- Wednesday, March 15, 2017

release-team@conference.openafs.org

Wednesday, March 15, 2017< ^ >

Room Configuration

Room Occupants

GMT+0
[00:50:19] mvita joins the room
[01:01:54] mvita leaves the room
[01:06:00] mvita joins the room
[02:13:48] mvita leaves the room
[12:53:07] mvita joins the room
[13:50:14] meffie joins the room
[14:01:06] jhg joins the room
[14:01:25] <meffie> greetings
[14:01:46] <jhg> hallo
[14:02:42] <kadukoafs@gmail.com/barnowl538BDB78> $timeofday
[14:03:23] <meffie> i believe europe is still on standard time, so we are an hour early
[14:04:01] <meffie> or an hour late. one of those.
[14:06:30] <kadukoafs@gmail.com/barnowl538BDB78> Early, I think.
[15:00:10] wiesand joins the room
[15:00:47] <wiesand> sorry, it escaped me that you're living on that braindamage called dst already
[15:01:14] <kadukoafs@gmail.com/barnowl538BDB78> No worries.
[15:01:18] <mvita> np
[15:01:24] <kadukoafs@gmail.com/barnowl538BDB78> Our brains haven't recovered from the transition yet, anyway.
[15:03:10] <wiesand> My major open question w.r.t. 1.6.20.2 is still whether the chnages to make PAGs wort on  Solaris 11 are likely to break Linux 2.4.
[15:05:07] <wiesand> Ans some of those touch half a dozen files w/ common code but don't have a second +1 yet. I'm hesitant to accept them in this state.
[15:05:07] <kadukoafs@gmail.com/barnowl538BDB78> That's 12525, mostly?
[15:05:32] <wiesand> 112524..7
[15:05:53] <wiesand> s/11/1/
[15:06:14] <kadukoafs@gmail.com/barnowl538BDB78> I think linux 2.4 will continue to function as it does.
[15:07:37] <kadukoafs@gmail.com/barnowl538BDB78> (From looking just now)
[15:07:42] <wiesand> 12525 still carries an objection by Mark
[15:07:49] <wiesand> Thanks Ben
[15:08:21] <kadukoafs@gmail.com/barnowl538BDB78> The issue Mark raises would need to be addressed on master, too, if it
is addressed at all.
[15:09:06] <mvita> I've since learned that very old tests like this (from 2.4/2.6 era) are valid when done by version
[15:09:06] <meffie> maybe it makes more sense for the PAGs refactoring to go in 1.6.21, instead of a "point" release
[15:09:08] <wiesand> Sure, but if it's serious we should wait for that to have happened before shipping the change in a stable release.
[15:09:54] <wiesand> Mike: maybe.
[15:10:06] <mvita> there are a few others scattered throughout the code
[15:10:09] <kadukoafs@gmail.com/barnowl538BDB78> Well, I am not worried about it, myself; if Mark is not, either, then
maybe there is not a real issue.
[15:10:30] <mvita> but all modern tests need to test by feature (autoconf), not by version
[15:11:20] <meffie> yes, but this predates that.
[15:11:30] <meffie> that would need to be a different commit
[15:11:31] <mvita> that's what I'm saying.
[15:11:35] <meffie> ah
[15:12:10] <mvita> I will withdraw my objection.
[15:12:26] <wiesand> Good point. But then2.6.4 is so old…
[15:12:53] <kadukoafs@gmail.com/barnowl538BDB78> 2.6.4 is pre-git, right?
[15:12:58] <meffie> yes
[15:13:18] <wiesand> Ok, looks like we're going to ship it with 1.6.20.2 as planned. It doesn't break my systems (in any obvious way) anyway ;-
[15:13:31] <meffie> great
[15:14:26] <wiesand> RHEL5 is being retired after 10 years of service this month, and it had 2.6.18 kernels…
[15:14:40] <mvita> +1
[15:14:53] <mvita> (on12525)
[15:15:26] <wiesand> Thanks.
[15:15:27] <kadukoafs@gmail.com/barnowl538BDB78> I could probably find someone still running RHEL4 without too much
effort, of course...
[15:15:56] <wiesand> I think the last .2 candidate needing review is 12567 then
[15:16:13] <meffie> i'm sure they would be thrilled to update libafs on rhel4 :)
[15:16:21] <wiesand> Ben: one regularly being update to the latest openafs-1.6 release?
[15:16:30] <kadukoafs@gmail.com/barnowl538BDB78> No, of course not.
[15:16:54] <mvita> thank you for the backport on 12567.  I meant to review that Monday but didn't get to it.
[15:17:02] <mvita> and yesterday was a "snow day"
[15:17:20] <kadukoafs@gmail.com/barnowl538BDB78> What kind of numbers were you seeing?
[15:17:36] <kadukoafs@gmail.com/barnowl538BDB78> Supposedly Boston was going to get 12-18", but I'm not sure I heard
actual measured values.
[15:17:39] <wiesand> ah, here's my new vocabulary to acquire today :)
[15:18:33] <mvita> I am in Allentown PA area - we got 8" - less than the expected 12-18 due to some rain instead of snow
[15:18:38] <wiesand> 12567 was not a clean cherry-pick
[15:18:55] <mvita> Stephan - understood, looking at it now
[15:19:26] <kadukoafs@gmail.com/barnowl538BDB78> "There are more -x03 stanzas than I remember"
[15:20:02] <mvita> yes
[15:20:05] <wiesand> "there are more sunx_*) cases than on master too"
[15:20:19] <mvita> I would take out the ones for older releases
[15:20:30] <wiesand> This was my best guess - could be complete nonsense.
[15:20:50] <mvita> since untested
[15:21:22] <wiesand> Looks they'll all be affected on master though?
[15:22:08] <wiesand> Please feel free to push over my change, or push a new one and I'll abandon mine.
[15:24:01] <wiesand> But if removing those two -x03 lines is all it takes, I can push a new patchset.
[15:24:14] <mvita> oh, you're right, it does apply to them all on master
[15:24:24] <mvita> <gulp>
[15:25:18] <mvita> I still stand by my opinion for 1.6 stable - let's only change it for 5.11
[15:27:37] <wiesand> Does this warrant an addition to the commit message?
[15:28:12] <mvita> I don't think so.  What do the others think?
[15:28:31] <wiesand> (stating the semantic difference between the change and the one it was "cherry picked" from)?
[15:29:09] <mvita> I'm waffling - now I think it's a good idea
[15:29:35] <wiesand> Alternatively, we can merge it as-is and add a 1.6-only change reverting those two lines…
[15:29:42] <meffie> well a line that says "backported from master" should be enough.
[15:29:58] <mvita> ("weisand: restrict -xO3 to 5.11 only")
[15:30:22] <wiesand> It's not a backport. It's a different change.
[15:30:38] <wiesand> (w/o those two additional -X03's)
[15:31:50] <wiesand> I think this one's for the guardian to decide ;-)
[15:31:58] <mvita> oops, misspelled your name, sorry
[15:32:07] <wiesand> np happens all the time
[15:32:33] <mvita> We have a local grocery chain "Weis"
[15:32:56] <wiesand> Not a common name here
[15:33:20] <wiesand> Nor is mine btw.
[15:35:22] <wiesand> Ok, I want the guardian's rule on 12567, and then we could release 1.6.20.2 (after the changes Mark suggested for the NEWS change).
[15:35:39] <wiesand> Looks like the Linux 4.11 changes on master will take more time.
[15:37:48] <mvita> yes
[15:38:01] <mvita> don't hold 1.6.20.2 for those
[15:38:13] <meffie> yes, i think andrew finally convinced jhg we cant check against version numbers
[15:38:48] <meffie> i like bens description, "willy nilly"
[15:38:56] <jhg> yeah, sorry for the delay
[15:39:27] <wiesand> np, we'll have those in 1.6.21 then (or 1.6.20.3)
[15:40:15] <wiesand> Mike: yes, I like that too
[15:40:22] <meffie> wiesand: do you want to be added to the email notifications for the linux-rc builders?
[15:40:39] <wiesand> depends on the traffic ;-)
[15:40:50] <meffie> 4 emails a day.
[15:41:08] <wiesand> ok let's try. thanks
[15:41:15] <meffie> ok, great.
[15:41:33] <wiesand> 4 more don't really matter…
[15:41:56] <mvita> That's the spirit!
[15:42:02] <mvita> ;-)
[15:42:02] <meffie> lol
[15:43:04] <wiesand> Ok, I think we're mostly go regarding 1.6.20.2 "within days"
[15:43:12] <mvita> yay
[15:43:32] <meffie> sweet.
[15:43:35] <kadukoafs@gmail.com/barnowl538BDB78> Whoops, got called away.
[15:43:58] <wiesand> welcome back - there's a question fo you in the scrollback
[15:44:22] <kadukoafs@gmail.com/barnowl538BDB78> Can you use studio 12.5 to build for older things?
[15:44:42] <mvita> I don't know.
[15:46:43] <kadukoafs@gmail.com/barnowl538BDB78> I'm leaning towards only changing 5.11's behavior on the stable
branch, and leaving the older solarisen with unoptimized builds.
[15:47:03] <meffie> seems prudent
[15:47:07] <mvita> aye
[15:48:27] <kadukoafs@gmail.com/barnowl538BDB78> Having Stephan do it with the [wiesand: limit change to solaris 5.11
for stable branch] in the commit message sounds like a fine plan to
me.
[15:48:58] <wiesand> Ok, will push a new ps like this.
[15:50:16] <kadukoafs@gmail.com/barnowl538BDB78> Thanks
[15:50:18] <wiesand> Maybe I'll just work on ti while you discuss 1.8/other matters?
[15:50:39] <kadukoafs@gmail.com/barnowl538BDB78> I don't know that we have much news on 1.8 this week :(
[15:51:17] <kadukoafs@gmail.com/barnowl538BDB78> I have been having trouble trying to set up a test environment with
AFS homedir, as the sshd/etc. in the freebsd base system is being
really reluctant to give me kerberos tickets after login.
[15:51:42] <kadukoafs@gmail.com/barnowl538BDB78> Doesn't seem to matter whether I authenticate via password or GSSAPI;
there aren't credentials stored to a ticket cache.
[15:52:12] <kadukoafs@gmail.com/barnowl538BDB78> Maybe I should break down and (1) use debian, or (2) install sshd and
pam_krb5 from ports
[15:53:13] <mvita> I wish I could help but I have zero experience with the BSDs
[15:53:44] <mvita> I should probably rectify that at some point
[15:53:46] <kadukoafs@gmail.com/barnowl538BDB78> But to reiterate, I think that we need to investigate this issue in RT
about (apparently) system:authuser issues with bogus permissions
getting stuck in the cache, and after that it may be time for a
1.8.0 beta1 (aka pre2).
[15:54:06] <kadukoafs@gmail.com/barnowl538BDB78> This particular issue isn't BSD-specific; I just wanted to get a BSD
test machine up so that I could later on do other testing...
[15:54:18] <mvita> <nod>
[15:54:19] <meffie> can you remind me of the ticket number?
[15:56:01] <kadukoafs@gmail.com/barnowl538BDB78> 133654
[15:56:43] <meffie> thanks
[15:58:00] <kadukoafs@gmail.com/barnowl538BDB78> I still am uncertain what issues Ted Creedon is actually seeing, but
as far as I can tell the scenarios he's talking about have been tested
by us and work.
[15:58:10] <wiesand> 12567 ps2 pushed
[15:59:50] <mvita> did you forget the modified commit msg?
[16:00:04] <kadukoafs@gmail.com/barnowl538BDB78> No?
[16:01:04] <meffie> "[wiesand: limit change to solaris 5.11 for stable branch]"
[16:01:24] <mvita> hmm - first browser refresh didn't pick it up
[16:01:25] <meffie> thanks stephan.
[16:01:52] <kadukoafs@gmail.com/barnowl538BDB78> Yes, thanks you.
[16:01:56] <meffie> ok, have to run off to the next meeting. i can write up notes for -devel
[16:02:15] <kadukoafs@gmail.com/barnowl538BDB78> Thanks, Mike!
[16:02:18] <mvita> +`
[16:02:21] <mvita> +1
[16:02:22] <kadukoafs@gmail.com/barnowl538BDB78> I think we can probably adjourn
[16:02:36] <kadukoafs@gmail.com/barnowl538BDB78> Unless there are other topics?
[16:02:47] <mvita> nothing from me
[16:04:15] <wiesand> new ps for the news change pushed too
[16:04:36] <wiesand> Please have a look, and then we should be able to get out 1.6.20.2 r.s.n.
[16:05:15] <wiesand> Let's adjourn then. Thanks a lot everyone, and sorry again for the late start.
[16:05:33] wiesand leaves the room
[20:23:39] wiesand joins the room
[20:25:54] <wiesand> and btw, even RHEL4 had a 2.6.9 kernel
[20:26:32] <kadukoafs@gmail.com/barnowl538BDB78> Heh.
[20:26:42] <kadukoafs@gmail.com/barnowl538BDB78> "I won't ask about RL4."
[20:27:41] <wiesand> Well that was 2.4 or even earlier...
[20:31:47] <wiesand> 2.4.19 was a very good Linux kernel...
[20:32:17] <wiesand> 1.4.7 was a very good openafs release...
[20:34:04] wiesand leaves the room
[21:21:43] meffie leaves the room