On Mon, Dec 16, 2013 at 06:52:23AM -0200, Rafael Weingartner wrote: > What's a "RAID NAS controller"? Details, please, or we can't help > > you. > > > Maybe I am not expressing my self clearly. That is what I meant: > http://www.starline.de/produkte/raid-systeme/infortrend-raid-systeme/eonstor/es-a08u-g2421/ > > It is a piece of hardware that we use to apply RAIDx (normally 1 or 5) over > physical disks instead of plugging them on the storage server and applying > RAID via software or something else. It exports the volumes using an SCSI > channel. The devices are seen on the server as normal sd*, as if they were > normal physical devices. > > So, hardware RAID5, lost a drive, rebuild on replace, filesystem in > > a bad way after rebuild? > > > That is exactly what happened, the RAID5 array lost a drive, and after we > replaced and rebuild it, the filesystem was not mounting anymore. > Teorically, this should not affect the filesystem since the RAID5 would > have recovered any lost information. In theory, yes. But we hear about "successful" hardware raid rebuilds ifrom situations like yours that result in corrupt filesystems. i.e. the hardware RAID rebuild didn't actually recover properly... So, what's the contents of the primary XFS superblock on that device? What does this give: # dd if=<dev> bs=512 count=1 | hexdump -C Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
