- Wednesday, January 21, 2015

release-team@conference.openafs.org

Wednesday, January 21, 2015< ^ >

Room Configuration

Room Occupants

GMT+0
[12:00:21] wiesand joins the room
[14:40:29] meffie joins the room
[14:45:24] kaduk joins the room
[15:00:16] <kaduk> $timeofday
[15:00:46] <wiesand> Hi
[15:01:53] <wiesand> I’m afraid I was just too late to make today’s meeting a success. Sorry.
[15:03:11] <kaduk> I had commented on 11669.  Any thoughts about that?
[15:03:16] <wiesand> The only good news is that I finally managed to run test builds on my EL5/6 test systems.
[15:03:29] <kaduk> Test builds is good.
[15:03:36] <kaduk> (Er, mostly asking Daria about 11669)
[15:04:20] <wiesand> I saw the comment. Sorry, I just have no clue.
[15:04:32] <wiesand> EL7 is still todo though.
[15:04:51] <wiesand> And I don’t have anything more recent to run on right now.
[15:05:38] <wiesand> For the record, test build is 7308aab, + 11616 11656/11645 11658..11662 M11643
[15:05:52] <wiesand> And the extra iput *en*abled.
[15:08:29] <kaduk> That's shaping up to be what the release will be, it seems.
[15:09:35] <wiesand> Likely.
[15:10:01] <kaduk> I think you are right that 11656, 11645, and 11658..11662 are good to go.
[15:10:06] <wiesand> A pre2 w/o 11616 would probably make sense. But not w/o 11643, I think.
[15:10:58] <wiesand> Ok, I’m going to merge those after checking for acincludes hassles.
[15:11:03] <kaduk> I don't have particularly strong feelings about whether 11643 should require the user to pick a flavor of the configure knob.
[15:11:29] <kaduk> (I actually didn't even check that it does, since I tried to scope such silliness to linux and failed to fire up a linux VM to do a test build.)
[15:11:44] <kaduk> Maybe I can do that when I get in to the office...
[15:12:20] <wiesand> I can check that. Give me a second
[15:14:20] <kaduk> Back to 11669, thinking about it more, I think the check should probably be conditional on 'locked'.
[15:14:37] <kaduk> I guess I should just push that update, since Daria is unlikely to bite my head off if I do.
[15:15:42] <shadow@gmail.com/barnowlE5B64A04> sorry i'm late
[15:15:57] <shadow@gmail.com/barnowlE5B64A04> it was time for my every few days computer crash
[15:16:08] <kaduk> nevermind, it's already inside a block which checks locked; I am dumb
[15:16:36] <shadow@gmail.com/barnowlE5B64A04> i would not bite, but yeah, it should do that
[15:17:44] <kaduk> I wonder if gerrit will grow (has grown?) the feature where I can click somewhere in the patch display and get more context
[15:18:43] <kaduk> Any thoughts about 11643 and mandatory configure options, Daria?
[15:19:16] <wiesand> Build on EL6 succeeded without specifying --enable-linux-d_splice_alias-extra-iput :-(
[15:19:23] <meffie> you can get more context with the dropdown at the top
[15:20:05] <kaduk> Oh hey, you can.  I think I even knew about that, once. :(
[15:20:28] <kaduk> I guess I should revisit 11643 when I get in, then.
[15:20:57] <shadow@gmail.com/barnowlE5B64A04> merge now fix overtop was already my opinion; but yeah, i think fixing
overtop is a good plan
[15:21:29] <wiesand> Of course it would be really nice if Michael were right...
[15:22:40] <meffie> it works for me after i hit "update"
[15:24:17] <kaduk> So, anything else to say about 1.6?
I guess there is this RT ticket you mentioned
[15:24:21] <wiesand> merge now what?
[15:25:27] <wiesand> That one and a very recent mail to -info from Markus Koeberl, maybe
[15:26:50] <kaduk> I'm not sure that 131991 is our fault.  I think we'd have to be very wrong in the page-handling vops to cause that type of crash.
[15:27:51] <kaduk> Markus's mail looks like the same thing we're in the process of fixing, if I remember correctly.
[15:28:33] <kaduk> I guess 11645 is what we think is actually fixing it.
[15:29:05] <wiesand> Great. We may get *one* test report for pre2 ;-)
[15:30:52] <kaduk> :)
[15:31:54] <kaduk> I don't think I have any particular discussion topics up for 1.8; I was just going to call out a few changes from the bottom of my dashboard (i.e., old ones) that are probably fine but could use gatekeeper review.
[15:32:21] <wiesand> Please, go ahead. I think we’re done with 1.6 toady.
[15:32:43] <kaduk> okay, I am pulling them up.
[15:33:22] <kaduk> 11599 hasn't had any review at all and thus may not strictly speaking fit into the category I described.
[15:34:09] <kaduk> 11432, 9348, and 8904 are the sort of thing I had in mind.
[15:34:32] <kaduk> I did note some potential nits with 8904, but they seem unimportant from here.
[15:34:51] <kaduk> I can't tell if 9709 is still relevant/useful or not.
[15:35:17] <kaduk> Since I recall that there were some fairly substantial changes in the IH_OPEN/IH_REALLYCLOSE area.
[15:36:31] <kaduk> 5405 is sitting in the dashboard but may not be worth spending time on.
[15:36:33] <Jeffrey Altman> Mike, does 11599 potentially write a CellServDB with a loopback address in it?
[15:36:55] <meffie> looking
[15:37:49] <kaduk> 8204 only has Andrew's own +1 on it, and I seem to have never managed to pull it up and look at it while I'm not tired, so I haven't reviewed it yet.
[15:37:50] <shadow@gmail.com/barnowlE5B64A04> 9709 is still real, there's no... it doesn't yet that i can tell
guarantee it's done at the return of REALLYCLOSE (and so merged)
[15:38:28] <shadow@gmail.com/barnowlE5B64A04> it it does, you would also rxbind 127.0.0.1 so it's useless anyway
[15:38:34] <kaduk> 2594 is probably fine as-is (well, it probably needs a rebase...)
[15:38:36] <shadow@gmail.com/barnowlE5B64A04> (and that's pre-existing)
[15:38:55] <meffie> Jeffrey Altman: i recall checking for such in 11599, those functions called actually avoid returning a loopback
[15:39:08] <meffie> i will double check to be sure.
[15:39:39] <kaduk> 11516 is related to our previous volume header times discussions, but it seems to mostly have the jhutz seal of approval.
[15:40:30] <kaduk> For 11620, enough time has passed that it's probably okay to just push the button.
[15:41:30] <kaduk> 11313 is probably fine (but is a bit of a large chunk to review/process)
[15:42:21] <kaduk> 11161 had gotten some comments from Chas, but I can't decide how important the issues raised are.
[15:42:46] <kaduk> [I think I'll stop at this point in the dashboard, since that's a decent number of changes]
[15:46:50] <wiesand> You got 3 merges and lots of review :)
[15:48:03] <meffie> yay
[15:48:33] <wiesand> So, progress on 1.8. Loose ends regarding 1.6. are 11669 and 11643.
[15:49:03] <kaduk> 11669 shouldn't really be a loose end anymore; I was just being dumb
[15:50:01] <wiesand> So it “just” takes merging it.
[15:51:00] <wiesand> Anything else to discuss today?
[15:51:33] <Jeffrey Altman> as a head up, the YFS team is will not be available next week
[15:51:55] <wiesand> Bad news, but thanks.
[15:51:58] <kaduk> I could go further up my dashboard, if people still have energy...
[15:52:34] <Jeffrey Altman> I have asked Marc to review 11669
[15:52:41] <wiesand> Thanks!
[15:53:57] <kaduk> I guess I have these quickstart guide deorbits sitting around in 11590..11592 that aren't doing much.
[15:55:42] <Jeffrey Altman> I will raise the following question here before I raise it elsewhere.   If we have no ability to build AIX in buildbot, should AIX be supported for 1.8 and beyond?
[15:56:09] <kaduk> Chas has 11634/11635 for some irix cleanup
[15:56:41] <kaduk> If we don't have anyone testing a given architecture, we probably shouldn't claim to support it.
[15:57:11] <Jeffrey Altman> quick start guide stack needs rebasing
[15:58:45] <wiesand> I haven’t come close to AIX for more than a decade. No idea whether users may be left.
[15:59:23] <Jeffrey Altman> YFS has no AIX customers
[15:59:29] <kaduk> We won't find out if there are users left unless we threaten to (or do) make it go away.
[16:00:06] <kaduk> So, I nominate Jeff to inform -info that AIX will not be supported in 1.8 and handle any backlash.
*ducks*
[16:00:25] <wiesand> Seconded. *ducks too*
[16:01:02] <Jeffrey Altman> I nominate Ben to submit a patch removing AIX from the quck start guides
[16:01:22] <kaduk> ;)
[16:02:20] <wiesand> Adding a statement like “AIX is unsupported and will be removed in the next release” seems appropriate?
[16:02:24] <meffie> i believe there are AIX users still in the wild.
[16:02:46] <Jeffrey Altman> there definitely are
[16:02:58] <Jeffrey Altman> are any of them SNA customers?
[16:03:18] <meffie> not at this time.
[16:03:41] <Jeffrey Altman> the reason I ask is that if one of them are unwilling to pay for a buildbot builder they shouldn't expect the developers to do so
[16:03:54] <meffie> i could see if we still have AIX available for a buildbot.
[16:04:42] <Jeffrey Altman> I think it should come from the users
[16:05:22] <meffie> ideally, yes
[16:09:06] <meffie> i'd imagine most of the AIX sites have moved to linux.
[16:10:17] <Jeffrey Altman> there is at least one in Germany that IBM wrote a whitepaper about last year describing how their new Power architecture tripled OpenAFS performance
[16:11:05] <meffie> interesting!
[16:11:24] <kaduk> huh.
[16:11:42] <Jeffrey Altman> and I believe one of the NC schools is still AIX based
[16:22:36] <meffie> ok, have a good day/evening
[16:23:13] <Jeffrey Altman> so long
[16:26:12] <shadow@gmail.com/barnowlE5B64A04> ciao
[16:26:25] <wiesand> Thanks a lot everyone.
[16:26:39] <wiesand> I’ll be gone too. Bye.
[16:26:40] wiesand leaves the room
[16:28:23] wiesand joins the room
[16:28:37] wiesand leaves the room
[16:38:28] kaduk leaves the room
[20:54:22] meffie leaves the room
[20:57:31] kaduk joins the room
[23:12:22] kaduk leaves the room