From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.2) Gecko/20040803 Description of problem: if a pid on gfs was holding posix range locks and the process died from a signal, the posix range lock was not released. Later, other processes that might try to get the same lock would fail or block forever. A cluster reset was required to get out of this. Version-Release number of selected component (if applicable): cvs head and rhel4 branch How reproducible: Always Steps to Reproduce: plock, kill, damn. Additional info:
Easy enough fix. actually just had the state inverted. Wasn't calling the upside posix lock update function when I should have been. Do that right, and this bug is gone.
I believe this was fixed about 5 years ago, but was never added to any erratum.