SG test_return ??

[MikeL_4]

Within our package start/stop script we start our applications and then do a test_return 56 and within that we execute:
56)
print "\tERROR: Application Start "
print "\tERROR: Failed to start/stop customer commands"
halt_services
customer_defined_halt_cmds
disown_dtc
remove_ip_address
verify_ha_nfs stop
umount_fs
deactivate_volume_group
deactivate_disk_group
to_exit=1
;;

If at this point I don't want to have SG start the halt package process, but rather just send out a notification that errors occured, can I change this to something like:
54)
print "\tERROR: Application Start "
print "\tERROR: Failed to start/stop customer commands"
wall message_file
to_exit=0
;;

This would allow the already started databases to remain up and then the applications group can fix the application startup problem.

[Kent Ostby · Accepted solution]

Mike -- That should work.

[MikeL_4]

thanks

[Kent Ostby]

Mike --

The actual key to the script stopping is "exit_value".

So if some OTHER aspect of the start up fails and to_exit is set (which will trigger exit_value being set) or exit_value is set directly, then the package will still fail.

However, your change should keep the script from failing within that particular part of the package startup.