Home
release-team@conference.openafs.org
Friday, July 26, 2019< ^ >
Room Configuration
Room Occupants

GMT+0
[12:01:15] mbarbosa joins the room
[12:48:41] yadayada joins the room
[12:55:41] kaduk@jabber.openafs.org/barnowl joins the room
[13:56:58] meffie joins the room
[14:11:37] yadayada joins the room
[14:27:31] yadayada leaves the room
[14:58:43] <meffie> good morning/afternoon/evening
[14:58:55] <meffie> hello yadayada
[14:59:21] <yadayada> Hello Mike
[15:00:01] <kaduk@jabber.openafs.org/barnowl> Greetings everyone!
(Hi Yadav!)
[15:00:10] <yadayada> Hi Ben
[15:00:41] wiesand joins the room
[15:00:49] <kaduk@jabber.openafs.org/barnowl> Uhoh, am I in "send but not receive" mode
[15:01:38] <wiesand> Hello. And a special "welcome!" to yadayada
[15:01:58] <yadayada> Hello Wiesand
[15:02:02] kaduk@jabber.openafs.org/barnowl leaves the room
[15:02:43] kaduk@jabber.openafs.org/barnowl joins the room
[15:03:13] <kaduk@jabber.openafs.org/barnowl> I guess even if I am I can still just keep refreshing the log
[15:04:22] <kaduk@jabber.openafs.org/barnowl> I'm in Montreal this week for the IETF, but don't need to be giving
full focus to the session I'm in.
[15:04:25] <wiesand> I'm on a rather slow link today. I'm also being mobbed at work, so expect me not to think all that straight and react all that promptly today.
[15:05:19] <wiesand> re 1.84.pre, I think we're really close now
[15:05:29] <kaduk@jabber.openafs.org/barnowl> Do you have a list of remaining tasks for 1.8.4pre1?
[15:05:37] <kaduk@jabber.openafs.org/barnowl> ("not in front of me" is a fine answer)
[15:05:43] <wiesand> merged (after smoke testing) a lot of changes yesterday
[15:05:54] <wiesand> I do though ;-)
[15:06:25] <wiesand> 13769 is already cleared
[15:07:12] <wiesand> on 13692, 13676 at least one more +1 would help me accept it
[15:08:18] <wiesand> Mike +1'ed the NEWS change (dealing with the "TBD"s it's what's left to do)
[15:08:38] <yadayada> 13692 I can help in reviewing
[15:08:57] <meffie> thank you yadayada!
[15:09:24] <wiesand> Once again, don't be shy to criticize my contributions (most of this is from Mike anyway) - , wording, speeling, grammar, style, whatever
[15:10:15] <wiesand> The last open change for 1.8.x is then 17324 - how urgent is that one?
[15:11:12] <kaduk@jabber.openafs.org/barnowl> I get EPERM for 17324 (or ENOENT?)
[15:11:30] <wiesand> And are there any other changes that should urgently make 1.8.4pre1?
[15:11:38] <wiesand> Sorry, that's 13724
[15:11:45] <meffie> ok
[15:12:38] <meffie> it's not urgent, but we need to to fix these warnings eventually
[15:13:00] <wiesand> well, "review review review" ;-)
[15:13:03] <kaduk@jabber.openafs.org/barnowl> It's in libadmin, which is maybe a bit of a second-class citizen in
practice
[15:13:14] <wiesand> thanks
[15:13:55] <kaduk@jabber.openafs.org/barnowl> That said, it's also pretty easy to review :)
[15:15:04] <wiesand> so, any objections to get 1.8.4pre1 out the door after:
[15:15:49] <wiesand> 1) merge 13692, 13676, 13679 and - maybe - 13724
[15:16:08] <wiesand> 2) finalize 13643
[15:16:17] <wiesand> ?
[15:17:14] <meffie> looking up those numbers :)
[15:18:00] <meffie> 13692 afs: Avoid non-dir ENOENT errors in afs_lookup
13676 aklog: Avoid misleading AFSCELL message
13679 auth: make PGetTokens2 work with 3-char cellnames
13724 libadmin: overlap warning in strcpy with gcc9
13643 Update NEWS for 1.8.4pre1
[15:18:08] <wiesand> I really need some bot to turn those into URLs…
[15:18:45] <wiesand> "where is all that AI when you need it?!"
[15:18:50] <meffie> LOL
[15:19:16] <meffie> those look good too me.
[15:20:02] <wiesand> asking again for objections…
[15:20:03] <wiesand> start…
[15:20:14] <meffie> no objections
[15:20:16] <wiesand> stop…pity…
[15:20:17] <kaduk@jabber.openafs.org/barnowl> no objections here either
[15:21:12] <wiesand> Looks like we're "go" for 1.8.4pre1. Will get it ready to be tagged etc. over the weekend. Thanks!
[15:21:34] <wiesand> On to master/1.9?
[15:22:02] <kaduk@jabber.openafs.org/barnowl> Not too much going on; I tried to look at some of the latest new
changes this morning, but this week has been very crazy for me.
[15:22:56] <kaduk@jabber.openafs.org/barnowl> I want to look at per-chroot sysnames and the rx listener stack "soon"
but have no ETA, and look forward to seeing a revamped rxgk-phase2 to
look at as well
[15:24:02] <yadayada> We tried some testing on Rx listener patches from Andrew. Found some deadlock issue and updated Andrew about it. He has found the fix and will push soon
[15:24:15] <kaduk@jabber.openafs.org/barnowl> good to know; thanks
[15:24:53] <wiesand> My personal favourite is "per-chroot sysnames". Would make containers so much more useful.
[15:26:31] <kaduk@jabber.openafs.org/barnowl> sigh, network stalled for a bit there
[15:26:56] <kaduk@jabber.openafs.org/barnowl> I don't think I have other topics for master.
[15:27:09] <meffie> i think the automount patches are ready for more testing/user as well.
[15:27:18] <meffie> s/user/use/
[15:27:36] <kaduk@jabber.openafs.org/barnowl> Oh, yes, there was a new upload for the automount stuff, too; thanks
[15:28:00] <meffie> some fixes this week for those. more feedback would be helpful.
[15:28:10] <wiesand> I see some "linux-native-mounts" changes were merged recently too - that's a topic wanted badly here, in the hope it will solve those getcwd() problems. It's top on my personal 1.8.5 wish list.
[15:28:45] <meffie> would it be for 1.9.1 first?
[15:28:52] <kaduk@jabber.openafs.org/barnowl> I'm trying to remember if there was anything in that stack that's
pretty invasive/risky
[15:29:12] <kaduk@jabber.openafs.org/barnowl> I think most of it is just adding a new optional codepath that could
just be disabled at runtime if found to be problematic
[15:29:19] <meffie> good point!
[15:29:39] <kaduk@jabber.openafs.org/barnowl> But yeah, would probably be for 1.9.1 first
[15:29:55] <wiesand> If it's ready/suitable for 1.8.x, I really really want it. If just in a pre1 and removed in pre2.
[15:30:08] <meffie> ok
[15:31:11] <meffie> in other news, cheyenne is testing with linux 5.3rc
[15:31:25] <wiesand> and…?
[15:31:29] <meffie> he found a problem during shutdown and is looking into it.
[15:31:41] <kaduk@jabber.openafs.org/barnowl> During *shutdown*, interesting.
[15:31:55] <kaduk@jabber.openafs.org/barnowl> I note that Andrew fixed a few memory leaks in the shutdown path while
looking at freebsd
[15:32:08] <meffie> yes, the signal handling api changed, that we use to kill the listener
[15:32:35] <wiesand> that's both good and bad, but it's very good to hear he's working on it
[15:32:43] <kaduk@jabber.openafs.org/barnowl> Ah
[15:33:26] <meffie> it's not too bad i hope
[15:33:36] <wiesand> "memory leaks in the shutdown path" sounds like a bit of an academic problem to me…
[15:34:45] <meffie> for userspace, true.
[15:35:09] <kaduk@jabber.openafs.org/barnowl> leaks in the kernel module, sorry
[15:36:45] <wiesand> the clients on our systems here are very rarely ever stopped except during system shutdown
[15:37:40] <wiesand> i.o.w. "I don't care about memory being leaked during shutdown"
[15:38:09] <kaduk@jabber.openafs.org/barnowl> I don't really, either (the amounts are pretty tiny), but freebsd
prints some ominous looking warnings to the console
[15:38:49] <wiesand> ok, "i do care about warnings on consoles"
[15:39:04] <wiesand> even if they're not serious
[15:40:13] <wiesand> Anything else to discuss today? In particular: Yadav, do you want to bring something up?
[15:40:29] <yadayada> I was working on one deadlock issue on client side
[15:40:50] <meffie> oh, that sounds serious.
[15:40:51] <yadayada> mainly dput can call dentry_iput which will call OpenAFS afs_dentry_iput
[15:41:28] <yadayada> now if Global lock is taken before dput we retry taking lock in afs_dentry_iput and that can result in deadlock
[15:41:35] <yadayada> Told andrew about it
[15:41:43] <yadayada> and have a fix
[15:41:52] <yadayada> will create gerrit review shortly on this
[15:41:53] <wiesand> nice :)
[15:41:57] <kaduk@jabber.openafs.org/barnowl> > have a fix
that's what I love to hear :)
[15:42:20] <meffie> :)
[15:42:29] <yadayada> that all from my side at present
[15:42:48] <kaduk@jabber.openafs.org/barnowl> Okay, thanks for letting us know
[15:43:26] <kaduk@jabber.openafs.org/barnowl> It sounds like we don't have any other topics for today, so should we
adjourn?
[15:43:34] <wiesand> There was that question regarding the time slot for the meeting….
[15:44:07] <kaduk@jabber.openafs.org/barnowl> Ah, good point
[15:44:38] <wiesand> I think it currently starts 8am for Ben, 11am for Mike and Mark, 5pm for me and 10:30pm for Yadav?
[15:44:43] <meffie> oh, buildbot update?
[15:44:52] <yadayada> for me it is 8:30 PM
[15:45:12] <wiesand> yadav: ah, better :)
[15:45:46] <yadayada> yes this time suits to me
[15:46:27] <wiesand> it's clear that this is isn't perfect for everyone but seems at least close to to a good compromise
[15:46:51] <kaduk@jabber.openafs.org/barnowl> Does anyone want to propose a change?
[15:47:05] <meffie> not i.
[15:47:34] <wiesand> that's what I wanted to ask, maybe with s/propose/desire/
[15:48:53] <kaduk@jabber.openafs.org/barnowl> (I am not inconvenienced by the current time)
[15:51:19] <wiesand> It seems we're fine at least for the time being. Things will become even more interesting than in the past when the US and then Europe change back from DST  successively (while I think India  wisely doesn't participate in that nonsense at all)… let's chat about this when it comes up.
[15:51:45] <wiesand> Mike: please go ahead with buildbot news
[15:52:12] <meffie> ok. 2 items.
[15:53:13] <meffie> 1. the master has been crashing recently. ben took a quick look, thank you ben. for now i'm using bosserver to run the master so if it crashes it will automatically restart
[15:53:46] <kaduk@jabber.openafs.org/barnowl> Our current best hypothesis is that an updated python3 caused a bad
interaction with buildbot
[15:54:16] <meffie> 2. cheyenne  finished his patches for us to do "forced builds" by gerrit number! i'll be trying those when you guys say it is ok to make changes on the buildbot.
[15:54:51] <meffie> i can do it today if you like, or wait until after 1.8.4?
[15:54:55] <wiesand> ok
[15:54:57] <kaduk@jabber.openafs.org/barnowl> I don't personally have strong time demands on the buildbot setup;
stephan, do you?
[15:55:35] <wiesand> being able to do forced rebuilds would have been beneficial yesterday
[15:55:46] <meffie> yes, indeed.
[15:56:59] <wiesand> and if we push out 1.8.4pre1, I'm not going to need buildbot before next week - so "go" for doing it today
[15:57:07] <meffie> ok, thanks!
[15:57:37] <meffie> that's all i have for now. thanks.
[15:57:51] <kaduk@jabber.openafs.org/barnowl> Okay.
[15:57:52] <wiesand> Anything else to discuss today?
[15:58:28] <kaduk@jabber.openafs.org/barnowl> Let's adjourn, then.  Thanks everyone!
[15:58:42] <yadayada> Thanks
[15:58:48] <wiesand> If not, let's adjourn. Yadav, thanks again for joining us. You're most welcome.
[15:59:12] <wiesand> Thanks a lot everybody. Bye.
[16:00:08] wiesand leaves the room
[16:19:25] meffie leaves the room
[16:35:35] yadayada leaves the room
[16:46:06] yadayada joins the room
[17:29:30] yadayada joins the room
[17:46:46] yadayada leaves the room
[17:55:07] yadayada joins the room
[18:03:25] yadayada leaves the room
[18:18:44] mbarbosa leaves the room
[18:18:47] mbarbosa joins the room
[19:51:14] kaduk@jabber.openafs.org/barnowl leaves the room
[19:57:45] mbarbosa leaves the room
[20:00:13] mbarbosa joins the room
[21:04:02] yadayada joins the room
[21:20:01] yadayada leaves the room
[21:25:38] yadayada joins the room
[21:32:34] mbarbosa leaves the room
[21:36:02] yadayada leaves the room
[21:40:12] yadayada leaves the room