[01:38:42] --- abo has left [01:38:42] --- abo has become available [05:39:22] --- Marc has become available [05:39:39] --- Marc has left [05:56:20] --- mvita has become available [07:15:25] --- wiesand has become available [07:28:04] ping marc [07:29:56] Is anyone here who could kick Marc Dionne out of this room? [07:30:16] I doubt it. That would need to be russ and he isn't here [07:33:06] --- meffie has become available [07:49:40] --- Marc has become available [07:50:01] We'll try this [07:50:29] Hi Marc. [07:50:49] Hi again, just switched to a jabber.openafs.org account for this [07:52:41] so we were discussing the issue brought up in https://lists.openafs.org/pipermail/openafs-devel/2013-December/019681.html [07:53:23] See https://bugzilla.redhat.com/show_bug.cgi?id=1038315 [07:54:04] Poposed fix in http://gerrit.openafs.org/10545 [07:54:59] --- deason has become available [07:55:25] thanks for the recap [08:18:58] Interestingly, there's a knowledge base article from April about openafs related panics on RHEL5 looking very similar: https://access.redhat.com/site/solutions/267593 [08:21:08] i think there's a link to that in the BZ [08:51:09] No, that's a new one specifically for RHEL 6.5. They have a couple of those ;-) [08:52:07] I think all it takes to be affected is to have auditd installed. [08:52:24] I also think that's not the default. [08:52:34] Hmm, I wonder if that's why my rhel5 machine was panicing yesterday. [08:53:12] (The other contender is that my kernel-devel package doesn't quite match my running kernel.) [08:53:25] i think auditing is installed by default and the service activated. but to trigger i think you need to actually have auditing rules in place [08:53:44] Where would I look to see if there are auditing rules in place? [08:53:51] auditctl -l [08:54:15] Hmm, "No rules". Maybe it's the -devel package, then. :-/ [08:54:42] (This machine is not entirely "mine", so I am limited about installing kernels/etc.) [08:54:51] The service is activated when the package is installed. [08:57:29] yes but an audit context is attached to a process only if auditing is currently ongoing for that task [08:58:07] and that's what triggers the special name handling [09:05:12] trying something much more simple with just kmalloc and kfree [09:24:34] Anyway, the kernel-devel package that I have appears to be for 2.6.18-238.9.1.el5-x86_64 but I am running 2.6.18-348.el5. Would you expect that to make things sad? [09:27:15] yes. :) We use many things that are not part of the KABI that they try to be careful not to break. the differences between kernels can be substantial, esp. if you go between minor releases [09:27:36] Okay. I will see if I can get someone to install a -devel package and running kernel that match. [09:28:27] (As it happens, the kernel module loads and I was able to start afsd at least once, but I can't stop it cleanly as some threads hang around, and it panic'd at some point, too.) [09:28:30] worst case, rpm2cpio | cpio -id into /tmp and point configure to that [09:28:37] for fun I did a diff of the diffrences between the last RH 6.4 kernel and the 6.5 one... 19.5MB [09:29:00] Yikes [09:32:20] (This is me trying to get rxgk-bos running on our test cell, by the way.) [09:34:21] Gotta leave, or I'll need an ice axe to get into my car. I'll be back later. [09:34:26] --- wiesand has left [10:12:05] --- kaduk@jabber.openafs.org/barnowl has left: Disconnected [10:13:46] --- kaduk@jabber.openafs.org/barnowl has become available [10:14:03] --- stephan.wiesand has become available [10:38:41] --- meffie has left [10:44:55] --- meffie has become available [11:15:25] --- meffie has left [11:25:00] --- deason has left [11:25:01] --- deason has become available [11:46:02] --- meffie has become available [11:53:16] --- stephan.wiesand has left [13:54:06] --- meffie has left [15:18:34] --- deason has left [17:12:05] --- jaltman/FrogsLeap has left: Replaced by new connection [17:12:06] --- jaltman/FrogsLeap has become available [18:27:46] --- mvita has left [18:52:18] --- Marc has left: Disconnected [19:21:55] --- ballbery@sinenomine.net has left [19:24:20] --- ballbery has become available [21:59:28] Well, my panics were also partially because I was installing afsd into a different path than my init script was calling it. [21:59:28] --- Marc Dionne has left: Lost connection [22:01:04] But, hey, the rxgk-enabled bosserver is running now.