Serviceguard Issue

[JOHN TURNER_2]

we are using service guard for mounting and unmounting volume groups on different machines using a shared array - DS2300, we unmounted the volume group from one machone did a disk copy and when we try and mount up the volume group on another machine we get the error
"script failed with no restart: wps1dbmis should not be restarted."

this was however working last week. can anyone shed any light on the subject or know of any documents / white papers that provide solutions to serviceguard errors

[RAC_1 · Accepted solution]

are you mounting it with script?

Seems to be problem with script.

Activate the vg on the node where you want mount the file. and try mounting manuly.

Post the error if any.

[John Meissner]

I agree with Anil - do a

vgchange -ay

This should fix your issue.

[David_246]

Hi John,

When importing VG (manualy, you'd do a "fsck -m", this does a quick check on the FS that should be ready for mount. This tool will report if your FS is corrupt or needs a check.
After you discovered a fS that is corrupt or just needs checking do a fsck (-y) FS

Be carefull with the -y option.

If even importing the VG is a problem, this can be very easily caused by an incorrect exporting of the VG. Therefor on the other node a "vgchange -a n" was necesary to export it correct and terefor a "vgchange -a y" should work on the other node. A "vgchange -a r" could be possible as well for read-only mounting.
Please show us any error messages that the above commands will report.

Best Regs David

[Duncan Edmonstone]

Instead of all these workarounds why not try to get to the root of the problem?

The script that failed is the package control script, (which in your case is probably called /etc/cmcluster/wps1dbmis/wps1dbmis.cntl or something like that... This script writes to a log file, usually called .log (so in your case it's probably called /etc/cmcluster/wps1dbmis/wps1dbmis.cntl.log) - Have a look in there for the reason why the script falied.

HTH

Duncan
You should have a log file

[Michael Steele_2]

(* Assuming LVM and not VXVM. *)

1) Was this a disk copy to another machine not in the cluster?

2) Use 'cmcheckconf -k -v -C /etc/cmcluster/cmclconf.ascii' to verify the connectivity to all disks within the cluster.

3) hat is the value of the 'service_restart' field in package.cntl? Should be, "-r n" -or- "-R" and not blank.

4) Any core dumps in /var/adm/cmcluster or /etc/lvmconf?