[00:30:44] --- shadow@gmail.com/barnowl29570F23 has left [00:30:56] --- shadow@gmail.com/barnowl29570F23 has become available [01:07:54] --- Simon Wilkinson has become available [01:23:50] --- Simon Wilkinson has left: Lost connection [02:59:51] --- sxw has become available [03:21:02] --- sxw has left [06:23:13] --- reuteras has left [07:12:44] --- mfelliott has become available [08:02:47] --- deason has become available [08:23:34] --- sxw has become available [08:34:58] --- sxw has left [09:34:42] --- phalenor has become available [09:49:32] --- rra has become available [10:32:43] --- jaltman/FrogsLeap has left: Replaced by new connection [10:32:44] --- jaltman/FrogsLeap has become available [15:10:19] jaltman: ping [15:10:19] --- steven.jenkins has left: Lost connection [15:10:32] --- steven.jenkins has become available [15:56:12] the machine hosting the buildbot master will be going done for maintenance tonight around 8pm for four hours. [15:56:19] EDT [15:58:40] so, -settime on master is pretty broken at the moment... should I even be bothering to try to fix this, or can we just get rid of it? (on master only) [16:00:33] do you know what broke it? [16:01:41] my personal feeling is that afsd should not be setting the clock of the system and that the option should go away in the next major release. However, I would like to know why it broke before ripping it out. [16:01:51] and of course I would like Derrick's opinion [16:01:56] the introduction of the caps stuff I think was the first thing, and also the further refactoring of the checkserver-y loops [16:02:28] certainly if gettime rpcs are not being issued the time can't be set [16:02:36] (it needs to be fixed on 1.6, but I was going to on master first if it's going to be fixed there at all) [16:03:18] if -settime is set, then afsd will have to issue gettime rpcs even if the capability rpcs returned [16:03:27] well, they get issued; I just mean the client-side processing code is just wrong [16:05:19] I mean, the breakage is very local to the -settime stuff, if you're worried about a deeper issue or something [16:06:07] if the fix is the same for both master and 1.6, apply to master. we will rip it out later [16:06:30] if the fix is different, then we will have to think about it [16:07:01] it's different [16:07:05] 1.6 fix is trivial [16:07:54] then fix 1.6 for now [16:08:17] btw, thanks for checking klog.krb5 [16:11:21] okay; and np, sorry that took a few days [16:11:26] --- deason has left [16:12:04] --- deason has become available [16:12:30] --- andersk has left [18:39:37] --- rra has left: Disconnected [18:57:57] --- Russ has become available [19:12:28] > so, -settime on master is pretty broken at the moment it should be ditched on master, imo, but i guess we do need to keep it on 1.6 for now [19:40:21] buildbot master is back up [19:43:52] --- jaltman/FrogsLeap has left: Disconnected [19:44:09] --- jaltman/FrogsLeap has become available [19:58:17] Derrick, did you get a chance to retest your ukernel client for 'all buffers locked' with Simon/Marc's patch? [19:58:34] it didn't reproduce, but that proves nothing [20:41:55] --- jaltman/FrogsLeap has left: Disconnected [21:25:06] --- jaltman/FrogsLeap has become available [21:29:23] --- jaltman/FrogsLeap has left: Disconnected [21:39:54] --- deason has left [21:47:30] --- jaltman/FrogsLeap has become available [22:09:31] --- steven.jenkins has left [22:09:55] --- shadow@gmail.com/barnowl29570F23 has left [22:11:19] --- shadow@gmail.com/barnowl29570F23 has become available [22:55:42] --- steven.jenkins has become available [23:35:29] --- reuteras has become available