- Wednesday, April 19, 2017

release-team@conference.openafs.org

Wednesday, April 19, 2017< ^ >

Room Configuration

Room Occupants

GMT+0
[07:07:22] mvita leaves the room
[08:24:42] kadukoafs@gmail.com/barnowl4AB31BCF leaves the room
[08:25:13] kadukoafs@gmail.com/barnowl4AB31BCF joins the room
[13:15:55] meffie joins the room
[13:28:17] mvita joins the room
[13:59:16] wiesand joins the room
[14:00:29] <wiesand> Hello
[14:00:57] <kadukoafs@gmail.com/barnowl4AB31BCF> Greetings.
[14:01:27] <wiesand> looks like we still have buildbot woes
[14:01:35] <meffie> oh noe
[14:02:21] <wiesand> is "woes" the wrong term?
[14:02:22] <meffie> yes, i will need to upgrade the buildbot master
[14:02:44] <kadukoafs@gmail.com/barnowl4AB31BCF> Your English is fine.
[14:03:13] <meffie> some people have started using a new version of the client, which is not compatible with our ancient buildbot master version.
[14:03:44] <mvita> transitioning
[14:04:03] <meffie> the new version of master should work with old clients though.
[14:05:19] <wiesand> Let me briefly go through the candidate list for 1.6.21…
[14:05:42] <meffie> great!
[14:06:16] <wiesand> omitting some trivial doc changes - a +1 would still be nice for those
[14:06:32] <wiesand> Id like to give "shake harder" another try
[14:06:38] <meffie> yay
[14:06:50] <wiesand> that's 12448..12451 + 12513
[14:07:05] <wiesand> ubik fix in 12516
[14:07:18] <wiesand> dafs fixes in 12568 and 12569
[14:08:01] <wiesand> still open: 12328 - if nobody reviews it, I'll abandon it - after all we're living w/o it
[14:08:40] <kadukoafs@gmail.com/barnowl4AB31BCF> I have a terminal open reminding me to look at it.
[14:09:01] <wiesand> then there's still 11151 (more details in vos rele -verb output), but it needs 12455 on top and that;s hard to rebase
[14:09:33] <meffie> oh, did i not rebase that one yet?
[14:10:06] <wiesand> arm64_linux26 is on the wish list, but I'd like to prioritize the other stuff highr a.t.m.
[14:10:40] <wiesand> 10245 would be a nice fix to have, but isn't easy to rebase
[14:11:10] <wiesand> and it touches the same file as 11855, which is from Mike's wish list and hard to pull up too
[14:11:56] <meffie> i can help pull stuff up, and put then in my github before gerrit if that helps.
[14:12:08] <wiesand> again the same file: some nice simple fixes in 12346 and 12347 - I'll pull those up if nothing happens w.r.t. 10245 and 11855
[14:12:48] <meffie> (speaking of arm64_linux26, looks like i need to restart my raspberry pi builder)
[14:13:05] <wiesand> Mike: feel free to pull up 10245 and 11855
[14:13:47] <meffie> ok, i'll take a look.
[14:14:19] <wiesand> then there's new one from Marcio: 12602 - doesn't look entirely solaris specific though
[14:14:42] <wiesand> and of course Linux 4.12 changes when they come up
[14:15:11] <meffie> i think 12602 is for macos
[14:15:24] <wiesand> and I remember Mike saying we'll want something like 12592
[14:16:09] <meffie> correct.
[14:16:12] <wiesand> and that's about what I have on my 1.6.21 radar right now - anything missing? anything we shouldn't bother including?
[14:16:39] <mvita> that's plenty!
[14:17:06] <kadukoafs@gmail.com/barnowl4AB31BCF> Only thing that comes to mind is if I ever get around to actually
merging ted's freebsd stack.
[14:17:13] <meffie> i think you covered it all. it would be nice to get the dvhints sorted out at some point.
[14:17:19] <kadukoafs@gmail.com/barnowl4AB31BCF> But that should be pretty orthogonal to everything else.
[14:17:33] <kadukoafs@gmail.com/barnowl4AB31BCF> I'm still unconvinced by the dvhint locking strategy.
[14:18:15] <wiesand> ben: sure, once the fbsds are ready they shouldn't be a problem to include - given sufficient review...
[14:18:45] <wiesand> re dv hints I completely have to rely on you folks…
[14:19:35] <kadukoafs@gmail.com/barnowl4AB31BCF> That is, whether the access to dp->d_time needs to take (e.g.) the
dentry spinlock or something, for example.
[14:21:05] <meffie> ok, sounds like dvhints is post 1.6.21 then.
[14:21:29] <kadukoafs@gmail.com/barnowl4AB31BCF> I mean, it could be that RCU mumble makes it all fine; I just don't
know and no one has claimed that it's okay.
[14:21:31] <wiesand> Was that Andrew's alternative approach to speeding up those "stat() like crazy" workloads?
[14:21:48] <mvita> yes, I would like to take a look at the dvhints and provide another ps, but I can't get to that any time this week
[14:21:59] <kadukoafs@gmail.com/barnowl4AB31BCF> It started out as Andrew's change, yes, though some other people have
been modifying it since then.
[14:22:54] <meffie> yes, we are looking at linux_access_permissions next. it's the next bottleneck.
[14:24:06] <wiesand> About timing: In a perfect world we'd have a pre1 including Linux 4.12 changes at around linux-4.12rc4
[14:24:52] <wiesand> That's in about a month
[14:25:31] <wiesand> That's what I have on 1.6.x for today
[14:25:37] <meffie> ok, thanks.
[14:26:08] <meffie> i'll post notes. to -info or -devel?
[14:26:22] <kadukoafs@gmail.com/barnowl4AB31BCF> I once again failed to get a testbed set up to work on the 1.8 issue
from RT.
[14:26:29] <wiesand> T think -devel is the right list for those.
[14:26:38] <meffie> ok
[14:26:49] <wiesand> Thanks!
[14:27:00] <kadukoafs@gmail.com/barnowl4AB31BCF> (#133654)
[14:27:19] <kadukoafs@gmail.com/barnowl4AB31BCF> So I don't know that there's any 1.8 news for the agenda.
[14:27:48] <meffie> sigh, yes, i've not made much progress on 133654 yet.
[14:28:22] <meffie> i've been communicating with tedc out of band. should be able to sort out his issues and close his tickets.
[14:28:31] <kadukoafs@gmail.com/barnowl4AB31BCF> Yay
[14:28:52] <kadukoafs@gmail.com/barnowl4AB31BCF> I'm glad that there has been successful communication. Hopefully it
was not too frustrating on your end.
[14:29:56] <wiesand> 133654 is not a tedc ticket though?
[14:30:06] <kadukoafs@gmail.com/barnowl4AB31BCF> Yes.
[14:30:12] <kadukoafs@gmail.com/barnowl4AB31BCF> tedc has a lot of other tickets
[14:30:50] <wiesand> right ;-)
[14:31:59] <wiesand> sorry I can't help - I'm hardly managing to stay on top of 1.6.x currently
[14:32:15] <kadukoafs@gmail.com/barnowl4AB31BCF> Understood.
[14:32:22] <wiesand> NB how's the foundation doing?
[14:32:35] <wiesand> could they help?
[14:32:43] <meffie> i hear they are still active and meeting.
[14:33:34] <meffie> i gave them some updates to their website text and they took it.
[14:33:45] <wiesand> hmm, http://www.openafsfoundation.org/about/minutes/ looks sad
[14:34:27] <meffie> oh, dear yes
[14:35:25] <wiesand> > could they help?
[14:35:31] <wiesand> unlikely :-(
[14:36:23] <meffie> perhaps they could fund a less-spammy replacement for rt?
[14:36:49] <kadukoafs@gmail.com/barnowl4AB31BCF> Perhaps. But I am skeptical that they could also find volunteer time
to run it.
[14:37:04] <wiesand> Right.
[14:37:08] <kadukoafs@gmail.com/barnowl4AB31BCF> (The current RT being "free", of course.)
[14:37:53] <meffie> we could make a request?
[14:40:46] <meffie> the time saved in cleaning up spam would pay for itself.
[14:40:48] <wiesand> Frankly, if there are any resources to spend they'd be better invested in getting 1.8.0 done IMHO.
[14:41:43] <meffie> good point.
[14:43:27] <wiesand> But that's for others to decide of course.
[14:43:34] <wiesand> Anything else to discuss today?
[14:43:44] <meffie> motion to adjourn?
[14:44:20] <wiesand> That's just another way to ask the same question...
[14:44:46] <wiesand> And I guess the answer to yours is "sustained".
[14:45:01] <mvita> cool
[14:45:13] <wiesand> Thanks a lot everyone!
[14:45:42] <kadukoafs@gmail.com/barnowl4AB31BCF> Thanks!
[14:45:59] <meffie> thanks! notes out shortly
[14:46:05] wiesand leaves the room
[14:47:52] meffie leaves the room
[18:19:12] kadukoafs@gmail.com/barnowl4AB31BCF joins the room
[22:31:38] kadukoafs@gmail.com/barnowl4AB31BCF leaves the room
[22:31:46] kadukoafs@gmail.com/barnowl4AB31BCF joins the room
[22:45:04] kadukoafs@gmail.com/barnowl4AB31BCF leaves the room
[22:45:11] kadukoafs@gmail.com/barnowl4AB31BCF joins the room