Hardware

Hi ,

I see an occessional (once in a month) error sequence in the messages like following on a solaris 9 SPARC host

mpt5: fault detected in device; service still available
mpt5: Connected command timeout for Target 0
WARNING: /pci@1e,600000/scsi@3,1 (mpt5): Target 0 disabled wide SCSI mode
mpt: [ID 554818 kern.warning] WARNING: ID[SUNWpd.mpt.sync_wide_backoff.6012]
scsi: [ID 107833 kern.warning] WARNING: /pci@1e,600000/scsi@3,1 (mpt5): Target 0 reverting to async. mode
mpt: [ID 675377 kern.warning] WARNING: ID[SUNWpd.mpt.sync_wide_backoff.6013]

scsi: [ID 107833 kern.warning] WARNING: /pci@1e,600000/scsi@3,1/sd@0,2 (sd218): Error for Command: write Error Level: Retryable

The MPT fault but doesn't say who cause it and what fault it is. Does anyone know how can I troubleshoot this problem?

I also see following messages appearing in the messages often

scsi: [ID 365881 kern.info] /pci@1e,600000/scsi@3,1 (mpt5):
target0-scsi-options = 0x41ff8

I know it shows the conf file setting but why it pops up often?

SE3320 with RAID module, single bus configuration and Dual port PCI-X HBA.
RAID 1
sccli - 2.3.0
SE3320 f/w - 415H
14x 72GB drives ( 13x 72GB MAW3073NCSUN72G Disk firmware 1703
1x 72GB DK32EJ72NSUN72G with f/w - PQ0B ( global standby) )
MPT driver - v1.23
PCI-X HBA Fcode - 1.3.27.0
SAF-TE- 1180

I am suspecting some unknown mpt fault most likely a h/w problem is occuring while high IO occurs on SE3320 and maybe a hard drive or something is not able to handle it. There is nothing in the sccli event log.

Thanks in advance,