- Wednesday, June 15, 2016

release-team@conference.openafs.org

Wednesday, June 15, 2016< ^ >

Room Configuration

Room Occupants

GMT+0
[00:11:35] mvita leaves the room
[01:28:16] kadukoafs@gmail.com/barnowl4B0829E7 leaves the room
[01:28:28] kadukoafs@gmail.com/barnowl4B0829E7 joins the room
[13:26:47] mvita joins the room
[13:57:01] meffie joins the room
[13:58:08] <meffie> hello
[13:58:23] Daria joins the room
[14:01:41] <mvita> on my way in a few minutes, bear with me
[14:03:34] wiesand joins the room
[14:03:49] <wiesand> Hi
[14:06:46] <wiesand> Sorry for being late.. We had a presentation on gravitational wave detection that was just too fascinating to leave.
[14:07:48] <wiesand> And sorry for not preparing anything for today. But I think the agenda is fairly obvious.
[14:08:26] <meffie> that sounds way more interesting than openafs!
[14:08:53] <wiesand> Do we want to push out a 1.6.18.1 with just the Linux 4.5 changes really soon now, or wait for 12297 and/or 12239?
[14:08:58] <Daria> yeah, can you just reshare that to us?
[14:10:28] <wiesand> It wasn’t recorded. But I’ll send a pointer to the slides as soon as I have it.
[14:11:18] <kadukoafs@gmail.com/barnowl4B0829E7> I am in a training session and thus not really here.
I did merge the 4.5 bits; sorry that took so long
[14:11:31] <meffie> thanks!
[14:12:03] <wiesand> Thanks. I already pulled them up, and also tested they don’t break EL5/6/7.
[14:13:20] <mvita> now or wait?   I'm of two minds
[14:13:43] <mvita> but I guess I'm leaning toward wait as long as the wait is less than a week
[14:14:01] <wiesand> So am I. If we don’t wait, who would benefit a lot?
[14:14:04] <mvita> just to avoid the overhead of a .2 release
[14:14:11] <wiesand> exactly
[14:14:13] <Jeffrey Altman> I believe Fedora switches to 4.5 kernels tomorrow
[14:14:55] <mvita> then we should probably go with what we have
[14:16:41] <wiesand> is 12274 actually required?
[14:17:05] <mvita> I don't know
[14:17:09] <wiesand> I remember Marc saying "it’s just a warning"
[14:17:21] <mvita> and that's what the commit msg says too
[14:18:20] <mvita> well. okay, it's a bug fix
[14:18:29] <wiesand> So Marc himself wasn’t in favour of pulling it up. But Ben may have had a good reason to do it, and under the "linux45" topic.
[14:18:39] <mvita> if bdi_init failed then afs_fill_super has a silent failure
[14:19:06] <mvita> so the fix is important to pull up, but don't know if it belongs in a point release
[14:19:19] <mvita> maybe the "warning" aspect made it sound like a build issue
[14:20:05] <mvita> definitely either 1.6.19 or 1.6.18.1, if Ben wants it there
[14:20:54] <wiesand> Under what circumstances would bdi_init fail?
[14:21:15] <mvita> (_could_ be a silent failure, depending on the answer to your question)
[14:21:34] <mvita> looking at bdi_init now, stand by
[14:27:23] <mvita> all I see in a quick look is ENOMEM
[14:27:31] <Jeffrey Altman> if bdi_init() fails it is likely that future memory allocation requests will fail as well.   It can only return success or -ENOMEM
[14:28:07] <wiesand> So, unlikely unless the system is wedged anyway
[14:28:16] <mvita> yeah, I think so
[14:28:50] <wiesand> In that case it’s rather for 1.6.19 IMO
[14:28:51] <mvita> worst case is SIGSEGV on an uniitialized block
[14:29:03] <mvita> sorry, unallocated block
[14:29:19] <wiesand> Hmm, that’s a kernel panic, right?
[14:29:24] <mvita> aye
[14:30:50] <wiesand> Would anyone be able to test the four patches on a 4.5 kernel?
[14:31:00] <mvita> that's already been done
[14:31:08] <mvita> fvvo "test"
[14:31:24] <wiesand> including 12274?
[14:31:40] <mvita> oh, don't know about that one, let me ask Joe...
[14:33:23] <mvita> I believe billings was gonna pull the 4.5 fixes too
[14:34:07] <mvita> don't know if he ever did or what his results were (billings)
[14:34:42] <mvita> joe seems to be afk right now, but I'll have him post in 12274
[14:34:55] <wiesand> Thanks
[14:36:37] <meffie> 12274 seems ok
[14:38:48] <wiesand> Thanks
[14:39:45] <wiesand> so we agree on releasing 1.6.18.1 with https://gerrit.openafs.org/#/q/status:open%20project:openafs%20branch:openafs-stable-1_6_x%20topic:linux45
[14:40:09] <mvita> yes.
[14:40:19] <wiesand> (+ 12303 and a finalized NEWS change)
[14:41:23] <meffie> thank you.
[14:41:46] <wiesand> I’ll get that done a.s.a.p. then.
[14:42:44] <wiesand> NB on that other, more interesting, topic: LIGO is going to announce "something" later today. https://aas.org/aas-briefing-webcast
[14:43:00] <mvita> oooo
[14:43:52] <mvita> at one time I wanted to be a physicist when I grew up.   My GRE physics score put an end to that.
[14:44:36] <mvita> but at least I got the BS degree.
[14:47:01] <mvita> re: joe's tests - he tested those 4 on master, not on 1.6.x
[14:47:25] <wiesand> I’ll test on EL5/6/7 anyway.
[14:47:52] <wiesand> If Fedora switches to 4.5 tomorrow, I’ll also whip up a VM and test on that.
[14:47:56] <mvita> I believe he tested Debian and Fedora, don't know which releases
[14:49:13] <mvita> asking now
[14:49:41] <mvita> Debian 8, 64-bit
[14:51:23] <mvita> oh, he didn't get to Fedora yet
[14:51:29] <mvita> my mistake
[14:52:06] <meffie> ok, thanks.
[14:52:28] <wiesand> A few more +1s on the changes would still be good.
[14:52:53] <wiesand> But I won’t insist on them when I actually manage to test on Fedora with 4.5.
[14:53:17] <meffie> ok
[14:54:02] <wiesand> BTW, 12297 doesn’t need a configure test?
[14:54:32] <meffie> hmm.
[14:54:53] <mvita> ah, yes.
[14:54:58] <mvita> it should have one
[14:55:46] <mvita> I think I remember reading that PAGE_SIZE and PAGE_CACHE_SIZE have always been identical
[14:55:49] <mvita> but I'm not sure
[14:56:12] <mvita> if that were true, technically you wouldn't need AC
[14:56:34] <wiesand> Yes, maybe it’s really not needed.
[14:58:59] <wiesand> I’ll test that one on EL5/6/7 too.
[14:59:24] <wiesand> But it’s unlikely to make 1.6.18.1.
[14:59:36] <wiesand> Which we’ll hopefully get out this week.
[14:59:45] <wiesand> Anything else to discuss today>?
[15:01:21] <wiesand> Thanks for the comment on 12297
[15:03:36] <wiesand> If there’s nothing more, let’s adjourn?
[15:03:39] <meffie> i dont have anything else yet. some patches / fixes will be coming soon.
[15:03:47] <mvita> okay
[15:03:56] <meffie> i guess for 1.6.19 or later.
[15:04:25] <wiesand> NB I cherry picked some 1.6.19 candidates yesterday. Feel free to -1 them if you think they don’t belong into 1.6.
[15:05:06] <wiesand> But I have to run. Thanks a lot everyone!
[15:05:26] <mvita> thank you!
[15:05:30] <wiesand> BYe
[15:05:31] wiesand leaves the room
[15:06:30] meffie leaves the room
[15:20:58] <kadukoafs@gmail.com/barnowl4B0829E7> I thought I remembered seeing a build error about bdi_init() very
early on when I was looking at building 1.6 on a 4.5 kernel, but I
don't think I ever reproduced it when I was looking more carefully.
It may not be needed to pull up that change, but it should also be
harmless to do so.
[18:28:12] kadukoafs@gmail.com/barnowl4B0829E7 leaves the room
[18:28:23] kadukoafs@gmail.com/barnowl4B0829E7 joins the room