ideas what u on about [22:42] if you want to stay on topic that is hiring girls go ahead [22:45] plus I think if company expands and our hiring policy is public we can get pretty cool candidates who love to work in such enviroment [22:49] hehehe: what you are describing is illegal in several countries. Please take it elsewhere, as it absolutely not ontopic [22:49] nacc: then why do u comment [22:49] if its not on topic [22:49] :) [22:50] plus surely business can decide who they hire [22:50] not only illegal, but wrong.....but anyway /ignore does wonders [22:51] powersj: something seems up: https://jenkins.ubuntu.com/server/job/git-ubuntu-ci/24/console, appears to need a proxy setup? [22:51] (snap failed to build) [22:51] hehehe: if your metric for being ontopic is someone telling you that youare *offtopic*, you are 100% a troll and I would politely ask you to leave the channel and not come back [22:52] nacc: "git.sigxcpu.org"?? is that a new dependency being pulled in? [22:52] powersj: yeah, gbp built from src [22:52] (in the snap) [22:52] ahhh [22:52] and while i am hosting it on github, the submodules come from upstream [22:52] then yes I'll need to use proxy and hope we can get out [22:52] heh [22:53] nacc: I've got a shell script that builds the snap and installs it on xenial, only need to figure out how to invoke the test script now so it finds git-ubuntu [22:54] powersj: this is the same PATH thing you were hitting before? [22:54] yeah [22:54] powersj: i think you can do something like `ssh bash -l -c ...` [22:54] heh -l... too easy [22:55] powersj: i'm not 100% -- i think either way, it's possibly becuase you need to be running in a shell. The login shell may be overkill (it might be sufficient to just be in bash as PATH expansion to include /snap/bin is from /etc/profile.d/) [22:57] nacc: ok I'll play with both for now I've got -l in there thanks! [22:57] here is what I got btw https://github.com/canonical-server/test-scripts/blob/master/git-ubuntu/integration_test [22:58] powersj: great, thanks [23:34] * Epx998 waits for saltstack repo email...