[00:02] rbasak: there's something off with the snap, but i'm not sure what yet, will deubg in the AM. since it's classic, it's workaroundable by installing the deps [00:33] if i did a sudo command and it wants me to type in my password but i want to cancel the command, is that possible? [00:36] ctrl-C [01:20] madLyfe: indeed === JanC is now known as Guest61806 === JanC_ is now known as JanC [03:33] ahasenack: i've [mostly] figured out my issue with systemd and filesystems not unmounting/shutdown hanging [03:34] coincidentally, i ended up solving another problem at the same time. in fact, the reason i figured out the answer to the reboot problem was because i was troubleshooting the other problem, and i had an aha moment, realizing they were related [05:33] good morning [07:45] how do I dig a DMARC record? It doesn't show with 'dig domain TXT' [07:45] albech: _dmarc.example.com [07:49] andol: it doesnt show, so i am guessing its not a valid string i have entered at the DNS [07:50] dig txt _dmarc.blah.com [07:50] works for me (tm) [07:50] RoyK: yeah nothing shows and I entered it 10+ hours ago. [07:51] albech: try querying the server directly, dig @servernameoraddress [07:51] RoyK: did that too [07:52] which domain is this? I can check from here [07:52] ipat.dk [07:53] looks fine [07:53] hmm [07:53] erm - no [07:53] wait [07:53] its the spf you see [07:54] this is me dig txt _dmarc.karlsbakk.net [07:54] what sort of dns server? [07:55] RoyK: no idea.. its some hosting company [07:55] gratisdns.dk... [07:56] will im seing that my dmarc string ends with a ; and yours doesnt.. [07:59] albech: Hmm, according to https://web.gratisdns.dk/domaener/dns/ it looks like gratisdns.dk pushes updates in batches, five times per day. If you did your change 10 hours ago, perhaps they are simply running a bit behind? [07:59] albech: If nothing else you could always make another change, and see if that at some point later results in a newer SOA serial. [07:59] andol: possibly [08:00] andol: good idea.. will do that as well, just to verify when the push has run [08:00] If nothing else the current SOA serial *suggests* that nothing has been pushed the last 10 hours [08:00] 2017082102 [08:02] andol: could be why its not showing.. didnt even think of that [08:03] Might also be that it didn't push/reload due to a syntax error in the dmarc record. On the other hand, ideally the web-ui ought to have complained about that to you. [08:04] v=DMARC1; rua=mailto:postmaster@nixmail.net; ruf=mailto:postmaster@nixmail.net; p=none; sp=none [08:08] albech: Strictly you should have quotes surruning that, but may be that the web-ui taxes care of that for you. Your spf txt record does have those quotes, so unless you had to enter them then, that is probably not the issue here. [08:08] s/taxes/takes/ [08:09] andol: that is taken care of by the webui [08:09] i checked that myself [08:11] Well, unless you are in a hurry the easiest way forward is probably to remove the potentially problematic _dmarc TXT record, create some other simple/safe test entry, and see if that propagtes during the day. [08:12] rbasak: you and nacc just "did" the git ubuntu upload tagging - I want to update the wiki and then do my first one [08:13] rbasak: if you could take a look if you think http://pad.ubuntu.com/XD5ESDSBne is right that would be great === led2 is now known as led1 [10:21] cpaelzer: yeah that's right. So before upload, git ubuntu tag --upload, push that, and then dput as normal. [10:27] rbasak: thanks for the confirm [12:07] rbasak: could you re-imort virt-manager or teach me how to do it? [12:07] it has the upload tag from the merge two weeks ago but not imported since then [12:08] I might be able to work on the base of the upload tag thou [12:08] that would give us time to check why it wasn#t reimported automatically [12:08] cpaelzer: "mkdir -p /tmp/import; git ubuntu import -v -d /tmp/import/virt-manager virt-manager" is what I use. [12:08] maybe it isn't in the default set that is synced [12:09] rbasak: ok let me run that and check if it stumbles somewhere [12:09] cpaelzer: default set: https://git.launchpad.net/usd-importer/tree/gitubuntu/import-cron-packages.txt [12:14] rbasak: the import is failing me on the pristine-tar of 1.4.0 [12:14] rbasak: do you mind running the same import and checking if that is a local issue to me? [12:16] Trying [12:17] 08/30/2017 13:16:54 - ERROR:stderr: pristine-tar: Unknown subcommand "verify" [12:17] 08/30/2017 13:16:54 - ERROR:stderr: pristine-tar: Unknown subcommand "verify" [12:17] 08/30/2017 13:16:54 - ERROR:stderr: pristine-tar: Unknown subcommand "verify" [12:17] Aargh. [12:17] Anyway, I think I'd hit that already - my pristine-tar is too old now. [12:17] (on Xenial) [12:18] IIRC nacc meant some things only work on zesty [12:19] yep artfuls pristine tar has the verify command [12:19] but I thought the snap would encapsulate such things === petevg_ is now known as petevg [12:20] I'm not using the snap. [12:20] maybe it only does so with the git ubuntu libs but not external executables [12:20] hmm, maybe I currently don't either [12:20] let me check [12:20] It's kinda hard to develop with the snap directly :) [12:20] well I do use it and it fails more or less the same [12:20] let me run it in an artful container [12:21] rbasak: well maybe the snap does encapsulate it [12:21] rbasak: I don't have it failing for missing "verify" [12:21] rbasak: the actual verification fails for me [12:22] rbasak: nacc: http://paste.ubuntu.com/25431785/ [12:22] cpaelzer: there is a known pristine-tar bug [12:22] 08/30/2017 14:12:29 - ERROR:stderr: fatal: ambiguous argument '6c61fcfe2c55a9269c94b6ec2fbeac09759bf80b^{tree}': unknown revision or path not in the working tree. [12:22] yep [12:23] That seems like a bug in git ubuntu though [12:23] nacc: ^ [12:23] I'll work on the upload tag for now and let nacc take a look later then [12:39] coreycb: doing the ceph l release right now plus some associate ceph-deploy updates [12:40] jamespage: awesome [12:41] jamespage: looks like qemu backport is fixed up [12:41] coreycb: which one? [12:42] cpaelzer: the build failure from rc3 was preventing it from backporting [12:42] yeah I handled that (I thought) - so -rc4 is building for you? [12:42] cpaelzer: yep [12:42] in artful it is still on dep8 [12:42] coreycb: nice, good to get that confirmed [12:42] cpaelzer: thanks [12:43] yw++ [12:52] jamespage: i'm starting to work through final uploads of pike packages [13:27] coreycb: do you want to split the list, do it in half the time? [13:27] coreycb: I need to help admcleod with something and then I can be all yours... [13:28] jamespage: sure, i've started from the top of the alphabet if you want to start from the bottom [13:34] jamespage/coreycb: come on chop chop ;) [13:34] hey zul :) [13:34] coreycb: okies [13:34] coreycb: zaqar it is then [13:38] coreycb: zaqar uploaded - I'm using lescina branch builds to de-risk build failures btw (not blind uploading) [13:38] coreycb: watcher next [13:39] no-op [13:39] moving on [13:39] jamespage: ok yep [13:39] coreycb: swift next [13:40] coreycb: I feel like mr no-op today [13:40] swift - no-op [13:40] jamespage: :) [13:41] coreycb: moving onto openstack-trove [13:42] jamespage: ok i'm almost done through the murano's. i'll stop at networking-bgpvpn. [13:44] coreycb: I'll pick sahara next [13:46] jamespage: oh btw, i dropped the db sync's from the postinstall scripts where it made sense. basically any package that wasn't inherited from zigo. [13:46] jamespage: that does affect some services in that they won't start without user running the db sync commands. so i had to update some of the autopkgtests to run the cmds. [13:46] coreycb: we should sync with wendar on that front - she's decommissioning all of the upstream git repos and working on moving things back to debian infra [13:47] coreycb: ack - I think that's OK [13:47] coreycb: I'd be tempted to go further and not auto-start services as well - but maybe that's to late this cycle :-)( [13:48] coreycb: build-testing sahara - no branch build to look at [13:48] jamespage: well, i don't think some will start without a configured db [13:48] coreycb: I think you're right [13:48] coreycb: that was some of the original rationale for defaulting to sqlite, but its really quite useless and does not help anyone IMHO [13:50] jamespage: makes sense, and I agree. there are docs that describe how to set up openstack if you want to do it manually. [13:50] coreycb: its the approach we took with ceph - nothing starts automatically on install - ditto on restarts for package updates [13:51] jamespage: ok [13:51] coreycb: plus there are some top level systemd targets for restarting all ceph daemons on a machine [13:51] that might be neat - openstack.target for example [13:51] as part of a common package [13:52] coreycb: ok sahara done [13:52] * jamespage looks for next target [13:53] coreycb: nova? [13:53] jamespage: go for it [13:54] actioning nova... [13:56] coreycb: nova done [13:57] coreycb: neutron next? [13:57] jamespage: yep go for it, i'm on networking-bagpipe/bgpvpn [13:59] ack [14:00] jamespage: ok those are uploaded. i'll get networking-odl/ovn [14:00] coreycb: picking the neutron-* ones next [14:05] coreycb: ok neutron-* done [14:05] jamespage: great, one left. want to get networking-sfc and we're done? [14:05] coreycb: neutron-vpnaas is dead right? [14:05] jamespage: i think so. let me check the repo though. [14:06] coreycb: hmm they did push an rc1 [14:06] jamespage: yeah looks like it [14:07] coreycb: ok done [14:08] coreycb: lemme get networking-sfc unless you already did? [14:08] jamespage: go for it [14:09] coreycb: doing so now [14:10] coreycb: done [14:11] jamespage: great \o/ [14:11] jamespage: anything else to do? [14:11] jamespage: could use a new pylxd i think [14:12] coreycb: yes and I also need to cut the nova-lxd release tag upstream [14:12] coreycb: but that's looking good as well [14:12] https://review.openstack.org/#/c/499140 [14:13] jamespage: great, i assume you'll upload nova-lxd and pylxd together? === ossurayynot is now known as tonyyarusso [14:17] coreycb: I will [14:17] jamespage: ok === tonyyarusso is now known as Guest13718 === Guest13718 is now known as tonyyarusso === led2 is now known as led1 [14:56] cpaelzer: can you do the import with --no-fetch first [14:56] cpaelzer: i'm 95% sure that it's because of us chaning how imports work === led2 is now known as led1 [14:57] cpaelzer: although you are indicating a future bug is going to be hit [14:57] cpaelzer: pristine-tar can't be used directly, because of copmonent tarballs [15:11] nacc: the --no-fetch fails at pristine-tar verify still [15:11] nacc: just at another step of it I think [15:11] * cpaelzer comparing logs [15:12] nacc: http://paste.ubuntu.com/25432445/ [15:12] nacc: that is the error now [15:14] cpaelzer: that's using the snap? [15:15] nacc: my case is snap, rbasak'S was from git [15:15] cpaelzer: right, from git is probably basically broken in principle, we now need new gbp versions, new pristine-tar functionality. Hence the snap [15:15] cpaelzer: i'm setting up a fresh VM to test the snap, one moment [15:15] Yeah mine is known broken for a different reason. [15:15] nacc: I'm ok without you hurrying [15:16] i think there's something more fundamental with the snap [15:16] something wonky with PATH and possibly PYTHONPATH [15:16] nacc: let me know til tomorrow morning if the import worked and is there or not [15:16] cpaelzer: ack [15:16] nacc: then I'll upload "classic" or do a new MP [15:16] nacc: that gives you time to do it right [16:09] hi, anybody knows what's going on with python and openssl? [16:09] I was running pip and got an error: AttributeError: 'module' object has no attribute 'SSL_ST_INIT' [16:09] googling that showed a whole bunch fo reports all over the place about incompatibilities between python and older ssl lib versions [16:09] which happens to be what's default in ubuntu xenial [16:11] I'm not sure why I'm seeing this problem now and didn't see it a while back, some of the reports are from dec 2016 [16:13] Have you tried sudo pip install -U pyopenssl [16:20] preseed question: trying to install a package from universe during a preseed install. It appears that "d-i pkgsel/include" cannot find it. [16:20] I then tried using d-i preseed/late_command to do an apt-get update; apt-get install -y pkg; which appears to have worked. [16:21] Is this the best way to install a package from universe? [16:21] powersj: i think there is a wrapper for the latter 'apt-install' ? [16:22] nacc: ah ok - I'll look into that one [16:22] thx [16:28] genii: yeah, I get the same problem [16:28] so right now I'm basically stuck as far as I can see [16:28] I need the openssl working to fetch the packages so I can't update openssl [16:31] powersj: d-i pkgsel/include string package-name1 pkgN [16:31] however ime it doesn't work despite being in the docs [16:31] so I do all my touch-ups, including installing some packages, in a post-inst script [16:32] drab: I used that to install packages, but only works with things found in main or on the CD [16:32] for example "d-i pkgsel/include string openssh-server" works [16:32] powersj: d-i preseed/late_command string wget -O /target/var/tmp/postinst http://... ; in-target sh /var/tmp/postinst [16:32] ah ok [16:32] oh, ok, I couldn't get it to do that reliably either for some reason [16:42] weird, it works just fine on a clean ubuntu container [16:43] I wonder what's wrong with pip on my desktop... [16:45] ah, weird, had to apt remove --purge python-openssl [16:46] Strange [16:46] yeah, don't get it [16:49] rbasak: around? [16:49] (ideally not) [17:20] lol [17:21] anybody around happens to know if and how it's possible to get an AD compatible system to integrate a few windows desktop on the lan? [17:21] a bunch fo googling seems to indicate a world of pain is ahead for anybody going that direction [17:44] is there an existing bug for the next round ocata stable releases? (like this - https://bugs.launchpad.net/ubuntu/+source/swift/+bug/1696139) [17:44] Launchpad bug 1696139 in neutron-fwaas (Ubuntu Zesty) "[SRU] ocata stable releases" [Undecided,New] === sammyg is now known as Guest79034 === hakra is now known as sammyg === moonligh- is now known as moonlight [19:01] oh, I see we likely need to get Pike released first [20:14] !info pike [20:14] Package pike does not exist in zesty [20:16] I mean Pike release of openstack === drab_ is now known as drab === Epx998- is now known as Epx998 [22:13] so my boss gets me a new version of the intel x550 10gb nic and of course the latest ixgbe driver doesnt seem to see it [22:19] Epx998: doh [22:20] 10G X550T hmm ok why am I not seeing its interface [22:27] there we go [22:34] so, i've installed node-less, but npm list -g doesn't include the less module as it should. [22:35] is this a case of 'abandon hope all ye who enter here', or is there a chance of actually fixing/bringing to the attention of someone who will care to fix it? [22:36] IOW, should i invest the effort to find out why ubuntu's pkgs are broken, or should i just do what i have to do to fix it? [22:37] I know nothing about node or npm but quite often the approach of cpan / pip / etc kinds of tools is that they manage something separate from distribution-provided tools [22:38] so it's possible that node-less works for what it is intended for, and npm isn't expected to work with those packages -- I stress that it's -possible- that this is working as intended, I don't know.. [22:38] yeah. i could go via the manual/npm approach, but it seems a bit silly to provide a node-less package if it simply won't work. [22:38] i guess that's plausible [22:39] however, it's still broken from the pov of node and anything that relies on canonical node procedure [22:40] here's the files that the package installs http://paste.ubuntu.com/25434873/ -- does npm look in those locations for information? [22:41] good question. i don't know enough about npm's internals to guess how it works. i'm looking into that now. [22:51] this is weird, I created two extra files in interfaces.d, one for 10g and another for 1g (different interfaces) when I just enable the 10G, I cannot reach the server at all. If I enable the 10G and the 1G - I can get to the server via either interface. [22:51] why would the 10g only be pingable when the 1g is enabled?