Bug 6140 - taging response handlers destroy transfers if they are still active
: taging response handlers destroy transfers if they are still active
Status: RESOLVED FIXED
: GRAM
wsrf managed execution job service
: alpha
: Macintosh All
: P3 blocker
: 4.2
Assigned To:
:
:
:
:
  Show dependency treegraph
 
Reported: 2008-06-09 12:02 by
Modified: 2008-06-09 12:16 (History)


Attachments


Note

You need to log in before you can comment on or make changes to this bug.


Description From 2008-06-09 12:02:33
All handlers for staging responses (StageInResponseStateHandler,
StageOutResponseStateHander, FileCleanUpResponseStateHandler) don't
behave properly if the staging is not yet done: They destroy a transfer 
nonetheless. The problem shows up rarely, because normally a job resource
is only queued in the state processing executor in a response state after
a staging is done. However, it happens in rare situations, and then the
transfer is destroyed even if it's not yet finished.

Solution:
If a job is processed in a staging response state, and the transfer is
not registered in the StagingListener, then the transfer is polled for
state. And only if the transfer is in a final state we go and destroy it.