Bugzilla – Bug 5986
New fault type for job resource expiration
Last modified: 2008-04-17 15:45:38
You need to
before you can comment on or make changes to this bug.
If a job resource lifetime expires, then a failed job notification is sent.
Currently, a client (like globusrun-ws) does not know from the fault type that
for this failed job, the job resources has already been destroyed. If a client
tries to destroy the job, they will get a resource not found exception.
With a new "jobResourceExpired" fault type, a client would be able to know that
they should not attempt to terminate the job resource.
Added. If a resource expires a failed notification is sent after
cleanup that contains this fault type
The type is JobResourceExpiredFaultType and it's an extension of
the existing type FaultType. globusrun-ws is probably capable of
handling it in the fault element of the notification message.
Might be good if globusrun-ws does not explicitly call terminate()
on such a resource