DRD Clone Error: Selection of the target disk fails

[Stephen Wallace]

Cloning an Itanium (rx2660) boot disk to an EMC CLARiiON LUN fails with:
"ERROR: Selection of the target disk fails.
- The device special file "/dev/dsk/c7t1d3" cannot be identified in the system configuration information."

The LUN device is visible with ioscan:
# ioscan -fnkC disk /dev/dsk/c7t1d3
Class I H/W Path Driver S/W State H/W Type Description
==========================================================================
disk 47 0/2/1/0.1.1.0.0.1.3 sdisk CLAIMED DEVICE DGC CX700WDR5
/dev/dsk/c7t1d3 /dev/rdsk/c7t1d3

[Marco A.]

Hi,

If that disk was used previously for data, etc, you need to overwrite (overwrite=true)the information first as follows.

# drd clone -x overwrite=true -p -v -t /dev/dsk/c7t1d3

If not, well, another option could be check the necessary patches in this location:

http://docs.hp.com/en/DRD/patch.html

[Moderator note: Above link is no longer valid. Please refer to https://support.hpe.com]


Regards,

Marc'o

[Stephen Wallace]

Hi Marco,
Thanks for the quick response but...
the SAN LUN is brand new and unused, and the unsafe patch list is up-to-date (Mar. 24, 2008).

[Stephen Wallace]

Can anyone tell me exactly what is meant by "the system configuration information"?
I haven't been able to determine which system configuration file(s) are being referred to.

[Sameer_Nirmal]

As fas as I know, when "drd" runs, it scan the system's h/w and creates a system configuration file under /var/opt/drd/tmp/drd*/sconf*. This file is parsed to get the disk details (h/w, configuration etc.) besides other information.

Even "ioscan" sees the disk but for some reason, the drd is not able to see the disk.I would run following commands and check their output.

# /opt/drd/lbin/drd_scan_hw_host
# /opt/drd/lbin/disklist -c /tmp -f -p

I would also run drd clone with verbosity level 5 and check the log /var/opt/drd/drd.log

[Stephen Wallace]

Thanks for the info Sameer.

I checked the /var/opt/drd/tmp/drd*/sconf* file and found that there are entries for the new disk (SAN LUN):

# grep c7t1d3 /var/opt/drd/tmp/drd*/sconf*
io_info += "diskid: iux-0000-001a [] [\] 188 31 0x71300 /dev/rdsk/c7t1d3 /dev/dsk/c7t1d3 1 47 sdisk DGC_CX700WDR5 [\] [\] [\] [\] [\] [\] disk [\]"
...
io_info += "diskp: iux-0000-001a [] 188 31 0x71300 /dev/rdsk/c7t1d3 /dev/dsk/c7t1d3 5 47 sdisk 0/2/1/0.1.1.0.0.1.3 fibre_channel 9 [\] sdisk 0/2/1/0.1.1.0.0.1.3 47"
...
hw_info += "disk: 0/2/1/0.1.1.0.0.1.3 47 sdisk 188 31 71300 DGC_CX700WDR5 0 /dev/rdsk/c7t1d3 /dev/dsk/c7t1d3 -1 -1 0 1 "

The drd_scan_hw_host command lists only the current boot disk:

# /opt/drd/lbin/drd_scan_hw_host
* Scanning system for IO devices...
* Boot device is: 0/1/1/0.0.0.0.0

... and when I run 'disklist' against the LUN hardware path, it does see it:

# /opt/drd/lbin/disklist -c /tmp -f -p 0/2/1/0.1.1.0.0.1.3
Tue Nov 11 10:26:32 2008
0/2/1/0.1.0.0.0.1.3,0/2/1/0.1.1.0.0.1.3,0/4/1/0.1.0.0.0.1.3,0/4/1/0.1.1.0.0.1.3|c6t1d3,c7t1d3,c8t1d3,c9t1d3|LVM|vgclone1|65552|4|scanning|DGC CX700WDR5

So it seems that DRD sees the disk, but the clone process doesn't like what it sees. In fact, when I use max. verbosity level, the drd clone commands lists all the other LUNs on the server as 'bad disks' (even though none of these disks are in the root volume group!) :

/opt/drd/bin/drd clone -p -v -t /dev/dsk/c7t1d3 -x verbosity=5 -x logfile=/var/opt/drd/d>

======= 11/11/08 10:33:02 GMT BEGIN Clone System Image Preview

* Reading Current System Information
WARNING: Device (0/2/1/0.1.1.0.0.0.1) is a bad disk.
WARNING: Device (0/4/1/0.1.0.0.0.0.1) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.0.3) is a bad disk.
WARNING: Device (0/4/1/0.1.0.0.0.0.3) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.0.5) is a bad disk.
WARNING: Device (0/4/1/0.1.0.0.0.0.5) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.0.7) is a bad disk.
WARNING: Device (0/4/1/0.1.0.0.0.0.7) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.1.0) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.1.4) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.1.3) is a bad disk.
WARNING: Device (0/4/1/0.1.0.0.0.1.3) is a bad disk.
* Selecting System Image To Clone
WARNING: Device (0/2/1/0.1.1.0.0.0.1) is a bad disk.
WARNING: Device (0/4/1/0.1.0.0.0.0.1) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.0.3) is a bad disk.
WARNING: Device (0/4/1/0.1.0.0.0.0.3) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.0.5) is a bad disk.
WARNING: Device (0/4/1/0.1.0.0.0.0.5) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.0.7) is a bad disk.
WARNING: Device (0/4/1/0.1.0.0.0.0.7) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.1.0) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.1.4) is a bad disk.
WARNING: Device (0/2/1/0.1.1.0.0.1.3) is a bad disk.
WARNING: Device (0/4/1/0.1.0.0.0.1.3) is a bad disk.
* Selecting Target Disk
ERROR: Selection of the target disk fails.
- Selecting the target disk fails.
- Validation of the disk "/dev/dsk/c7t1d3" fails with the following error(s):
- The device special file "/dev/dsk/c7t1d3" cannot be identified in the system configuration
information.
* Selecting Target Disk failed with 1 error.

======= 11/11/08 10:33:19 GMT END Clone System Image Preview failed with 1 error.



This only leaves me more confused!!

[Sameer_Nirmal]

When you run the drd_scan_hw_host, it creates 3 files under /tmp named host.info, hw.info and io.info

The information retured by the grep command on the configuration files returns the similar information in hw.info and io.info

The hw.info in this case shows LUN size as 0 and io.info doesn't show size at all. Maybe that's why the LUN is being decleared as " a bad disk ".

The disklist reults shows some wonderful results. It looks there are 4 paths to the LUN? Moreover it shows LVM header maybe it's LVM VG clone disk?

Are you using powerpath?

[Stephen Wallace]

No, Powerpath isn't being used.

There are 4 paths to each LUN - the server has 2 HBAs connected to separate switches, and each switch is connected to each of the 2 CLARiiON SPs. In this case, SP-B is the default.

And you're right about the LVM info - I'd initially run the 'drd clone' command on the vanilla LUN, but when that didn't work I tried adding the LUN to a volume group (vgclone1). Still got the same result though!

Since then, I've removed the disk from LVM control (vgreduce/vgremove/pvremove) and rerun the 'drd clone' command - again to no avail.