ff> I have check the drives with smartctl:
ff> ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE
UPDATED WHEN_FAILED RAW_VALUE
ff> 1 Raw_Read_Error_Rate 0x000f 115 075 006 Pre-fail
Always - 94384069
ff> 5 Reallocated_Sector_Ct 0x0033 100 100 036 Pre-fail
Always - 0
ff> 195 Hardware_ECC_Recovered 0x001a 065 056 000 Old_age
Always - 173161329
ff> 199 UDMA_CRC_Error_Count 0x003e 200 200 000 Old_age
Always - 0
ff> But with no UDMA_CRC_Errors I believe the disks are fine.
no, UDMA_CRC_Errors counts checksum errors on PATA cables. I cannot
confirm/deny if it counts CRC errors on SATA cables (and even if it
did this is complicated because there are weird scsi-emulation
proprietary drivers, port multipliers, u.s.w.) so, if you are having
problems, and that parameter is increasing, then it's probably cabling
problems not drive problems.
The other three values I quoted are the ones that matter. The VALUE
is scaled by constants defined by the manufacturer and used for the
``overall health assessment'', but the constants they use are always
way too forgiving, so it's worthless. The RAW_VALUE looks bigger than
I'm used to, but this may also be meaningless. The only way I know to
get information out of the report is: How do the RAW_VALUE's of the
three parameters I quoted compare with other drives of the same model,
or to this drive before it started failing?
There is another section of the smartctl -a report that logs the last
5 or so errors the drive has reported to the host. IIRC you will see
errors called 'ICRC' or 'UNC' on failing drives.
this experience is all PATA/SATA-specific.
pgpUhcxtASNbS.pgp
Description: PGP signature
_______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
