Home
release-team@conference.openafs.org
Wednesday, June 14, 2017< ^ >
Room Configuration
Room Occupants

GMT+0
[03:00:30] mvita leaves the room
[03:29:29] mvita joins the room
[03:54:52] mvita leaves the room
[12:30:55] meffie joins the room
[13:25:30] mvita joins the room
[14:59:38] wiesand joins the room
[15:00:57] <wiesand> Good morning
[15:01:03] <meffie> good day.
[15:01:18] <kadukoafs@gmail.com/barnowl2B662C26> hi
[15:01:33] <meffie> running 1.6.21pre1 on my desktop now.
[15:03:06] <wiesand> Great.
[15:03:06] <wiesand> Any other test results yet?
[15:03:17] <wiesand> Ben, thanks for the tag.
[15:03:30] <meffie> not yet
[15:03:40] <mvita> oops, sorry
[15:04:26] <wiesand> I figure there's a good chance that we'll release 1.6.21 as is, in time for Linux 4.12
[15:04:58] <wiesand> Provided that we're still good with the latest rc - any news non that front?
[15:05:14] <meffie> wiesand: in other news we have some "mock on kvm/qemu" setup now to build kmods for all the known versions of rhel6, rhel7 we can find.
[15:05:26] <meffie> e.g. http://download.sinenomine.net/openafs/bins/1.6.20.2/rhel6/x86_64/
[15:05:53] <wiesand> Once more, great!
[15:06:03] <meffie> i can do these for pre-releases too if you want.
[15:07:17] <kadukoafs@gmail.com/barnowl2B662C26> Yes, that's really exciting
[15:07:36] <wiesand> It won't harm. The question is how to make folks find them
[15:08:14] <meffie> well, we could copy them to g.c.o or just publish a link on the wiki.openafs.org?
[15:09:12] <mvita> post to openafs-info mailing list?
[15:10:17] <wiesand> Mark: sounds good
[15:10:19] <meffie> (fyi, those are in /afs, of course.)
[15:10:37] <wiesand> of course ;-)
[15:11:21] <wiesand> NB do you still have those linux-daily/rc builders running?
[15:12:19] <wiesand> I liked those two mails a day telling me that we're still good, but I seem not to receive those any longer
[15:12:31] <wiesand> Are they eaten by our spam filter?
[15:13:42] <mvita> yes, they are running.
[15:13:45] <meffie> yes they are running.
[15:13:56] <mvita> http://buildbot.openafs.org:8010/builders/linux-daily-x86_64-builder
[15:14:33] <meffie> hmm, maybe i broke the email notices while working on the buildbot. sorry! i will look into it.
[15:14:33] <wiesand> Next topic: Provided that no more work is required to cope with Linux 4,12 and we do release 1.6.21 without further changes in ~3 weeks...
[15:15:09] <wiesand> … the plan is to release an an early 1.6.22pre1 with those ubik fiixes
[15:15:22] <meffie> great
[15:15:29] <kadukoafs@gmail.com/barnowl2B662C26> Why wait a whole 3 weeks? ;)
[15:15:57] <wiesand> Looks like there hasn't been much movement there since jhutzs comments?
[15:16:07] <kadukoafs@gmail.com/barnowl2B662C26> Sorry :(
[15:16:17] <kadukoafs@gmail.com/barnowl2B662C26> It was a busy week for me
[15:16:18] <wiesand> Ben: the next 3 weeks will fly by even faster than the last 3 ;-)
[15:17:13] <wiesand> What are the obstacles to having those ubik changes ready by the time of the 1.6.21 release?
[15:18:33] <wiesand> NB I filed a bug in RT today. Nothing serious, but something is clearly fishy.
[15:18:38] <mvita> the logging changes are relatively independent and could go as soon as they are merged on master.
[15:18:48] <kadukoafs@gmail.com/barnowl2B662C26> I think, given jhutz and Andrew's review, I just need some time to
actually look at the changes in depth/context and do my review.
[15:19:23] <wiesand> Ben: thanks for your judgement.
[15:19:28] <meffie> we have several large/busy sites running with 12592. one for about 3 weeks now on about a dozens db servers.
[15:20:02] <mvita> Stephan's RT https://rt.central.org/rt/Ticket/Display.html?id=134320
[15:20:21] <meffie> > NB I filed a bug in RT today. Nothing serious, but something is clearly fishy.
yes, saw that.
[15:20:37] <mvita> Mike, do you get notified on new tickets?
[15:20:39] <meffie> i wonder if this is another reason for afsd --volume-ttl
[15:21:04] <meffie> use F5 for my notifications
[15:21:40] <wiesand> Mark: no objections to including other changes in 1.6.22 (in particular, we'll likely need something for Linux again), but the major objective is releasing a pre1 with those ubik changes really early.
[15:22:00] <mvita> ah
[15:22:43] <mvita> the logging changes would come in handy if there _were_ problems with the other ubik changes
[15:23:55] <wiesand> Noted
[15:24:13] <mvita> re: your ticket - I wonder why the fs get-caps isn't timing out
[15:24:38] <wiesand> interesting, isn't it?
[15:24:57] <wiesand> (not that I'd really know what get-caps does)
[15:25:10] <mvita> I also wonder why the clients didn't see the callbacks revoked (ICBS3, I think?) when the fileservers were shut down
[15:25:40] <wiesand> [mumbles] that were 1.4 servers
[15:25:49] <mvita> oh.
[15:25:53] <meffie> heh
[15:25:56] <mvita> that might make a difference.
[15:26:17] <kadukoafs@gmail.com/barnowl2B662C26> Is there a NAT or firewall in the way that might be timing out an
association?
[15:26:25] <wiesand> no
[15:26:34] <mvita> how about the clients?  what release are they (in general)
[15:26:53] <wiesand> 1.6.20 as said in the ticket
[15:26:55] <mvita> oh, drr, you said in the tx
[15:26:57] <mvita> sorry
[15:27:45] <wiesand> they have a local firewall, but they also have those "ipt_recent" iptables rules I presented in Rome
[15:28:44] <wiesand> (when they send a packet to 7000/udp, the recipient can send back to 7001/udp for 24h)
[15:30:22] <wiesand> As I said, it's not a big deal
[15:31:01] <wiesand> Unlike in the 1.6.0 days when those ping storms were a very real problem
[15:31:33] <meffie> i wonder if fs checkv and/or fs checks would clear the afs_servers ?
[15:31:35] <wiesand> I think that's all I had on the agenda regarding 1.6 for today
[15:32:22] <wiesand> Mike: I'll try
[15:32:47] <meffie> (although, yes, it is odd)
[15:33:55] <wiesand> Ben, want to talk about 1.8?
[15:34:39] <meffie> i tried again to duplicate 133654, no luck.
[15:35:00] <meffie> perhaps an fstrace would help?
[15:35:54] <mvita> stephan, re: your ticket:  fileservers have sent ICBS3 since IBM 1.0, so that's not it.
[15:36:30] <wiesand> ICBS3?
[15:36:51] <kadukoafs@gmail.com/barnowl2B662C26> I don't have anything new for 1.8 other than Mike's comment in RT.
[15:37:14] <mvita> InitCallBackState3 - it's an RPC that fileservers send to clients
[15:37:39] <mvita> clears all callbacks
[15:38:10] <wiesand> ah right, that "give up callbacks on shutdown" we introduced after quite a bit of discussion was the other way around - thanks
[15:38:12] <meffie> i'm not sure if we will be able to get fstraces. i think i will try to ask more people to try the alpha client.
[15:38:37] <wiesand> I'm feeling guilty...
[15:38:44] <mvita> ?
[15:40:40] <wiesand> of not testing the 1.8 alpha
[15:40:52] <meffie> ah, dont feel bad.
[15:41:28] <wiesand> Looks like we're finished for today?
[15:41:39] <kadukoafs@gmail.com/barnowl2B662C26> I think so.
[15:41:46] <kadukoafs@gmail.com/barnowl2B662C26> Thanks everybody!
[15:41:57] <meffie> thanks, have a good day
[15:42:02] <wiesand> Let's adjourn then. Thanks folks!
[15:42:12] wiesand leaves the room
[21:08:19] meffie leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!