DRD spits up error on disk not in vxvm disk group

[Steven E. Protter]

root@dxd22hxd# /opt/drd/bin/drd clone -t /dev/disk/disk27 -x overwrite=true

======= 02/21/11 14:51:16 CST BEGIN Clone System Image (user=a121360)
(jobid=dxd22hxd)

* Reading Current System Information
* Selecting System Image To Clone
* Selecting Target Disk
NOTE: There may be LVM 2 volumes configured that will not be recognized.
* Selecting Volume Manager For New System Image
* Analyzing For System Image Cloning
* Creating New File Systems
* The command "vxrootmir" is being invoked. This step will take some
time......
ERROR: Clone file system creation fails.
- Creating the target file systems fails.
- Command "/usr/lib/vxvm/bin/vxbrk_rootmir" fails. The entire output
from the command is given below:
- Start of output from /usr/lib/vxvm/bin/vxbrk_rootmir:
- VxVM vxbrk_rootmir INFO V-5-2-4023 15:23: Checking specified disk(s)
for presence and type
VxVM vxbrk_rootmir INFO V-5-2-4025 15:23: DA disk27_p2, DM
rootdisk02 is a valid root mirror
VxVM vxbrk_rootmir INFO V-5-2-4052 15:23: Saving configuration data
for later restoration
VxVM vxbrk_rootmir ERROR V-5-2-4596 More than 1 plex on DM
rootdisk02 for optvol (optvol-02
varoptvol-02)

root@dxd22hxd# vxdisk -o alldgs list
DEVICE TYPE DISK GROUP STATUS
disk25_p2 auto:hpdisk rootdisk01 rootdg online
disk29 auto:cdsdisk disk29 vg01 online
disk30_p2 auto:hpdisk disk30_p2 swapvg online
disk31_p2 auto:hpdisk disk31_p2 swapvg1 online
disk32 auto:cdsdisk - - online
disk33 auto:cdsdisk - - online
disk34 auto:cdsdisk - - online
disk35 auto:cdsdisk - - online
disk36 auto:cdsdisk - - online
disk37 auto:cdsdisk - - online
disk38 auto:cdsdisk - - online
disk27_p2 auto:hpdisk rootdisk02 rootdg online
root@dxd22hxd# vxdisk -o alldgs list
DEVICE TYPE DISK GROUP STATUS
disk25_p2 auto:hpdisk rootdisk01 rootdg online
disk29 auto:cdsdisk disk29 vg01 online
disk30_p2 auto:hpdisk disk30_p2 swapvg online
disk31_p2 auto:hpdisk disk31_p2 swapvg1 online
disk32 auto:cdsdisk - - online
disk33 auto:cdsdisk - - online
disk34 auto:cdsdisk - - online
disk35 auto:cdsdisk - - online
disk36 auto:cdsdisk - - online
disk37 auto:cdsdisk - - online
disk38 auto:cdsdisk - - online
disk27_p2 auto - - error

This may be a vxvm issue.

What I'd like to do is completely remove all partitioning information from disk27 and try again.

I am currently trying dd but think there may be a better answer.

SEP

[Steven E. Protter]

Support has identified this as a vxvm support issue.

SEP

[Rita C Workman]

It won't help, but I came across this the other day when checking on something...I thought it might be related to your issue, since DRD is making a clone that is bootable.

Just an FYI:

From setboot(1m) in HP-UX 11i v3 document, page 12:

VxVM
VxVM does not support the agile view for boot devices. VxVM converts any lunpath hardware path to a legacy hardware path before using it.
Please refer to the â Impacts to Other Subsystemsâ in the whitepaper The Next Generation Mass Storage Stack: HP-UX 11i v3 for more information.
12


Was thinking the work-around might be to change the legacy reference and for setboot try it this way (also found in this document) for legacy:

Example: force the use of the lunpath hardware path provided as input as alternate boot path
# setboot â a 0/5/1/0. 0x50001fe15001eea9.0x4001000000000000
Alternate bootpath : 0/5/1/0. 0x50001fe15001eea9.0x4001000000000000
Please see Table 1 below for a summary of setboot behaviors when used with agile addressing.

...which is a bit different than the usual legacy h/w path.

Kindest regards,
Rita

[Rita C Workman]

Boy...c&p messed up:

Should read:

setboot -a 0/5/1/0. 0x50001fe15001eea9.0x400100000000

I'm thinking if you change from using agile to this revised legacy approach, it might be something to try.

Shalom,
Rita

[Steven E. Protter]

Update:

What is actually broken is:
SR: QXCR1000970687) SYMANTEC Incident Number : 1737228(1274204) The names of the plexes to be mirrored are obtained by using the 'grep' command for "volume_name-". The 'grep' command returns the plexes of the other volume, which has the same end name as the volume that is being mirrored. This leads to incorrect plex names, thereby causing vxbrk_rootmir to fail.
I suspect that this problem has been ported to 5.2 and not fixed, although I would need to confirm a few things.



The vxbrk_rootmir which is actually just a script is broken.

Patches may fix it.

HP support actually sent me a hack for the script which I am testing.

SEP

[Steven E. Protter]

PHCO_41434/
PHCO_41046/
PHKL_41678/

These patches fix the problem.

Kind of tough to get them into a production environment.

Moving on to other issues, testing my integrated patch scenario.