scsi error

[Carlos_172]

Hello.
We have a DS2100 attached to L1000 through FAST WIDE card. First of all, it had 2 U160 SCSI (SEAGATE ST3146707LC) and 2 U320SCSI disk (SEAGATE ST3146807LC) 2+2 mirrored. We detected scsi errors in dmesg in two disks U320. We change one of this by a U160 disk. Seems to be success, but now the u320disk and the new U160 are in dmesg again. Is a good idea change last u320 disk?
U160 -> c0t8d0
U320 -> c0t10d0

thanks a lot.

SCSI: Unexpected Disconnect -- lbolt: 69182643, dev: 1f00a000, io_id: c62dab
LVM: vg[1]: pvnum=1 (dev_t=0x1f00a000) is POWERFAILED

SCSI: Write error -- dev: b 31 0x00a000, errno: 126, resid: 8192,
blkno: 140081064, sectno: 280162128, offset: 1709088768, bcount: 8192.

SCSI: Write error -- dev: b 31 0x00a000, errno: 126, resid: 49152,
blkno: 140081088, sectno: 280162176, offset: 1709113344, bcount: 49152.

SCSI: Write error -- dev: b 31 0x00a000, errno: 126, resid: 16384,
blkno: 140081072, sectno: 280162144, offset: 1709096960, bcount: 16384.

SCSI: Write error -- dev: b 31 0x00a000, errno: 126, resid: 262144,
blkno: 1102464, sectno: 2204928, offset: 1128923136, bcount: 262144.
LVM: Failed to automatically resync PV 1f00a000 error: 5


LVM: vg[1]: pvnum=0 (dev_t=0x1f008000) is POWERFAILED
vmunix: SCSI: Async write error -- dev: b 31 0x008000, errno: 126, resid: 4096,
vmunix: blkno: 113914056, sectno: 227828112, offset: 683876352, bcount: 4096.
vmunix: blkno: 110308680, sectno: 220617360, offset: 1286938624, bcount: 4096.
vmunix: blkno: 33087816, sectno: 66175632, offset: 3817152512, bcount: 4096.
vmunix: SCSI: Write error -- dev: b 31 0x008000, errno: 126, resid: 75776,
vmunix: SCSI: Async write error -- dev: b 31 0x008000, errno: 126, resid: 4096,
above message repeats 2 times
vmunix: blkno: 2326, sectno: 4652, offset: 2381824, bcount: 75776.
vmunix: SCSI: Write error -- dev: b 31 0x008000, errno: 126, resid: 2048,
vmunix: blkno: 6133, sectno: 12266, offset: 6280192, bcount: 2048.
vmunix: LVM: Recovered Path (device 0x1f008000) to PV 0 in VG 1.
vmunix:
above message repeats 4 times
vmunix: LVM: Restored PV 0 to VG 1.

[Carlos_172]

In ds2100/2110 user's guide I have found the following:

"Both (DS2100/DS2110) units support Ultra3 (U160) drives and the new U320 drives. However, the U320
drives are supported at a maximum of 160 MB/s speeds. The Disk System 2100 does
not support the U320 73-GB drive (part number A7286A) or the U320 146-GB drive
(part number A7287A)".

Does it means DS2100 not support U320 146-GB and this can rely on scsi errors we have?

Any help will be appreciated. Thanks.

[RAC_1]

As you check in the document, when document says those are not supported means HP wont't service a call if you were to log one. But that does not mean that they would not work.

SCSI errors could be on account on many things like-loose cable/connectors, scsi patches not up to date, terminators etc. Your disks seem to be failing. You should be replacing current ones with the one that are supported.

[Carlos_172]

We already replace one of this and sometimes appear in dmesg as scsi write error

[Andrew Rutter · Accepted solution]

hi carlos,

It is a good idea to have the disks matching, especially if there mirrored. As you know you wont get any benefit from having u320 disks in there.

Also the reason that those disks are not supported is they are 15k rpm disks and the max supported is 10k rpm disks. These will get overheated as there isnt enough cooling in the box for them. This could be a reason why they are powerfailing.

Andy

[Carlos_172]

Disk u320 that contains ds2100 is 10 rpm. It is SEAGATE ST3146807LC ...