Hi,
I understood that you want to get a mail from the cluster when one node is dead, but not from the notifyer, because this mail is not very useful. You want to make your own mechanism.
Ok, the cluster itself has no trigger you can fire in start sequence. But there is the postonline trigger of service groups which you can use for those things. Initially each service group fires this trigger after going online on a system so that you can get the start information of each group per mail outside the cluster if you send this mal.
Another nice thing happens when one node leaves the cluster (for whatever reason). One of the surviving nodes fires the sysoffline trigger. But I'm not able to say if you can use this thing to do things, I have no systems available this days (but will very soon :)
A very strange and crude solution would be in configuring a failover service group for each node (with only a phantom in so that this group is not switching because of ressource failures, only when the node is dead) and setting an preonline trigger that checks if the actual node this group is running on is the first in SystemList, and if not it sends you a mail ...
Anyway, you can use the postonline trigger with every service group to send you a mail when the node on which the group is starting is not the first in SystemList. Then it is not the preferred node (if you use failover policy Priority) and the service group probably has faulted. Just an idea ...
Roger
