3. if a disk dies suddenly, the system is gonna crash regardless of raid
because the kernel can no longer communicate with /dev/sdx, it just
disappears.  go ahead, set up software raid with hot swap disks then yank
one out while the system is running, see what happens.  the data itself is
probably ok (you'll have a degraded array upon reboot) but availability is
shot.  plus your fstab may no longer be accurate once a disk is removed.


Not sure this is true in all circumstances, it may well be true for
non-hotswap capable hardware, but I had a drive die the other day in
an array where the only results were:

1. A lot of messages in log about SCSI timeouts.
2. A polite email from mdadm telling me that a drive had died.
3. Slightly lower performance as the hotspare was rebuilt.
4. Visit to colo to replace the bad drive and re-add it the the array
as the new hotspare.
--
TriLUG mailing list        : http://www.trilug.org/mailman/listinfo/trilug
TriLUG Organizational FAQ  : http://trilug.org/faq/
TriLUG Member Services FAQ : http://members.trilug.org/services_faq/

Reply via email to