- Friday, February 21, 2020

release-team@conference.openafs.org

Friday, February 21, 2020< ^ >

wiesand has set the subject to: wow, what a crowd today

Room Configuration

Room Occupants

GMT+0
[14:27:23] mbarbosa joins the room
[16:52:55] yadayada joins the room
[16:58:04] <kaduk@jabber.openafs.org/barnowl> [it's time for "has my client silently disconnected" roulette...]
[16:58:54] wiesand joins the room
[17:00:14] meffie joins the room
[17:00:16] <wiesand> Hello
[17:00:20] <meffie> good day
[17:00:25] <yadayada> Hello All
[17:00:31] <kaduk@jabber.openafs.org/barnowl> greetings!
[17:01:32] <wiesand> Looks like we could finalize 1.8.6pre1 today?
[17:01:58] <kaduk@jabber.openafs.org/barnowl> Did I miss a mail requesting a tag?  I am ... very behind on email.
[17:02:12] <wiesand> No.
[17:02:17] <kaduk@jabber.openafs.org/barnowl> phew
[17:02:45] <wiesand> There are two changes left to merge: 14070 and 14047
[17:03:29] <wiesand> Mike +1'ed both, but I wanted to give others the chance to comment, in particular on NEWS
[17:04:52] <wiesand> So, if you'd like to review, please either do it now or ask for more time
[17:05:35] <yadayada> I need to go through 14070, I am reviewing it now
[17:05:45] <wiesand> Thanks
[17:05:57] <meffie> thank you yadav
[17:06:34] <wiesand> Also, if you think something is missing for 1.8.6pre1 at this point, this would be a good time to bring it up ;-)
[17:08:14] <wiesand> Thank you Ben
[17:12:46] <yadayada> 14070 LGTM
[17:13:08] <wiesand> Thanks a lot.
[17:13:48] <wiesand> Last call: any objections to merging 14070 and 14047 and asking Ben for the tag?
[17:14:07] <meffie> none here
[17:14:31] <yadayada> 14047 LGTM
[17:14:38] <yadayada> we can go with merge
[17:15:16] <wiesand> I probably won't upload etc. today since I'm too tired, but once we have the tag I can anytime, probably tomorrow.
[17:15:30] <wiesand> Ok, thanks. Merging now…
[17:21:01] <wiesand> Changes arrived on git.openafs.org. Ben, could you please tag 1045bcc97941e2788fb1d625b38f7727dd5a6813 as openafs-stable-1_8_6pre1 ?
[17:21:23] <kaduk@jabber.openafs.org/barnowl> stand by...
[17:22:37] <kaduk@jabber.openafs.org/barnowl> ...done.
[17:23:01] <meffie> yay!
[17:24:33] <wiesand> Looks good, thanks!
[17:24:55] <kaduk@jabber.openafs.org/barnowl> Thank *you*!
[17:25:10] <wiesand> I'll get it out the door and announced a.s.a.p.
[17:27:33] <kaduk@jabber.openafs.org/barnowl> [looks like we got a kernel NULL deref bugreport w/ 1.8.5 on RHEL7
just now, during the meeting, but it's probably better to talk about
(if needed) next week after we've had a chance to look at it properly]
[17:27:44] <wiesand> If no showstoppers are found during pre1 testing, I think we should aim for releasing 1.8.6 ~ when Linux-5.6 is release.
[17:28:13] <wiesand> RT or mailing list?
[17:28:19] <kaduk@jabber.openafs.org/barnowl> RT
[17:28:28] <kaduk@jabber.openafs.org/barnowl> #135069
[17:30:31] <meffie> ok
[17:32:34] <kaduk@jabber.openafs.org/barnowl> I assume there's not more on 1.8 for today since the release is
underway
[17:32:53] <wiesand> Right
[17:32:53] <kaduk@jabber.openafs.org/barnowl> Is there enough time before linux-5.6 to think much about 1.6.x?
[17:33:37] <wiesand> Once I have 1.8.6pre1 out I'll "think" about it.
[17:34:18] <wiesand> It still vexes me that I haven't understood what went wrong before Mike rebased everything.
[17:35:40] <wiesand> So that's still the first thing I'll do about 1.6.25.
[17:37:06] <kaduk@jabber.openafs.org/barnowl> I guess gitk would probably be more useful than `git log --oneline
--graph --decorate origin/changes/NN/MMMNN/X ...` for all the
"problematic" changes/revs
[17:37:21] <kaduk@jabber.openafs.org/barnowl> Not entirely clear whether it would actually be useful, though.
[17:37:54] <meffie> i think the key is the commit message note in 9407
[17:37:58] <meffie> (backport note: Remove the 1.6.x specific -sync 'delayed' option.)
[17:38:00] <wiesand> And as mentioned a couple of weeks ago, I'm inclined to pick the most important changes only for the time being, rather than aiming for getting the full backlog merged. It's just too long.
[17:38:28] <meffie> -sync delayed was 1.6.x specific, so 1.8.x patches will not apply cleanly.
[17:39:12] <wiesand> I don't think we should remove options from 1.6.x at this point, nor should we introduce major features.
[17:39:35] <kaduk@jabber.openafs.org/barnowl> +1
[17:39:36] <meffie> that's sounds good.
[17:40:08] <wiesand> 1.8.x should be sufficiently mature now to become really conservative with 1.6.x.
[17:40:15] <meffie> agreed
[17:41:01] <wiesand> Also note that if that bug report mentioned above is valid, there's a serious regression in 1.8.5 w.r.t. 1.6.24.
[17:41:29] <wiesand> If it exists, chances are it's in that 1.6.x backlog…
[17:44:10] <kaduk@jabber.openafs.org/barnowl> The way I'm interpreting the description is something that sounds like
a "can't happen" scenario, so we'll need to dig in a bit more, I
think.
[17:44:18] <kaduk@jabber.openafs.org/barnowl> Anyway, shall we move on to master/1.9?
[17:44:29] <kaduk@jabber.openafs.org/barnowl> (not that I have a whole lot to talk about there...)
[17:45:04] <wiesand> Yes, let's move on
[17:45:42] <kaduk@jabber.openafs.org/barnowl> I'm starting to get dug out from the backlog that accumulated while I
was on vacation, and got a few things reviewed last night/this
morning.
[17:46:08] <wiesand> BTW the 1.9.0 NEWS was helpful. Alas, it also became clear to me how outdated it is already.
[17:46:18] <kaduk@jabber.openafs.org/barnowl> Non-review work on my agenda still focuses on the 1.9 NEWS and getting
a 1.9.1 out (skipping 1.9.0?)
[17:46:36] <wiesand> Why skip .0?
[17:46:49] <kaduk@jabber.openafs.org/barnowl> It's outdated, yes, but it has planted a post in the sand to start
from when fixing the outdated-ness
[17:47:27] <kaduk@jabber.openafs.org/barnowl> 1.9.0 sounds like a "big-announcement" thing, but the idea is just to
have a number and tarball for "some stuff" and have it change
frequently.  But maybe that's just me.
[17:48:23] <kaduk@jabber.openafs.org/barnowl> I'm very open to feedback
[17:49:09] <kaduk@jabber.openafs.org/barnowl> I did end up bulk-deleting a bunch of mail from gerrit that came in
while I was gone, so if you asked me specifically a question and
didn't get a response, you should probably ping me again. :(
[17:50:31] <kaduk@jabber.openafs.org/barnowl> I see there was some activity on the pthread-bos topic, but the change
that I thought was the first one (10286) still has a -1 on it for
unresolved comments.  So I'm not sure I understand properly what work
was done in that space -- was it just adding more changes to the top
of the stack?
[17:50:31] <yadayada> Sorry for little late in reply. looks this NULL deref has something to do while we read cellservdb in /proc/fs/openafs. I have 3.10.0-1062.9.1.el7.ppc64le and I do not see any issue with 1.8.5 code. Since this issue happens while opening  CellSerDB  to get in kernel CellSerDB content not sure something to do with CellServDB entry. Will it be possible for us to see dump ? Or get CellServDB they are using ?
[17:50:46] <meffie> i think 1.9.0 is fine. it think the risk is people will automatically try the biggest version number, so we need to be clear is is a non-stable release
[17:51:15] <kaduk@jabber.openafs.org/barnowl> okay, thanks for thinking about it Mike!
[17:51:36] <kaduk@jabber.openafs.org/barnowl> Yadav: I would not expect to get a kernel dump, but asking about the
CSDB in use seems wise.
[17:52:26] <kaduk@jabber.openafs.org/barnowl> I also see that Mark updated the afs-dir-pkg stuff, which is exciting
(not that I've gotten to look at it yet...)
[17:52:27] <yadayada> sure, I think CSDB if we get we can try it and see if we can replicate it. If you get CSDB can you pass it along, will try it on my cluster
[17:52:27] <meffie> regarding pthread-bos, i rebased the commits so they will build against the current master. just minor changes need for that. we were looking at want needs to be done for these commits.
[17:52:51] <kaduk@jabber.openafs.org/barnowl> Yadav: sounds good; thanks
[17:53:00] <kaduk@jabber.openafs.org/barnowl> Mike: ah, okay.
[17:53:21] <meffie> i think some reordering and cleanup would be helpful. can we propose a revised stack?
[17:53:32] <kaduk@jabber.openafs.org/barnowl> Sure
[17:54:13] <meffie> thanks!
[17:54:19] <kaduk@jabber.openafs.org/barnowl> I mean, there's a bunch of stuff in that stack that has +1s and I
forget if they are non-trivial ones, so it would be a little annoying
to have lost that review results, but not the end of the world if it
ends up working better
[17:55:30] <meffie> also, mark and i started a tests/bozo/bos-t.c TAP test suite for bosserver
[17:55:46] <kaduk@jabber.openafs.org/barnowl> Stephan, Yadav: did either of you look at 14046 much?  I was not
really planning to stay in the driver's seat for that one
[17:55:57] <kaduk@jabber.openafs.org/barnowl> Cool, more tests are always good
[17:56:41] <meffie> also, i added a "make check" step to the ubuntu18 buildbot (just that one for now).
[17:57:31] <yadayada> 14046 at first glance looks ok, but I want to test it once. I will review and test it and share my findings on gerrit
[17:57:32] <meffie> the plan was to do a make check on just the master branch, but it seems to be working everywhere, just not many tests to run on the old branches.
[17:58:01] <kaduk@jabber.openafs.org/barnowl> Yadav: thanks!  It was not entirely clear to me from the report in RT
whether there was going to be anything else needed.
[17:58:30] <kaduk@jabber.openafs.org/barnowl> meffie: indeed, there are more tests on the newer branches
[18:01:16] <kaduk@jabber.openafs.org/barnowl> I don't think I have much else for today; are there other
topics/ongoing work on master that I should be prioritizing?
[18:02:29] <meffie> my afsconf_SetCellInfo() bug on master is super annoying. sorry. i hope we can get that fix in soon.
[18:03:01] <meffie> gerrit 14076  and friends
[18:03:02] <wiesand> 14046 looks ok. No clue whether it's sufficient though.
[18:03:03] <kaduk@jabber.openafs.org/barnowl> Yeah, I will try to not drop that.  You are probably right about the
lock thing, too, since the function is now mostly just "printf"
[18:04:36] <wiesand> Anything else to discuss today?
[18:04:54] <kaduk@jabber.openafs.org/barnowl> Not from here
[18:05:06] <yadayada> Nothing from my side
[18:05:18] <meffie> i updated 14076, so it is called under the lock in the "extended" function.
[18:06:17] <kaduk@jabber.openafs.org/barnowl> Do you mean 14077?
[18:06:29] <kaduk@jabber.openafs.org/barnowl> Also, is 14061 going to be abandoned given 1407[67]?
[18:07:56] <meffie> hmm...
[18:08:00] <kaduk@jabber.openafs.org/barnowl> Anyway, we don't need to hold up the meeting for the
SetExtendedCellInfo discussion, I think.
[18:08:19] <meffie> i'll check the numbers later, have a good weekend!
[18:08:26] <kaduk@jabber.openafs.org/barnowl> You, too.  Thanks!
[18:09:07] <wiesand> Ok, let's adjourn. Thanks a lot everyone!
[18:09:10] <yadayada> Thanks !!
[18:09:54] meffie leaves the room
[18:09:56] wiesand leaves the room
[18:10:52] <kaduk@jabber.openafs.org/barnowl> Thanks everyone!
[18:14:44] mbarbosa leaves the room
[19:26:19] mbarbosa joins the room
[19:34:50] mbarbosa leaves the room
[19:35:01] mbarbosa joins the room
[19:35:22] mbarbosa leaves the room
[19:42:19] yadayada leaves the room