15:05:33 <gwd> #startmeeting Virt SIG
15:05:33 <centbot> Meeting started Tue Oct 31 15:05:33 2017 UTC.  The chair is gwd. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:05:33 <centbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:05:36 <gwd> #chair sbonazzo
15:05:36 <centbot> Current chairs: gwd sbonazzo
15:06:17 <gwd> rafaelmartins carlwgeorge alynpost strigazi ^
15:06:37 * lsm5 here
15:06:40 * jirkade here
15:06:57 <gwd> #chair lsm5 jirkade
15:06:57 <centbot> Current chairs: gwd jirkade lsm5 sbonazzo
15:08:08 <gwd> OK, let's start.  lsm5, you want to go first?
15:08:21 <lsm5> sure
15:08:27 <lsm5> #topic Container update
15:08:40 <lsm5> #info still waiting on bug syncing for centos-release-container
15:08:43 <lsm5> #link https://bugs.centos.org/view.php?id=13780
15:08:54 <lsm5> #info cri-o updated to v1.0.2
15:09:05 <lsm5> that's about it for me, carlwgeorge you got stuff to add?
15:09:29 <lsm5> (if not, it's back to you gwd)
15:09:36 * kbsingh here
15:09:51 <gwd> #chair kbsingh
15:09:51 <centbot> Current chairs: gwd jirkade kbsingh lsm5 sbonazzo
15:10:09 <gwd> kbsingh, who needs to be prodded to make the container repo on mirrors? (see above)
15:11:40 <gwd> OK, we'll discuss that in a bit. :-)
15:11:44 <gwd> #topic Xen update
15:12:02 <gwd> #info Built packages for XSAs 236-245
15:12:46 <gwd> #info Working on trying to tie CentOS xen package testing into our XenProject testing infrastructure
15:13:07 <gwd> #info Interesting meeting in CERN a a week and a half ago
15:13:26 <gwd> ...with the other SIG leaders.  I'll have a separate topic on this later in the meeting.
15:13:46 <carlwgeorge> sorry i'm late, here now.  but i have nothing to add.
15:13:46 <gwd> I think that's all for me.  sbonazzo, you want to go next?
15:13:55 <sbonazzo> gwd ok
15:14:01 <sbonazzo> #topic oVirt updates
15:14:22 <sbonazzo> #info upstream 4.2.0 beta released a few hours ago, just in time for halloween
15:14:36 <sbonazzo> #info 4.1.7 RC5 to be released upstream today / tomorrow
15:14:52 <sbonazzo> #info Virt SIG packages getting built for 4.1.7 GA
15:15:31 <sbonazzo> #info Virt SIG packages getting published on mirrror.centos.org for altarch too after CentOS Dojo at CERN
15:15:50 <sbonazzo> I think that's it on my side
15:16:10 <gwd> OK -- jirkade, any updates?
15:16:14 <jirkade> #topic libvirt update
15:16:37 <jirkade> #info upstream will release 3.9.0 soon (expected on Thursday)
15:17:24 <jirkade> #info but libvirt (not even 3.8.0) can't still be built in centos because i686 is not ready :-/
15:17:53 <gwd> jirkade: I think it would be OK to build libvirt first for x86_64-only
15:18:00 <jirkade> that's it, unfortunately
15:18:40 <gwd> For CentOS 7, i686 is basically an altarch -- it should probably have a separate set of tags (like arm64 does)
15:18:44 <jirkade> gwd: I guess so, but it would need some tweaking in the virt7-libvirt-latest target
15:19:08 <gwd> jirkade: You can also add some rune or other to the spec file; that's what I do for the Xen packages
15:19:31 <gwd> ExclusiveArch: x86_64 aarch64
15:19:51 <gwd> Don't know if that's the best "official" way of doing that or not
15:20:04 <jirkade> gwd: it won't help, because the build root repository can't be built as it depends on the i686 arch
15:20:51 <gwd> What do you mean?
15:21:41 <gwd> Evolution: I think you're in charge of the CentOS 7 i686 stuff, right?
15:21:53 <Evolution> gwd: nope. that's hughesjr. I do aarch64 stuffs
15:22:26 <TrevorH> i686 7.4 was released recently, is that the bottleneck?
15:22:26 <jirkade> http://cbs.centos.org/koji/taginfo?tagID=1132 the repo is currently broken since it was last rebuilt on Sep 6 and the packages it points too were replaced by newer ones since then
15:22:46 <gwd> Evolution: Oh sorry; somehow got the impression you were doing all the altarch stuff
15:23:08 <Evolution> hehehe, no worries.
15:23:44 <jirkade> alphacc: hmm, do you know why http://cbs.centos.org/koji/taginfo?tagID=1132 is still old when i686 was already released as TrevorH says?
15:25:07 <gwd> OK, well let's come back to that.
15:25:16 <gwd> Anyone else have any updates?
15:28:23 <gwd> #topic Follow-ups from SIG meeting @ CERN dojo
15:28:51 <gwd> Unfortunately I haven't prepared a "report", but there was one thing I thought I'd mention and give people the opportunity for feedback
15:29:15 <gwd> And that was about having the CBS build from git rather than having to upload SRPMS
15:30:20 <gwd> The basic issue is that we have to have a way for CBS to get access to large binaries (like source tarballs, &c) outside of "normal" git
15:31:14 <gwd> The classic way of doing this would be with a "look-aside" cache: you upload large binaries to a separate location indexed by hash, and then store a hash of the binaries you want in the git tree
15:31:46 <gwd> Another way that was discussed was using the git "large file" support.  This is available on github, and would also be available on gitlab.
15:32:21 <gwd> Arrfab sent an e-mail to centos-devel with the various options last week.  Anyone have any opinions / questions?
15:32:23 <kbsingh> gwd: re: container repo, just the usual bugs.centos.org issue, assign to bstinson or Arrfab ( who's on pto this week )
15:32:27 <TrevorH> i686 announcement email -> https://lists.centos.org/pipermail/centos-announce/2017-October/022568.html
15:33:19 <gwd> kbsingh: The bug was opened the beginning of Sept and last touched the beginning of Oct; so the question was whom to poke to get the ticket moved forward. :-)
15:33:27 <kbsingh> gwd: +1 for look-aside, its the only thing that works consistently, this specific work has been trialled many times by many people.
15:33:48 <kbsingh> gwd: humm i would say join the cbs meeting on Monday and poke there as an escalation point
15:34:04 <gwd> kbsingh lsm5 ^
15:34:12 <kbsingh> everyone who can do anything with this, is typically at that meeting
15:35:27 <gwd> OK, well re the cbs/git issue, if anyone has opinions reply to the mail on the list. :-)
15:35:32 <gwd> #topic AOB
15:35:34 <gwd> Anything else?
15:35:48 <alynpost> Did the meeting time move?
15:36:05 <kbsingh> DST in the UK / EU did
15:36:10 <alynpost> ah.
15:36:20 <alynpost> tricky; I can make my calendar behave now.
15:36:22 <gwd> alynpost: Sorry, I should have sent an email abou tit.
15:36:27 <gwd> about it
15:36:57 <alynpost> I had one topic, excuse my tardiness, even if I'm less tardy than I expected.
15:37:19 <gwd> #chair alynpost
15:37:19 <centbot> Current chairs: alynpost gwd jirkade kbsingh lsm5 sbonazzo
15:37:21 <gwd> Please go aheard
15:37:29 * gwd can't type today
15:37:56 <alynpost> Some time ago now I had offered to help on the testing end.  I am characteristically slow but have a report.
15:38:26 <alynpost> The recent round of XSAs included an initial blurb saying that linear page tables weren't used by the BSDs.
15:39:21 <alynpost> Chris Brannon was testing at prgmr and found that NetBSD was impacted.  His effort was part of our general push to improve testing.
15:39:32 <gwd> alynpost: That was my mistake; I think we fixed that in the final version
15:39:41 <alynpost> That's correct, it was.
15:40:10 <gwd> (I looked at FreeBSD and assumed NetBSD would be similar)
15:40:50 <alynpost> My report then is that while we haven't work to publish yet the testing we did manage detected a problem.  If you'll be patient with my slowness I do tend to aim well.
15:41:30 <alynpost> and hope to build on our work that uncovered the NetBSD linear page table issue.
15:41:40 <gwd> Yes, thanks for that -- I'm glad we caught that before it went public.
15:42:13 <alynpost> I'll adjust the timezone in my calendar as well, thank you kbsingh.
15:42:40 <gwd> The testing you did based on the CentOS packages, then, and you think this may translate eventually to being able to test things out of the CBS?
15:43:40 <alynpost> That is my hope, yes.
15:43:49 <gwd> Great, thanks for the updaet.
15:44:01 <gwd> OK, anything else?
15:45:08 <gwd> OK, thanks everyone!
15:45:10 <gwd> #endmeeting