Home
release-team@conference.openafs.org
Friday, June 28, 2019< ^ >
Room Configuration
Room Occupants

GMT+0
[01:05:15] mvita leaves the room
[01:05:15] mvita joins the room
[12:04:38] mbarbosa joins the room
[14:59:15] <kaduk@jabber.openafs.org/barnowl> Greetings (a minute early, right?)
[14:59:33] <mvita> hi
[14:59:45] <mvita> thanks for the ping, I was heads down
[15:00:04] <kaduk@jabber.openafs.org/barnowl> I think you may be the popular guy today
[15:00:11] <mvita> ?
[15:00:22] <kaduk@jabber.openafs.org/barnowl> everybody wants to know about clearing the call_queue_lock (maybe)
[15:00:30] <mvita> oh
[15:00:43] <mvita> that old thing
[15:00:53] wiesand joins the room
[15:00:56] <mvita> maybe I shouldn't have pushed it
[15:01:53] <wiesand> no, I'm a minute late ;-)
[15:02:04] <wiesand> Hello
[15:02:06] <mvita> hello Stephan
[15:02:18] <kaduk@jabber.openafs.org/barnowl> good afternoon!
[15:02:21] <mvita> I'm glad you got to catch some of the workshop
[15:02:27] <kaduk@jabber.openafs.org/barnowl> I mean, mostly I'm just remembering how timezones work :)
[15:03:56] <wiesand> Yes, being able to follow much of it was very valuable
[15:04:28] <kaduk@jabber.openafs.org/barnowl> That's good to hear
[15:04:52] <wiesand> Though there was overlap with a local workshop, so at times I was following two presentations at the same time. Or rather trying to.
[15:05:37] <wiesand> And frankly, on the second evening I fell asleep sitting in front of my notebook, so I missed some of that day's content ;-)
[15:06:47] <kaduk@jabber.openafs.org/barnowl> Following two presentations at the same time is ... mostly doable, but
leaves no hope of doing anything else (email, chat, etc.)
[15:08:34] <kaduk@jabber.openafs.org/barnowl> Stephan, do you want to get started with any 1.6/1.8 topics?
[15:08:52] <wiesand> It depends - if most info is actually in the slides, yes, doable. Otherwise it's very hard for me. Listening to two persons talking at the same time is not one of my strengths.
[15:08:57] <wiesand> Yes,
[15:09:09] <wiesand> Mike continued to work on the NEWS for 1.8.4pre1 - some feedback would be nice
[15:10:13] <wiesand> 13599 seems to be making good progress, so I'm still hoping to get that in
[15:11:17] <kaduk@jabber.openafs.org/barnowl> Hopefully Mark has some time to make the requested tweaks
[15:11:27] <mvita> I should get to it today
[15:11:34] <kaduk@jabber.openafs.org/barnowl> Thanks!
[15:12:05] <mvita> my current system for keeping up w/ gerrit comments is not working
[15:12:32] <kaduk@jabber.openafs.org/barnowl> "keeping up with" in terms of seeing when they come in and changing
the code accordingly, or something else?
[15:12:47] <kaduk@jabber.openafs.org/barnowl> I am getting a crash course in the ssh review interface, myself.
[15:12:51] <mvita> I think I need to leave them in my inbox instead of putting them in a separate folder (that I always forget to read)
[15:12:59] <wiesand> That aside, what's open for 1.8 is either reviewed or so trivial that I'm willing to merge w/o.
[15:14:03] <wiesand> And Mark calling 13641 "that old thing" probably means we shouldn't hold 1.8.4pre1 for it?
[15:14:15] <mvita> no, not at all
[15:14:35] <wiesand> ok, so we're really close
[15:14:48] <mvita> I need to update that one as well
[15:15:08] <wiesand> barring unforeseen circumstances, I should get 1.8.4pre1 ready on Sunday
[15:15:25] <kaduk@jabber.openafs.org/barnowl> Yay
[15:15:52] <wiesand> re 1.6.x, work on that is still postponed until "after 1.8.4pre1"
[15:16:03] <mvita> ok
[15:16:24] <kaduk@jabber.openafs.org/barnowl> Eminently defensible
[15:17:30] <wiesand> (I have to find out how the mess came about that made it impossible to rebase changes I think should be trivial to rebase - so there's some real threshold)
[15:18:37] <wiesand> luckily, the seems to be no pressure from the linux kernel side yet
[15:19:15] <wiesand> OTOH I tend to need pressure to get anything done these days…
[15:20:02] <wiesand> Unless you have something to bring up, I think we're finished with the stable series for today.
[15:20:48] <wiesand> On to master/1.9 then?
[15:21:36] <kaduk@jabber.openafs.org/barnowl> We got a report (and patch) about the RPM packaging systemd unit, see
13654 (and, I guess, 13658)
[15:21:40] <wiesand> BTW thanks for the hint how to operate gerrit entirely over ssh. Seems to work fine, though quoting is a little tricky.
[15:21:55] <kaduk@jabber.openafs.org/barnowl> Yeah, I messed up the quoting the first time, too
[15:22:26] <kaduk@jabber.openafs.org/barnowl> About three minutes after I sent the note to -devel, I found the part
of the documentation that says how to do inline comments.
But it involves writing a lot of JSON, so I don't plan to actually do
so.
[15:22:29] <wiesand> Yes, I saw the report and the changes
[15:23:23] <kaduk@jabber.openafs.org/barnowl> It seems reasonable to me, so I was planning to merge it.
I just wanted some eyes on it from someone who actually uses RPM
systems, first.
[15:24:29] <mvita> not I  ;-)
[15:24:35] <wiesand> 13654 is from the author, so in principle that one should be preferred IMO. But it needs some polish…
[15:25:13] <kaduk@jabber.openafs.org/barnowl> Do you think I should apply the polish myself?
[15:25:40] <wiesand> And frankly I'm neither sure whether it really solves the problem nor how normal/usual/frequent that problem is at all.
[15:26:37] <kaduk@jabber.openafs.org/barnowl> I exchanged some IRCs with the submitter; I guess it can cause issues
on first boot after a new client version, since systemd doesn't wait
for the module to be built before starting the client (which then
fails)
[15:27:23] <wiesand> Good question. Some people, like me, don't mind when others push over their changes. Others do.
[15:29:18] <kaduk@jabber.openafs.org/barnowl> Something like " For us, everything is fine until we reboot the
cluster, when systemd tries to mount afs before dkms has finished
compiling the kernel.  A minor annoyance with a simple fix." is what I
learned on IRC
[15:29:39] <wiesand> Understood. But (1) shouldn't the module be built upon update of openafs-dkms, and (2) I doubt that the "After" declaration in the unit file is strong enough to guarantee that the new module is in place in time.
[15:30:52] <kaduk@jabber.openafs.org/barnowl> I don't remember whether the RPMs are set up to build the module at
openafs-dkms install time or not (I thought not, though?), and I'd
have to double-check about the "After" as well, I suppose.
[15:30:58] <kaduk@jabber.openafs.org/barnowl> But if he thinks it works for them...
[15:31:50] <wiesand> IOW I'm not sure it's right. It may well be, and shouldn't do harm, but I'm not going to +1 it w/o having understood.
[15:32:04] <kaduk@jabber.openafs.org/barnowl> I wouldn't ask for anything else :)
[15:32:59] <kaduk@jabber.openafs.org/barnowl> In terms of other news on master, I'm trying to keep up with the
flurry of stuff that came in recently (which is good to see!) and
merged some stuff this morning already.
[15:33:28] <kaduk@jabber.openafs.org/barnowl> It sounds like Andrew is going to rework the linux-automount stuff to
have per-volume superblocks, so that's unlikely to make 1.9.0
[15:33:57] <wiesand> pity!
[15:34:09] <kaduk@jabber.openafs.org/barnowl> Well, there's always 1.9.1!
[15:34:30] <wiesand> ok
[15:34:30] <kaduk@jabber.openafs.org/barnowl> I'm also not super-inclined to wait for the rx listener stuff, though
I should probably at least look at it all first.
[15:35:12] <kaduk@jabber.openafs.org/barnowl> So I think we just need a NEWS and a version bump, and then could roll
out a 1.9.0
[15:35:20] <wiesand> I fully understand your urge to get 1.9.0 out.
[15:35:29] <mvita> yup
[15:36:10] <wiesand> ok, new priorities: 1) 1.8.4pre1 2) work on NEWS for 1.9.0 3) 1.6.x
[15:36:39] <kaduk@jabber.openafs.org/barnowl> Oh, you would be able to put some time into 1.9.0 NEWS?
(I can't remember if Mike was planning to do so as well.)
[15:37:06] <wiesand> we both promised to several weeks ago…
[15:37:11] <kaduk@jabber.openafs.org/barnowl> Ah.
[15:37:29] <mvita> Mike's out today - I think he's just focused on 1.8.4-pre1 news for now
[15:37:38] <mvita> I'm taking a first review look today
[15:37:46] <kaduk@jabber.openafs.org/barnowl> That's a fine thing to prioritize
[15:37:52] <wiesand> Thanks.
[15:38:39] <mvita> Is it okay if I review on the web using Anonymous Coward (1000130)?  That's what I get with my Yahoo openid
[15:39:09] <mvita> I'm not sure what the implications would be of changing that account to read Mark Vitale as well.
[15:39:12] <kaduk@jabber.openafs.org/barnowl> I should be able to figure it out.  Though I could try to merge
accounts if you want
[15:39:28] <kaduk@jabber.openafs.org/barnowl> Changing it to read literally "Mark Vitale" will confuse some things,
IIRC
[15:39:34] <mvita> okay
[15:39:46] <mvita> then I will refrain from doing that
[15:39:56] <kaduk@jabber.openafs.org/barnowl> At least in our previous gerrit version, trying to add you as a
reviewer would show two indistinguishable entries and picking either
one would just give an error
[15:39:59] <mvita> Elativ Kram it is
[15:40:01] <wiesand> if you could add something like "- mark" at the end of the comment, to make it easy to spot who made it, fine y me
[15:40:05] <kaduk@jabber.openafs.org/barnowl> But, "Mark Vitale (alternate)" would be okay
[15:41:37] <wiesand> Getting things reviewed at all is more important than avoiding the string "Anonymous Coward" IMHO.
[15:42:25] <mvita> Ben, in regard to the openid woes:   I also found a report that lack of SNI support on the hosting gerrit webserver can also cause the observed symptoms
[15:42:47] <kaduk@jabber.openafs.org/barnowl> Well, we definitely have SNI support in the apache reverse proxy
[15:43:15] <mvita> okay
[15:45:14] <kaduk@jabber.openafs.org/barnowl> I don't think I have a whole lot else for today.
[15:45:38] <kaduk@jabber.openafs.org/barnowl> I did merge the ticket5 update from heimdal -- tell Andrew thanks for
refreshing that quickly
[15:46:13] <wiesand> You mean my weekend is really close now? :)
[15:46:21] <kaduk@jabber.openafs.org/barnowl> Yes.
[15:46:54] <mvita> I will
[15:47:55] <wiesand> Fine, let's adjourn. Thanks a lot Ben & Mark!
[15:48:02] <mvita> Thank you!
[15:48:09] <kaduk@jabber.openafs.org/barnowl> Thanks everyone!
[15:48:21] wiesand leaves the room
[15:58:00] <mvita> kaduk:  I see the "let's encrypt" cert I get for gerrit.openafs.org is actually labeled for buildbot.openafs.org - could that be a problem?
[15:58:45] <mvita> (stepping away for a bit)
[16:36:56] <kaduk@jabber.openafs.org/barnowl> I don't think that's the problem (and  in fact buildbot, gerrit, git,
jabber, and conference all point to the same machine and use SNI to
dispatch appropriately).
The issue is when gerrit goes forward to the openid provider and acts
as a TLS client -- the java keystore doesn't have the root needed to
verify the cert that launchpad presents to us, so we get the "provider
not properly configured" error.
[20:36:42] mbarbosa leaves the room
[20:36:44] mbarbosa joins the room
[21:19:57] kaduk@jabber.openafs.org/barnowl joins the room
[21:20:55] mbarbosa joins the room
[21:39:14] mbarbosa leaves the room
[22:05:19] mvita joins the room