Look through the script and make sure it is calling the same path you do when you manually run prtdiag. I have had issues with uname -i resolution on some hardware (sun4u vs sun4us on fujitsu hardware running solaris.)
You can echo out the critical vars to a file when it runs to see what gets set. Think script debugging to output files.
BTW - for the group - I have been working on the meta probe Galen wrote originally for my environment - I got it to go yellow on resyncs of the mirrors (SDS mirrored system volumes and some soft partitions), Today I got it to look for orphan meta devices, singletons that are not mounted, that would signal a broken mirror. I chose to look for singletons, as mirrors can be 3 way or more. I set a yellow and tell the viewer of a possible broken mirror , then list the single devices. I am thinking of putting in an exclusion file, so if a device is known and wanted to be this way for whatever purpose - it could be listed as an ignore.
Any thoughts of other functions the Meta might look for? (we use default 2 way mirroring in SDS (aka SVM) so that I know best!
Brian
lurch at inorbit.com -------Original Message-------
From: Craig Cook Date: 8/31/2009 11:05:27 AM To: hobbit at hswn.dk Subject: RE: [hobbit] bb-prtdiag.sh - Sun SPARC Enterprise M9000 Server
I am having truble with the bb-prtdiag script on "Sun SPARC Enterprise M9000 Server " and also "SUNW,UltraAX-i2" hardware. Always saying something is wrong when I cant see anything wrong?
The M9000 has not explicitly been added for support.
Check the readme with the bb-prtdiag tar file for how to run the built in debug. There is also information in the comment section at the top of the script.
Support for UltraAX-i2 has been added, run debug to find out why/where it is complaining.
bb-prtdiag is a very simple test. It looks at the return code of the prtdiag command. No. bb-prtdiag does not look at the return code from running prtdiag. If you want a simple prtdiag check there is another one on deadcat.
Craig
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk