Home
release-team@conference.openafs.org
Friday, July 10, 2020< ^ >
kaduk@jabber.openafs.org/barnowl has set the subject to: openafs release team
Room Configuration
Room Occupants

GMT+0
[11:31:07] meffie joins the room
[12:01:09] mbarbosa joins the room
[13:54:03] mbarbosa leaves the room
[13:59:25] mbarbosa joins the room
[15:58:18] kaduk@jabber.openafs.org/barnowl joins the room
[15:59:30] cwills joins the room
[15:59:39] yadayada joins the room
[16:00:19] <meffie> hello all
[16:00:25] <yadayada> Hi All
[16:00:30] <kaduk@jabber.openafs.org/barnowl> greetings
[16:02:59] meffie leaves the room
[16:03:14] <cwills> Morning
[16:03:21] wiesand joins the room
[16:03:31] <kaduk@jabber.openafs.org/barnowl> Let's give Stephan another couple minutes
[16:03:38] <kaduk@jabber.openafs.org/barnowl> ooh, magic
[16:04:07] <wiesand> It sometimes takes a while to get into the room, not sure why.
[16:05:30] <wiesand> I started working on 1.8.7pre1.
[16:06:00] <wiesand> I'm currently at the bookkeeping & resolving path conflicts stage.
[16:06:23] <wiesand> Quite a few changes in the queue also could have a bit more review for my taste.
[16:07:08] <yadayada> I am looking at the NAT ping issues, as you might have seen in my last mails, there are couple of issue related to NAT ping. This week didn;t get much time to work on fix, I think will get time next week
[16:07:39] <meffie> Thank you yadayada.
[16:07:44] <kaduk@jabber.openafs.org/barnowl> Okay.  This week didn't have much time for me, either ...
[16:07:56] <wiesand> I have no explicit list of changes where I'll ask for more review yet. https://gerrit.openafs.org/#/q/status:open+project:openafs+branch:openafs-stable-1_8_x is not that long right now - you can help ;-)
[16:08:30] <meffie> ok, i'll add some reviewers to them.
[16:08:31] <kaduk@jabber.openafs.org/barnowl> That's "help with review", not "help fix the problem that the list is
not longg", right? ;)
[16:08:57] <cwills> oohh.. I can fix that one :)
[16:09:39] <wiesand> Noo, I actually meant "help with review".
[16:10:42] <wiesand> Anything pulled up now, creating more path conflicts while the current ones haven't been resolved, will make it onto my "blind spot"list…
[16:11:33] <meffie> understood.
[16:12:51] <wiesand> Anyway, I was serious about trying to get 1.8.7(pre1) out a bit more swiftly, and the essential things warranting a new release are already present in the queue.
[16:13:38] <wiesand> If there's anything else (significant) you think should make 1.8.7pre1, please bring it up here (before pulling it up) and I'll put it on the wish list.
[16:14:39] <meffie> i'll have to look at the master branch to see if there are any other important changes for 1.8.7
[16:15:17] <cwills> I believe that what I have in there allows building with the latest linux kernel (5.8) and gcc-10 (which picks up fedora 32)
[16:15:28] <meffie> since we did hold off on pushing to gerrit's 1.8 branch while 1.8.6 was in progress
[16:15:49] <wiesand> If there really is a shortage of changes queued up for 1.8.7 I'll have that look myself. Hasn't been the case for the last few releases though…
[16:16:32] <meffie> yes, but we really put the brakes on pushing to the stable branch this time.
[16:17:12] <meffie> so we should check to see if there is anything critical on master that should be pulled up?
[16:17:24] <wiesand> right, and thank you for that
[16:18:05] <yadayada> https://gerrit.openafs.org/#/c/14239/ https://gerrit.openafs.org/#/c/13827/
[16:18:39] <kaduk@jabber.openafs.org/barnowl> Those look like good candidates, yes
[16:19:07] <meffie> agreed.
[16:19:51] <meffie> i could look and if i see others post a list to the release-team mail list next week.
[16:20:24] <meffie> we can precheck for conflicts and such.
[16:20:25] <wiesand> put on the wish list, thanks
[16:21:38] <yadayada> I am not getting gerrit number, but recently Andrew has submitted gerrit to avoid InlineBulkStatus in case cache is full, as it will result in cache thrashing
[16:22:05] <kaduk@jabber.openafs.org/barnowl> IIRC I merged that last night
[16:22:12] <wiesand> avoiding conflicts in the first place is even better - i.o.w. reviewing changes already pulled up is more helpful than finding new ones to pull up
[16:22:12] <meffie> topic:bulkstat-vcache-pressure
[16:22:29] <kaduk@jabber.openafs.org/barnowl> 13256
[16:22:46] <meffie> i'll ask others to review while i make the list ;)
[16:23:24] <meffie> i dont think it will be a long list actually.
[16:24:02] <yadayada> right 13256 will improve performance since we will avoid cache thrashing. I have seen in my test where client really becomes slow due to cache thrashing
[16:24:48] <kaduk@jabber.openafs.org/barnowl> the whole topic's worth of changes would probably help in those cases
[16:26:48] <wiesand> You should anticipate my growing resistance to accept mere optimizations, unless trivial, into the stable series.
[16:27:58] <kaduk@jabber.openafs.org/barnowl> I'm hearing that I should actually make time and push the button on
1.9.0, then
[16:28:39] <wiesand> It is my belief that effort is better spent there at this point in time.
[16:29:37] <kaduk@jabber.openafs.org/barnowl> I'm not sure I have a compelling argument to the contrary.  (Though I
haven't had my morning tea yet...)
[16:29:55] <wiesand> Fixes, sure. Enabling platforms and compilers, of course. Trivial stuff, yes. Anything else: probably makes more sense to spend the resources on the next stable series.
[16:32:26] <kaduk@jabber.openafs.org/barnowl> Should I take that as a sign to switch to that agenda topic?
[16:33:29] <meffie> heh
[16:34:08] <kaduk@jabber.openafs.org/barnowl> (through no fault of anyone, it seems that my part of the agenda will
resemble "pick on Mike day"
[16:34:24] <meffie> lol
[16:35:08] <kaduk@jabber.openafs.org/barnowl> Mostly just that there are a few gerrits that I want to sync up on the
next steps for
[16:35:19] <wiesand> go ahead (with switching topics, *not* picking on Mike)
[16:35:45] <kaduk@jabber.openafs.org/barnowl> In 14151 especially, the state is not fully clear to me; I think there
some changes left to be made, but wanted to check.
[16:37:11] <meffie> sorry i've not been following up in gerrit.
[16:37:39] <kaduk@jabber.openafs.org/barnowl> Not a problem (it's easier to keep track of "I need to do something"
when there's a -1, but there isn't...)
[16:38:06] <kaduk@jabber.openafs.org/barnowl> Also, while we're all here, on 14266 just let me know if I should
merge or you want to put a configure log snippet in the commit
message.
[16:38:46] <meffie> well, i guess i'm not happy about making that new script, i'd rather fix/change the make-release thing, but it means changing the value we store in the '.version' file to the actual version string, not the git tag name.
[16:39:25] <wiesand> ping (there seems to be a serious latency from my client to the server and back today, sorry)
[16:39:29] <meffie> or if we cant live with that, make a new .version-something file that actually has the version number.
[16:39:41] <wiesand> 10s
[16:40:15] <kaduk@jabber.openafs.org/barnowl> We might have to do .version-git and .version-release or something,
yeah.  It's rather unfortunate that we changed the .version semantics
already
[16:40:19] <meffie> that is the issue blocking that whole silly stack. the new 'make-tarballs' was my workaround.
[16:41:40] <meffie> ok, i'll propose a change in gerrit.
[16:41:43] <kaduk@jabber.openafs.org/barnowl> Thanks
[16:42:21] <kaduk@jabber.openafs.org/barnowl> And, last one for Mike, in 14154 I think I agree with Andrew that we
can simplify the last part of the commit message.  I am happy to make
the edit, but didn't know if you were attached to it as-is.
[16:42:23] <meffie> thanks ben.
[16:43:43] <meffie> ah, i'm not attached at all! whatever changes we need to make to the commit message is ok with me :) feel free to make them if you have the time. i can do it today otherwise.
[16:43:56] <meffie> thank you!
[16:44:10] <kaduk@jabber.openafs.org/barnowl> Great, thanks
[16:44:32] <meffie> (i guess i was rambling on that one ;)
[16:45:18] <kaduk@jabber.openafs.org/barnowl> I don't have a whole lot else for today, I think.  (I'm still catching
up, a bit, after a busy week.)
[16:45:38] <meffie> hope you are feeling better.
[16:45:45] <kaduk@jabber.openafs.org/barnowl> I do want to check if anyone is currently looking at the "recent" RT
tickets -- 135117 and 135122
[16:46:02] <kaduk@jabber.openafs.org/barnowl> Yeah, I'm feeling healthy again, there's just a lot to do.
[16:46:50] <meffie> I think mark vitale may be interested in 135122. i'll ask him.
[16:46:51] <wiesand> I'm having those in mind, but I'm afraid I can't help a lot.
[16:47:45] <meffie> not sure about 135117. i did see some kafs changes for silly-renames recently.
[16:47:47] <kaduk@jabber.openafs.org/barnowl> I think I'm comfortable enough with the code to investigate 135117,
but I'd want to reproduce it and haven't had the time to make such a
setup
[16:48:01] <yadayada> I will try 135117 next week once done with NatPing
[16:48:16] <meffie> cool, thanks yadayada!
[16:48:22] <kaduk@jabber.openafs.org/barnowl> Thanks Yadav!
[16:49:39] <kaduk@jabber.openafs.org/barnowl> Any other topics for today?
[16:49:56] <wiesand> Anything else to discuss today?
[16:50:04] <wiesand> grr, that latency
[16:50:05] <kaduk@jabber.openafs.org/barnowl> I still have to ask MIT to open the firewall for the new buildbot
coordinator
[16:50:18] <meffie> fyi i pushed my buildbot.openafs.org ansible playbooks to github.
[16:50:44] <meffie> they are in https://github.com/openafs-contrib/afsbotcfg
[16:50:53] <kaduk@jabber.openafs.org/barnowl> cool
[16:51:16] <meffie> i'm using this now to push changes to both the new and old buildbot masters, to keep them in sync :)
[16:52:04] <meffie> fyi, the buildbot worker passwords are encrypted with AES using ansible-vault.
[16:52:22] <kaduk@jabber.openafs.org/barnowl> nifty
[16:52:37] <meffie> is there a safe place to share the vault key?
[16:53:53] <meffie> i'm trying to avoid buses in the meantime...
[16:54:22] <wiesand> er, software or real ones?
[16:55:01] <kaduk@jabber.openafs.org/barnowl> Does it make sense to have it on disk on the main buildbot machine, or
are you thinking some sort of off-site escrow?  "I have a pgp key" but
that's not really a great answer
[16:55:06] <meffie> i mean, in case "i get hit by a bus", i want to make sure some else can has the vault key
[16:55:31] <kaduk@jabber.openafs.org/barnowl> wiesand: software key
[16:55:32] <meffie> i really dont know. i could put in on the buildbot machine.
[16:55:36] <wiesand> ah, thanks
[16:55:56] <meffie> it's just in my password manager at the moment.
[16:56:31] <meffie> i'll put in on the new buildbot master machine in the buildbot users home dir.
[16:56:41] <kaduk@jabber.openafs.org/barnowl> Okay.
[16:57:16] <kaduk@jabber.openafs.org/barnowl> Long-term I'll probably move it to a root-only location, in case
there's some vulnerability in buildbot itself that gives filesystem
access.
[16:57:49] <meffie> Maybe MIT has something for storing secrets?
[16:57:59] <yadayada> I need to leave little early today, Thanks All
[16:58:08] <meffie> thanks, have a good weekend.
[16:58:08] <kaduk@jabber.openafs.org/barnowl> I expect they have something in the puppet config, yes
[16:58:16] <kaduk@jabber.openafs.org/barnowl> Thanks, have a good weekend!
[16:58:56] <meffie> that's all i have.
[16:59:21] <kaduk@jabber.openafs.org/barnowl> Sounds good.
Adjourn?
[16:59:34] <meffie> second.
[16:59:43] <wiesand> Adjourn. Thanks a lot everyone!
[16:59:51] <meffie> have a good weekend all.
[16:59:54] <kaduk@jabber.openafs.org/barnowl> Thanks everyone!
[16:59:59] meffie leaves the room
[17:01:48] <cwills> have a good weekend everyone
[17:15:10] yadayada leaves the room
[17:20:35] mbarbosa leaves the room
[17:22:59] mbarbosa joins the room
[18:00:32] wiesand leaves the room
[18:40:11] mbarbosa leaves the room
[18:40:34] mbarbosa joins the room
[18:43:57] cwills leaves the room
[20:48:05] kaduk@jabber.openafs.org/barnowl leaves the room
[21:43:19] mbarbosa leaves the room