Re: Serious ext3 file corruption in Taroon?

[Chris Adams <cmadams hiwaay net>]

Once upon a time, Brett Morrow <brett morrow noaa gov> said:
> I would like to also find out if other are seeing this also.  We run
> many raid systems here (Raid 0 and Raid 5 depending on servers usage)
> and have seen the data corruption with both adaptec and LSI
> controllers.  I get the feeling that ext3 is not really ready for very
> high traffic loads yet.

What do you mean by "high traffic loads"?  I run ext3 on a lot of
servers, some that are quite busy (for example, one fs is on a MegaRAID
with 6 10K U160 SCSI drives striped and mirrored and it runs 200-400 I/O
transactions per second continuously 24x7) without any trouble.

I also have a few Sun Cobalt RaQ550 servers that use XFS, and I've seen
a couple of them eat a filesystem worse than I've seen ext2/ext3 screw
up in 11+ years of using Linux (okay, only 8.5 years of using Linux on
servers); the kernel hung when trying to mount the FS, when I mounted it
without playing the journal I could get to some stuff, but when I ran
xfs_repair, it completely blew away the filesystem (good thing it
happened about an hour after nightly tape backups finished).  I won't
touch XFS for anything I build for a long time.

-- 
Chris Adams <cmadams hiwaay net>
Systems and Network Administrator - HiWAAY Internet Services
I don't speak for anybody but myself - that's enough trouble.