Do I need a reboot after SCSI disk replacement

[Ratzie]

I had a SCSI disk fail, on a mirrored system.
Do I need a reboot?
HP-UX B.11.00 U 9000/800
System is L3000-7X

Hot swap drives.

How would I clear the counters in CSTM?
cstm > ru > rs
Summary of: /var/stm/logs/os/log2.raw.cur
Date/time of first entry: Wed Feb 9 04:18:37 2005
Date/time of last entry: Wed Feb 9 09:24:52 2005

Number of LPMC entries: 0
Number of System Overtemp entries: 0
Number of LVM entries: 0
Number of Logger Event entries: 0

Number of I/O Error entries: 498

[Patrick Wallek · Accepted solution]

No, no reboot needed.

If it was a non-root disk:

replace disk
# vgcfgrestore -n vg?? /dev/dsk/c?t?d0
# vgchange -a y vg??
# vgsync vg??

That will do it. And I think your cstm counters should be cleared then as well.

[B. Hulst]

Hi,

And for a boot disk replacement see this thread...

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=808464

Regards,
Bob

[A. Clay Stephenson]

No, I have not shutdown to replace a disk in well over 5 years and I have replaced many, many disks in that time.

[Geoff Wild]

This depends...

1) If the disk is completely "dead", such as if you run ioscan and status is "no_hw" then you can hot swap the disk online.

2) However in circumstances where the disk has not fully failed please do one of the following to avoid data corruption :

a) reduce mirror before replacing the disk
b) deactivate VG before replacing the disk
c) shutdown system to replace the disk

Rgds...Geoff

[Steven E. Protter]

On hot swaps the answer is no.

There will be an lbolt in the syslog.log file until you do reboot. That is not serious.

SEP

[Ratzie]

Here is the issue. I have a new HP tech here to work on a critical system.

The drive is part of a software mirror.
He states that when he ran
lvdisplay there were some stale and some current extents. If we pop a drive in and there are still current extents, it will not rebuild those extents. IS THIS CORRECT?

It is not part of the OS

ioscan -fnC disk --> hung the console

What would be the correct procedure to replace that disk. I thought if I pop out that drive and put a new one in it would recreate the drive.


There is no hardware raid on the drive just software.

LE PV1 PE1 Status 1 PV2 PE2 Status 2
00000 /dev/dsk/c1t0d0 00000 current /dev/dsk/c2t0d0 00000 stale
00001 /dev/dsk/c1t0d0 00001 current /dev/dsk/c2t0d0 00001 stale
00002 /dev/dsk/c1t0d0 00002 current /dev/dsk/c2t0d0 00002 current
00003 /dev/dsk/c1t0d0 00003 current /dev/dsk/c2t0d0 00003 stale
00004 /dev/dsk/c1t0d0 00004 current /dev/dsk/c2t0d0 00004 current
00005 /dev/dsk/c1t0d0 00005 current /dev/dsk/c2t0d0 00005 stale
00006 /dev/dsk/c1t0d0 00006 current /dev/dsk/c2t0d0 00006 current
00007 /dev/dsk/c1t0d0 00007 current /dev/dsk/c2t0d0 00007 current
00008 /dev/dsk/c1t0d0 00008 current /dev/dsk/c2t0d0 00008 current
00009 /dev/dsk/c1t0d0 00009 current /dev/dsk/c2t0d0 00009 current
00010 /dev/dsk/c1t0d0 00010 current /dev/dsk/c2t0d0 00010 stale
00011 /dev/dsk/c1t0d0 00011 current /dev/dsk/c2t0d0 00011 stale
00012 /dev/dsk/c1t0d0 00012 current /dev/dsk/c2t0d0 00012 current

[Geoff Wild]

To be safe, I would reduce /dev/dsk/c2t0d0 from the lvols and the vg - then hot swap it.

Then recreate your mirror...(which in itself is a good test on a hard drive)...

Rgds...Geoff

[James George_1]

Hi

After the HP CE replaces the disk, do the following.

# vgcfgrestore -n /dev/vg_name /dev/dsk/cXtXdX

# vgchange -a y /dev/vg_name
# vgsync vg_name

If you reduce the lv and vg then u will not be able to do the vgcfgrestore . Thenyou will have to manually do it manually with vgextend and lvextend.

Best of Luck

James

[A. Clay Stephenson]

For "flaky" hot-plug disks my procedure is as follows:

Do an lvdisplay -f for each LVOL that is housed by the disk in question. Make sure that on the good disk all extents are current. Now we are going to really clobber the bad disk by sliding it out a few centimeters and letting it spin down while still on the rails. Leave it that way for a minute or so and now you are in the NO_HW state. This will trigger a SCSI bus reset but that is considered a normal event.

Now fully remove the bad disk and insert the replacement.
Next, vgcfgrestore -n /dev/vgxx /dev/rdsk/cXtYd0

vgchange -a y /dev/vgxx
vgsync /dev/vgxx

The sync operation will take a few minutes and then all is well.

[Ratzie]

Thanks guys