Hello,
Some of my disks get inaccessible after a while of operation, with this
kernel.log:
------------
Jan 4 16:46:33 csone vmunix: AdvFS I/O error:
Jan 4 16:46:33 csone vmunix: Volume: /dev/rzg48c
Jan 4 16:46:34 csone vmunix: Tag: 0xfffffffa.0000
Jan 4 16:46:34 csone vmunix: Page: 1
Jan 4 16:46:34 csone vmunix: Block: 48
Jan 4 16:46:34 csone vmunix: Block count: 16
Jan 4 16:46:34 csone vmunix: Type of operation: Write
Jan 4 16:46:34 csone vmunix: Error: 5
Jan 4 16:46:34 csone vmunix:
Jan 4 16:46:34 csone vmunix: bs_osf_complete: metadata write failed
Jan 4 16:46:34 csone vmunix: AdvFS Domain Panic; Domain cs_central Id
0x3a0bdcab.000d7ddd
Jan 4 16:46:34 csone vmunix: An AdvFS domain panic has occurred due to
either a metadata write error or an internal inconsistency.
This domain is being rendered inaccessible.
Jan 4 16:46:34 csone vmunix: Please refer to guidelines in AdvFS Guide
to File System Administration regarding what steps to take to recover
this domain.
------------
When I restart the system, with a power off, the disks become accessible
again and verify shows no
errors at all. I have tried to restart the system without a power off,
which resulted in still having problems.
Other disks on the same HSG60 are still accessible.
Anybody having an idea? It looks like a faulty FC controller or driver
to me.
My cluster configuration is like this:
Hardware
- 2x DS20e with FibreChannel controller
KGPSA-CA : Driver Rev 1.22 : F/W Rev
_3.02A1(1.11) : wwn
_1000-0000-c922-58d0
emx0: Using console topology setting
_of : Fabric
- 1x SAN Switch
- 1x HSG60 controller
Software
- UNIX 4.0G with t64v40gas0001-20000802
- TruCluster 1.6
Greetings,
Maurice Akkermans
Tellabs Denmark A/S
Management Systems
Lautrupbjerg 7-11
DK-2750 Ballerup
Denmark
tel.: +45 44732244
mail: maurice.akkermans_at_tellabs.com
Received on Fri Jan 05 2001 - 08:38:13 NZDT