| |||||||||||||
Ext3 and RAID: silent data killers?Ext3 and RAID: silent data killers?Posted Sep 1, 2009 1:35 UTC (Tue) by flewellyn (subscriber, #5047)Parent article: Ext3 and RAID: silent data killers?
I can't believe a kernel developer thought that a degraded RAID5 having data corruption problems if it suffered power loss (or another fault) while still degraded, or rebuilding, is at all noteworthy.
It's right in the definition of RAID5: you cannot expect to maintain data integrity after a double-fault. The filesystem used on top of it is irrelevant.
(Log in to post comments)
Ext3 and RAID: silent data killers? Posted Sep 1, 2009 3:22 UTC (Tue) by zlynx (subscriber, #2285) [Link]
I think it's noteworthy.
Most people, myself included, expect a degraded RAID array to fail **only if another drive fails**. I do NOT expect losing 256KB of data because a single 4KB write failed.
And in fact, the array didn't lose it. It just can't tell which 4KB went bad, which it could if MD did good write-intent logging.
Ext3 and RAID: silent data killers? Posted Sep 1, 2009 7:51 UTC (Tue) by job (guest, #670) [Link]
It's the fact that it's silent that is the problem. Writes fail under certain circumstances, and that's acceptable, but when a failed write perhaps affects other data silently and now your system reports healthy but you don't know if it really is... That's just .. ouch.
|
|||||||||||||