serviceguard won't vgchnage volume groups

[Paul Wasik]

Hi, I have a two node cluster that exhibits a failure to operate the volume groups with commnads like 'vgchnage -a n' and 'vgchnage -a e'. this occurs durring package starts and stops. I can manually run the "-a n" and "-a e" on approriate nodes resulting in a successfull package start/failover. the actual failure logged is fsck not able to access device (because the vg was not changed).

my thoughts are on an issue in some portion of the "sg/volume_groups" or it's "operation_sequence".

[Matti_Kurkela]

Please attach your package configuration (or at least the relevant parts) and copies of the package log and syslog entries created when trying to start/stop a package.

Identifying your Serviceguard version would be very useful too.

Without some hard facts, you're not likely to get anything but sympathy and wild guesses. We are sysadmins, not oracles or seers...

MK

[Paul Wasik]

I've come to find a missing log entry. in the package log there is no mention of the "volume_groups.sh" script being run.

I also removed several variables from the conf file (for cvm vxdg ...)

see attached.

[Paul Wasik]

here's a typical log of a failure.

[Paul Wasik]

I created and solved my own issue. note in the conf file there is no "vg" variables. therefore no colume_group.sh need run ....

I replaced my "vg" and all seems fine now.

[Paul Wasik]

close