- Wednesday, June 21, 2017

release-team@conference.openafs.org

Wednesday, June 21, 2017< ^ >

Room Configuration

Room Occupants

GMT+0
[00:11:45] mvita leaves the room
[12:39:50] meffie joins the room
[12:59:13] mvita joins the room
[14:45:29] wiesand joins the room
[14:46:16] <wiesand> /!\ slow link, short fuse today
[15:02:52] <mvita> hello
[15:03:05] <meffie> greetings
[15:03:11] <wiesand> Hi
[15:03:48] <wiesand> Any Linux news?
[15:04:16] <mvita> not that I have heard
[15:04:19] <wiesand> (other than "rc6 is out and 4.12 will probably released in ~10 days)
[15:04:27] <wiesand> Good.
[15:04:41] <wiesand> Any test results for 1.6.21pre1?
[15:05:10] <meffie> i've not heard any news
[15:05:39] <jhg> greetings
[15:06:00] <jhg> the buildbot has not failed, so the bat-signal remains quiet =)
[15:06:14] <wiesand> I think our plan is then still to release 1.6.21 w/o further changes, ~ in time for Linux 4.12.
[15:06:40] <meffie> great
[15:06:42] <wiesand> Which is probably the meeting a fortnight from now.
[15:07:08] <wiesand> And then have an early 1.6.22pre1 if possible.
[15:07:41] <wiesand> Main objective: those ubik fixes. I see there's some progress, but it seems slow.
[15:08:19] <mvita> Thorough.  It just _looks_ like slow ;-)
[15:08:49] <wiesand> Any ideas why the configure step failed on that Solaris builder for 12639?
[15:09:03] <mvita> looking
[15:09:33] <wiesand> Mark: thanks for the explanation. I like "thorough" in the context of potentially high impact changes.
[15:10:18] <mvita> hmm - the Makefile plumbing sprung a leak
[15:10:45] <wiesand> Sounds bad
[15:10:45] <mvita> it will take some digging, hold on
[15:10:51] <wiesand> Thanks
[15:11:48] <meffie> that's odd.
[15:11:57] <mvita> spurious?  but why now?
[15:13:10] <meffie> dont know. thanks for pointing it out wiesand.
[15:13:16] <wiesand> Doesn't look spurious to me
[15:13:46] <wiesand> I'll submit some throwaway change to check that
[15:14:05] <mvita> well, just resubmit that build
[15:14:16] <wiesand> ok
[15:14:27] <mvita> I got it.
[15:14:34] <wiesand> Ah?
[15:14:34] <meffie> i've not changed that slave.
[15:14:53] <mvita> I submitted the rebuild
[15:15:06] <wiesand> good, let's see
[15:15:13] <wiesand> And then I looked at https://gerrit.openafs.org/#/q/status:open+project:openafs+branch:master+topic:linux-enoent-caching-fixes
[15:16:08] <wiesand> I tried to create a test build, but the last one of those is too hard for me to backport. Too much code skew...
[15:17:08] <wiesand> Are there opinions on that stack? Is it fodder for 1.6.22>
[15:17:44] <mvita> I've not looked at it before now
[15:18:21] <kadukoafs@gmail.com/barnowl2B662C26> I've not looked at it yet, either.
[15:19:02] <mvita> oh, this one
[15:19:10] <meffie> it's for the fast-path stuff
[15:19:11] <mvita> papering over an old Linux kernel issue
[15:19:35] <jhg> I have got to run.
[15:20:04] <mvita> solaris config error again
[15:20:04] <jhg> I'll follow up on any Linux issues I see in the minutes or the log here when I get back
[15:20:18] <mvita> same thing, bleah
[15:20:20] <mvita> bitrot?
[15:20:24] <wiesand> solaris configure issue is unlikely to be spurious
[15:20:45] <wiesand> some update to m4/autoconf/automake on that builder?
[15:21:02] <mvita> meffie says no changes - hence my bitrot hypothesis
[15:21:18] <meffie> i'm on the buildslave now. i'll take a look there.
[15:22:11] <wiesand> Joe: Thanks
[15:22:55] <meffie> regarding the linux-enoent-caching-fixes, i think those will probably be needed on 1.6.x, and definitely for 1.8.x
[15:23:30] <kadukoafs@gmail.com/barnowl2B662C26> I read the commit messages just now, and they do look like good things
to pull in, yes.
[15:24:46] <wiesand> Given how different the code looks, and how far back the changes on master go and what their scope is: Time to get 1.8 ready, really. Or abandon it. (Did I mention I'm on a short fuse today)
[15:25:53] <wiesand> Hi Ben BTW
[15:25:57] <mvita> you may have alluded to it, yes
[15:26:42] <meffie> my understanding is the dvhint patches makes the enoent bugs more visible
[15:26:45] <kadukoafs@gmail.com/barnowl2B662C26> Yes, it's time to get 1.8 ready.
[15:27:12] <kadukoafs@gmail.com/barnowl2B662C26> Hmm, my memory is trying to tell me that I left some questions on the
dvhint patches and they haven't been answered yet.
[15:27:42] <wiesand> The solaris builder woes aside, I think we have a plan and a clear path for 1.6. So on to 1.8?
[15:27:48] <kadukoafs@gmail.com/barnowl2B662C26> meffie, did you by any chance get to look at the cellful mounts thing?
[15:27:58] <wiesand> (the dvhint stuff is not in 1.6, is it?)
[15:28:02] <kadukoafs@gmail.com/barnowl2B662C26> I got as far as ... remounting the homedir of my test user, and no
farther :(
[15:28:07] <meffie> working on that now, ben
[15:28:23] <kadukoafs@gmail.com/barnowl2B662C26> dvhint was still in gerrit targetting master IIRC
[15:28:55] <meffie> yeah, it's running experimentally is some places.
[15:30:18] <meffie> the enoent topic is a result of an old linux bug, if i understand correctly
[15:30:36] <kadukoafs@gmail.com/barnowl2B662C26> Anyway, once we have a story for the cellful mount thing, I think it's
time to issue a pre2 and call it beta quality.
[15:31:09] <kadukoafs@gmail.com/barnowl2B662C26> Probably it would have the ubik fixes and linux-enoent stuff we've
been discussing as well, but I don't know that they are absolute
requirements.
[15:31:22] <meffie> ok, great. now that we have a good test case, that should be possible to find a root cause.
[15:31:32] <kadukoafs@gmail.com/barnowl2B662C26> Right.
[15:32:02] <wiesand> Ben: In doubt, include it. My 2c.
[15:32:40] <kadukoafs@gmail.com/barnowl2B662C26> Hmm, probably better to get early testing, yeah.
[15:33:12] <wiesand> BTW, just food for thought: do we have a mid-term vision of release handling?
[15:34:10] <wiesand> Like "don;t allow that much skew from one stable release to the next ever again"?
[15:34:37] <kadukoafs@gmail.com/barnowl2B662C26> I think the current claim is "release new major versions more often
than every 5^W6 years".
[15:34:58] <kadukoafs@gmail.com/barnowl2B662C26> Which is a pretty similar claim, really.
[15:35:07] <wiesand> If we agree on that, that's a start.
[15:35:59] <mvita> <wishes he had a portrait aspect monitor to review this commit>
[15:36:49] <kadukoafs@gmail.com/barnowl2B662C26> xrandr takes --rotate :)
[15:37:22] <meffie> (was just going to say, xrandr is your friend :)
[15:38:09] <wiesand> Then just tilt your notebook to the side - at a 120 deg opening angle it will stand quite stably. ;-)
[15:38:16] <meffie> ha!
[15:38:27] <kadukoafs@gmail.com/barnowl2B662C26> external keyboard recommended.
[15:38:35] <mvita> you guys are a scream
[15:38:56] <meffie> motion to adjourn? food trucks have arrived.
[15:39:20] <kadukoafs@gmail.com/barnowl2B662C26> Fine by me; I think the outstanding tasks are well understood and just
need implementation.
[15:39:30] <wiesand> Who could deny that motion under these circumstances...
[15:39:48] <wiesand> Thanks a lot everyone!
[15:40:02] <mvita> bye
[15:40:12] <mvita> or should I say:
[15:40:14] <mvita> b
[15:40:15] <mvita> y
[15:40:16] <mvita> e
[15:40:18] <meffie> 1. i'll ping folks for pre1 testing. 2. work on gaja's bug
[15:40:31] <wiesand> thanks mie
[15:40:38] <meffie> have a good day
[15:40:51] <wiesand> but as usual, the actual tesing will happen after the release...
[15:40:53] <kadukoafs@gmail.com/barnowl2B662C26> Thanks mike!  (And thanks everyone for being here!)
[15:40:53] <wiesand> bye
[15:41:07] <kadukoafs@gmail.com/barnowl2B662C26> I have this evil idea to put pre2 in debian which will get it pulled
into ubuntu.
[15:41:17] <kadukoafs@gmail.com/barnowl2B662C26> But maybe that is a little too evil.
[15:41:32] wiesand leaves the room
[18:35:58] kadukoafs@gmail.com/barnowl2B662C26 leaves the room
[18:36:05] kadukoafs@gmail.com/barnowl2B662C26 joins the room
[19:51:36] meffie leaves the room
[23:11:26] mvita leaves the room
[23:11:26] mvita joins the room