Disk at hardware path 0/1/1/0.0.0.0.0 : Media failure EMS

[Syedfurquan]

Hello,

I have a RX2660 server with hp-ux 11.23. From past more than 2 weeks I am getting everyday twice EMS in my /var/adm/syslog/syslog.log stating

Disk at hardware path 0/1/1/0.0.0.0.0 : Media failure

But When I am doing diskinfo it looks fine, in ioscan o/p all disks are claimed and I tried reading from the disk also and it shows succesful read.

Now what could be the reason for this warning and how to fix this as it keep coming everyday.

Thanks for any help in advance.

Attaching logs from the server for reference.

[Bill Hassell]

The media error is very real and the disk needs to be replaced. Your dd command is looking at just a tiny part of the disk. You need to specify a count large enough to cover the entire disk surface. The block size (1024k) is reasonable but the count must be large enough to cover the disk:

dd if=/dev/rdsk/c1t0d0 of=/dev/null block=1024k count=143374744

You'll probably see an I/O error before the dd finishes.

The good news is that the disk appears to be mirrored. Check that all the lvols are synced:

vgdisplay -v vg00 | grep -e LV.Name -e LV.Status

 Then you can replace the bad disk and go through all the steps needed to remirror a boot disk on Itanium hardware.

[Syedfurquan]

Hello Bil,

Thaks for the reply. As per your suggestion I tried reading from the full disk and it also completed successfully.

I checked the LV status and all LV shows in synced state.

Attaching the op for your reference.

[Bill Hassell · Accepted solution]

From the EMS message:

Command Data Block Fields (10-byte fmt):
          Command Operation Code...(0x28)..: READ
          Logical Unit Number..............: 0
          DPO Bit..........................: 0
          FUA Bit..........................: 0
          Relative Address Bit.............: 0
          Logical Block Address............: 152871202 (0x091CA122)
          Transfer Length..................: 64 (0x0040)

The error appears to be somewhere in the middle of the disk space (152,871,202 blocks), assuming that a block is 512 bytes. However, since the full dd read completed without error, this appears to be transient. You might use cron to repeat the dd read, possibly using an offset to start towards the beginning of the disk area and reading about 20 GB, something like this:

dd if=/dev/rdsk/c1t0d0 of=/dev/null bs=1024k skip=60000 count=20000

 You could schedule this to run during off hours so it does not affect production work. Be sure to log the results of the dd. Otherwise the messages will go to root's email.

[Syedfurquan]

Hi Bill,

Once again thanks for your response.

Since my full disk dd command also completed successfully & now I will schedule the dd command in crontab also.

So now in this scenario what you recommend to do to stop this EMS messages coming as we are using HP Open view and mails keep coming to higher management also with so criticality. Any options to fix this issue.

Regards,

Syed

[Bill Hassell]

The disk is going bad. The best thing to do is to replace it with new one.

[Syedfurquan]

Hi Bill,

Thank you so much for all your support on this post.

I have already an open case with HP but as per ioscan o/p, diskinfo and dd command successfull HP L1 support engineer asked me to ignore this message. Now I will try to reopen this case and ask them to provide the disk so that I can replace it.

Regards,

Syed

[Syedfurquan]

Hello Experts,

can anyonep please provide me the steps to replace the faulty disk for vg00 which is LVM mirrored.

[Torsten.]

Get the disk position:

# sasmgr get_info -D /dev/sasd0 -q raid

then follow this guide

When_Good_Disks_Go_Bad_WP
http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c01911837/c01911837.pdf