[00:08:19] --- shadow@gmail.com/barnowlECEA8F00 has left
[00:08:45] --- shadow@gmail.com/barnowlECEA8F00 has become available
[01:09:58] --- wiesand has become available
[05:47:56] --- meffie has become available
[06:57:18] --- Marc Dionne has become available
[07:01:17] <wiesand> Happy New Year everyone
[07:02:50] <Marc Dionne> happy new year to you and everyone
[07:03:03] --- deason has become available
[07:04:07] <Jeffrey Altman> happy 2014.  may it be healthy and prosperous for all.
[07:04:23] <deason> hello from the frozen u.s.
[07:05:31] <wiesand> Hello from the springtime Europe.
[07:06:01] <wiesand> Marc, any Linux news?
[07:06:41] <Marc Dionne> nothing new - still ok with current mainline as far as i can tell
[07:07:05] <Marc Dionne> ETA for 3.13 would be around the weekend of jan 18
[07:07:22] <Jeffrey Altman> the most interesting news that I heard is that CentOS is now a Red Hat community
[07:07:37] <wiesand> Good. This late in the release cycle, there's hope we'll need no further changes for Linux 3.13.
[07:08:09] <wiesand> Yes, it seems they hired the core CentOS team.
[07:09:01] <wiesand> Probably a good thing. No impact on OpenAFS though, I believe.
[07:09:30] <wiesand> Does anyone have new pre2 test results?
[07:10:54] <wiesand> Probably not... I have the client running on EL7.0 beta on an xfs cache. Works fine.
[07:11:48] <wiesand> Andrew, thanks for bringing up the server side NAT pings on the mailing list a while ago.
[07:12:01] <wiesand> Was there any more feedback on this?
[07:12:30] <deason> there are a couple of people who contacted me about it, but we probably won't see any resuts soon enough, so I'm fine with not waiting for pulling it out for now
[07:13:08] <wiesand> Er, you're in favour of pulling it out now?
[07:13:42] <kaduk@jabber.openafs.org/barnowl> No pre2 testing from me, alas; I'm still recovering from having my
laptop stolen, but I think I'm mostly done.
[07:13:42] <wiesand> (failing to decode "fine with not waiting for pulling it out for now")
[07:14:06] <wiesand> Sorry to hear that.
[07:14:10] <deason> I mean, depending on results it may make sense to keep it in, but I don't think we should wait for those; it'll take too long
[07:14:14] <deason> so sure, we can take it out
[07:15:16] <wiesand> It seems to behave reasonably on the fileserver here.
[07:17:10] <wiesand> At this point we probably want to wait for Linux 3.13 anyway. So ripping it out - if we really want to - can wait a few more days.
[07:17:55] <wiesand> Is anyone strongly in favour of reverting server side NAT pings?
[07:19:28] <wiesand> What I think we want in 1.6.6 are the spec file changes for F20 and EL7.
[07:19:46] <wiesand> How do we get those merged on master?
[07:20:52] <kaduk@jabber.openafs.org/barnowl> We need more +1s, right?
[07:20:52] <deason> did anyone try them? I pushed the 1.6 versions so RPMs could be built on the place that actually makes the rpms that will get used
[07:21:01] <Jeffrey Altman> I don't think strongly is the right word but I still do not believe that the pings can help in the majority of cases and will result in the fileserver spewing packets on the wire unnecessarily.   
[07:22:38] <deason> f20/el7 stuff is 10618 and 10620, btw
[07:22:44] <Jeffrey Altman> I would like to see an approval from Ken Dreyer on spec file changes
[07:24:07] <wiesand> Seems reasonable. Maybe we'll get that ball rolling again later.
[07:25:16] <wiesand> These do not warrant a 1.6.6pre3 IMO.
[07:25:37] <wiesand> Is there any other reason for a pre3?
[07:25:57] <Marc Dionne> i think i had tested the packaging on f20 before xmas and that it was ok.  but i'm not a regular user of the packaging stuff
[07:27:00] <wiesand> I can test the EL7 changes. No F20 here yet.
[07:28:52] <wiesand> btw, if anyone knows how to unbreak the kerberos credentials cache on EL7 beta, I'd love to hear about it...
[07:29:15] <Marc Dionne> hmm, is it using the keyring: cache type?
[07:29:18] <kaduk@jabber.openafs.org/barnowl> This has a new keyring cache type? 
[07:29:23] <shadow@gmail.com/barnowlECEA8F00> zephyr was talking about that yesterday
[07:29:44] <wiesand> yes, by default it uses the keyring cache.
[07:29:59] <Marc Dionne> and what's the breakage
[07:30:00] <wiesand> and the same for any session
[07:30:12] <wiesand> and I haven't found how to change that.
[07:30:12] <kaduk@jabber.openafs.org/barnowl> This is maybe not the best forum for it, but what's the breakage?
[07:30:39] <wiesand> log in on the console. then try to log in with ssh.
[07:31:08] <wiesand> for selinux, those are different keys the other login process can't access.
[07:31:44] <kaduk@jabber.openafs.org/barnowl> Let me pull up a pointer to where to change the default.
[07:31:52] <Jeffrey Altman> If I remember Simo's design it is per-user keyring (one per machine)
[07:32:23] <kaduk@jabber.openafs.org/barnowl> sshd has a restricted selinux policy, though, which gets applied for
the ssh logins, apparently.
[07:33:13] <wiesand> You ene can't log in on a VC once you logged in under gdm or through ssh...
[07:33:48] <kaduk@jabber.openafs.org/barnowl> http://web.mit.edu/kerberos/krb5-latest/doc/admin/conf_files/krb5_conf.html#libdefaults
, look for default_ccache_name.
"DEFCCNAME" is set at configure time, and ISTR that the fedora/red hat
packages set it to the keyring type there.
[07:34:39] <wiesand> Without SELinux, things kind of work.
[07:34:42] <kaduk@jabber.openafs.org/barnowl> FILE:/tmp/krb5cc_blah is still the lowest common denominator, alas.
[07:35:43] <wiesand> Thanks Ben. I tried that yesterday, and managed to get different Caches for ssh and the others (because this seems not to be in effect for ssh). But still the same for every ssh login, and the same for every VC/GDM login.
[07:36:06] <wiesand> The good news is that AFS PAGs work fine.
[07:36:18] <kaduk@jabber.openafs.org/barnowl> Simo said they were looking at fixes on the selinux side, but I don't
have any details.
[07:37:38] <wiesand> I'd much rather have one cache per session.
[07:38:21] <wiesand> But enough of my whining. Thanks for the input.
[07:39:29] <wiesand> So I think the plan for 1.6.6 is to release it right after Linux 3.13, with the spec cahnges, and possibly without server side NAT pings.
[07:40:19] <wiesand> Any chance we could get binaries for OS X for people to test?
[07:42:10] <wiesand> Ok, probably not :-(
[07:42:49] <wiesand> Does anyone see a need for 1.6.5.3?
[07:43:22] <wiesand> (maybe including the LINUX_USER_NS change?)
[07:44:09] <wiesand> (but then Russ already has packaged 1.6.6pre2 for debian unstable, and that's probably the only case?)
[07:45:31] <wiesand> I'll take the silence as "no need for 1.6.5.3".
[07:45:44] <wiesand> Anything else to discuss today?
[07:45:54] <deason> if debian-based stuff is the only practical thing that needs it... debian builds their own packages, so they don't need a release from us
[07:47:14] <wiesand> It's the only one I know about.
[07:47:49] <wiesand> Is there any news on the 1.8/1.9 branches?
[07:49:43] <wiesand> Taking this as a no too.
[07:50:15] <wiesand> I think we're done for today. Thanks a lot everyone.
[07:50:43] <Jeffrey Altman> my recent focus has been on the 70+ windows changes sitting in gerrit.   no time for anything else over ht eholidays
[07:50:55] <wiesand> Seen them ;-)
[07:55:26] <meffie> thank you.
[07:57:41] <wiesand> Bye.
[07:57:41] --- Marc Dionne has left
[07:57:43] --- wiesand has left
[07:57:47] <Jeffrey Altman> bye
[07:58:05] --- Marc Dionne has become available
[09:19:05] --- Marc Dionne has left
[12:18:28] --- deason has left
[12:18:29] --- deason has become available
[15:16:02] --- meffie has left
[16:02:30] --- deason has left