SCSI read error

[Radhakrishnan Venkatara]

hi ,

I was getting the following error in the server.

SCSI: Read error -- dev: b 31 0x070000, errno: 126, resid: 2048,
blkno: 8, sectno: 16, offset: 8192, bcount: 2048.

SCSI: Read error -- dev: b 31 0x070000, errno: 126, resid: 2048,
blkno: 8, sectno: 16, offset: 8192, bcount: 2048.

SCSI: Read error -- dev: b 31 0x070000, errno: 126, resid: 2048,
blkno: 8, sectno: 16, offset: 8192, bcount: 2048.

SCSI: Read error -- dev: b 31 0x070000, errno: 126, resid: 2048,
blkno: 8, sectno: 16, offset: 8192, bcount: 2048.

SCSI: Read error -- dev: b 31 0x070000, errno: 126, resid: 2048,
blkno: 8, sectno: 16, offset: 8192, bcount: 2048.

The harddisk /dev/dsk/c7t0d0 was not sensing in ioscan.

Later today the output of dmesg is
LVM: Recovered Path (device 0x1f070000) to PV 0 in VG 2.
LVM: Restored PV 0 to VG 2.

and the harddisk is getting sensed in the ioscan and filesystem is up.

what could be possible reason that have happened to make harddisk not available.
Details :-
hard disk was in fc10 disk array N class server

thanks
radhakrishnan

[Balaji N]

hi
make sure you have you all the backups ready. time to repalce the hdd i guess.
-balaji

[T G Manikandan]

Looks like the disk has some defective blocks.

The option you can try is to re-create the file system on that disk using newfs and check it.
It still it gives errors then you should go in replacing the disk.

//Make sure that you backup//

[Eugeny Brychkov]

Agree you have to have a good recent backup. But before disk replacement you can try reseating disk within FC10 slot (offline) and checking disk surface with 'dd if=/dev/rdsk/c7t0d0 of=/dev/null bs=4096k'.
If there're another disks installed within the same FC10 and you get this error only for this one (considering that you use other FC10 disks on this host) then replace disk
Eugeny

[KCS_1]

hello~

I have seem this message before.

Above posted message are good stuff for solving your problem.

In addition,

Check your h/w(disk : /dev/dsk/c0t8d0) :

# ioscan -funCdisk

look at all disk which are returned correct status or not

# diskinfo -v /dev/dsk/c0t8d0

Returned correct informations from the disk or not

# dd if=/dev/dsk/c0t8d0 of=/dev/null bs=512k

testing of read/write

are they say normal status??
if, all of things are good,
in my opinion, that's ok!

but, one of things returned not ok, swap in/out and then gathering the disk information,again such as above,again.

you have a problem,may it's bad disk!! i think
Call your CE of HP.

finally, look at under thread about wwwName Fc disk replace procedure:

http://www1.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&admit=-682735245+1047980445589+28353475&docId=200000062908600



have a good day!