Disk replacement question

[Waqar Razi]

I am a remote unix administrator and we have a customer whose vg00 hard disk (c0t6d0) failed. This disk is mirrored to another disk (C3t2d0)on the server and the disk is hot swapable.

The server is RP 7410 and the disk is hot swapable.
The problem I am having here the server is not properly labelled and the customer is unable to find the failing disk (c0t6d0).

Is there any way that you can point out physically which disk is failing by sending some commands or signals. THe disk does not have yellow light on. I dont want that some one pulled the wrong disk (working disk ) out and crush the whole system.

Any help in this regard is highly appreciated.

[R.K. #]

Hi Waqar,

# ioscan -funC disk

Check the hardware path.
See if it helps as vg00 mostly have internal disks.

Regds,
R.K.

[24 7]

hi,

you can give a dd command as

dd if=c0t6d0 of=/dev/null bs=1024

then tell the local person to look for the bhysical disk LED activity .. so that he can identify the most blinking HDD.

Rgds.
24/7

[Patrick Wallek]

>>then tell the local person to look for the
>>bhysical disk LED activity .. so that he can
>>identify the most blinking HDD.

And how is this supposed to work if the disk is dead? If the disk isn't working dd will likely just hang.

[Henry Fauni]

If the disk is not completely dead, you can try the dd command:

dd if=/dev/dsk/c0t6d0 of=/dev/null bs=2048k

This will light up the LED solid on the disk in question. Ctrl C to get out.

lssf /dev/rdsk/c0t6d0 will give you H/W path. HP CE's should be able to figure out where it is physically if you give them H/W path.

If they're both internal disks, and if you can't do a "dd" on the bad disk. Your option is to do the "dd" on the good disk, then tell them not to touch it, and replace the other dead disk instead! :)

[24 7]

Patrik,
then you give a resolution!

If the local engineer not able to identify by looking on the LEDs , probable the disk is not fully dead. So he can try the command what I suggested.

[Vikas Thorat]

Hi Waqar,

Just, follow below steps :

1) # lvlnboot -v
This command tells you the primary boot disk and secondary boot disk.
In your case :
Pri - c0t6d0
Sec - C3t2d0

2) # ioscan -fnC disk
Try to find here your above discs and locate there H/W Path it may help you to identify location.

3) # olrad -q
olrad - command for OnLine Addition/Replacement/Deletion of PCI I/O cards and Online Addition/Deletion of I/O chassis.

4) # dd if=/dev/dsk/c0t6d0 of=/dev/null bs=1024
As soon as you execute this command the disk LED will start glowing constantly. Here you have to ask your onsite person to confirm the blinking and after get successfull confirmation ask him to label it first. Do "ctrnl + c" to come out from command.

Do the above command for both Primary and secondary disk.

Note : Give more attention to "if" and "of" values, if you exchange these value by mistake it will scrap your entire data. Hence, it is most dangerous command if used in wrong way.
After labeling you can proceed with Disk replacement. Recommanded to have HP Engg at onsite to replace the disk to avoid other problems.

regards,

Vikas

[Michal Kapalka (mikap)]

Hi,

the best way to identify the disk is to ask hp hardware support technician, or use dd, and check the LED on the disk.

I found a document from RP7410, about HW path's. But I'm not able to garantie if it's correct, in our company we don't have this server.

RP7410 HW paths is in attachment of this response.

if you are use "ioscan -fnC" disk you should be able to identify which disk is failed with the HW path.

mikap