[03:53:55] --- wiesand has become available
[03:55:02] <wiesand> test
[04:10:36] --- wiesand has left
[04:40:53] --- wiesand has become available
[05:35:23] --- meffie has become available
[06:53:14] --- Marc Dionne has become available
[07:01:26] <wiesand> Good morning
[07:01:46] <Marc Dionne> good morning
[07:02:11] --- deason has become available
[07:02:19] <wiesand> I see Jeff, Marc and Mike in my client. Is that correct? Oh, and Andrew.
[07:02:59] <deason> I also see ben and 3 shadows, but I don't know if they're there :)
[07:03:09] <kaduk@jabber.openafs.org/barnowl> mornin'.
[07:04:03] <wiesand> Hello Ben. Don't see you (nor the shadows) - glad you're here.
[07:04:35] <wiesand> 1st topic on the agenda: Linux. Marc, any news?
[07:05:06] <Marc Dionne> tested current 1.6 head against current kernel mainline this morning - no problem
[07:05:36] <wiesand> Thanks. Next topic: problem reports. Any news, or any new ones I've missed?
[07:06:05] <wiesand> (mostly a question to Andrew I believe)
[07:06:17] <deason> tracy has a workaround patch for the nbsd thing
[07:06:33] <deason> she said it's been working for... a few days now, I think
[07:06:53] <wiesand> Good. Something we might want for 1.6.8?
[07:07:22] <deason> I'm still not completely sure what was triggering it, but it'll need a more robust patch for 1.6
[07:07:37] <deason> I wouldn't wait for it, but I'll try to get it in time if I can
[07:07:53] <wiesand> Ok.
[07:08:15] <wiesand> No news on the multiple mounts issue I guess?
[07:09:34] <deason> I was looking into reproduction cases a bit more; I can get the newer kernels to crash easily, but not with an 'rm' instead of an 'rmdir'
[07:10:05] <deason> but I don't really know if that matters, since we know this stuff needs to change; so I started implementing the bind-mount thing, but it's not done yet
[07:10:33] <wiesand> Ok, let's see how it goes.
[07:11:24] <wiesand> And I see Mark is on that vos release -stayonline problem, but I think I haven't seen anything in gerrit for that yet.
[07:11:25] <deason> but I think I am supposed to be mentioning the various possible solutions on -devel and -info, and just say that I'm trying "bind mounting" for now
[07:11:51] <wiesand> I think that would be good. -devel may suffice?
[07:12:54] --- shadow has become available
[07:13:17] <wiesand> Hello Derrick.
[07:13:24] <deason> I think both are helpful; they're kinda two different things: talking about actual technical issues, and opinions on user-visible differences
[07:13:44] <wiesand> Just in time for the next topic: the 1.6.7 security release. is there an ETA when that will be ready?
[07:13:50] <shadow> I can't send from my other client 
[07:14:07] <wiesand> Reading you now.
[07:14:33] <wiesand> Andrew: fine.
[07:15:09] <shadow> The security patch exists, so basically it's just a matter of getting the application in place when you are ready 
[07:15:46] <wiesand> Me?
[07:16:24] <wiesand> So far I haven't been involved in security releases at all.
[07:16:38] <shadow> You. When you decide 1.6.8 is otherwise final
[07:16:49] <deason> I thought 1.6.7 was coming before 1.6.8
[07:16:55] <shadow> It is
[07:17:17] <shadow> This would be easier if I had a real keyboard. Let me try something else
[07:17:19] <kaduk@jabber.openafs.org/barnowl> "in close succession"
[07:17:26] <deason> I don't see how 1.6.7 could be blocking on stephan
[07:17:30] --- Daria has become available
[07:17:50] <wiesand> Right. Why wait with 1.6.7? I expected we just pull up the change afterwards, and go on with 1.6.8 as usual.
[07:17:59] <Daria> the presumption being that we release 1.6.7 (with just the security fix) and 1.6.8 (with everything) at the same time
[07:18:32] <deason> yeah, I understand that's what you were saying, but I don't understand why
[07:18:32] <Daria> we've done so previously. i assumed we were close enough to 1.6.8 that it didn't make sense to offer two rapidfire stable releases
[07:19:01] --- shadow has left
[07:19:11] <deason> I thought we had estimated 1.6.8pre1 for a _month_ from now
[07:19:13] <Daria> if you would prefer we get 1.6.7 done, we can do it that way
[07:19:20] <wiesand> I believe 1.6.8 is at least two months away.
[07:19:37] <Daria> ok. then i guess we should just get on it. i'll talk to simon later today
[07:19:46] <Daria> (since he's not online now)
[07:20:12] <Daria> (please tell me i'm the only one getting everything i say 4 times)
[07:20:13] <wiesand> Ok. I don't mind doing them almost simultaneously, but see no reason why that's important.
[07:20:35] <deason> I'm only seeing your messages once
[07:20:35] <wiesand> you're the only one getting everythig yous say 4 times
[07:20:55] <wiesand> BTW, who are you? ;-)
[07:21:02] <deason> it's "shadow"
[07:21:10] <wiesand> daria = shadow?
[07:21:48] <deason> yeah
[07:21:53] <Daria> there. does that help?
[07:21:56] <wiesand> Ok.
[07:22:00] <deason> (you are showing up as "daria"; not sure if you knew that)
[07:22:06] --- Daria has left
[07:22:15] --- shadow has become available
[07:22:30] <deason> now you're "shadow"
[07:22:33] <meffie> (nick is shadow)
[07:22:45] <shadow> wait, who's nick?
[07:22:57] <kaduk@jabber.openafs.org/barnowl> nick's on first?!
[07:23:01] <deason> ha
[07:23:04] <deason> moving on?
[07:23:04] <meffie> heh
[07:23:05] <wiesand> Those questions are my part here.
[07:23:20] <wiesand> Ok, let's move on. Changes...
[07:23:35] <shadow> anyway, like i said, i'll talk to simon. so, we can move on to whatever's next
[07:24:16] <kaduk@jabber.openafs.org/barnowl> "open changes"
[07:24:44] <deason> actually wait, hold on
[07:24:47] <wiesand> Right. The 3 GUACB changes seem ok now, but need review. 
[07:24:52] <deason> just going back to "linux mtpt" for just a sec
[07:25:16] <wiesand> Fine, let's go back to that.
[07:25:19] <deason> are we going to issue any kind of announcement or warning or anything? like, a warning if people upgrade their RHEL clients to the relevant kernel version
[07:26:08] <wiesand> What were the  relevant kernel versions?
[07:26:15] <deason> obviously we don't have a quick fix right now, and I believe it's not even certain if it will be remedied for 1.6.8
[07:26:32] <wiesand> even unlikely ;-)
[07:26:53] <deason> 2.6.18-367+ and 2.6.32-408+
[07:27:06] <deason> I'm not sure at the moment what update numbers those are, but I'd want to mention that, too
[07:27:18] <deason> (i.e. RHEL5.8, 5.9, or whatever)
[07:27:51] <shadow> i think we should
[07:28:37] <wiesand> That would be 5.10 and 6.5, but the releases you mention seem to be from internal betas?
[07:29:05] <deason> those numbers come from red hat changelogs
[07:29:12] <wiesand> 5.10 = -371, 6.5 = -431
[07:29:28] <deason> so yeah, I assume they were inbetween actual red hat releases
[07:29:51] <wiesand> Ah. So it's 5.10 and 6.5. Or: any current security update for the kernel.
[07:30:21] <wiesand> And yes, I think that's a good idea. Can you do it?
[07:30:57] <deason> where does it go? or I suppose I send a draft to release-team?
[07:31:29] <wiesand> Draft is fine. Right list is -announce IMO.
[07:31:54] <deason> okay, you'll see a draft sent to release-team; anything further can probably be discussed there
[07:31:57] <deason> so, moving on
[07:32:07] <deason> (sorry for the interruption :)
[07:32:28] <wiesand> np. Ok: changes... (open changes...)
[07:33:58] <wiesand> Any objections to the GUACB ones? Any new thoughts on the issue altogether? (*new*)
[07:34:40] <deason> I still don't agree with it, but I agree with what you said about not wasting more time on it :)
[07:34:45] <kaduk@jabber.openafs.org/barnowl> Haven't looked at the GUACB changes at all.
"Gosh, those numbers are far from each other."
[07:35:55] <wiesand> Yes, that's what we get for letting changes linger for years.
[07:36:58] <wiesand> Anyway, any review much appreciated.
[07:37:24] <wiesand> About the vnodeunique rollover: how urgent is this?
[07:37:37] <meffie> it is not urgent
[07:38:03] <wiesand> Risk?
[07:38:06] <shadow> … unless it's happening to you
[07:38:25] <meffie> it's been lingering for years :)
[07:38:55] <deason> if it's broken, it only affects volumes that have a large number of vnodes
[07:39:03] <deason> and the behavior is already broken
[07:39:59] <wiesand> OK, I think I'll take them unless a negative review comes up in the next week.
[07:40:31] <deason> (or I guess technically not a large number of vnodes, but a large number of vnodes that have ever been created, whatever)
[07:40:38] <wiesand> 10744 is the last of the Mavericks build slave changes I believe.
[07:41:18] <wiesand> Getting that reviewed would be good.
[07:42:17] <wiesand> About the logging improvements in 10756/7... Ben, why's the latter one a problem but not the former?
[07:42:32] <kaduk@jabber.openafs.org/barnowl> <one moment while I quote things>
[07:42:41] <wiesand> (any thoughts from others?)
[07:44:27] <shadow> sorry. web browser wedged.
[07:44:28] <deason> I could change the messages to maybe be more similar to what's already there, but these messages have not stayed the same before anyway
[07:45:40] <kaduk@jabber.openafs.org/barnowl> 10756 is taking a ~sentence and adding some extra information onto the
end of it, with details of where/what things are happening.
10757 is taking something like "CallPreamble: Couldn't get CPS for
client] from host %s:%d, railing request]" (where the square brackets
part is new from 10756), and turning it into "Cannot get CPS for
client while handling request from host %s:%d viceid %d [...]", in
particular "CallPreamble:" is gone, and "Couldn't" turns into
"Cannot".
I expect that there are scripts scraping the logfile for
"CallPreamble: Coudln't get CPS", because I've seen sites run into
that with nasty consequences.
[07:46:02] <shadow> the first patch adds additional information at the end. the second changes things entirely
[07:46:20] <kaduk@jabber.openafs.org/barnowl> I guess I don't know the history of how often these messages change,
so if Andrew is sure that they've changed a lot already, I can
withdraw the objection.
[07:47:08] <deason> I don't mind altering them to be more similar, but I don't think they can stay that way forever
[07:47:47] <deason> (that is, with a new change; submitted to master and pulled back, but it's simple, just changing constant strings)
[07:48:05] <kaduk@jabber.openafs.org/barnowl> Anyway, we've done some agonizing about changing logged messages from
the MIT KDC, for similar reasoning.
[07:49:31] <wiesand> Sigh. One more of those...
[07:50:32] <wiesand> Wait for something less likely to break existing scripts on master? Bring it up on -info?
[07:51:10] <wiesand> I do understand the reservations.
[07:51:19] <deason> I'll submit a change to make them more similar, it's fine
[07:52:16] <deason> and anything more with that can be discussed there, so moving on... :)
[07:52:31] <wiesand> Ok, let's get back to it then. For the time being, 10756 seems ok to accept.
[07:53:26] <wiesand> So will have to rebase 10758 (move quota calculation), bot otherwise that one seems ok too.
[07:54:33] <wiesand> 10759: not even reviewed by Andrew yet, who pulled it up. Not important?
[07:56:01] <wiesand> If there are objections to 10810/11, please speak up.
[07:56:35] <shadow> *crickets*
[07:56:42] <kaduk@jabber.openafs.org/barnowl> 10759 seems on the surface to be a minor performance aid.
[07:57:16] <wiesand> Ah, dumb question time. Circkets?
[07:57:49] <deason> well, it can be less "minor" depending on the situation, but for "most" sites not noticeable
[07:58:06] <meffie> crickets means "silence" 
[07:58:12] <deason> but yeah, I've got nobody yelling at me that it needs to go in _now_
[07:58:14] <shadow> crickets: you hear them chirping when there is silence.
[07:59:20] <wiesand> Thanks :)
[08:00:01] <wiesand> Andrew: do you still want to review it before it goes in? And the same question for 10812/3/4
[08:01:26] <deason> you don't have to wait for me if you think you have enough +1s in general
[08:01:48] <deason> I'll get to them sometime, though
[08:02:10] <deason> the "error codes" ones I think are more important, though, because that fixes confusing/wrong error messages for end-users
[08:02:32] <wiesand> Ok. Just wasn't sure what to make of the fact that you tend to +1 almost everything you pull up.
[08:03:19] <deason> it just indicates approval after review, same as (or similar to) anyone else
[08:04:22] <wiesand> Ok, I think they can go in. Had test builds with those running for a while (like for most others).
[08:05:14] <wiesand> 6266 is another pretty old one where a second +1 would be good.
[08:05:26] <wiesand> Then the whole feature could be merged.
[08:05:52] <deason> yeah, I will get to it
[08:06:31] <wiesand> The last ones that kind of block things for the moment are 10770/1/2.
[08:10:30] <wiesand> Thanks a lot for all review that happened during the meeting.
[08:10:51] <wiesand> So, anything else to discuss today?
[08:12:22] <shadow> i have nothing.
[08:12:54] <meffie> nothing
[08:13:34] <wiesand> So, probably not. Thanks a lot everyone!
[08:14:25] <wiesand> Bye for now.
[08:14:27] --- Marc Dionne has left
[08:14:33] --- wiesand has left
[08:14:34] --- shadow has left
[08:14:45] --- meffie has left
[08:14:59] --- Marc Dionne has become available
[08:15:26] --- Marc Dionne has left
[09:30:07] --- deason has left
[09:30:26] --- deason has become available
[13:27:00] --- deason has left
[13:32:01] --- deason has become available
[14:28:34] --- deason has left
[22:11:18] --- Jeffrey Altman has left: Disconnected
[22:18:16] --- Jeffrey Altman has become available