- Friday, July 27, 2018

release-team@conference.openafs.org

Friday, July 27, 2018< ^ >

Room Configuration

Room Occupants

GMT+0
[03:35:00] mvita joins the room
[05:49:18] mvita leaves the room
[11:55:19] Marcio Barbosa joins the room
[12:55:58] meffie joins the room
[12:58:53] mvita joins the room
[12:59:51] wiesand joins the room
[13:00:25] <mvita> hello
[13:00:33] <kaduk@jabber.openafs.org/barnowl> greetings
[13:00:46] <wiesand> Hello
[13:01:17] <wiesand> Looks like Ben is working on 1.8.1 (or pre2?)
[13:01:28] <kaduk@jabber.openafs.org/barnowl> That's the question, yes
[13:01:49] <meffie> good morning
[13:01:59] <kaduk@jabber.openafs.org/barnowl> Hi Mike
[13:02:37] <kaduk@jabber.openafs.org/barnowl> I'm still looking at 13191, but even 13235 might be grounds for a pre2
(so I held off merging it until we could talk about it)
[13:02:37] Marcio Barbosa leaves the room
[13:03:25] <kaduk@jabber.openafs.org/barnowl> I'll also ask if anyone wants to get the ubik logging/cleanup stuff
into 1.8.1, or if 1.8.2 is fine for those.
[13:04:31] <mvita> 1.8.2 is fine for those
[13:04:34] <meffie> i feel 1.8.2 would be fine. what do you thing mark?
[13:04:44] <mvita> (unless we have an ubik problem in 1.8.1)
[13:04:47] <kaduk@jabber.openafs.org/barnowl> haha, "race condition"
[13:04:49] Marcio Barbosa joins the room
[13:04:49] <mvita> ;-)
[13:04:51] <meffie> heh
[13:05:05] <meffie> (and s/thing/think/)
[13:06:13] <mvita> full disclosure:  I get to close a ticket when they reach 1.6.x
[13:06:28] <mvita> 8-)
[13:06:59] <kaduk@jabber.openafs.org/barnowl> Before we get too excited about 1.8, Stephan, did you have any 1.6
topics for us?
[13:07:37] <wiesand> mumblemumble
[13:08:34] <kaduk@jabber.openafs.org/barnowl> Indeed.
[13:08:52] <wiesand> So, no. Or maybe 13234 - how critical is that?
[13:09:56] <kaduk@jabber.openafs.org/barnowl> I don't *think* it's critical.  Presumably Auristor clients don't like
some of our error returns, but hopefully error returns are not common.
[13:10:36] <mvita> it could have been kafs as well...
[13:10:36] Marcio Barbosa leaves the room
[13:10:49] <wiesand> It looks fairly innocuous though.
[13:11:44] <wiesand> Let's get excited about 1.8 then?
[13:12:21] <meffie> looks like the openafs windows client also checks for InterfaceVersion == 0x1
[13:12:28] <kaduk@jabber.openafs.org/barnowl> heh
[13:12:33] <kaduk@jabber.openafs.org/barnowl> Woo!  1.8!!!!!
[13:12:40] Marcio Barbosa joins the room
[13:12:41] <meffie> :)
[13:12:42] <wiesand> Is there anything left to do on our side regarding linux-next once that interface we use i snon-GPLONLY again?
[13:13:08] <mvita> not that we know of
[13:13:35] <wiesand> nb any news on this?
[13:13:55] <meffie> that patch hit linux-next, not sure who to bug next
[13:14:27] <mvita> The Linux upstream process is opaque to me, but I believe that patch should be in -rc7 this weekend
[13:15:11] <wiesand> Let's hop it's not too late in the cycle…
[13:15:48] <meffie> yes. at least it is clearly a regression.
[13:15:50] <wiesand> hope
[13:15:53] <mvita> I did get an update on it yesterday, but I didn't understand what new information (if any) the note was providing
[13:15:56] <meffie> so that helps our cause.
[13:16:20] <meffie> oh, was the info on one of the linux mailing lists?
[13:17:38] <mvita> it looks like mm-commits was the list
[13:17:59] <kaduk@jabber.openafs.org/barnowl> link?
[13:18:31] <mvita> oh, look, there's a link here:
[13:18:34] <mvita> Link: http://lkml.kernel.org/r/153128611970.2928.11310692420711601254.stgit@dwillia2-desk3.amr.corp.intel.com
[13:19:13] <mvita> I think that means it is in an mm branch for eventual merging to the mainline linux tree
[13:19:14] Marcio Barbosa leaves the room
[13:19:46] <mvita> also, it looks like the patch numbering has changed again, but that's not important
[13:20:52] <kaduk@jabber.openafs.org/barnowl> Hmm, so looks like Andrew Morton is supposed to take it and give it to
Linux.
[13:21:14] <mvita> yes, akpm is one of the main merge sources to linus
[13:21:21] Marcio Barbosa joins the room
[13:22:47] <mvita> ah, I see, that's a link to the previous correspondence
[13:24:21] <meffie> yes, that's right, changes flow up to linus from andrew morton (and a few others)
[13:24:39] <mvita> found it in the mm-commits archives:
[13:24:41] <mvita> https://marc.info/?l=linux-mm-commits&r=1&b=201807&w=2
[13:24:45] <meffie> i dont see the fix in linus's tree yet.
[13:25:23] <mvita> right, it wasn't in linus' tree when I checked yesterday
[13:25:42] <mvita> and you just saved me the trouble of checking today...
[13:25:42] <kaduk@jabber.openafs.org/barnowl> Hmm, so mostly we are just waiting, I guess
[13:25:47] <mvita> aye
[13:26:19] <mvita> we have asked some of our contacts to ping RH for us and IIRC they have done so.
[13:26:31] <mvita> meffie, is that right?
[13:26:46] <meffie> yes
[13:27:05] <mvita> so yes, just waiting
[13:27:15] <wiesand> ok, thanks
[13:27:39] <wiesand> makes me wonder whether auristor has this issue
[13:27:43] <kaduk@jabber.openafs.org/barnowl> Okay, so pre2 or 1.8.1-final, time.  Thumbs-up/down on 13191 and
13235?
[13:28:43] <kaduk@jabber.openafs.org/barnowl> I could probably be convinced to put both in and no pre2, but I am
currently on the fence.
[13:28:43] Marcio Barbosa leaves the room
[13:29:16] <wiesand> (actually, pre3)
[13:29:26] <meffie> ah, yes, pre3.
[13:29:39] Marcio Barbosa joins the room
[13:29:40] <mvita> 13191 thumbs up for 1.8.1
[13:29:43] <kaduk@jabber.openafs.org/barnowl> (thanks)
[13:29:48] <meffie> because pre1 was "skipped".
[13:30:32] <wiesand> I frankly don't understand 13235
[13:30:51] <meffie> um, i think 13191 is important because it addresses the "crashes" people see on rhel 7.5 due to the signals from systemd.
[13:31:59] <kaduk@jabber.openafs.org/barnowl> And you (both) think we wouldn't need a pre3 after taking 13191?
[13:32:19] <wiesand> I'm not seeing crashes on EL7.5
[13:32:46] <meffie> it was reported on the list and by a site andrew gave this patch to.
[13:33:35] <meffie> i guess it depends on timing or something, because it happens randomly
[13:34:01] <mvita> I believe (based on a quick perusal) that 13235 handles the case when we alloc a temp volume for a move, then get VNOVOL when we clean it up.
[13:34:33] <wiesand> how can that happen?
[13:34:34] <mvita> instead of treating that as an error, we leave any possible remnants for the next salvage to clean up
[13:34:55] <mvita> looking
[13:36:27] <mvita> vos move <> -live
[13:36:42] <mvita> sets the RV_NOCLONE option
[13:36:45] <kaduk@jabber.openafs.org/barnowl> My understanding matches Mark's read.
[13:37:06] <wiesand> ah, I never did that
[13:38:51] <mvita> hmm
[13:38:59] <mvita> something funny here, still reading in vos
[13:39:29] <kaduk@jabber.openafs.org/barnowl> We can probably hold that to 1.8.2 and make people more comfortable
[13:39:54] <mvita> ah, nm, I'm off by one due to parms being 0-based
[13:39:58] <mvita> all is well
[13:40:08] <mvita> yes, vos move -live option
[13:40:27] <meffie> kaduk@jabber.openafs.org/barnowl: so for my notes, the following are expected in 1.8.0pre3 ? plus maybe 13191 and 13235?
[13:40:35] <meffie> 4b42b49 afs: Make afs_osi_Free(NULL) a no-op
6fed867 LINUX: Detect NULL page during write_begin
9b04eac vldb_check: write mh entry header flags in network order
38a6a62 redhat: parallel builds
b70c8e1 redhat: speed up userland-only rpm builds
d5def58 redhat: package new file include/opr/lock.h
[13:41:05] <kaduk@jabber.openafs.org/barnowl> 1.8.*0*?  None of them ;)
[13:41:23] <meffie> heh, er 1.8.1pre3
[13:41:30] <mvita> (-live is also available for vos copy and vos shadow)
[13:41:57] <kaduk@jabber.openafs.org/barnowl> I just merged those today, yes.  I don't know that any of those are
big enough to force a pre3, though -- I was considering just going
straight to 1.8.1 final
[13:42:08] <mvita> I'm okay with both 13191 and 13235 in 1.8.1 final
[13:42:55] <wiesand> 1.8.1 final is what will actually get tested out there…
[13:43:21] <kaduk@jabber.openafs.org/barnowl> Exactly
[13:43:29] <wiesand> NB we're starting to roll out 1.8.1pre to production use.
[13:43:37] <mvita> I think we have had better-than-expected field uptake of all the 1.8.x field releases so far
[13:43:55] <mvita> better than _I_ expected, anyway
[13:44:04] <kaduk@jabber.openafs.org/barnowl> I think the "surpise!  Your Ubuntu update took it for you" may have
"helped" on that front
[13:44:43] <meffie> LOL. yes. i think what ben is saying is do we want to delay the 1.8.1 final for these late fixes, right?
[13:44:44] <mvita> sure, but we even had a number of good reports on the various -preX releases
[13:45:32] <wiesand> Skipping pre3 seems reasonable to me.
[13:45:57] <mvita> brb
[13:47:28] <mvita> back
[13:47:56] <meffie> technically 13191 and 13235 dont fix regressions since 1.8.1pre2, but we are so early in the stable series maybe we could relax that rule ?
[13:48:23] <kaduk@jabber.openafs.org/barnowl> I don't know that anyone was clamoring for 13235 right away, so
postponing it to 1.8.2 is probably best.
[13:48:49] <meffie> that sounds reasonable.
[13:49:18] <mvita> d'accord
[13:49:21] <kaduk@jabber.openafs.org/barnowl> We're early in the stable series, so fixing obvious bugs with minor
fixes after a preN (that are not strictly regressions) seems
reasonable to me.
[13:50:01] <meffie> ok, sounds sensible.
[13:50:02] <wiesand> agreed
[13:51:14] <kaduk@jabber.openafs.org/barnowl> Okay.  Anyone think they could prep NEWS and version number changes
for 1.8.1 final today, or should I plan to do that tomorrow?
(I assume you are all responsible and actually take your weekends
off...)
[13:51:15] Marcio Barbosa leaves the room
[13:51:34] <mvita> <ahem>
[13:52:12] <wiesand> I'll give it a go while waiting for the lunar eclipse this evening
[13:52:29] <kaduk@jabber.openafs.org/barnowl> Thanks!
[13:52:35] <mvita> nice!  No views of it in the US
[13:52:40] <kaduk@jabber.openafs.org/barnowl> *jealous* -- no eclips for me.
[13:52:55] <kaduk@jabber.openafs.org/barnowl> Though I guess having gotten a few minutes of solar totality is
probably adequate compensation.
[13:52:55] <wiesand> the longest this century
[13:53:09] <meffie> yayh! yeah, my daughter will be enjoying it tonight (in italy)
[13:53:13] <wiesand> yeah, that's what I missed last time
[13:53:20] Marcio Barbosa joins the room
[13:54:08] <kaduk@jabber.openafs.org/barnowl> (My apartment is just outside the band that got totality, but we were
flying to Canada that afternoon, and the route to the airport does
pass through the totality zone, so we just stopped partway.)
[13:54:09] <meffie> since we have some spec file fixes, we might want to take marcios fix for the libuafs thing.
[13:55:02] <meffie> 13257 redhat: provide full path of unused man page
[13:55:19] <kaduk@jabber.openafs.org/barnowl> Oh, I was thinking of 13258 -- do not ignore -march
[13:55:20] <meffie> some commit messages improvements would be nice
[13:55:51] <mvita> "If it was hard to write it should be hard to understand"
[13:55:52] <meffie> 13258 is not ready yet.
[13:56:05] <kaduk@jabber.openafs.org/barnowl> > commit messages improvements would be nice
"make it appear today"
[13:56:19] <kaduk@jabber.openafs.org/barnowl> Feel free to just push a new version yourself, if you want.
[13:56:26] <meffie> everything hard for *me* to write. so i must not understand anything
[13:56:43] <meffie> i can do, since he fixed my bug :(
[13:57:18] <kaduk@jabber.openafs.org/barnowl> My latest "hard to debug" resulted in
https://github.com/openssl/openssl/commit/45a2353056da3f357a924131578ad0a4a2e5fbb7
though hopefully the commit message is not too hard to understand
[13:58:05] <kaduk@jabber.openafs.org/barnowl> Thanks, Mike
[13:58:28] <mvita> I have another mtg in a few minutes
[13:58:38] <kaduk@jabber.openafs.org/barnowl> I should have some time to put in tomorrow, try to clear some stuff on
master and maybe work on <mumble>
[13:58:42] <mvita> I just have a few other items when we are ready
[13:58:48] <kaduk@jabber.openafs.org/barnowl> please go ahead
[13:58:53] <mvita> https://rt.central.org/rt/Ticket/Display.html?id=134577
[13:59:16] <mvita> I pushed a 1.6.x pullup of an existing master commit
[13:59:28] <mvita> the site is going to pull it from gerrit and try it out
[13:59:49] <mvita> https://gerrit.openafs.org/#/c/13261/1
[14:00:23] <mvita> so that's a pending 1.6.x item if it fixes the issue for him
[14:00:52] <mvita> however, that fix _could_ be considered a workaround for busted syscall-probing code
[14:01:13] <kaduk@jabber.openafs.org/barnowl> Seems generally worth taking that commit on 1.6.x
[14:01:34] <wiesand> it should be pulled through 1.8.x though ;-)
[14:02:01] <mvita> apparently it has ben broken in various ways for a long time (that is, it started bit-rotting shortly after keyring support made syscall-probing unnecessary for Linux)
[14:02:07] <mvita> s/ben/been/
[14:02:17] <kaduk@jabber.openafs.org/barnowl> It's already in 1.8 from before 1.8 diverged from master, no?
[14:02:31] <mvita> ^this
[14:02:37] <kaduk@jabber.openafs.org/barnowl> heh, I was all like "wait, what did I break this time?"
[14:03:03] <mvita> yes, Ben, I think I told you there was nothing on master to fix that issue, but I overlooked one.
[14:03:33] <kaduk@jabber.openafs.org/barnowl> Ah.
[14:03:39] <kaduk@jabber.openafs.org/barnowl> Don't worry about it :)
[14:04:18] <mvita> that's all for issue 1)
[14:04:41] <kaduk@jabber.openafs.org/barnowl> We could fall back to email if you need to make your other meeting
[14:05:00] <mvita> issue 2)   https://rt.central.org/rt/Ticket/Display.html?id=134574
[14:05:13] <mvita> any disposition, Ben?
[14:05:22] <mvita> or thoughts?
[14:05:45] <kaduk@jabber.openafs.org/barnowl> I haven't really taken any action there, but the described approach
and reasoning seems correct to me.
[14:05:56] <mvita> okay
[14:06:19] <mvita> I intend to look for one more class of issues of the same ilk
[14:06:26] <mvita> but haven't gotten back to that yet
[14:06:30] <kaduk@jabber.openafs.org/barnowl> Okay.  Thank you again for doing that work!
[14:07:00] <mvita> okay, dropping (half attention)
[14:07:17] <kaduk@jabber.openafs.org/barnowl> It was sounding like we were wrapping up anyway?  Any other topics?
[14:07:21] <meffie> kaduk@jabber.openafs.org/barnowl: my turn?
[14:07:29] <kaduk@jabber.openafs.org/barnowl> Go for it, Mike!
[14:07:50] <meffie> buildbot woes. so it seems i *really* need to upgrade to buildbot 1.x
[14:08:44] <kaduk@jabber.openafs.org/barnowl> What can we do to help?
[14:08:45] <meffie> i think i was blocked because we had an old python on that machine. i could try to build a python3 there, but i feel it would be better if someone with sudo could install a python
[14:08:57] <kaduk@jabber.openafs.org/barnowl> On the buildbot master?
[14:09:03] <meffie> yes
[14:09:12] <kaduk@jabber.openafs.org/barnowl> Okay, I can take a look.
[14:09:22] <meffie> i think we need python3.x, but that is from memory, so i need to check.
[14:09:53] <wiesand> maybe miniconda would help?
[14:10:02] <meffie> oh, maybe.
[14:10:23] <kaduk@jabber.openafs.org/barnowl> I think I hae root on that machine; I just need to coordinate so as to
not step on the toes of the people donating it to us.
[14:10:31] <meffie> we are currently using python virtualenv for each version.
[14:11:01] <wiesand> you want to replace /usr/bin/python with a 3.x one where it's 2.x now?
[14:11:14] <wiesand> ah, ok
[14:11:29] <meffie> i'll look to see if we can setup a buildbot master 1.x that is using different ports, so we can run them side by side for a transition?
[14:12:18] <meffie> we can have both python2 and python3 there, and i'll setup the python virtualenv to use the right interpreter for the env.
[14:12:31] <meffie> thanks ben
[14:13:13] <meffie> so what did we decide for 1.8.1 (for my notes)?
[14:13:13] Marcio Barbosa leaves the room
[14:13:40] <meffie> are we just doing a 1.8.1-final ?
[14:13:58] <kaduk@jabber.openafs.org/barnowl> 13191, and backport of 13257 if it's ready, then straight to
1.8.1-final
[14:14:18] <meffie> (after the lunar eclipse)
[14:14:28] <wiesand> right
[14:15:21] Marcio Barbosa joins the room
[14:15:32] <wiesand> (still have to decide where to watch the eclipse - old home or new one….)
[14:19:07] <wiesand> Looks like we're finished for today?
[14:19:08] <meffie> heh
[14:19:14] <meffie> yes
[14:19:44] <kaduk@jabber.openafs.org/barnowl> Okay.
[14:19:48] <kaduk@jabber.openafs.org/barnowl> Thanks everyone!
[14:20:05] <wiesand> Let's adjourn. Thanks a lot everyone!
[14:20:10] wiesand leaves the room
[14:34:43] mvita leaves the room
[15:21:59] Marcio Barbosa leaves the room
[15:24:44] Marcio Barbosa joins the room
[16:56:54] meffie leaves the room
[17:04:25] mvita joins the room
[17:17:51] mvita leaves the room
[20:02:56] mvita joins the room
[20:14:47] <mvita> Merged this morning by Linus to his master branch as a 4.18 fix from akpm:
[20:14:50] <mvita> 31c5bda3a656 mm: fix exports that inadvertently make put_page() EXPORT_SYMBOL_GPL
[20:14:54] <mvita> YAY!
[20:15:29] <kaduk@jabber.openafs.org/barnowl> Yay!
[22:19:52] Marcio Barbosa leaves the room
[22:24:04] mvita leaves the room