Are there any fundamental changes in v3.0.5? Especially in scheduling?
I have a project that works well in v2.5. If GPE goes down, gadmin stop gpe and gadmin start gpe will always bring it back.
However, in v3.0.5, If GPE goes down, a stop/start brings it back momentarily (1~2 seconds) and then it would go down again. It takes lots of stop/start to bring it back eventually. Is there a process that keeps submitting a query to GPE if the query crashes GPE? If yes, can I disable that?