- Friday, February 16, 2018

release-team@conference.openafs.org

Friday, February 16, 2018< ^ >

Room Configuration

Room Occupants

GMT+0
[01:04:42] kaduk@jabber.openafs.org/barnowl leaves the room: Replaced by new connection
[01:04:43] kaduk@jabber.openafs.org/barnowl joins the room
[11:54:50] Marcio Barbosa joins the room
[13:33:56] meffie joins the room
[13:34:21] <meffie> yay
[13:56:47] wiesand joins the room
[14:00:19] <meffie> good afternoon wiesand
[14:00:34] <kaduk@jabber.openafs.org/barnowl> greetings
[14:00:39] <wiesand> good morning mike & ben
[14:01:07] <meffie> good morning dr kaduk
[14:01:25] <wiesand> wow, there's been a lot of churn this week
[14:01:46] <Marcio Barbosa> good morning all
[14:01:51] <wiesand> hello
[14:02:18] <wiesand> Looks like we're getting closer to 1.6.23pre1
[14:03:11] <meffie> great news
[14:03:25] <wiesand> The -1 on one of the ubik changes is correct, but a number of those have to be rebased anyway, and that should take care of it.
[14:03:39] <wiesand> I'll take care of the rebases.
[14:04:02] <meffie> ok, thanks
[14:04:36] <wiesand> Most of the other stuff had sufficient review. A few open questions mainly about the rpm spec stuff. No big deal.
[14:05:05] <kaduk@jabber.openafs.org/barnowl> > sufficient review
I know some review requests came in to me that I haven't touched yet.
Am I in the critical path for any of the outstanding 1.6.x reviews?
[14:05:06] <wiesand> And I have a few on my list yet to be pulled up. I'll take care of that as well.
[14:05:54] <wiesand> Well, I value your +1s very highly
[14:06:02] <kaduk@jabber.openafs.org/barnowl> *blush*
[14:06:07] <meffie> lol
[14:07:26] <wiesand> But then I know you don't have much time. Let's say I'm ok with merging unless you make it clear that you would rather review a change before.
[14:07:44] <kaduk@jabber.openafs.org/barnowl> Okay, understood.
[14:08:32] <wiesand> And the good news of the week is that maybe we'll get away w/o having to target Linux 4.16 :)
[14:08:51] <wiesand> Which means there's a bit more time for the 1.6.23 release than usual
[14:08:56] <meffie> hooray
[14:09:28] <wiesand> But let's not forget the RHEL7.5 kernel problem - any news on that?
[14:10:12] <meffie> working with redhat here
[14:10:38] <meffie> got access to sources but not commits yet
[14:10:46] <kaduk@jabber.openafs.org/barnowl> Is there a release date known for RHEL7.5 not-beta?
[14:11:09] <meffie> i dont know, i'll ask
[14:11:15] <wiesand> "a few months after the beta"
[14:12:19] <wiesand> 7.4 beta was announced May 24, GA August 1
[14:12:59] <wiesand> 7.3 beta August 16, GA November 3
[14:14:10] <wiesand> 7.5 beta was announced January 18
[14:14:35] <kaduk@jabber.openafs.org/barnowl> Hmm, so if we get 1.8.0-final out before mid-March, then we still
support all supported RHEL versions.  Hmm...
[14:15:15] <meffie> heh
[14:15:18] <wiesand> But only for a couple of days…
[14:16:27] <meffie> and there's a solaris 11.4 beta, but no one cares.
[14:16:38] <meffie> no love for solaris
[14:16:50] <kaduk@jabber.openafs.org/barnowl> Wait, there is?  I thought solaris was dead and Oracle had fired the
entire dev team.
[14:16:53] <wiesand> not here...
[14:17:44] <meffie> yes, they continue to limp along. and there's a new version of solaris studio out too. we will need to cope with that probably.
[14:18:37] <wiesand> Anyway, my plan is to continue merging 1.6 changes I consider ready. That's quite a few right now. Any objections?
[14:18:58] <meffie> sounds great!
[14:19:03] <kaduk@jabber.openafs.org/barnowl> Go for it!
[14:20:19] <wiesand> 12905 is not considered required before the current ubik stack can go in?
[14:20:48] <kaduk@jabber.openafs.org/barnowl> I believe 12905 is totally orthogonal to the other ubik stuff
[14:20:56] <meffie> that's an entirely different thing as far as i can tell
[14:21:32] <wiesand> wasn't the "labeling 0.0" introduced in one of the other changes?
[14:21:38] <meffie> no
[14:22:16] <meffie> not related to those changes. this is something that has always existed.
[14:22:51] <meffie> marcio is setting up some tests to better explain this and propose some perhaps better solutions.
[14:23:42] <kaduk@jabber.openafs.org/barnowl> Cool
[14:23:44] <wiesand> speaking of ubik changes, I still have two older ones on my watch list: 12283 and 12290
[14:24:44] <wiesand> not important? or forgotten?
[14:24:59] <kaduk@jabber.openafs.org/barnowl> both :)
[14:25:15] mvita leaves the room
[14:25:26] <kaduk@jabber.openafs.org/barnowl> I can't tell if Marcio intentionally removed code-review +1 on 12283 in
the 17 sept 2016 comment.
[14:25:47] <kaduk@jabber.openafs.org/barnowl> (Perhaps only the verified+1 was intended?)
[14:25:59] <Marcio Barbosa> oh
[14:26:04] <Marcio Barbosa> 12283 is important
[14:27:04] <kaduk@jabber.openafs.org/barnowl> needs-to-be-in-the-next-release important?
[14:27:30] <Marcio Barbosa> hm
[14:28:00] <Marcio Barbosa> iirc, without this patch, a transaction may works even if we do not have quorum
[14:28:24] <meffie> that sounds suboptimal.
[14:28:48] <kaduk@jabber.openafs.org/barnowl> That would indeed be suboptimal, yes
[14:28:57] <Marcio Barbosa> data can be lost
[14:29:07] <wiesand> I take it that all testing of the ubik changes currently under discussion for the next stable (pre)releases have been tested with 12283/12290 included?
[14:29:31] <wiesand> er, you know what I meant to say I hope
[14:29:33] <kaduk@jabber.openafs.org/barnowl> But I thought we convinced ourselves that the other change (already
merged) would prevent the data loss
[14:30:01] <Marcio Barbosa> > I take it that all testing of the ubik changes currently under discussion for the next stable (pre)releases have been tested with 12283/12290 included?
[14:30:02] <Marcio Barbosa> no
[14:30:35] <kaduk@jabber.openafs.org/barnowl> Because we don't increment okcalls if we have no conn for the server
[14:31:35] <kaduk@jabber.openafs.org/barnowl> And ContactQuorum_rcode will not return 0 if we don't hvae a big enough
okcalls
[14:33:38] <wiesand> So we're safe w/o 12283?
[14:35:02] <meffie> seems it's a separate bug from the others, but it's still a bug?
[14:35:49] <meffie> sounds like we want it in 1.8.0 and 1.6.23?
[14:35:59] <meffie> it == the fix, not the bug
[14:36:00] <kaduk@jabber.openafs.org/barnowl> My recollection was that the current version of 12283 is now just some
internal bookkeeping and cosmetic stuff, but should not affect anything
wire-visible.  The initial version of it was much more critical.
[14:36:05] <wiesand> what's the risk that it will be triggered with a higher probability due to the other changes?
[14:36:41] <wiesand> it == the bug ;-)
[14:36:45] <meffie> heh
[14:36:51] <Marcio Barbosa> >And ContactQuorum_rcode will not return 0 if we don't hvae a big enough
okcalls
yes, make sense.
[14:37:07] <Marcio Barbosa> was looking at the code. sorry for the delay
[14:38:17] <meffie> stephan do you have any other gerrits on your internal watch list?  is there a way for gerrit to show your list?
[14:38:41] <kaduk@jabber.openafs.org/barnowl> It looks like Marcio left a comment on 12290 that the commit message is
out-of-date.  It's probably okay for you to just upload a fixed version,
after this long of a delay.
[14:38:41] <wiesand> Yes. No.
[14:39:07] <meffie> sad, sadder
[14:39:45] <wiesand> The last change I wanted to ask about is 12906
[14:39:58] <Marcio Barbosa> > It's probably okay for you to just upload a fixed version, after this long of a delay.
understood.
[14:41:08] <kaduk@jabber.openafs.org/barnowl> So far, 12906 makes sense to me, but I'd like an additional review.
[14:41:23] <kaduk@jabber.openafs.org/barnowl> It's also unclear to me whether it merits inclusion in 1.8.0-final
[14:42:54] <wiesand> Ok, let's wait for more review
[14:43:33] <wiesand> There's plenty of homework for me to do, and hopefully I'll have a bit more time than this week to get it done.
[14:43:43] <wiesand> On to 1.8/master?
[14:44:24] <kaduk@jabber.openafs.org/barnowl> I've held off on issuing pre5 to get some more input on whether 12906
and/or 12905 merit inclusion.
[14:44:41] <kaduk@jabber.openafs.org/barnowl> But the consensus seems to be that 12905 is not a regression, and
therefore I do not feel a strong need to wait for it.
[14:45:28] <meffie> seems reasonable
[14:45:35] <kaduk@jabber.openafs.org/barnowl> Also I posted a new version of the NEWS update, if anyone wants to
review
[14:45:49] <meffie> ok, thank you.
[14:45:52] <kaduk@jabber.openafs.org/barnowl> So maybe someone is off looking at 12906 during the meeting...
[14:46:12] <kaduk@jabber.openafs.org/barnowl> I couldn't have done it without you ;)
[14:46:41] <wiesand> 12908 and 12909 seem straightforward, but shouldn't block anything
[14:47:00] <kaduk@jabber.openafs.org/barnowl> *nods*
[14:47:17] <wiesand> I'll pull those up
[14:47:57] <kaduk@jabber.openafs.org/barnowl> thanks
[14:48:33] <kaduk@jabber.openafs.org/barnowl> I guess I can also ask/confirm that we are okay calling the
soon-to-be-released pre5 a "release candidate".
[14:48:57] <meffie> i think so
[14:49:50] <wiesand> +1
[14:50:06] <kaduk@jabber.openafs.org/barnowl> Yay
[14:53:02] <kaduk@jabber.openafs.org/barnowl> Anyway, that's about it for 1.8
[14:53:27] <wiesand> NB it looks like the pre4 announcement went to -devel and -info only. Was that intentional?
[14:53:37] <kaduk@jabber.openafs.org/barnowl> I don't know that there's a huge amount of news for master
[14:53:51] <kaduk@jabber.openafs.org/barnowl> I didn't bother sending pre4 to -announce, yes
[14:54:00] <kaduk@jabber.openafs.org/barnowl> I don't remember exactly why I made that decision though
[14:54:41] <kaduk@jabber.openafs.org/barnowl> Stephan noted the minor changes from Andrew that landed, and I hope to
look at some of the vos zap cleanup soon.  (It's a three-day weeknd, so
I should have time, right?)
[14:55:52] <meffie> it's a 3 day weekend? did i forget about a holiday?
[14:56:21] <kaduk@jabber.openafs.org/barnowl> President[']s['] day, though not everybody gets it off.
[14:56:39] <meffie> oh. ok.
[14:57:22] <kaduk@jabber.openafs.org/barnowl> Hmm, looking at 12283 again it seems that Marcio also found a nit in the
commit message there.  Do you want to fix it or should I try to come up
with something, as he-who-touched-it-last?
[14:58:03] <Marcio Barbosa> i can work on that
[14:58:14] <meffie> thanks marcio
[14:58:54] <kaduk@jabber.openafs.org/barnowl> Thanks!
[14:59:19] <meffie> can i mention some stuff about buildbot?
[14:59:20] <wiesand> Looks like my weekend is approaching?
[14:59:30] <wiesand> but go ahead
[15:00:22] <meffie> buildbot 1.0 was released. so i think i should probably try to upgrade to 1.0 and skip 0.9 (on master)
[15:00:32] <meffie> er, the build master
[15:00:47] <meffie> sorry i failed getting 0.9 to work.
[15:01:34] <meffie> also, i should have the arm64 builders ready soon (to slow for gerrit builders, but will be nightlies)
[15:01:58] <kaduk@jabber.openafs.org/barnowl> Thanks for looking at it -- I know it can be tiresome work to do.
[15:02:10] <meffie> thats all i have today about buildbot.
[15:02:38] <wiesand> thanks for all your buildbot work
[15:03:09] <meffie> can i ask a gerrit question here before the weekend starts?
[15:03:27] <kaduk@jabber.openafs.org/barnowl> I also pushed an updated rxgk_crypto_rfc3961 to (hopefully) address
more of Andrew's concerns.  But that's been slow going.
[15:03:28] <wiesand> sure...
[15:03:32] <kaduk@jabber.openafs.org/barnowl> sure, go ahead
[15:04:33] <wiesand> oh oh
[15:05:29] <meffie> i guess i should write an email; but briefly; i see gerrit supports "restores" of abandoned changes, so they are not really lost. do we thing we should be more aggressive about abandoning changes?
[15:05:46] <kaduk@jabber.openafs.org/barnowl> Possibly.
[15:06:10] <meffie> or have some other way to have more accurate dashboards in gerrit?
[15:06:14] <kaduk@jabber.openafs.org/barnowl> I have this plan in the back of my head that once 1.8.0 goes final I
will magically have time to go through all the commits I marked -2 not
for 1.8, and see if they still make sense.
[15:06:34] <meffie> ok
[15:06:40] <kaduk@jabber.openafs.org/barnowl> And if they don't make sense, or would take lots of work to rebase, or
whatever, they could get marked abandoned.
[15:06:59] <meffie> i think i want to find a way to make it easier for people to find gerrits that need attention.
[15:07:08] <kaduk@jabber.openafs.org/barnowl> That sounds good.
[15:07:21] <wiesand> IMHO "abandoned" is for changes that (no longer) make sense
[15:07:45] <wiesand> We shouldn't mix those up with ones we just hadn't the resources to address yet
[15:07:54] <meffie> right.
[15:08:01] <meffie> ok
[15:08:15] <kaduk@jabber.openafs.org/barnowl> Sounds like we're generally in agreement, then.
[15:08:21] <kaduk@jabber.openafs.org/barnowl> Is it time to wish Stephan a happy weekend?
[15:08:45] <wiesand> Happy President's Day ;-)
[15:08:46] <kaduk@jabber.openafs.org/barnowl> (Everyone else gets a happy weekend, too, just not for a few more
hours...)
[15:09:30] <wiesand> Thanks a lot everyone!
[15:09:58] <kaduk@jabber.openafs.org/barnowl> Yes, thanks!
[15:10:26] <Marcio Barbosa> thanks
[15:10:38] wiesand leaves the room
[15:27:40] mvita joins the room
[16:43:54] meffie leaves the room
[21:39:29] Marcio Barbosa leaves the room
[23:35:40] mvita leaves the room
[23:36:14] mvita joins the room