No luck so far with disabling the service. That is, calling svcadm
disable at the end of the script does what it should but the service
will then not be restarted when OOo is updated. In other words the
restart_fmri associated with one of the files will not make the
service restart when it is disabled. This is similar to the behavior
of 'svcadm restart', which only works if the service is enabled. So
I assume this is the intended behavior.
A member of the IPS team just told me that the SMF actuators have no
impact on services on an alternate BE. In that case, disabling the
service at the end of the start method will prevent you from running
your script on update. The service should not disable itself here.
Actually I tried it on one and the same BE. That is, I first installed
OOo with the service which installed properly all OOo extensions. Then I
added a "new" version of OOo to my local IPS repository, and called 'pkg
install ooo_test' (ooo_test is the name of my selfmade IPS Package). The
service is the same in both versions. When the service disabled itself,
that is, at the time of installing the new version of OOo the service is
disabled, then the service will not be restarted.
Everything happens on the same BE. So could it be that this the normal
behavior independent of what BE is used?
Let me know if you have more questions.
I will :)
Thanks,
Joachim
Antonello
Best Regards,
Joachim
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
