4

I am in the process of replacing some failing SATA disks with new SAS disks. The raid array that one of the failing disks is part of is in the process of resilvering. As that progresses, I can see the write error correction rate on the new disk slowly but steadily increase. This is my first exposure to SAS hardware, and I am unfamiliar with the smartctl output, since I'm used to the big table that SATA disks produce.

The disk is an HGST Ultrastar 2TB, attached through an IBM M1015 HBA, in passthrough mode.

How exactly should I interpret the SMART data below? Should the rising corrected-error rate worry me? What about the Correction Algorithm Invocations field?

  Device: HITACHI  HUS723020ALS640  Version: A222
Serial number:         xxxxxxxx
Device type: disk
Transport protocol: SAS
Local Time is: Thu Jul 11 15:42:30 2013 MDT
Device supports SMART and is Enabled
Temperature Warning Enabled
SMART Health Status: OK

Current Drive Temperature:     44 C
Drive Trip Temperature:        85 C
Manufactured in week 29 of year 2012
Recommended maximum start stop count:  50000 times
Current start stop count:      2 times
Elements in grown defect list: 0
Vendor (Seagate) cache information
  Blocks sent to initiator = 149833037905920

Error counter log:
           Errors Corrected by           Total   Correction     Gigabytes    Total
               ECC          rereads/    errors   algorithm      processed    uncorrected
           fast | delayed   rewrites  corrected  invocations   [10^9 bytes]  errors
read:          0        0         0         0          6          0.565           0
write:         0      400         0       400         51       1034.246           0
verify:        0        0         0         0         37          0.000           0

Non-medium error count:        0

1 Answers1

4

Delayed errors are errors that slow down other requests. ECC corrected errors aren't much of a worry on SCSI/SAS drives, we have drives deployed with hundreds of millions of them and they still run fine. Correction algorithm invocations is a bit more serious, those may require rereading/rewriting the disk, and retrying the ECC calculation.

It's hard to say whether the number of errors you have is worrisome or not. There is no one standard, each manufacturer has their own and they keep it proprietary. You can usually download a utility from the manufacturers website to get extra information.

For Hitachi the link is: http://www.hgst.com/support/downloads#WINDFT

You can read up more on the different errors here: http://smartmontools.sourceforge.net/smartmontools_scsi.html

Gregg A.
  • 41