System Administration
cancel
Showing results for 
Search instead for 
Did you mean: 

Determine bad disk from syslog error message.

SOLVED
Go to solution
Steven E. Protter
Exalted Contributor

Determine bad disk from syslog error message.

Believe it or not I've never had to do this before.

I/O error: dev 08:10, sector 0

various sectors.

I need to determine exactly which disk is having problems.

/dev/cciss/c0d0p3 1008M 139M 818M 15% /
/dev/cciss/c0d0p1 463M 20M 420M 5% /boot
/dev/cciss/c0d0p5 5.9G 2.0G 3.6G 35% /home
/dev/cciss/c0d0p7 3.9G 689M 3.0G 18% /opt
none 1005M 0 1005M 0% /dev/shm
/dev/cciss/c0d0p10 2.0G 33M 1.8G 2% /tmp
/dev/cciss/c0d0p8 3.9G 1.5G 2.3G 39% /usr
/dev/cciss/c0d0p6 3.9G 87M 3.6G 3% /var
/dev/cciss/c0d0p9 2.0G 34M 1.8G 2% /var/dtv
/dev/sda5 9.8G 160M 9.1G 2% /oraback/DB2
/dev/sda3 20G 12G 7.4G 61% /oradata/DB1

HP/Compaq Proliant servers.

lspci and dmesg have not been very helpful.

fdisk -l

Disk /dev/sda: 255 heads, 63 sectors, 8855 cylinders
Units = cylinders of 16065 * 512 bytes

Device Boot Start End Blocks Id System
/dev/sda1 1 3 24066 83 Linux
/dev/sda2 4 6 24097+ 83 Linux
/dev/sda3 7 2617 20972857+ 83 Linux
/dev/sda4 2618 8855 50106735 5 Extended
/dev/sda5 2618 3923 10490413+ 83 Linux
/dev/sda6 3924 4054 1052226 83 Linux

Unable to read /dev/sdb

Our standard setup involves doing hardware raid with a smart card. Seems we have a dead disk. I need to know which filesystem to run fsck on.

TIA, as always generous points for good answers.

SEP
Steven E Protter
Owner of ISN Corporation
http://isnamerica.com
http://hpuxconsulting.com
Sponsor: http://hpux.ws
Twitter: http://twitter.com/hpuxlinux
Founder http://newdatacloud.com
12 REPLIES
David Claypool
Honored Contributor

Re: Determine bad disk from syslog error message.

Since the drive is protected by RAID, most likely your filesystem isn't affected and fsck won't show you anything. You should use the Array Diagnostic Utility.
Stuart Browne
Honored Contributor

Re: Determine bad disk from syslog error message.

The def 08:10 is your clue. Look for major-8-minor-10 in your /dev (usually /dev/sda10).
One long-haired git at your service...
Steven E. Protter
Exalted Contributor

Re: Determine bad disk from syslog error message.

Well Stuart, the system does not seem to have a /dev/sda10.

Seems the system is overwhelmed with I/O at the moment can't figure out why it can't read /dev/sdb when its not configured.

Other ideas welcome.

SEP
Steven E Protter
Owner of ISN Corporation
http://isnamerica.com
http://hpuxconsulting.com
Sponsor: http://hpux.ws
Twitter: http://twitter.com/hpuxlinux
Founder http://newdatacloud.com
skt_skt
Honored Contributor
Solution

Re: Determine bad disk from syslog error message.

brw-rw---- 1 root disk 8, 10 Jun 24 2004 sda10

Stuart Browne
Honored Contributor

Re: Determine bad disk from syslog error message.

Huh.. Weird..

Well, try using fdisk again, but changing the 'units' to sectors instead:

fdisk -lu /dev/sda

There is an 'sdb' in the system? Is it showing up in '/proc/scsi/scsi'?
One long-haired git at your service...
Court Campbell
Honored Contributor

Re: Determine bad disk from syslog error message.

/dev/sdb maybe your cdrom.
"The difference between me and you? I will read the man page." and "Respect the hat." and "You could just do a search on ITRC, you don't need to start a thread on a topic that's been answered 100 times already." Oh, and "What. no points???"
Court Campbell
Honored Contributor

Re: Determine bad disk from syslog error message.

what is the output from

# cat /proc/partitions

"The difference between me and you? I will read the man page." and "Respect the hat." and "You could just do a search on ITRC, you don't need to start a thread on a topic that's been answered 100 times already." Oh, and "What. no points???"
Steven E. Protter
Exalted Contributor

Re: Determine bad disk from syslog error message.

Shalom



[root@golan1 ~]# ll -u /dev/sdb*
0 brw-rw---- 1 root disk 8, 16 Jul 19 10:15 /dev/sdb
0 brw-rw---- 1 root disk 8, 17 Jul 18 16:32 /dev/sdb1
0 brw-rw---- 1 root disk 8, 21 Jul 18 16:32 /dev/sdb5
0 brw-rw---- 1 root disk 8, 22 Jul 18 16:32 /dev/sdb6


Clearly shows the major and minor number of the problem.

Shared storage is having read errors. If its bad disk on storage, we face downtime. If its just the fiber card or other local system issue, we'll fail over to the working node and correct the hardware problem.

Thannks.

I'll probably close this and hand out points very soon.

SEP
Steven E Protter
Owner of ISN Corporation
http://isnamerica.com
http://hpuxconsulting.com
Sponsor: http://hpux.ws
Twitter: http://twitter.com/hpuxlinux
Founder http://newdatacloud.com
Steven E. Protter
Exalted Contributor

Re: Determine bad disk from syslog error message.

Nice one Court.

I really like /proc/partitions

My output was not from the effected system, it was merely an example to document the case.

Thanks. I should have known that.

SEP
Steven E Protter
Owner of ISN Corporation
http://isnamerica.com
http://hpuxconsulting.com
Sponsor: http://hpux.ws
Twitter: http://twitter.com/hpuxlinux
Founder http://newdatacloud.com
Stuart Browne
Honored Contributor

Re: Determine bad disk from syslog error message.

Hrm.. You should have covered '/proc/partitions' and 'partprobe' in your RHCE.. Ah well ;)
One long-haired git at your service...
Steven E. Protter
Exalted Contributor

Re: Determine bad disk from syslog error message.

Partprobe yes use that all the time.

I must have forgotten /proc/partitions

Hopefully not again.

SEP
Steven E Protter
Owner of ISN Corporation
http://isnamerica.com
http://hpuxconsulting.com
Sponsor: http://hpux.ws
Twitter: http://twitter.com/hpuxlinux
Founder http://newdatacloud.com
Stuart Browne
Honored Contributor

Re: Determine bad disk from syslog error message.

:)

(I just re-did it last week, so it's still fresh)
One long-haired git at your service...