[01:01:41] --- Simon Wilkinson has become available [01:23:22] --- lars.malinowsky has become available [01:38:11] --- pod has left [01:41:38] --- lars.malinowsky has left [01:53:48] --- lars.malinowsky has become available [01:57:57] --- Simon Wilkinson has left [03:05:49] --- Simon Wilkinson has become available [03:18:24] --- Simon Wilkinson has left [03:21:05] --- Simon Wilkinson has become available [03:22:50] --- Simon Wilkinson has left [04:28:16] --- meffie has become available [05:43:29] --- jaltman/FrogsLeap has left: Disconnected [05:46:21] --- jaltman/FrogsLeap has become available [05:52:09] --- haba has become available [05:54:21] What do you make of the following in the FileLog: Fri Mar 25 12:30:08 2011 DT: inode=73458513593556606, name=a02c11n07.pdc.kth.se:mxrten:032507530102:20110325112717, errno=12 Fri Mar 25 12:30:08 2011 Volume 537086826 now offline, must be salvaged. [05:58:47] --- Simon Wilkinson has become available [05:59:24] file server errno 12 (ENOMEM????) [06:00:03] This is 1.4.14 on amd64 Centos5 [06:00:15] Sorry, Centos 4.2 [06:08:19] Looks to me like the link count of the IH_DEC () got wrong but the errno is just old garbage [06:10:54] /src/viced/afsfileprocs.c line 1337 [06:14:47] The question is if I should do something else before trying to salvage [06:21:18] Given that the IRIX builder had nearly 24 hours of build backlog, I've cleared it all out. [06:31:17] --- Simon Wilkinson has left [06:35:31] --- michs has become available [06:37:30] --- michs/PDC has become available [06:37:49] --- michs has left [06:42:32] --- jaltman/FrogsLeap has left: Disconnected [06:48:50] --- jaltman/FrogsLeap has become available [07:20:28] --- deason has become available [07:20:31] --- michs/PDC has left [07:53:59] --- Simon Wilkinson has become available [08:03:23] --- Simon Wilkinson has left [08:03:39] --- Simon Wilkinson has become available [08:12:27] --- Simon Wilkinson has left [08:54:12] --- reuteras has left [08:55:39] --- haba has left [10:31:22] --- jaltman/FrogsLeap has left [11:18:26] --- Simon Wilkinson has become available [11:46:21] --- Simon Wilkinson has left [12:35:44] what's with the 5 bit shift in the readdir stuff? [12:37:40] not looking at it atm, but the directory blob structure iirc uses 32-bit 'slots' for storing things; it's probably converting from bytes to slots or something like that [12:38:30] or "buckets"? I forget if there's a proper term for it [12:38:38] blob? [12:39:00] anyway [12:39:16] er, the afs in-memory and on-wire representation [12:39:33] generally treated as an opaque blob accessed by the afs_dir_* stuff [12:53:35] 32-byte blobs, yes [13:02:11] and i mave have fixed my issue [13:02:21] can't set the resid to zero [13:02:49] now, the many other issues [15:15:42] --- Russ has become available [15:33:39] --- jaltman/FrogsLeap has become available [15:46:05] --- deason has left [16:09:56] --- meffie has left