Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
/ 2006-03-13 16:25:22 +0100 \ Alessandro Ferrari: > DRBD 0.7.17 > > > Hi, > > in a previous version of DRBD there is a fsckcmd setting in drbd.conf. I can't find this command into 0.7.17. > So, how does the slave-server check filesystem ext3 error when a master-server crash? basically it does not. did you ever fsck a 3TB ext2 (or ext3, or anything else, for that matter)? you don't want that to happen on failover time. that is what journalled file systems are for. the heartbeat Filesystem resource agent script does /sbin/fsck -t $FSTYPE -a $DEVICE for any filesystem not in "ext3 reiserfs xfs jfs vfat fat nfs". for those in the list, which are known to be journalled or uncheckable, fsck is skipped. in case fsck is executed (e.g. for ext2), and the exit code is >= 4, the Filesystem resource agent script bails out. btw, * missing in that list: at least smbfs... * if /sbin/fsck does not find the proper fsck.$FSTYPE executable, it seems to return success (exit code 0) which may not be expected :) anyways, this boils down to: use one of the journalled file systems. if they do no longer mount, no automatic fsck will help you. for anything larger than say 200GB, I tend to recommend xfs. that is, meanwhile; I used to recommend against it, when it still caused lots of trouble on linux... that seems to have been sorted out, though. so you probably want to use a recent kernel with that. and make sure that kernel stack size is configured large enough! cheers, -- : Lars Ellenberg Tel +43-1-8178292-0 : : LINBIT Information Technologies GmbH Fax +43-1-8178292-82 : : Schoenbrunner Str. 244, A-1120 Vienna/Europe http://www.linbit.com : __ please use the "List-Reply" function of your email client.