- Friday, September 20, 2019

release-team@conference.openafs.org

Friday, September 20, 2019< ^ >

Room Configuration

Room Occupants

GMT+0
[04:22:18] mvita leaves the room
[14:01:57] meffie joins the room
[14:53:35] wiesand joins the room
[14:54:19] <meffie> good afternoon
[14:54:33] <kaduk@jabber.openafs.org/barnowl> Hi Mike
[14:59:24] <wiesand> Hello
[14:59:47] yadayada joins the room
[14:59:55] <yadayada> Hello all
[15:00:10] <wiesand> updated 13853 - I think I'll unleash 1.8.4pre2 today, with or without feedback…
[15:00:23] mvita joins the room
[15:00:29] <mvita> hi
[15:00:42] <wiesand> hi
[15:01:30] <wiesand> the next thing I'll tackle is then the 1.6 mess - I started looking into it, but haven't found out what went wrong yet
[15:02:26] <kaduk@jabber.openafs.org/barnowl> There's a 1.6 mess?
[15:02:29] <wiesand> 1.8.x already looks daunting too - dozens of changes already pulled up, and I bet there are path conflicts all over the place
[15:02:41] <wiesand> yes there is
[15:03:51] <wiesand> there is no straightforward path from the current HEAD to a 1.6.24pre1 release
[15:04:13] <wiesand> may take a couple of reverts and reworking changes
[15:04:33] <kaduk@jabber.openafs.org/barnowl> For master I can't get away with having a plan for what I want to
merge first, so I just take things that are ready and have to rebase
stuff multiple times sometimes
[15:05:55] <wiesand> that's unavoidable - but for 1.6 the rebases of the changes I had planned to merge next would be not trivial at all
[15:06:15] <wiesand> and that shouldn't happen on that branch IMO
[15:06:18] <kaduk@jabber.openafs.org/barnowl> Ah.
[15:06:35] <kaduk@jabber.openafs.org/barnowl> Feel free to poke people for help if things are too hairy
[15:07:00] <mvita> yes, absolutely
[15:07:23] <wiesand> once I have a better understanding of what went wrong (chances are I screwed up anyway) I will
[15:09:39] <wiesand> thanks for the +1s on 13853 - pre2 will be announced today then
[15:10:17] <wiesand> I think I'll start working on it right away, since I have no other stable release topics anyway
[15:10:48] <wiesand> if nobody else has, on to master/1.9 then?
[15:11:06] <yadayada> On RHEL8 I tried pam_sss.so + pam_afs_session.so for Integrated login and it works by making some config changes. So as we thought pam_sss.so will get us TGT and pam_afs_session will get us PAG based token.  I will create a doc for same and share early next week
[15:11:19] <kaduk@jabber.openafs.org/barnowl> Very cool; thank you!
[15:11:49] <wiesand> quite relieving, thanks!
[15:12:44] <yadayada> I will check on GNOME based session and how PAG based token works for RHEL8
[15:12:50] <yadayada> Now one query I had
[15:13:04] <wiesand> I just was about to say "though I still have doubts regarding PAG based tokens on EL8…"
[15:14:03] <yadayada> pam_afs_session always create PAG based tokens.  I was thinking to have GNOME/GDM PAM config to somehow create UID based tokens. Does pam_afs_session allows UID based token ? OR is there some other way we can get UID based token using PAM ?
[15:15:06] <wiesand> nopag
If this option is set, no PAG will be created. Be careful when using this option, since it means that the user will inherit a PAG from the process managing the login. If sshd, for instance, is started in a PAG , every user who logs in via ssh will be put in the same PAG and will share tokens if this option is used.
[15:15:24] <yadayada> I tried that but no luck with that
[15:15:41] <yadayada> I was getting PAG based token even by giving nopag option
[15:16:02] <wiesand> sigh
[15:16:24] <yadayada> Need to check if there is some issue with pam_afs_session on this.
[15:17:09] <wiesand> what kin(s) of login have you tried yet? gdm? ssh?
[15:18:14] <yadayada> I tried ssh till now, gdm is still pending .. but was thinking if gdm is started by systemd then PAG will not work, so try to get UID based token for it
[15:22:00] <yadayada> need to work more on gdm based login, will send some data next week on this
[15:22:22] <wiesand> thanks again!
[15:23:30] <meffie> thank you Yadav!
[15:23:35] <wiesand> 1.8.4pre2 web change is live; writing the announcement now
[15:23:51] <mvita> huzzah!
[15:24:13] kaduk@jabber.openafs.org/barnowl leaves the room: Replaced by new connection
[15:24:22] kaduk@jabber.openafs.org/barnowl joins the room
[15:24:49] <kaduk@jabber.openafs.org/barnowl> That was a really lousy time for a power blip
[15:25:20] <kaduk@jabber.openafs.org/barnowl> had to power-cycle my cable modem, as it was quite confused afterward
[15:26:09] <kaduk@jabber.openafs.org/barnowl> I got a few things done yesterday, though not much progress on the
"big project" patch series
[15:26:52] <kaduk@jabber.openafs.org/barnowl> Mark's solaris-rt135006 changes seem to have exposed some CFLAGS
duplication in lwptool, so we may have to think about the best way to
fix that
[15:27:09] <mvita> looking at that now
[15:27:18] <kaduk@jabber.openafs.org/barnowl> (I think making LWP_CFLAGS be very minimal, just things that are
LWP-specific, would be a reasonable approach)
[15:27:45] <mvita> they didn't seem duplicated to me in my tests - but I see where you are coming from
[15:27:53] <mvita> yes
[15:28:51] <mvita> flags + lwptools = self-induced hair loss
[15:29:04] <kaduk@jabber.openafs.org/barnowl> I'll have to remember that expression
[15:29:54] <mvita> maybe I'll get meffie to help me, he's good at mazes of twisty little passages, all different
[15:30:04] <meffie> heh
[15:30:12] <wiesand> announcement sent, moderators pestered to let it pass
[15:30:14] <kaduk@jabber.openafs.org/barnowl> MODULE_CFLAGS I would expect to frequently be empty, but COMMON_CLFAGS
would not
[15:30:27] <mvita> aye
[15:31:00] <mvita> an early version of my fix set CFLAGS_rxkad_server.o
[15:31:03] <meffie> (we should try to use our limited time on killing lwp)
[15:31:08] <mvita> but that was deemed ugly
[15:31:42] <meffie> just keep the changes for lwp small if we can :)
[15:31:48] <kaduk@jabber.openafs.org/barnowl> ugly, and not really the right fix either
[15:32:01] <mvita> it was a prototype
[15:33:18] <mvita> ok, what else?
[15:33:59] <kaduk@jabber.openafs.org/barnowl> jhutz pinged me on 2288 this week
[15:34:41] <mvita> so I noticed
[15:34:48] <meffie> ah yes. fast restarts would be nice (esp for testing)
[15:34:51] <kaduk@jabber.openafs.org/barnowl> So I took off my -1 but would welcome broader perspective
[15:36:04] <meffie> ok
[15:36:45] <meffie> i could see this as a really nice feature.
[15:37:13] <meffie> we can take some time to review and test it.
[15:37:33] <kaduk@jabber.openafs.org/barnowl> cool; thank you
[15:39:31] <meffie> i was wondering, if in a test cell we could use this to "simulate" a warm start. the startup of the db servers is the bottleneck in the ansible roles. so i'd i have a keen interest.
[15:39:58] <kaduk@jabber.openafs.org/barnowl> It seems probable that you could do so, sure
[15:40:05] <meffie> thanks.
[15:41:30] <meffie> mark, andrew, and marcio would be better at an actual code review. we will try to make time for that.
[15:41:49] <kaduk@jabber.openafs.org/barnowl> I'm failing to remember anything else I wanted to talk about
(distracted by the power issue, no doubt)
[15:42:36] <wiesand> power issue?
[15:42:37] <meffie> any feelings about the 'ipv6-prep-format-helper'
[15:43:25] <meffie> does that approach make sense?
[15:43:57] <kaduk@jabber.openafs.org/barnowl> wiesand: I had a power blip and lost my main VM and network for a few
minutes
[15:45:12] <kaduk@jabber.openafs.org/barnowl> I haven't had a terribly close look at ipv6-prep-format-helper, but it
seems roughly in line with what I expected, from what I've seen
[15:46:13] <meffie> ok thanks.
[15:47:05] <kaduk@jabber.openafs.org/barnowl> (I tried to send mail to release-team from my phone but it got
moderated, sigh)
[15:47:07] <meffie> once those are merged, we can look at the next stage of the int32 to sockaddr conversion.
[15:47:48] <kaduk@jabber.openafs.org/barnowl> Indeed
[15:48:05] <meffie> thanks.
[15:50:03] <kaduk@jabber.openafs.org/barnowl> I have the 1.9.0 NEWS open in a tab to go through, as well :)
[15:50:46] <kaduk@jabber.openafs.org/barnowl> Anyway, not much else from here.  Any other topics?
[15:51:08] <mvita> nothing from me
[15:51:26] <meffie> nothing else, thanks.
[15:52:00] <wiesand> nothing from me - adjourn?
[15:52:09] <mvita> second
[15:52:34] <kaduk@jabber.openafs.org/barnowl> Thanks everyone!
[15:52:40] <wiesand> Thanks a lot everybody. Have a nice weekend!
[15:52:41] <meffie> have a good weekend all.
[15:52:44] <mvita> bye
[15:52:59] wiesand leaves the room
[16:02:25] yadayada leaves the room
[16:06:32] meffie leaves the room
[16:14:49] kaduk@jabber.openafs.org/barnowl leaves the room
[16:49:22] kaduk@jabber.openafs.org/barnowl joins the room
[16:53:59] meffie joins the room
[16:54:03] meffie leaves the room
[17:03:09] mbarbosa joins the room
[17:04:08] mbarbosa leaves the room