ITEM: AD2057L

restoring from a clone/0506-324 cannot mount /

Question:

restoring from a clone/0506-324 cannot mount /dev/usr0 on usr0/file or directory in path name does not exsist

Response:

Note: The customer had an install problem however he called in outside of
his contract hours and will call us back later.

Response:

This customer is valid for 24 x 7 Monthly assistance. We received
their revised supplement on 10/21/94. Please go ahead and assist the
customer. Transferred call to Jon Miles. Then transfer item back to
info.

Response:

Env: AIX 3.2.5, Two systems connected twin-tail to a RAID array.
Desc: Customer has two systems hooked to a RAID array, and one of
them went down. When attempting to restore from a backup on the
machine, he found that the backup was bad, so he restored from a
mksysb of the other machine, and then restored the configuration files
from the backup of the first. After doing this, the system comes up
fine, but the email0vg volume group, which is located on the RAID
array, will not varyon.
Action: Conferenced in Dan Braden for help with the problem, and he
instructed the customer to do the following:
1. export the volume group
2. remove the configured hdisks from the RAID array
3. run cfgmgr
4. import the volume group from the RAID disks
5. varyonvg the volume group
The customer will try this, and will call back if he has any trouble.
I am re-queueing this item to the KERN queue as this is a problem with
the logical volumes and not an install/backup issue. The customer
will call back if he needs assistance, but it might be a good idea to
give him a callback by noon on 11/10 to make sure that he does not
require further assistance.
Next Action: Re-queueing to KERN.

Response:

Desc: Bruce Friedman, the IBM System Engineer-OSS for UNC called.
Two Risc R20s with 7135 RAID between them. The RAID contains
4 disks, hdisk2-5. Both systems contain 2 internal disks in rootvg
called hdisk0 and hdisk1. hdisk2,3 make up volume group emailvg01
and hdisk3,4 emailvg02. The two volume groups are NFS mounted across.
. The system serves over 16,000 student email accounts and contains
over 10GB of data.
.
Act: At this point, machine B is powered off and they are trying to varyon
emailvg01 on system A. This fails stating that 'lqueryvg ailed because
hdisk2 is not a volume group member.' We tried to run
'lqueryvg -Atp' on hdisk2-5, but they all return a message that they
are not a volume group member. Either the disks are not being accessed
properly, or they have lost the VGDA info.
. We tried removing the disks with 'rmdev -l hdisk\# -d' and then running
cfgmgr to reconfigure the disks. At that point, hdidsk2-5 were not
listed at all in the 'lspv' or 'lsdev -Cc disk' output. Cfgmgr -v did
not report any problems though. Bruce checked the cabling, but found
nothing wrong. The strange thing, is that: [smit readi, list all disk
arrays], lists hdisk2-5
. I then found out that machine A was reinstalled yesterday evening with
a mksysb backup created on machine B! This could have led to the disk
problems because of the identical ODM copies on different machines.
I recommended shutting down machine A and rebooting B. When B booted,
it had adapter problems and listed a new disk, hdisk6 as available and
using the same slot as hdisk5(unavailable).
. Bruce then ran the steps in dsksync.rvg.krn to rename disks properly.
Upon reboot- the correct disk names were shown, but email01vg still
would not varyon. 'lqueryvg' could not read the VGDA off any disk hdisk2-
5. It appears that the nonroot disks are empty and need to be restored
from backup.
. The problem is that the backups they have seem to be corrupt. They
have tried 2 different Sysback tapes, but always get errors:
. import of PCLs failed
do not import existing file
do not specify existing file
restore 0511-119 and 0511-139

Response:

Act: Bruce was preparing to send the Sysback tapes in for testing when
he decided to install one on a different RISC System- an R25T and
guess what? The tape restores fine!
. He is running Sysback 3.1.0.18 on both the R25T and the R20s. Same
tape drives. I found item 8X612,b137 in Retain which describes the
same problem and states that the cause may be with the SBREAD and
SBWRITE routines. The >3250 C library fix U434513 should fix the
problem.
. Bruce then mentioned that they do have a modified libc.a file on the
R20 machines because of the large number of users.
.Next: They will probably reinstall AIX 3.2.5 on one of the R20s, Sysback 6000,
and then restore the Sysback tape.
Just in case- I also opened pmr 6x171,b352 to order the ptf for
him.
. closing call

Response:

Action: Assigning to Paul C. for closure.

Support Line: restoring from a clone/0506-324 cannot mount / ITEM: AD2057L

Dated: November 1995 Category: N/A

This HTML file was generated 99/06/24~13:30:30

Comments or suggestions?
Contact us