Hi,
I want to create a software RAID5 array out of a few old drives.
I need RAID5 in case a drive fails and I don't want to invest in a
hardware raid controller.
I am planning on using 4 disks in the array and an additional one to
hold the log file.
The single point of failure wil be the disk that holds the log.
If the log disk fails, will I be able to create a new log on another
disk and continue to use the RAID array ?
I asume that the data will be lost that has not being commited at the
time of the failure.
One way to do it (I think) would be to unmount the array and do a "dd
if=/dev/logdisk of=backup_log"
and store backup_log some where.
I can then "dd if=backup_log of=/dev/new_logdisk" and run a xfs_repair
for good measure, when disaster strikes.
Is the procedure that I described here a valid way of doing it ?
Is there a more elegant way ?
Paul
|