topic Re: cmlocklund errors in Operating System - Linux

cmlocklund errors

Samer Dahr — Tue, 12 Aug 2008 11:02:30 GMT

Hello,

I have serviceguard installed on redhat linux 2.6.9. the quorom disk is on eva4000 , (/dev/sda1), from some days i am getting the follosing errors on the servers ,and /dev/sda1 is not accessible.

Aug 11 06:19:11 ab-app1 cmlocklund[8354]: Warning: Unable to determine local domain name for ab-app1 Aug 11 06:19:11 ab-app1 cmlocklund[8354]: Port number returned by locklund_setup: 38132 Aug 11 06:19:11 ab-app1 cmcld[3512]: Lock LUN resumed (port = 38132).
Aug 11 06:19:11 ab-app1 cmlocklund[8354]: Cannot access device "/dev/sda1" (No such file or directory).
Aug 11 06:19:11 ab-app1 cmlocklund[8354]: Error binding block device (/dev/sda1) to a char device.
Aug 11 06:19:11 ab-app1 cmcld[3512]: Service cmlocklund terminated due to an exit(1).
Aug 11 06:19:11 ab-app1 cmlocklund[8355]: Total allocated: 540672 bytes, used: 0 bytes, unused 540672 bytes Aug 11 06:19:11 ab-app1 cmcld[3512]: Automatically restarted service cmlocklund for the 3715th time after failure.

the corresponding vdisk on the eva is healthy and presented on the servers , and the fc cards are detected by the servers.

what could the problem be ?

Thanks in advance.
Samer

Re: cmlocklund errors

Paul McCleary — Tue, 12 Aug 2008 14:33:28 GMT

Hi,

What version of ServiceGuard are you using? - as there have been some lock lun issues in the past. There are patches available to fix these.

How do you fix the problem? Do both nodes have the problem?

You don't mention the SG version you're using or your O/S version. But you can go to the ITRC

http://itrc.hp.com

Select 'patch database' from the 'maintenance and support (hp products)' section.

Then select 'Linux' from the 'find individual patches' section.

Then select your Redhat version and how you want to search, could do it by selecting patch or keyword and entering SGLX_ or just do a browse of the patch list.

Hope this helps,
Paul

Re: cmlocklund errors

Steven E. Protter — Tue, 12 Aug 2008 14:33:42 GMT

Shalom,

You have a storage problem. Your cluster can not access its lock disk.

Therefore the cluster can not function at all.

You need to make sure both nodes have access to the lock disk on shared storage.

SEP