+++ This bug was initially created as a clone of Bug #440896 +++ This was cloned so I can crosswrite the fix to RHEL5. Description of problem: The problem described in bug #436383 (RHEL4) and bug #434962 (RHEL5) introduces corruption into a gfs file system. Today's gfs_fsck is unable to repair that damage, and I think it needs to. Version-Release number of selected component (if applicable): RHEL5 How reproducible: Easily (now that I know the trick). Steps to Reproduce: 1. Follow the steps listed in bug #436383 to corrupt a file system 2. Run gfs_fsck on the corrupt file system Actual results: gfs_fsck's resource group analysis says "(passed)" and no corruption is detected. [root@exxon-01 /home/bob]# gfs_fsck /dev/exxon_vg/exxon_lv Initializing fsck Can't seek to last block in file system: 4309132281 Unable to determine the boundaries of the file system. Expected results: gfs_fsck should detect and repair the damage gracefully. Additional info: I've got a prototype patch that does this. The RHEL4 version is attached to bug #436383. It still needs some testing.
Created attachment 302272 [details] Patch to fix the problem This patch to gfs_fsck detects and repairs the rindex corruption caused by gfs_grow in bug #436383. This patch does NOT include any of the other bugs I found and fixed in gfs_fsck related to rindex repair. The right thing to do is open a new bugzilla record for those fixes, which I will do shortly. I do have that set of fixes for gfs_fsck written, debugged, tested and in-hand as well, but it's better to do that with a different bugzilla record.
This fix passes my hours-long "gfs_fsck_hellfire" test with flying colors. I pushed the fix to the RHEL5, STABLE2 and master branches of the cluster git tree. Now I need some ack flags from the appropriate parties so I can mark this modified.
Tested on roth-01. Changing status to modified.
An advisory 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 therefore 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-2009-0059.html