[00:29:07] --- Russ has left: Disconnected [00:44:38] --- haba has become available [03:44:00] So what is my volserver doing if it is in a futex(0x1b342ea4, FUTEX_WAIT_PRIVATE, 235553, NULL) = 0 .... futex(0x1b342ea4, FUTEX_WAIT_PRIVATE, 235557, NULL ..,.. [03:44:05] most of its time? [04:09:08] No, that was just one thread. Hmm. Digging more into strace logs. [04:25:11] When doing a volume dump, each file is fstat() _twice_ between open() and read()? [04:25:41] Like: [04:25:44] [pid 4054] open("/vicepa/AFSIDat/i/iMs+U/B0/X4/qRoM=Ax=53", O_RDWR) = 5 [pid 4054] fstat(5, {st_mode=S_IFREG, st_size=90, ...}) = 0 [pid 4054] fstat(5, {st_mode=S_IFREG, st_size=90, ...}) = 0 [pid 4054] read(5, "3bs0A 14 16 14\n1t16A 14 14 13\n2q"..., 90) = 90 [pid 4054] close(5) = 0 [04:34:22] you're not AIX, are you? [04:36:38] Linux [04:37:21] I am trying to find my way through the ifdefs. dumpstuff.c, isn't it? [04:37:27] ah, i guess aix uses fstat and then fstatfs [04:37:48] but lookking at volser/dumpstuff.c, it's not obvious why we'd fstat twice [04:38:00] ah, wait, hang on [04:38:54] yeah, this is dumb [04:38:55] * haba has this $USER who has all his data in files of size around 100 bytes. [04:39:30] (which is probably dunb as well) [04:44:22] 5398 [04:47:25] ? [04:47:35] gerrit [04:51:19] Ok. yes. nicely hidden between ifdefs [04:53:11] is there any reason to use "size" in that function? s/size/howBig/ in function? [04:54:09] If I read this correct, size and howBig are set once and then only read. Correct? [04:55:02] size could die. that was the minimum amount of work. of course if we have no easy way to try it i will just finish the patch now [04:57:17] test: But I could apply patch, kill volserver and bosserver will start the new one for me, without salvage, right? What was now the correct signal till kill it now again. [04:57:42] for volserver? kill -9 is fine [04:58:45] Ok, I will have to drive to KTH, patch a little, kill a little and will then come back to you with how the new volserver is doing. [04:58:53] 3h [04:59:58] ok. i'll try to contain my excitement [05:00:11] :-) [05:01:20] --- jaltman/FrogsLeap has become available [05:16:13] --- haba has left [06:03:32] --- Simon Wilkinson has become available [06:15:34] --- jaltman/FrogsLeap has left: Disconnected [06:23:32] --- Simon Wilkinson has left [06:28:50] --- jaltman/FrogsLeap has become available [06:42:55] --- Simon Wilkinson has become available [06:47:44] --- Simon Wilkinson has left [07:13:09] --- mfelliott has become available [07:14:43] --- Simon Wilkinson has become available [07:24:36] --- jaltman/FrogsLeap has left: Disconnected [07:27:11] --- jaltman/FrogsLeap has become available [07:35:55] --- deason/gmail has become available [07:43:52] --- Simon Wilkinson has left [08:11:31] --- haba has become available [08:30:52] --- meffie has left [08:37:02] --- meffie has become available [09:15:03] shadow: And at my work all the other interrupts took these 3h. [09:38:25] --- haba has left [09:49:19] --- jaltman/FrogsLeap has left: Disconnected [09:52:04] --- jaltman/FrogsLeap has become available [10:18:11] --- jaltman/FrogsLeap has left: Disconnected [10:22:23] Now that 1.6 is out, will 1.4.x cease to be supported? [10:23:07] --- jaltman/FrogsLeap has become available [10:27:09] there's going to be at least one more 1.4 bugfix release [10:27:40] --- Russ has become available [10:27:54] Also the “1.4.x Maintenance Release” link on navbar.html ought to be updated. [10:30:47] I’m trying to decide what to do with the Ubuntu PPAs. Currently https://launchpad.net/~openafs/+archive/stable is 1.4.14 and https://launchpad.net/~openafs/+archive/master is 1.6.0. [10:33:00] openafs considers 1.6.0 to be latest greatest stable [10:33:11] we would appreciate it if upstreams did as well [10:41:41] Is it useful to move 1.4.x to a stable-1.4 PPA, or should I just delete it? [10:42:31] at this point you should probably move it [10:42:35] tho i would imagine its life will not be long [10:42:59] there are probably sites that would prefer to stay on 1.4. If they are on ubuntu, they will appreciate packages with additional security fixes [10:49:58] Alright. Now https://launchpad.net/~openafs/+archive/old-1.4 is 1.4.14 (I should probably update it more if anyone’s going to use it, though), https://launchpad.net/~openafs/+archive/stable is 1.6.0, and https://launchpad.net/~openafs/+archive/master is 1.6.0. [10:51:40] --- meffie has left [10:53:16] thank you [10:57:29] --- meffie has become available [11:52:44] --- Simon Wilkinson has become available [11:57:57] --- steven.jenkins has left [11:58:21] --- steven.jenkins has become available [12:30:35] --- haba has become available [12:33:06] Well that's funky. My history of this room arrived in entirely reversed order. [12:47:32] --- jaltman/FrogsLeap has left: Replaced by new connection [12:47:34] --- jaltman/FrogsLeap has become available [12:49:10] --- Russ has left: Disconnected [13:01:20] --- Russ has become available [13:09:34] --- jaltman/FrogsLeap has left: Disconnected [13:21:09] --- jaltman/FrogsLeap has become available [13:22:35] --- meffie has left [13:26:42] --- meffie has become available [14:09:38] There still no vos endtrans in openafs? [14:10:00] $ vos help endtrans [14:10:01] vos endtrans: end a volserver transaction [14:10:15] been there for a while I think [14:11:51] Ah, it's a 1.5/1.6 feature [14:12:13] I had a 1.4.14 vos here [14:13:20] * haba seems to have wrong version on wrong computer (always) [14:13:42] I find that cfengine can help alleviate those issues [14:27:18] --- haba has left [16:05:43] --- deason/gmail has left [17:09:49] --- meffie has left [17:53:19] --- summatusmentis has left [19:50:08] --- abo has left: Lost connection [19:50:08] --- mho has left: Lost connection [19:50:08] --- reuteras has left: Lost connection [23:06:51] --- reuteras has become available [23:06:51] --- reuteras has left [23:07:02] --- abo has become available [23:10:36] --- reuteras has become available [23:24:45] --- jaltman/FrogsLeap has left: Disconnected