Crash while resyncing /dev/md0
 

Hi folks.

I am running a heartbeat vluster on two rhel3 mashines. It ist an FTP-Server, which uses a san-device an data-partition. This san-Device is on two EMC mashines, so i mirror the one partition (from one EMC) onto the other using software raid level 1.

Now, one cluster node crashed and the other made a takeover, as it should be. the mirror was out of sync, so it startet a resync on /dev/md0. This is very fine.

Now, both nodes "like" to crash while resyncing /dev/md0. I am trying it for 6 hours now, but on every try the one performing the resync crashes someweher in this action. Heartbeat notices the crash and the other node makes a takeover, initiating another resync, because it was not finished. So, this node crashes again while resyncing, waht makes the first node, i rebooted it in the meantime, to make another takeover. I notices too thet the mirror ist out of sync and everything begins again from the start.

Has anybody similar problems using rhel3 with 2.4.21-32.ELsmp, heartbeat 1.2.3?

Thanks a lot for help.

Hela Wann

Anytime I had one mirror keep crashing during a resync, it usually indicated bad blocks, bad or failing hard disk.