- Friday, November 8, 2019

release-team@conference.openafs.org

Friday, November 8, 2019< ^ >

Room Configuration

Room Occupants

GMT+0
[13:51:41] meffie joins the room
[15:15:13] wiesand joins the room
[15:22:30] <wiesand> Today's meeting will "officially" start 38 minutes from now.
[15:38:02] <meffie> thanks wiesand
[16:00:54] yadayada joins the room
[16:01:17] <wiesand> Good morning/day/afternoon/evening
[16:01:22] <yadayada> Hello All
[16:01:40] <wiesand> Welcome back yadav
[16:01:43] <meffie> hello
[16:02:42] <wiesand> So there was a lot of churn on the 1.6.x branch in gerrit lately
[16:03:03] <wiesand> I will admit that I still have to catch up
[16:04:10] <wiesand> I will also declare that I still have serious doubts regarding changes like "remove settime" and "remove -sync" this late in the 1.6 release cycle
[16:05:44] <meffie> understandable. if we mark them as -2 and abandon, i'll rebase without them in the staging stack.
[16:06:59] <wiesand> Sounds good, thanks. Let's give others a chance to chime in though.
[16:08:10] <wiesand> Anyway, thanks a lot for your work on 1.6.x!
[16:08:53] <wiesand> I'm not quite sure what to focus on next - 1.6.25 or 1.8.5?
[16:09:01] <meffie> andrew found a bug in one of the 1.6.x changes and pushed 13198 for the fix
[16:09:16] <meffie> that fix needs to be on master and 1.8.x
[16:11:29] <kaduk@jabber.openafs.org/barnowl> I just pushed 13933 to master a few minutes ago, which I think is the
fix you refer to
[16:11:42] <kaduk@jabber.openafs.org/barnowl> (13198 being the base change where Andrew noticed the bug)
[16:12:03] <meffie> ah, sorry i swapped the numbers in my notes. thanks.
[16:12:43] <yadayada> We tested 13928 with 1.6.* code and works fine in our env. I had couple of comments for Marcio on that. Since MacOS Catalina is out can we push changes for Catalina too ?
[16:12:49] <wiesand> ah, that makes more sense, thanks
[16:15:16] <wiesand> yadav: 1.8.5 should address Catalina
[16:15:36] <yadayada> sure
[16:15:37] <meffie> remote: New Changes:
remote: https://gerrit.openafs.org/13937 afs: Avoid giving wrong 'tf' to afs_InitVolSlot
[16:15:44] <meffie> pushed for 1.8.x
[16:18:00] <wiesand> sounds kind of serious?
[16:19:51] <meffie> not sure, it doesnt seem to effect linux that i know of
[16:20:25] <wiesand> ah, thanks :)
[16:21:26] <wiesand> Re Linux 5.4, it seems the workaround was implemented on the daily builders and we're fine?
[16:22:01] <meffie> yes, the workaround was "upgrade gcc to 8.x"
[16:22:10] <wiesand> Thanks!
[16:22:30] <meffie> gcc 7.x is the default on ubuntu still.
[16:23:46] <meffie> but we are using the ubuntu kernel team ppa so, we need to install a newer gcc.
[16:24:11] <meffie> lessons learned i guess.
[16:24:11] <wiesand> So with the Catalina changes in the pipeline we're rather close to a 1.8.6pre1?
[16:25:02] <meffie> yadayada: thanks for the catalina testing and help
[16:25:35] <wiesand> seconded
[16:25:36] <yadayada> Thanks to Marcio for quickly implementing the changes for synthetic.conf file.
[16:26:17] <yadayada> We did most of the testing on Catalina and I think we should be good to merge these in master branch
[16:26:27] <meffie> excellent news. thanks
[16:27:51] <meffie> yadayada: any interest in updating the macos preference pane for openafs? :)
[16:28:17] <wiesand> ok, looks like "plenty to work on for 1.6 and 1.8, with some more to come soon"
[16:28:24] <wiesand> on to 1.9/master?
[16:28:51] <kaduk@jabber.openafs.org/barnowl> I've effectively been zonked all week and have nothing to report :(
[16:29:20] <yadayada> During conference we discussed to make preference pane simple, and I think we can remove certain things which doesn;t seems to be working. I will resume on that in coming days.
[16:29:23] <meffie> one more bug fix to be added to the 1.8/1.6 wish list, marcio submitted a fix for the fileserver -readonly option
[16:31:53] <wiesand> do you have a gerrit number for that?
[16:32:02] <meffie> 13934 viced: refuse CHK_STORE* routines in RO servers
[16:33:21] <yadayada> for fileserver -readonly what is a gerrit number ?
[16:33:25] <meffie> the commit message should be updated to make it more clear what the bug is
[16:33:46] <meffie> https://gerrit.openafs.org/13934
[16:34:00] <wiesand> I wasn't even aware of readonly fileservers…
[16:34:21] <meffie> in some cases a user can bypass the checks and write to files
[16:34:28] <meffie> if they own the files i think.
[16:34:34] <wiesand> request noted though
[16:35:02] <meffie> yes, it is a feature that is used for data that should not be changed, even in read-write volumes.
[16:36:35] <meffie> fileserver -readonly
[16:37:16] <wiesand> you keep learning here…
[16:37:38] <meffie> never a dull moment.
[16:38:37] <wiesand> anything else to discuss today?
[16:38:48] <wiesand> if not, maybe the time slot?
[16:39:08] <yadayada> I was testing OpenAFS on Power9 systems, and sometimes I see "rcu stall messages" on 4.* kernel. I still think it is not related to AFS, but anyone has seen such messages ?
[16:39:56] <meffie> checking notes...
[16:40:39] <wiesand> I'm afraid I haven't ever touched a Power9 system… and I haven't seen such messages on other platforms.
[16:41:05] <meffie> i dont think i've seen that message either.
[16:41:14] <yadayada> sure, just thought of asking, since in stacks traces I do not see AFS. Thanks
[16:41:46] <kaduk@jabber.openafs.org/barnowl> IIRC we do use some RCU stuff for AFS, but so does ~everything else in
the kernel, so it's not necessarily pointing to AFS
[16:42:27] <yadayada> sure
[16:43:31] <meffie> thanks for testing on power9 though. good luck with your tests.
[16:44:23] <wiesand> Regarding the meeting time: Yadav indicated that "the later the better", Ben would probably welcome another shift to a later hour, for myself it wouldn't matter all that much (the second half of Friday is screwed up anyway) - thoughts?
[16:45:40] <kaduk@jabber.openafs.org/barnowl> I'm indifferent-to-positive about shifting an hour later, yes
[16:46:41] <meffie> i am flexible. i would gladly shift one hour for yadav and ben.
[16:46:58] <yadayada> I am ok with one hour later
[16:47:32] <wiesand> Mike: any idea what Mark's point would be?
[16:48:05] <wiesand> Yadav: are you "okay" or would you prefer it (or not)?
[16:48:43] <meffie> he has a conflict on friday mornings for the next couple of months, so an hour later may be helpful for Mark V.
[16:49:37] <yadayada> I would prefer it
[16:50:18] <wiesand> sounds a lot like we should try shifting the start time by another hour - fine, let's try!
[16:50:32] <meffie> excellent. thanks all.
[16:53:04] <wiesand> Ok, so it's 9am Pacific, 12am(?!) Eastern, 6pm Central Europe, 10:30pm India starting next week?
[16:53:17] <yadayada> sounds good
[16:53:43] <meffie> 12 noon eastern standard time
[16:54:09] <meffie> (aka lunchtime)
[16:54:23] <wiesand> I know that one ;-)
[16:54:43] <kaduk@jabber.openafs.org/barnowl> aka 12pm eastern
[16:55:21] <wiesand> 12pm sounds like midnight to me… but yes I'm just confused…
[16:56:01] <kaduk@jabber.openafs.org/barnowl> It is a very common point of confusion, yes.
[16:56:15] <wiesand> Looks like we agree on trying that new slot. Great!
[16:56:16] <meffie> boundary conditions
[16:56:16] <kaduk@jabber.openafs.org/barnowl> Sometimes 11:59 is easier to describe
[16:56:47] <wiesand> "have you ever heard of that weird invention called the 24h clock?"
[16:56:53] <meffie> :)
[16:58:26] <wiesand> Right after ridding the world of DST, we should rid it of am/pm…
[16:58:51] <kaduk@jabber.openafs.org/barnowl> +1
[16:59:18] <wiesand> Looks like we're finished for today. Adjourn?
[16:59:30] <yadayada> Thanks
[16:59:42] <meffie> thank you, have a good weekend
[16:59:53] <kaduk@jabber.openafs.org/barnowl> Thanks everyone
[17:00:07] <wiesand> Thanks a lot everyone, and have a good weekend (I surely will ;-)!
[17:04:19] meffie leaves the room
[18:43:37] yadayada joins the room
[18:59:27] yadayada leaves the room
[19:27:06] wiesand leaves the room