Hi,
I have a problem with a AdvFS-Domain on a RAID5-Set at a HSZ22 (D11V).
After a disk failure (and replacement/rebuild without any problems) in a RAID5-Set at our HSZ22 the kernel (DS20E, KZPBA-CB, Tru64
4.0F) reports some AdvFS I/O read errors of a AdvFS domain at this RAID-Set (messages, kern.log, uerf). The blocks where the error
occurs are only few blocks in the meta data of the domain (the error occurs while defragmenting the domain and all fsets can be
vdumped without errors). All Failure-LEDs of all disks at the controller are off and all normal operations of this domain are ok
with normal speed. But since the disk failure the eventlog from the HSZ22 reports "FORCED ERROR WRITTEN" of all disks at the channel
where replaced disk was failing.
One speculation is that a second disk fails before the first failed disk is rebuilded. Another speculation is a hardware defect in
the RA3000. But nothing of them seems to be happen.
Any further hints? Please!
What means a HSZ22 with "FORCED ERRORS"?
Joerg
Received on Mon Jun 11 2001 - 10:59:08 NZST