Conversation with #freeipa at Fri 07 May 2010 07:12:29 AM EDT on (irc)

(07:40:23 AM) sbose: sgallagh: I think if we remove the kdcinfo/kpasswdinfo files during shutdown we should create them at startup and not during the first kerberos operation. Do you agree? Shall I prepare a patch?
(07:43:26 AM) sgallagh: sbose: I agree.
(07:43:34 AM) sgallagh: sbose: Open a bug pleaswe
(07:43:42 AM) sbose: shanks: ok
(07:43:48 AM) sbose: shanks: sorry
(07:43:51 AM) sbose: sgallagh: ok
(07:44:26 AM) sgallagh: Right now I'm trying to figure out why several pieces of our test suite fail with MALLOC_PERTURB_ on :)
(07:44:39 AM) jhrozek: sgallagh: I have just fixes ipa timerules suite
(07:44:57 AM) jhrozek: sgallagh: apparently we had the same idea on enabling MALLOC_PERTURB :-)
(07:45:01 AM) sgallagh: hehehe
(07:46:14 AM) sgallagh: jhrozek: I'm also seeing this: (Fri May 7 07:42:10 2010) [sssd] [get_uid_from_pid] (7): Proc file [/proc/11360/status] is not available anymore, continuing.
(07:46:20 AM) sgallagh: In the find-uid suite
(07:47:48 AM) jhrozek: Hm, I'm not seeing that here..
(07:47:59 AM) sgallagh: Actually, that may have been a red herring
(07:48:06 AM) sgallagh: It only happened the first time
(07:48:08 AM) ***sgallagh sighs
(07:49:04 AM) sgallagh: sbose, jhrozek: Since you're both around, do you have time to discuss scoping for 1.2.1?
(07:49:16 AM) jhrozek: sure
(07:49:27 AM) sgallagh: We've had a number of late tickets opened up, but I don't want to push out 1.2 any further.
(07:49:56 AM) sgallagh: Since we're already doing a 1.2.1 for the pam_filter feature, I think we should consider including the bugs that Eugene opened as 1.2.1 fixes.
(07:50:25 AM) sbose: sgallagh: sounds good
(07:51:26 AM) jhrozek: fine by me.
(07:51:27 AM) sgallagh: My second thought was that this should be the *entire* scope of 1.2.1 (exceptions made for critical regressions or security bugs)
(07:52:55 AM) sbose: sgallagh: makes sense
(07:53:35 AM) sgallagh: Ok, so here are my thoughts on the schedule. Assuming we finish the reviews today, I want to release 1.1.91 as a preview of 1.2.0 immediately.
(07:54:05 AM) sgallagh: This will be string freeze, and I will start negotiating with the other interested distributions to test installation compatibility.
(07:54:21 AM) jhrozek: sgallagh: I think we will also see some bugreports once F13 goes live. Do we plan to release 1.2.1 around that time? Or have another point release?
(07:54:39 AM) sgallagh: We will release SSSD 1.2.0 on the 14th, unless we get a no-go from QE
(07:55:37 AM) sgallagh: Work on 1.1.1 should begin in parallel with the finishing touches of SSSD 1.2.0, with a targeted release of May 31.
(07:55:44 AM) sgallagh: s/1.1.1/1.2.1/
(07:56:50 AM) sbose: sgallagh: btw, the 'is not available' does not indicate an error, it just shows that the process is already gone
(07:57:36 AM) sgallagh: sbose: Yeah, I'm not going to worry about it any further.
(07:58:00 AM) jhrozek: I presume we also want to put all patches that are currently targeted for 1.2 into master soon - some will require rebasing on top of sysdb sync rewrite
(07:58:42 AM) sbose: sgallagh: shall sync sysdb be in 1.2.1 ?
(07:59:55 AM) sbose: sgallagh: I would vote for yes, because other fixing Eugene's bug would require twice the work
(07:59:56 AM) sgallagh: sbose: No, that is distinctly 1.3.0
(08:00:02 AM) sgallagh: oof
(08:00:26 AM) sgallagh: That is a valid point.
(08:00:28 AM) ***sgallagh thinks
(08:01:07 AM) sbose: sgallagh: I think there is no easy way to rebase a patch from 1.2.* to master which touches sysdb calls
(08:01:15 AM) sgallagh: You are correct
(08:01:40 AM) sgallagh: (part of me wishes that simo had left both interfaces available)
(08:02:02 AM) jhrozek: It's quite disruptive change, but it has proven to be very reliable for me
(08:02:19 AM) sgallagh: Yeah, I was using it for quite some time and haven't hit any issues
(08:02:29 AM) sbose: me too
(08:02:30 AM) sgallagh: But I still don't think it's 1.2.1
(08:03:11 AM) sgallagh: That's way too disruptive for a point release
(08:03:17 AM) sgallagh: Even if it seems stable
(08:03:40 AM) sgallagh: Ok, so now I'm having a different idea.
(08:04:08 AM) sgallagh: 1.2.1: just pam_filter + any high-priority Fedora bugs. Stays on sssd-1-2 branch
(08:04:29 AM) sgallagh: 1.3.0: Sync Sysdb + Eugene's bugs
(08:04:45 AM) sgallagh: Bump out the other tasks by a minor release.
(08:06:10 AM) sbose: sgallagh: and 1.3.0 will be released shortly after 1.2.1 ?
(08:06:44 AM) sgallagh: Probably end of June
(08:08:26 AM) simo: sgallagh, no you really don't :)
(08:09:06 AM) sgallagh: simo: No, I really don't. But this branch is brutal :)
(08:09:28 AM) sbose: sgallagh: shall the devel package still be a target for 1.3.0 or shall we move some of the trac items for 1.3 to 1.3.x or 1.4 ?
(08:09:47 AM) sgallagh: sbose: That's what I meant by bumping the tasks out.
(08:10:11 AM) sgallagh: I think we'll make the scope of 1.3.0 much smaller (since some of the big parts are already done)
(08:10:12 AM) sbose: sgallagh: ah, ok, sorry
(08:10:25 AM) sgallagh: Then what's in 1.3.0 now becomes 1.4.0, and so on
(08:11:03 AM) sbose: sgallagh: ok, then end of June sounds reasonable
(08:11:46 AM) simo: sgallagh, who is touching sysdb ? :)
(08:11:55 AM) sgallagh: simo: No one's changing sysdb
(08:12:29 AM) sgallagh: simo: But any fix written for the bugs Eugene reported will need to be rewritten (not simply ported) from 1.2 to 1.3
(08:12:41 AM) sgallagh: So we're going to opt to just do it once, on 1.3
(08:14:13 AM) sgallagh: jhrozek, sbose, simo: Are we all agreed on this course of action?
(08:14:43 AM) sbose: sgallagh: +1
(08:14:45 AM) jhrozek: sgallagh: OK
(08:15:16 AM) simo: whatever
(08:15:41 AM) sgallagh: simo: Thanks for that glowing endorsement ;-)
(08:15:49 AM) sgallagh: If you have reservations, please speak.
(08:16:26 AM) simo: sgallagh, sorry I don't know which patch you are talking about exactly so ...
(08:17:02 AM) sgallagh: simo: Tickets 465 - 469
(08:18:03 AM) simo: sgallagh, sorry connection too slow to browse
(08:18:27 AM) simo: sgallagh, mind summarizing in a few words ?
(08:19:59 AM) sgallagh: #465: When OFFLINE is detected in LDAP request, ERROR is returned
(08:19:59 AM) sgallagh: #466: LDAP backend pollutes event log when using GSSAPI authentication
(08:19:59 AM) sgallagh: #467: LDAP backend reconnects to LDAP server every 10 minutes when using GSSAPI authentication
(08:19:59 AM) sgallagh: #468: RFE: Track usage of LDAP connection by BE requests
(08:19:59 AM) sgallagh: #469: IPA access backend does not use failover retries
(08:20:30 AM) simo: ah ok
(08:21:20 AM) simo: yes I think we need to go on and just work on 1.3 and release that stuff there
(08:21:27 AM) simo: sgallagh, the only change I would make is to change the 10 minutes we force when acquiring credentials to the max available
(08:21:28 AM) simo: (in 1.2)
(08:21:41 AM) simo: it is a one line patch after all
(08:23:20 AM) sgallagh: simo: Yeah, true enough