Failing disk in vg00

[R.O.]

Hi,

We have been told to replace a failing disk belonging to vg00. But this disk seems to have dissapeared, because I cannot see it nor in ioscan, neither in stm, pvdisplay, diskinfo or vgdisplay. Attached is a "vgdisplay -v vg00", when I can see:

Cur PV 2
Act PV 1

I think I can directly remove the disk and insert a new one. Then I think I have to do a "vgcfgrestore -n vg00 /dev/vg00/c0t8d0"

With these commands, will vg00 have Act PV=2 again? Afterwards, I am going to make it bootable with

# mkboot /dev/rdsk/c0t8d0
# mkboot -a "hpux -lq" /dev/rdsk/c0t8d0

and lvextend -m 1 for each lvol in the new disk.

Is this procedure correct?

Regards,


P.S: I have no direct access to ths system, and now, checking the rc.log file they sent me, I have seen:

Virtual Partitions Initialization
Output from "/sbin/rc1.d/S105vparinit start":
----------------------------
Created VPAR device files
Created VPAR dump file (NORMAL)

.....

Output from "/sbin/rc1.d/S200CONFserver start":
----------------------------
WARNING: Inquiry failed for EXCLUDED disk /dev/rdsk/c0t8d0 Disk NOT Excluded!
Vpath: Configuring 6 devices (disks)
vgchange: Warning: Couldn't attach to the volume group physical volume "/dev/dsk/c0t8d0":
The path of the physical volume refers to a device that does not
exist, or is not configured into the kernel.
Volume group "/dev/vg00" has been successfully changed.
Volume group "/dev/vg01" has been successfully changed.
Volume group "/dev/vg02" has been successfully changed.

And in a "ps -ef" text I see the vpard daemon running, so I gess it is a vpar.

Do I need any change in the procedure above provided it is a vpar?

[Pete Randall]

The fact that it's a vpar should make no difference. This guide should give you all the necessary steps:

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


Pete

[Torsten.]

The only difference is the auto boot string written to the disk header:

# mkboot -a "hpux /stand/vpmon -a" /dev/rdsk/...


Note: it is "vpmon", not "vmunix"!

[R.O.]

Hi,

I am not sure of these steps:

Can I directly remove the missing disk and insert a new one?

Do I need to "pvcreate -B /dev/vg00/c0t8d0" before vgcfgrestoring ?

After doing a "vgcfgrestore -n vg00 /dev/vg00/c0t8d0" and "vgchange -a y /dev/vg00" will the mirror synchronize automatically or have I re-mirror again with "lvextend -m 1"?

Regards,

[R.O.]

Typos:

"pvcreate -B /dev/rdsk/c0t8d0"
"vgcfgrestore -n vg00 /dev/rdsk/c0t8d0"

Sorry

[Torsten.]

If you follow the PDF, you will know the steps.


No "pvcreate" needed, but a "vgchange -a n ..." before !!!

[Torsten.]

Sorry, typo :-(


If you follow the PDF, you will know the steps.


No "pvcreate" needed, but a "pvchange -a n ..." before !!!

[R.O.]

Hi Torsten,

"No "pvcreate" needed, but a "pvchange -a n ..." before !!!"

The problem is the bad disk is not seen with vgdisplay, ioscan, diskinfo,etc. So I guess I can pull it out directly....

Regards,

[Torsten.]

So the system was booted with this dead disk?

Or do you see NO_HW?

Try it.

You don't need the pvcreate, because you do the vgcfgrestore - it is the same and more at this point.

[R.O.]

Hi,

Yes (see my first post). This is from rc.log:

vgchange: Warning: Couldn't attach to the volume group physical volume "/dev/dsk/c0t8d0"

And in ioscan the device is missing, I mean it does not appear at all, even as "NO_HW". The same with diskinfo.
See the vgdisplay attached in the first post: the disk is missing and Cur PV 2 > Act PV 1

Regards,

[Torsten. · Accepted solution]

OK, now follow the steps from this point on

Step 2: Replacing the Faulty Disk - page 25

and remember "vpmon"!!!

[R.O.]

Hi,

We directly replaced the disk and after doing "vgcfgrestore" it started synchronizing.

Thank for your help!

Regards,