Home
release-team@conference.openafs.org
Wednesday, February 22, 2017< ^ >
Room Configuration
Room Occupants

GMT+0
[13:48:57] meffie joins the room
[14:21:26] wiesand joins the room
[14:21:52] <wiesand> I made it :)
[14:26:40] <meffie> good afternoon.
[14:34:56] <wiesand> Hello Mike
[14:58:52] <mvita> salutations
[15:01:03] <wiesand> Hello Mark
[15:01:38] <wiesand> Let's start?
[15:01:42] <meffie> yes
[15:01:51] <wiesand> Any news regarding Linux 4.10?
[15:02:27] <mvita> it's out, we work with 12506
[15:02:43] <kadukoafs@gmail.com/barnowlD058797B> Remind me whether we believe 12506 needs changes to be compatible with
2.x kernels?
[15:02:54] <mvita> I was mistaken about that
[15:03:07] <kadukoafs@gmail.com/barnowlD058797B> So, "no changes needed"?
[15:03:17] <mvita> in the gerrit review, I veriified w/ a 2.6.x kernel
[15:03:29] <mvita> no changes needed for 2.6.x afaik
[15:03:34] <kadukoafs@gmail.com/barnowlD058797B> I was pretty sure I remembered that being the case, but it is good to
be fully clear.
[15:03:51] <wiesand> Great, thanks.
[15:04:48] <wiesand> Ben, what else do you need to merge it?
[15:04:56] <kadukoafs@gmail.com/barnowlD058797B> I wanted to get to that last night but ran out of time :(
[15:05:38] <kadukoafs@gmail.com/barnowlD058797B> So, probably "just time to review it properly"
[15:05:51] <wiesand> Anyway, this means we'll stick with then 1.6.20.1 plan IMHO.
[15:05:58] <meffie> yes
[15:06:20] <wiesand> Good.
[15:06:36] <wiesand> I see a number of new changes for 1.6.x in gerrit...
[15:06:38] <meffie> 1.6.20.2
[15:07:08] <kadukoafs@gmail.com/barnowlD058797B> "you know, that next one"
[15:07:55] <wiesand> At least 12521 will "path conflict" with a number of others
[15:08:43] <wiesand> I don't have the list yet.
[15:09:14] <meffie> can i give some buildbot updates?
[15:09:30] <kadukoafs@gmail.com/barnowlD058797B> please do
[15:10:49] <meffie> ok, i've added two new semidaily builders: linux-rc-builder and linux-daily-builder. they will build master with the most recent linux rc and the ubuntu kernel team's daily linux image.
[15:11:27] <mvita> THANK YOU MIKE!
[15:11:33] <wiesand> Wait… 12521 does *not* clash with anything foreseen for 1.6.21. So, no problem.
[15:12:12] <wiesand> > THANK YOU MIKE!
[15:12:13] <meffie> they work by cloning a ubuntu 16.10 image to a new guest, install the linux images, reboots, build openafs, then do a smoke test. the guest is then deleted.
[15:12:15] <wiesand> seconded
[15:12:44] <wiesand> great stuff
[15:12:44] <meffie> at the moment, we need 12506,
[15:13:17] <kadukoafs@gmail.com/barnowlD058797B> Very exciting; thank you!
[15:13:32] <meffie> no problem.
[15:14:13] <meffie> the scripts to do this are in github, based on something called virt-utils, which are some contrib scripts for libvirt.
[15:14:14] <wiesand> What about those RT tickets I mentioned?
[15:14:46] <wiesand> https://rt.central.org/rt/Ticket/Display.html?id=134158 could become a fairly common issue I guess.
[15:15:16] <kadukoafs@gmail.com/barnowlD058797B> The debian bug got some draft patches from the submitter, let me find
a link...
[15:15:45] <mvita> oh, this my first time seeing this
[15:15:58] <kadukoafs@gmail.com/barnowlD058797B> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=855366
[15:17:07] <meffie> which debian version is this?
[15:17:53] <mvita> it says in the bug report
[15:18:42] <meffie> sorry, i just see the kernel version
[15:18:44] <wiesand> the debian ticket mentions "jessie's default kernel"
[15:18:49] <meffie> ah, ok.
[15:19:07] <kadukoafs@gmail.com/barnowlD058797B> > Found in version openafs/1.6.20-2
[15:19:28] <mvita> patch looks simple enough, but I'd like to take some more time.
[15:19:38] <meffie> /usr/src/linux-headers-3.16.0-4-common/include/linux/fs.h:2654:12: note: expected ‘struct dentry *’ but argument is of type ‘int’  extern int setattr_prepare(struct dentry *, struct iattr *);             ^
[15:20:09] <meffie> ah, this is the thing that is breaking our debian builders!
[15:20:57] <mvita> I think this is worth pushing to master, if it all checks out
[15:21:13] <mvita> and then backporting to 1.6.x
[15:21:36] <kadukoafs@gmail.com/barnowlD058797B> Seems reasonable at first glance, yeah.
[15:21:55] <meffie> and i'll standup an debian jessie amd64 build slave.
[15:22:00] <mvita> tx
[15:23:21] <mvita> so what's the etiquette here?   do we need to ask the author if we can commit it?
[15:23:37] <mvita> how do we attribute it to him?  author?
[15:23:56] <mvita> s/commit/submit
[15:24:21] <wiesand> "author" was done several times in the past
[15:24:37] <kadukoafs@gmail.com/barnowlD058797B> I think that we/I can reply to the debian bug and thank him for the
submission, noting that it will be submitted upstream.
[15:24:45] <kadukoafs@gmail.com/barnowlD058797B> And yeah, we can put the commit author as him.
[15:25:27] <wiesand> Sounds good. I don't expect him to object, but it's the polite way.
[15:26:25] <wiesand> And while FBSD11 is unlkely to become quite as common: Is https://rt.central.org/rt/Ticket/Display.html?id=134170 something we could address anytime soon?
[15:27:14] <mvita> so after Ben writes, we can submit this as-is  (I will do that) and see how it fares with all the kernels in our build farm.
[15:27:34] <mvita> we can do that in parallel with my review and research on correctness
[15:27:52] <wiesand> Thanks!
[15:28:01] <kadukoafs@gmail.com/barnowlD058797B> Making the freebsd port/packaging let you build just a server without
a client is probably worth doing in the interim, but that's mostly on
me.
[15:28:52] <wiesand> I'm inclined not to delay 1.6.20.1 for this...
[15:29:09] <kadukoafs@gmail.com/barnowlD058797B> Yeah, I don't think I can commit to being prompt enough about dealing
with it for you to delay the release.
[15:29:35] <kadukoafs@gmail.com/barnowlD058797B> Other FBSD news includes 12520, though; I don't know if anyone here
has enough history/familiarity with that part of the code to be able
to comment.
[15:29:38] <mvita> maybe Garrett Wollman could help w/ FBSD
[15:29:41] <mvita> he has in the past
[15:30:14] <kadukoafs@gmail.com/barnowlD058797B> Yeah, I tried adding him as a reviewer in gerrit, but there are two
users with his email address so it fails :-/
[15:30:30] <mvita> so has Tim Creech (dezgot)
[15:31:16] <wiesand> Going through my list for other 1.6.21 candidates...
[15:31:28] <wiesand> clearly 12512 (osx prefpane)
[15:31:41] <wiesand> 12499 (S390)
[15:32:33] <wiesand> maybe 12514 (Solaris ELF deps?)
[15:32:43] <meffie> yes please
[15:33:39] <meffie> the solaris pag fixes would be helpful
[15:34:12] <kadukoafs@gmail.com/barnowlD058797B> Sorry for jumping out of line, but I want to ask before I have to
disappear how many people are running master/1.8 clients regularly --
it would be good to be able to confirm/deny the maybe bug reports we
got about being unable to access certain files.
[15:35:09] <meffie> i run it every day, but on an older debian.
[15:35:29] <kadukoafs@gmail.com/barnowlD058797B> > older debian
does that mean oldstable?
[15:35:37] <meffie> yes
[15:35:39] <wiesand> [mumbles] I'm not running it at all
[15:36:02] <mvita> I run it all the time as well
[15:36:17] <mvita> on Solaris 11 and various Linux kernels
[15:36:40] <kadukoafs@gmail.com/barnowlD058797B> Okay.  It sounds like maybe I should just test some upgrade scenarios,
then; maybe tweaking the jhash endianness or something caused things
to get confused.
[15:36:52] <kadukoafs@gmail.com/barnowlD058797B> s/I/anyone who has time/, heh
[15:37:56] <mvita> what's the RT # for the client problems?
[15:40:27] <kadukoafs@gmail.com/barnowlD058797B> https://rt.central.org/rt/Ticket/Display.html?id=133654
[15:40:55] <wiesand> The solaris pag fixes are somewhat entangled with the earlier 11855 and 10245 (which are not trivial to backport - someone volunteered for doing that a couple of weeks ago) as well as the later 12346/12347.
[15:43:06] <meffie> oh, that's right, i offered to backport, but for some reason we wanted to wait.
[15:43:42] <meffie> let me revisit that.
[15:44:16] <wiesand> I haven't checked the code, but it tends to be less troublesome to include changes in roughly the order they were merged on master.
[15:44:36] <meffie> yes, agreed.
[15:44:39] <wiesand> Thanks
[15:47:57] <mvita> hmmm 133654 may be due to a "hole" in our autoconfig for that kernel version (Ubuntu 14.04 with Kernel 3.13.0-106-generic,)
[15:48:29] <mvita> that's what it smells like, anyways
[15:49:04] <mvita> just like the previously discussed Jessie patch
[15:49:30] <mvita> but 133654 is failing at runtime, not build time
[15:50:13] <mvita> Just to be clear, I'm not suggesting 133654 is the _same_ problem, just raising the possibility that it is in the same class of bug
[15:50:24] <wiesand> wasn't there some change addressing the issue which went in between alpha1 and pre1?
[15:50:42] <mvita> oh?
[15:51:46] <wiesand> 12363?
[15:52:03] <mvita> was there a tag for alpha?  I'm still looking for it
[15:52:40] <mvita> OH.
[15:53:48] <mvita> well, yeah, that could be it if D_INVALIDATE_IS_VOID was backported to stable 3.13
[15:54:04] <kadukoafs@gmail.com/barnowlD058797B> 1.8.0pre1 is stage-named alpha1
[15:55:26] <mvita> ok
[15:55:28] <wiesand> Ah, ok. Anyway, 12363 should have been included at that point.
[15:56:39] <wiesand> So wae may still have an issue with "shake harder" even with that fix? Or is it something else?
[15:56:51] <mvita> I don't know.
[15:57:13] <mvita> it's unlikely I can get to this before Friday
[15:59:38] <wiesand> it's been an hour - more to discuss today?
[16:00:27] <meffie> no thank you. have a good evening.
[16:00:31] <mvita> all I'm committing to do at this point is for the other problem (the Jessie patch) - I will write that up and prepare for pushing.  I won't actually push it until I hear from Ben that the author has consented or has been notified
[16:01:26] <kadukoafs@gmail.com/barnowlD058797B> The author has been notified.
[16:01:51] <kadukoafs@gmail.com/barnowlD058797B> (just got back from my other meeting)
[16:01:56] <mvita> okay, I'll be pushing it before noon
[16:01:58] <kadukoafs@gmail.com/barnowlD058797B> But I don't think I have more topics
[16:01:59] <wiesand> Fine. It would be great if this made 1.6.20.1.
[16:02:43] <wiesand> Let's adjourn then. Thanks a lot everyone!
[16:03:10] wiesand leaves the room
[16:10:39] meffie leaves the room
[18:28:17] kadukoafs@gmail.com/barnowlD058797B leaves the room
[18:28:26] kadukoafs@gmail.com/barnowlD058797B joins the room
[18:43:55] kadukoafs@gmail.com/barnowlD058797B leaves the room
[18:44:04] kadukoafs@gmail.com/barnowlD058797B joins the room