syslog messages

[Sridhar R]

Hi,

Below is the message that i get in the syslog.log file.

Feb 4 19:36:50 hrdvault cmcld: Health query for cluster lock disk /dev/dsk/c8t0d1 shows BUSY lock
Feb 4 19:36:50 hrdvault cmcld: Check device, power, and cables.
Feb 4 19:37:19 hrdvault vmunix: LVM: VG 64 0x010000: PVLink 31 0x090100 Failed! The PV is still accessible.
Feb 4 19:37:19 hrdvault vmunix: LVM: VG 64 0x010000: PVLink 31 0x090100 Recovered.


What does this mean? Anything serious problem with the server or is it negligible?

I get this message almost daily!

Kindly help!!

Thanks in Advance!!

Sridhar

[Torsten.]

The link to the disk in vg01 (c9t0d1) was failed and has been restored later.
If this is a SAN device the link (cable/switch port ...) was probably disconnected or at least not accessible for a moment.

[Anshumali]

If you are getting these messages almost daily, may be you would like to get it checked. One link is failing resulting in these messages.

[Prashanth.D.S]

Hi Sridhar,

Frequent messages means that this LUN/disk is failing intermittently.

Issue could be with the Hardware (c9t0d1) check by running dd command and diskinfo command.

Also check the LVM patch level upgrade to the latest.

Best Regards,
Prashanth

[Ramesh S]

Hi

You could also check the PV timeout value for that PV.

Hope this helps you!!

Ramesh

[Sridhar R]

Hi Prashant,

What would be the syntax to check the disk via dd command.

And moreover, how to check the PV timeout value?

Sridhar

[Ramesh S]

Hi

To check the disk with dd is

# dd -f=/dev/rdsk/cXtYdZ of=/dev/null bs=1024k count=1000

To check the PV timeout value

# pvdisplay /dev/dsk/cXtYdZ

and look for PV timeout value

Thanks & Regards

Ramesh

[Ramesh S]

Hi

To check the disk with dd is

# dd if=/dev/rdsk/cXtYdZ of=/dev/null bs=1024k count=1000

To check the PV timeout value

# pvdisplay /dev/dsk/cXtYdZ

and look for PV timeout value

Thanks & Regards

Ramesh

[Avinash20]

The "BUSY lock" messages logged once during the Health query for the cluster lock disk are only an indication of a temporary problem accessing the cluster lock device.

Serviceguard versions A.11.17 and earlier perform an hourly health check so an occasional "BUSY lock" message is not a concern. If there was an issue with the cluster lock disk or with the lock field, there would be an error logged every hour.

But if these messages begin to be logged regularly or they are also accompanied by any I/O failed messages for the cluster lock, a hardware case should be
logged to checkout the hardware.


NOTE: There is also a known issue for Serviceguard version A.11.15 where you may get the BUSY lock during a Health query for the cluster and this issue for A.11.15 was fixed in the PHSS_30370 Serviceguard A.11.15 cumulative patch.