- Wednesday, September 14, 2016

release-team@conference.openafs.org

Wednesday, September 14, 2016< ^ >

Room Configuration

Room Occupants

GMT+0
[00:07:10] Jeffrey Altman leaves the room
[00:14:45] Jeffrey Altman joins the room
[13:14:00] meffie joins the room
[13:52:18] stephan.wiesand joins the room
[14:00:58] mvita joins the room
[14:01:04] <stephan.wiesand> Hello crowd
[14:01:10] <meffie> hello
[14:01:11] <mvita> hi!
[14:01:46] <stephan.wiesand> Any news on linux 4.8? Are we good?
[14:02:12] <meffie> i think rc5 needs to be checked
[14:02:18] <mvita> oh, I need to do that.
[14:02:43] <mvita> hold on, I'll start it now while we talk, sorry, I forgot all about that
[14:03:27] <Jeffrey Altman> auristor is seeing significant client performance hits on linux 4.8 when xfs disk cache is used compared to 4.7.
[14:03:30] <stephan.wiesand> I figure nothing has happened regarding re-enabling splice?
[14:03:48] <stephan.wiesand> Thanks Jeff.
[14:04:39] <stephan.wiesand> Will Auristor be able to fix it?
[14:05:50] <stephan.wiesand> Other 1.6.19 topics:
[14:06:17] <stephan.wiesand> ubik (12283/12289) - seems stuck?
[14:07:10] <stephan.wiesand> re-enable "shake harder" (12362/12363) - seems stuck too?
[14:08:15] <meffie> on-going work for re-enabling splice, no patches yet.
[14:08:36] <stephan.wiesand> 12328 still seems interesting, but nobody is jumping on that one either
[14:08:48] <meffie> that one too.
[14:09:05] <stephan.wiesand> 12365/66 need review too
[14:09:29] <stephan.wiesand> I think these are ready to merge: 12339 12357 12308 12371
[14:09:44] <mvita> joe is scheduled to look at 12328
[14:09:51] <stephan.wiesand> If there are objections to those latter ones, please raise them today
[14:10:04] <stephan.wiesand> Mak: good
[14:10:56] <stephan.wiesand> 12387 is a brand new build fix, probably straightforward
[14:12:03] <stephan.wiesand> Did I miss anything else headed for 1.6.19?
[14:13:05] <meffie> i dont think so.
[14:17:08] <stephan.wiesand> I'll have a bit more time to spend on openafs the next two weeks.
[14:17:20] <stephan.wiesand> It would be nice to have pre1 done by then.
[14:17:23] <meffie> excellent news. thanks.
[14:17:36] <stephan.wiesand> (vacation...)
[14:17:36] kadukoafs@gmail.com/barnowl671FF1DB leaves the room
[14:17:36] kadukoafs@gmail.com/barnowl671FF1DB leaves the room
[14:18:08] kadukoafs@gmail.com/barnowlAEEFA31A joins the room
[14:18:15] <stephan.wiesand> NB right now my Mac's running on battery, so this meeting has a natural end ;-)
[14:18:24] <kadukoafs@gmail.com/barnowlAEEFA31A> Maybe I can send now...
[14:18:28] <kadukoafs@gmail.com/barnowlAEEFA31A> Mark did some review on the StaleVCache thing, which is probably in
good shape but I wasn't awake enough to do a final review last night.
[14:18:30] <stephan.wiesand> 75% cahrge right now
[14:18:38] <stephan.wiesand> Ben; yes you can
[14:19:12] <stephan.wiesand> Is the StaleVCache thing meant to land on 1.6.x eventually?
[14:19:24] <mvita> I would like that.
[14:19:34] <meffie> yes
[14:19:54] <kadukoafs@gmail.com/barnowlAEEFA31A> There was some other progress, too, though I forget exactly what.
[14:19:58] <stephan.wiesand> What does it doe, for dummies?
[14:20:36] <stephan.wiesand> Saw it. Thanks everyone.
[14:21:05] <mvita> re: the CVInit CVFlush issue
[14:22:08] <kadukoafs@gmail.com/barnowlAEEFA31A> So we should probably dig up the list of blockers and see where we are
;)
[14:22:10] <mvita> a version of StaleVCache patches is currently in use at an AFS site, and they discovered a bug
[14:22:29] <kadukoafs@gmail.com/barnowlAEEFA31A> re CVInit|CVFlushed, you saw my follow-up commit?
[14:22:53] <mvita> afsd went into 100% cpu because afs_lookup kept trying to use a vcache that was CVFlushed
[14:23:10] <mvita> it was not being removed from dnlc, so afs_lookup continued to use it
[14:23:28] <mvita> agreed that CVInit is moot
[14:23:51] <mvita> but I have seen a field example of CVFlush needing to e removed from dnlc
[14:24:04] <kadukoafs@gmail.com/barnowlAEEFA31A> Hmm, interesting.
[14:24:39] <kadukoafs@gmail.com/barnowlAEEFA31A> Thanks for testing!
[14:25:02] <mvita> I backported the latest stalevcache ps7 fixes for them and they will be testing soon - I'll let you know how it goes
[14:26:35] <stephan.wiesand> I guess we're already in 1.8 land
[14:26:42] <kadukoafs@gmail.com/barnowlAEEFA31A> And meffie saw my comment on the afsd -help crash fix
[14:26:48] <kadukoafs@gmail.com/barnowlAEEFA31A> Er, yeah, sorry.
[14:27:13] <stephan.wiesand> np
[14:27:24] <kadukoafs@gmail.com/barnowlAEEFA31A> What's the difference between 12387 and 12388?
[14:27:55] <meffie> yes, thanks ben. i'll fix the afsd_fuse one too. (anyone have an afs fuse howto handy ?)
[14:28:09] <mvita> "off by 1"?
[14:28:14] <stephan.wiesand> change-id
[14:28:29] <kadukoafs@gmail.com/barnowlAEEFA31A> (No fuse howto here, no)
[14:29:08] <stephan.wiesand> abandoned 12387
[14:29:17] <mvita> meffie - afsd.fuse yes
[14:29:43] <stephan.wiesand> this is what happens when I don't do this stuff every once in a while
[14:29:59] <meffie> i do use the debian afs fuse package somewhere...
[14:31:08] <stephan.wiesand> 69%
[14:33:50] <meffie> stephan.wiesand: can 11978 and 11382 be considered for 1.6.19?
[14:34:17] <stephan.wiesand> 978 or 897?
[14:34:19] <meffie> and 11855?
[14:34:50] <meffie> afs: Make ONEGROUP_ENV not Linux-specific
[14:35:13] <meffie> SOLARIS: Use AFS_PAG_ONEGROUP_ENV for Solaris 11
[14:35:24] <meffie> those would be nice for 1.6.x for solaris users.
[14:35:56] <stephan.wiesand> they touch a lot of generic code though
[14:36:32] <meffie> otherwise pags don't work on solaris 11.x
[14:37:05] <meffie> er, as of 11.2 (11.3 is out now)
[14:37:24] <stephan.wiesand> I'll check for clashes with what's already in the pipeline
[14:37:33] <meffie> ok thanks!
[14:37:43] <stephan.wiesand> so, "request noted"
[14:40:19] <meffie> thank you.
[14:40:21] <stephan.wiesand> 11855 is related to solaris pags?
[14:40:49] <meffie> not pags, but yes solaris.
[14:42:02] <jhg> stephan.wiesand: splice is getting closer to a patch
[14:42:31] <stephan.wiesand> good news, thanks
[14:43:16] <stephan.wiesand> 63%... anything else to discuss today?
[14:43:55] <meffie> oh, your battery is 63%. i get it now.
[14:44:09] <stephan.wiesand> 62%
[14:44:30] <stephan.wiesand> (had to turn up the brightness)
[14:46:15] <stephan.wiesand> Looks like we're done for today?
[14:46:33] <mvita> I have nothing further
[14:47:22] <stephan.wiesand> Let's call it a meeting then. Thanks a lot for participating!
[14:47:44] <stephan.wiesand> And to cite Ben: "review review review"
[14:47:51] <meffie> thanks. good day.
[14:47:52] <kadukoafs@gmail.com/barnowlAEEFA31A> :)
[14:47:58] <stephan.wiesand> Bye
[14:48:00] <meffie> will do.
[14:48:02] stephan.wiesand leaves the room
[14:58:41] meffie leaves the room