cmclconfd message "Probed disk /dev/dsk/cxtxdx, the 100th disk"

[UNIX Engr]

Hello,
I am receiving this message in syslog when I run the cmquerycl command on our cluster:
"Feb 1 08:52:58 dnux068 cmclconfd[7735]: Probed disk /dev/dsk/c57t1d2, the 100th disk."

We just put in an XP12000 and are in the process of building luns. Anyone have any idea if this is informational only or something to be concnerned about?

Thanks!
Steve

[Sameer_Nirmal]

Hi,

Eventhough the message looks as "information", need to know
SG version?
HP-UX version?
Patch status ?
Are you using Secure Path ?

As far as I know, the raw device should be used by "cmclconfd".

[Steven E. Protter]

Shalom Steve,

Nice name.

If the cluster is running its informational.

If you are also building a cluster now the results of cmapplyconf and cmcheckconf will determine definitively what this means.

SG is disk aware and if these disks should not be available to the SG node they should be configured if possible so that they don't show up on ioscan.

That may be possible.

SEP

[UNIX Engr]

to update a bit:
We are running SG on a superdome connected to an XP12000. We built 250 dual pathed luns when the XP went live two weeks ago. Antyime we run the cmclconfd command with our existing cluster config, we get that message.

SG version is 11.16
Unix Version is 11.11
Patches are current as of last week
and Secure path version 3.0f.

[Carsten Krege]

This message is pure informational. SG gives an additional message for every 100 disks that were probed. This message was added in SG A.11.16.

Carsten

[Sameer_Nirmal]

Hi Carsten,

Just trying to seek clarification here for understanding.

As per the latest patch details for SG 11.16, cmclconfd has been modified to use raw devices.

The informal message indicates using block device for probing.
Is it using block devices for probing only and raw devices for read etc..?

[Carsten Krege]

Hi Sameer,

you are right, starting with PHSS_33834 for SG A.11.16 (and SG A.11.17), cmclconfd does use raw devices. However, the message being logged still lists the block device and one could come to the conclusion that SG still uses the block device.
I can assure you that with the above patch level SG now uses the raw device whenever the device name starts with "/dev/dsk". The arguments that SG uses to output the message and to actually open(2) the device are different.

Well spotted though. :)

Carsten