Description of problem: On devices of certain sizes, gfs_mkfs will die and print "confused computing bitblock length". The problem is in a routine that takes a target size for a Resource Group and figures out how many blocks in the RG will be bitmap blocks and how many will be free blocks for allocation. It's possible that for certain RG sizes, x bitmap blocks is too few, but x+1 is too many. Confusing, huh? Version-Release number of selected component (if applicable): This affects all versions of gfs_mkfs. How reproducible: I believe this happens on about one out of about every 16000 randomly sized devices. Steps to Reproduce: 1. Run gfs_mkfs on randomly sized devices until it happens. 2. Or: Rip out the function with the problem and simulate running gfs_mkfs on randomly sized devices. 3. Or: Take my word for it. Actual results: gfs_mkfs dies with an error. Expected results: gfs_mkfs should run to completion. Additional info: As a temporary workaround, you should be able to avoid the problem by monkeying with the -r option to gfs_mkfs.
Fixed for 5.2.1, 6.0, and CVS head. Fixable for 6.1 upon request. Just throw away the one extra block that caused x bitmap blocks to be too few.
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 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-2005-466.html