replace bad root disk

[navin]

Hello -

I have an itanium box which has bad root disk.
I would lik eto know the answer for couple of questions - pls help.

1) setboot -v shows that there was alternate boot disk device is specified and autoboot ON in case of a disk failure. So i would assume in case of a bad root disk it automatically failed over to the mirror disk.But how do i make sure this had happened ? and how should i confirm currently i'm running ona mirror disk not on root disk(because still the drive has green lights)
2)How should we find from which disk the server boot up in the first place (before the failure)
3) This is lvm and vxfs - what is th eprocedure to replace a drive and do the mirroring.Also do i have to boot off th e mirror and do the replacement - or with out down time can it be done while system is running on the root mirror .

Server rx4640 - hpux 11.23

Please advice - much appreciated

[Raj D.]

Navin,


1) setboot -v shows that there was alternate boot disk device is specified and autoboot ON in case of a disk failure. So i would assume in case of a bad root disk it automatically failed over to the mirror disk.But how do i make sure this had happened ? and how should i confirm currently i'm running ona mirror disk not on root disk(because still the drive has green lights)

- If mirroring is in place and the alternate disk is getting mirrored data, it will switch automatically during a boot disk failure of primary , make sure -lq (low quorum )option is specified in the boot string.

To check -lq option is set in the boot string or not
# efi_ls -d/dev/rdisk/$TARGET_DSK'_s1' EFI/





2)How should we find from which disk the server boot up in the first place (before the failure)
- setboot will so which one is primary and which one is alternate.
- Keep an eye on the LED while the server is booting through the ISL.
- dd if=/dev/dsk/ctd of=/dev/null
# By seeing the LED activity you can figure out which disk it is refering.


3) This is lvm and vxfs - what is th eprocedure to replace a drive and do the mirroring.Also do i have to boot off th e mirror and do the replacement - or with out down time can it be done while system is running on the root mirror .

- http://www.docs.hp.com/en/5991-1236/When_Good_Disks_Go_Bad_WP.pdf



Hth,
Raj.

[Raj D.]

Navin Also,

> and how should i confirm currently i'm running ona mirror disk not on root disk(because still the drive has green lights)



- 1. If your disks are mirrored, use lvdisplay and you will see two disk .

# lvdisplay -v /dev/vg00/lvol1 # to...8




- 2. Verify the content of AUTO file on both the disks primary and in the mirror.


# efi_cp -d /dev/rdsk/c2t1d0s1 -u /EFI/HPUX/AUTO /tmp/AUTO.pri
# efi_cp -d /dev/rdsk/c3t2d0s1 -u /EFI/HPUX/AUTO /tmp/AUTO.alt

# cat /tmp/AUTO.pri
# cat /tmp/AUTO.alt


Hth,
Raj.

[Raj D.]

Navin ,
simple steps outlined to replace the faulty disk:

1. Identify the faulty disk.
# lvdisplay -v /dev/vg00/[lvol1 to lvol8 ]
# look for stale PE. That stale PE disk is faulty. If the disk is completely dead you will see NO_HW in ioscan.

2. Remove the mirror (separate the faulty disk):
LV reduce :
# lvreduce -m 0 /dev/vg00/lvol1
[ Do for all the lvols till lvol8 ]

3. Remove the Bad disk from the vg00:
# vgreduce /dev/vg00


4. # pvchange -a N

5. # Remove the faulty hotplug disk.

6. # Insert the NEW hotplug disk.

7. # Run ioscan , make sure it is visible.

8. # Partition the disk using idisk , (mentioned in the doc).

9. # insf -eC disk

10. # vgcfgrestore -n vg00

11. # pvchange -a y

12. # vgchange -a y vg00
# vgsync vg00

It will start resyncing the mirror with the primary disk. The process will take around 1 hour . Once the mirroring done you will see the messages and the command prompt will be back.


Hth,
Raj.

[navin]

Raj appreciate your replies - mine is itanium server - so i ahve to do the disk format and once the data sync is done - do i have to do the steboot etc?
Thanks

[Raj D.]

If your ctd number is not changing you don't have to do the setboot.

- In case ctd number changes it is simple:
After the vgsync just do ( setboot -a Alt_path ):

# setboot #Verify the listed Pri & Alt paths.

# Set alternate boot disk path:
# setboot -a ${ALT_HWPATH}


HW_Path can be found with the following commands.
# ioscan -fnC disk #Look for the Alt disk.
# vgdisplay -v vg00 | grep "PV Name"
# lvlnboot -v vg00


Hth,
Raj.

[navin]

Appreciate your time and expert solutions

[Raj D.]

Navin,
If you want to show appreciation when others
take time to reply to your questions, at a
minimum you should assign points.

I can see 930+ responses (in your profile) has not assigned any points and so called the appreciation.

"I have assigned points to 577 of 1512 responses to my questions."



Check it out:
http://forums11.itrc.hp.com/service/forums/questionanswer.do?&threadId=1377483

http://66.34.90.71/ITRCForumsEtiquette/after.html




I was checking your profile and found this:
Your 930+ Questions with unassigned points:


http://forums11.itrc.hp.com/service/forums/pageList.do?forumId=1&userId=CA712686&listType=unassigned&

Probably you can take a day off or do it in the weekend : appreciate them in terms of "assigning points" to these 930+ responses who deserve points to their valuable time & valuable answers. If you ignore this ll be reminded again soon.