/srv/irclogs.ubuntu.com/2016/05/25/#ubuntu-release.txt

pittiarges: thanks05:47
cariboupitti: yes ?07:12
pitticaribou: I meant I had a question for you in bug 153278907:14
ubot5bug 1532789 in multipath-tools (Ubuntu Trusty) "Trusty multipath-tools suffering seg faults" [High,Fix committed] https://launchpad.net/bugs/153278907:14
ckinghi there, stress-ng 0.6.04-1 has been stuck in autopkgtest testing for ppc64el for ~4+ days in  the "Test in progress" state (looking at the update_excuses page). I think it needs some kicking08:15
jamespagepitti, I think its important that we maintain interim Ubuntu release until the end of their life - one reason being that they are actually the primary source for Ubuntu Cloud Archive pockets, so if we stop maintaining Ubuntu two months prior to eol, then I have to start maintain the same packages directly in the UCA for 2 months extra :-)09:37
jamespagepitti, crmsh is a bit of exception to that rule as its not in the UCA, however ovs and ceph both are...09:38
jamespagepitti, for UCA updates, we push through the associated ubuntu release first (on the assumption its still supported)09:38
jamespageso for OpenStack Liberty UCA, Wily is still the point for entry for updates...09:38
pittijamespage: wow, we have production stuff running on wily?09:39
jamespagepitti, that's not quite what I said09:39
jamespagepitti, we have production stuff running on trusty based on the packages that are in wily09:39
jamespagepitti, UCA is just one huge backport09:40
jamespage:-)09:40
pittijamespage: ok, fair enough; I just wanted to ask if all that effort was actually useful, and it does seem to be a major risk for users on actual wily09:41
jamespagepitti, well for crmsh wily users are completely broken atm09:41
pittijamespage: right, so for the crmsh case: who will actually care09:41
jamespagepitti, I was all up for the 'no one is going to actually use wily for production' but then a load of people commented on that bug...09:42
pittijamespage: i. e. who will start using wily for crmsh now that xenial is out09:42
jamespagethats true now, but was not the case two months ago when I did the updates, testing and upload...09:42
jamespagepitti, tbh i think people where testing on wily in preparation for xenial09:46
pittiyeah, that seems plausible09:47
jamespagepitti, from my perspective, most of the work is already done in terms of time, so I'd rather see it through...09:47
pittijamespage: so you want ceph, curtin, maas, openvswitch09:47
jamespageand be happy stating that we did support wily for 9 months09:47
jamespagepitti, can't comment on maas and curtin, but yes please to ceph and openvswitch09:47
pittijamespage: the verification work of this is going to be substantial too, I suppose09:47
pittijamespage: ok, and I guess we bury  crmsh?09:48
jamespagepitti, for crmsh? no the testing is automated - we know its broken (wily-liberty is currently disabled...)09:48
jamespageinfact the testing is automated for ceph and ovs as well09:48
pittijamespage: I mean for ceph, curtin, etc.09:48
jamespagecurtin/maas - defer to server team on that front...09:48
pittijamespage: we have automated tests that check that an upgrade from a current wily deployment still works?09:48
pittias that's what actually matters09:49
pittiwe know that the new versions by themselves are fine (presumably)09:49
pittibut we are throwing that at existing users of the wily version09:49
jamespagepitti, yes09:49
jamespagedeploy/test/upgrade/test09:49
jamespagebasically09:49
jamespagepitti, I stand corrected - that's not fully automated, but is a cycle we go through for SRU's09:51
jamespagebut its still not resource intensive09:51
argespitti: i'm trying to understand your comment on the wily SRU review queue. isn't wily supported until july?11:27
pittiarges: it is, yes11:28
pittiarges: and we shuold certainly continue to supply major bug fixes11:28
pittiarges: but I wonder if "update a server package to a major new release" makes sense at this point11:28
pittithe testing and risk by far outweigh the benefit IMHO11:28
pittiarges: that said, see backscroll from jamespage a few lines up11:29
argespitti: fwiw some of the server packages go into wily then into cloud archive11:29
pittiyes, he pointed that out11:29
pittiso accepting these is fine11:29
pittiarges: but I think we should bury crmsh11:29
argesoh ok : ) yea i need to increase scrollback on weechat11:29
argesyea that seems like a bit jump anyway11:30
Odd_Blokeinfinity: I see you've commented again on https://code.launchpad.net/~daniel-thewatkins/livecd-rootfs/enable-backports/+merge/295059; have you had a chance to re-review it fully?13:48
Odd_Blokeinfinity: Oh, I see you're not working; don't worry about it!13:49
infinityOdd_Bloke: Nah, I'm still off work until tomorrow, that was just a side comment to ogra's comment.13:49
infinityOdd_Bloke: I'll get you all reviewed and merged tomorrow (or later today if this whole laying around on drugs thing threatens my sanity).13:49
Odd_Blokeinfinity: Cool, thanks!13:49
Odd_Blokeinfinity: Enjoy your stupor!13:50
infinityOdd_Bloke: Once this is tested and proven, is the intent to backport to xenial for 16.04.1?13:50
infinity(and maybe trusty too, though meh...)13:50
ogra_lol13:50
Odd_Blokeinfinity: Yeah, definitely for xenial.13:50
ogra_2011 ?13:50
ogra_woah13:50
infinityogra_: Yeah, it's okay, I was as surprised as you.  Old people don't deal well with change.13:51
ogra_lol, yeah13:51
infinityOdd_Bloke: Kay.  Let's make sure that ball gets rolling as soon as this is landed in yakkety and a few builds appear to produce sane results.13:52
infinityOdd_Bloke: For now, I'm high on Oxy and going to movie myself to sleep.13:53
ogra_"oxi brings movies to your sleep" :)14:03

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!