[
http://jira.jboss.com/jira/browse/JBESB-1590?page=all ]
Tom Fennelly closed JBESB-1590.
-------------------------------
Resolution: Done
Hardened the code around renaming/moving the input file to working such that the listener
will not attempt to process the working file unless it successfully acquired ownership of
it via the rename/move. Prior to this, the "setToWorking" code could return a
working file after failing to do the rename.
Review File lifecycle management in the FileGateway
---------------------------------------------------
Key: JBESB-1590
URL:
http://jira.jboss.com/jira/browse/JBESB-1590
Project: JBoss ESB
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Transports
Affects Versions: 4.2.1 CP1
Reporter: Tom Fennelly
Assigned To: Tom Fennelly
Fix For: 4.2.1 CP2
Task JBESB-1568 identified an issue thrown up by the scheduler code, but in doing so also
highlighted a secondary issue in the way that the gateway itself manages the lifecycles of
the files it is processing. Specifically... when a file "disappears" while that
instance of the gateway is processing it (or about to process it).
This could happen through a single FileGateway config before JBESB-1582 was fixed, but
can still happen when multiple gateways are configured to process the same fileset in the
same directory.
The gateway should handle such contention issue more cleanly.
Should it support multiple gateway instances processing a single fileset in a single
directory? This could be a head wrecker to get right *and prove right* on all platforms!
We could just document this as something that's not supported. It should def be able
to handle different filesets in the same dir using multiple gateway instances!
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira