hi... i wanted to confirm my reading of the raid1 code during a resync. from my reading, all writes go to all non-faulty disks even if the write is beyond the current sync point and the disks are being resync'd... is that correct? and writes are delayed when they overlap the current sync point... and issue after the sync passes by. i had a (2.4 kernel) box with a 3-way raid1 crash recently and on reboot it started resyncing from the 0th disk in the array... then 12h later that 0th disk failed before the resync finished (system was way too busy to handle the resync). at first i was worried i would have to go back to backup, but when i thought it through and inspected the code i convinced myself i probably had an almost perfect copy of the array on either disk 1 or disk 2... i.e. at crash disk 0,1,2 were at most a few thousand sectors out of sync... the usual crash situation. but during resync, if my reading of the code is right, all writes went to all disks, even those beyond the sync point... so after the disk 0 failure disks 1 and 2 had possibly the original few thousand sectors out of sync plus the sectors that were being sync'd at disk 0 failure time. anyhow even if my reading is wrong i got the system back on its feet without resorting to backups... and the only corrupted files discovered so far were in-flight at the time of the crash and/or disk 0 failure, so that's expected. thanks -dean - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
