Hot Swap Disk on D Class

[Jim VanMeter]

Our HP-UX system administrator has moved on to greener pastures and I'm trying to handle this problem. We currently have a hot swap disk on a HP D350 which is giving us I/O errors. Our hardware maintenance folks have recommended that we replace it. To the point, I have identified which disk device is bad, that is at least the hardware path according to Sam. What I don't know is how do I translate that to the physical hot swap? We have 5 hot swaps in that box and I don't know the physical location of the bad one. Also, is there anything I need to be aware of before pulling the disk. such as unmounting, etc. I have moved applications to another server and have backed up the disk. Thank you very much for any insight.

[John Palmer · Accepted solution]

Jim,

By default on the D class, the 5 drives in the front of the server have (from the top) SCSI addresses 5,11,10,9,8.

Provided the drive is still usable, you can confirm this by issuing the command:-

dd if=/dev/rdsk/c0t?d0 of=/dev/null bs=64k

and checking which drive light is on.

Before pulling the disk you need to:-

Remove it from LVM using vgreduce. You may have to unmount any filesystems and lvremove any logical volumes on the disk first. You can check what ios on the disk with the pvdisplay command - 'pvdisplay -v /dev/dsk/c0t?d0'

In my experience, you should also ensure that there is no activity on the SCSI bus (close or quiesce your applications) before pulling the disk and replacing it because both cause a bus reset.

Hope this helps,
John

[Jim VanMeter]

John,

I'm now a little confused. If you take a look at the attached word document, I have captured some Sam screens. You state that from top to bottom the SCSI addr's are 5,11,10,9,8. It has been determined from sysout logs that it is rdsk5 that is bad, and from Sam, that rdsk5 uses SCSI 11. According to your scenario that would make it the second one from the top. But, in our box we have 3 2G drives in the top three slots and 2 4G disks in the next two. rdsk5 is a 4G drive with SCSI addr of 11, so it has to be one of the two on the bottom. This doesn't match up with what you say. Any ideas. Thanks in advance.

[Jim VanMeter]

John,

I'm now a little confused. If you take a look at the attached word document, I have captured some Sam screens. You state that from top to bottom the SCSI addr's are 5,11,10,9,8. It has been determined from sysout logs that it is rdsk5 that is bad, and from Sam, that rdsk5 uses SCSI 11. According to your scenario that would make it the second one from the top. But, in our box we have 3 2G drives in the top three slots and 2 4G disks in the next two. rdsk5 is a 4G drive with SCSI addr of 11, so it has to be one of the two on the bottom. This doesn't match up with what you say. Any ideas. Thanks in advance.

[Jim VanMeter]

John,

I'm now a little confused. If you take a look at the attached below, I have captured a Sam screen. You state that from top to bottom the SCSI addr's are 5,11,10,9,8. It has been determined from sysout logs that it is rdsk5 that is bad, and from Sam, that rdsk5 uses SCSI 11. According to your scenario that would make it the second one from the top. But, in our box we have 3 2G drives in the top three slots and 2 4G disks in the next two,I can determin this because it is physically written on the hot swaps. rdsk5 is a 4G drive with SCSI addr of 11, so it has to be one of the two on the bottom. This doesn't match up with what you say. Any ideas. Thanks in advance.




#
# sam
+ === Disks and File Systems (langley) (1) +
|File List View Options Actions Help |
| Press CTRL-K for keyboard help. |
| |
| |
|------------------------------------------------------------------------------|
|Disk Devices 0 of 6 selected|
|------------------------------------------------------------------------------|
| Hardware Number Volume Total |
| Path of Paths Use Group Mbytes Description |
|+---------------------------------------------------------------------------+ |
|| 8/16/5.2.0 1 Unused -- 0 Toshiba CD-ROM SCSI ^ |
|| 8/4.10.0 1 LVM rdsk4 4095 SEAGATE ST34573WC |
|| 8/4.11.0 1 LVM rdsk5 4095 SEAGATE ST34573WC |
|| 8/4.5.0 1 LVM vg00 2033 SEAGATE ST32550W |
|| 8/4.8.0 1 LVM rdsk8 2033 SEAGATE ST32550W |
|| 8/4.9.0 1 LVM rdsk9 2033 SEAGATE ST32550W |
|| |
|| |
|| |
|| v |
|+< >+ |
| |
+------------------------------------------------------------------------------+

[Jim VanMeter]

John,

I appreciate your answer very much but something doesn't seem to match up with the information I've gathered to this point. When I use SAM I can see that the SCSI device # used is 11 for the bad disk and I know it is one of our 4G hot swaps. When I look at the physical placement of the hot swaps in the box I see, from top to bottom, 2G, 2G, 2G, 4G, and 4G. I can tell because it is written on the front of the hot swap. This information tells me that it is one of the two hot swaps on the bottom. Yet, the info that you gave me tells me that SCSI 11 is in the second slot from the top. ANy ideas on what the differences could be? Below is some info gathered from SAM.


Thanks,
Jim



|------------------------------------------------------------------------------|
|Disk Devices 0 of 6 selected|
|------------------------------------------------------------------------------|
| Hardware Number Volume Total |
| Path of Paths Use Group Mbytes Description |
|+---------------------------------------------------------------------------+ |
|| 8/16/5.2.0 1 Unused -- 0 Toshiba CD-ROM SCSI ^ |
|| 8/4.10.0 1 LVM rdsk4 4095 SEAGATE ST34573WC |
|| 8/4.11.0 1 LVM rdsk5 4095 SEAGATE ST34573WC |
|| 8/4.5.0 1 LVM vg00 2033 SEAGATE ST32550W |
|| 8/4.8.0 1 LVM rdsk8 2033 SEAGATE ST32550W |
|| 8/4.9.0 1 LVM rdsk9 2033 SEAGATE ST32550W |