On Monday 30 June 2003 13:23, Wayne Tucker wrote: > Greetings, > > I have a 120GB JFS volume on a busy server that is giving me > problems. I have a file that causes processes to permanently hang > when they attempt to read it. Interactive processes do not respond > to a ^C, nor do they respond to a SIGKILL or SIGHUP.
This sounds like a deadlock in the JFS code. > The system is a dual K7 running Linux kernel version 2.4.20, and the > version of JFS that came with it. (I could not find the specific > revision number in the source - where would I need to look?) Is this a vanilla 2.4.20 from kernel.org, or a vendor kernel? If vendor, which kernel? There have been a couple fixes since 2.4.20 that may fix your hang. I'm not aware of any similar hangs reported on 2.4.21. > The filesystem is also on an EVMS volume (engine version 1.2.0), but > I am not doing anything fancy with it at the moment. Also, since it > is just one file and I am unable to get any data from that file, I > believe/hope it to be at the filesystem level. I assume that it probably is. > This same problem happened once before, and was fixed by a reboot and > fsck (it needed a hard reboot since init could not kill some of the > processes). Unfortunately, the technician that handled the reboot > did not save the output from the FSCK, but there were a number of > errors, and the file that was causing the problem was deleted as part > of the FSCK. > > Does this sound like anything obvious that has been seen before? (I > didn't find anything on the archives, but it could be that I just > didn't know what to look for). If not, what other information should > I provide? A listing of any processes in the "D" state from "ps -el" may help me determine if it looks like a known problem or not. > Many thanks, > > Wayne Thanks, Shaggy -- David Kleikamp IBM Linux Technology Center _______________________________________________ Jfs-discussion mailing list [EMAIL PROTECTED] http://www-124.ibm.com/developerworks/oss/mailman/listinfo/jfs-discussion
