OpenSAF is today relying on external resources (TIPC link down) to control when a controller failover should be done. When TIPC is controlled outside opensaf a failover will not be triggered when OpenSAF is successfully closed down (e.g. opensafd stop). OpenSAF should trigger a failover using some internal mechanism, not that the TIPC link is lost. CLM could e.g. be expanded with a CLMND that can be used to detect when a node is lost (CLMND is down).
Hi Bertil,
The use case of '/etc/init.d/opensafd stop without OS reboot cycle' was not a requirement for OpenSAF nor does it ideally fits into the standard SAF architecture involving PLM.
However, based on some discussions during 4.2 i had raised ticket #220 to be able to support some active users. I have accepted the enhancement ticket #220, and will contribute it in the 4.4 release.
Please refer/track https://sourceforge.net/p/opensaf/tickets/220 for this topic.
Closing this as a duplicate of the enhancement ticket #220.