Description of problem: looping fsck giving a "unresolvable (cyclical) error" Version-Release number of selected component (if applicable): GFS 5.2.1 on RHEL3 How reproducible: unknown Steps to Reproduce: 1. mount gfs and run stuff 2. reboot storage device that contains GFS filesystem 3. run gfs_fsck Actual results: looping fsck Expected results: fsck fixing problems and exiting normally Additional info: Ask Erling for 594M logfile containing output from fsck
The problem was that two inodes were pointing to the same data and both inodes existed and were linked to from a directory. One inode would get removed, but later, because it was detected in the FTW, it would be readded. When the fsck was restarted, it would happen again. Now, when I remove the inode, I destroy the meta header as well, so that it is considered invalid during the FTW and does not get readded.
Created attachment 105909 [details] file containing a file system that has block conflicts Steps to test: 1. uncompress attachment 2 [details]. run fsck on it 3. mount the file 'mount -t gfs <file> <location> -o loop
Fixed in 5.2.1, 6.0, and 6.1
An errata has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2004-659.html