I have a pair of Western Digital RE4-GP (WD2002FYPS) drives in a software RAID1 configuration using Linux 2.6.30.3 on a LSISAS1068E controller. Within hours one of the drives was kicked out of the array with:
[ 4907.485324] end_request: I/O error, dev sdb, sector 3907028974
[ 4907.485543] md: super_written gets error=-5, uptodate=0
[ 4907.485546] raid1: Disk failure on sdb2, disabling device.
[ 4907.485547] raid1: Operation continuing on 1 devices.
[ 4907.499157] RAID1 conf printout:
[ 4907.499159] --- wd:1 rd:2
[ 4907.499162] disk 0, wo:0, o:1, dev:sda2
[ 4907.499164] disk 1, wo:1, o:0, dev:sdb2
[ 4907.503037] RAID1 conf printout:
[ 4907.503039] --- wd:1 rd:2
[ 4907.503041] disk 0, wo:0, o:1, dev:sda2
[ 6705.292961] sd 4:0:1:0: [sdb] Sense Key : Recovered Error
[current] [descriptor]
[ 6705.292967] Descriptor sense data with sense descriptors (in
hex):
[ 6705.292970] 72 01 00 1d 00 00 00 0e 09 0c 00 00 00 00
00 00
[ 6705.292978] 00 4f 00 c2 00 50
[ 6705.292983] sd 4:0:1:0: [sdb] Add. Sense: ATA pass through
information available
[ 6705.359497] sd 4:0:1:0: [sdb] Sense Key : Recovered Error
[current] [descriptor]
...
Subsequently, I disabled NCQ with:
$ echo 1 > /sys/block/sda/device/queue_depth
$ echo 1 > /sys/block/sdb/device/queue_depth
which rendered the system stable. Is there a better way of
implementing this work-around than a rcS.d or rc2.d script?
What is the right place to report this problem with NCQ?
[email protected] generated no response.
/Allan
--
Allan Wind
Life Integrity, LLC
<http://lifeintegrity.com>
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
