wingo - in flumotion/trunk: . flumotion/job

[flumotion-commit at lists.fluendo.com]

Author: wingo
Date: Thu Jan 11 13:52:44 2007
New Revision: 4386

Modified:
   flumotion/trunk/ChangeLog
   flumotion/trunk/flumotion/job/job.py
Log:
2007-01-11  Andy Wingo  <wingo at pobox.com>

	* flumotion/job/job.py (JobMedium.shutdownHandler): Wait for both
	calls to complete (either with success or failure), as they are
	independent, i.e. a failure on one does not mean the other will
	fail or fail to complete. Fixes components showing up as lost
	occasionally when a worker is control-C'd.



Modified: flumotion/trunk/ChangeLog
==============================================================================
--- flumotion/trunk/ChangeLog	(original)
+++ flumotion/trunk/ChangeLog	Thu Jan 11 13:52:44 2007
@@ -1,5 +1,11 @@
 2007-01-11  Andy Wingo  <wingo at pobox.com>
 
+	* flumotion/job/job.py (JobMedium.shutdownHandler): Wait for both
+	calls to complete (either with success or failure), as they are
+	independent, i.e. a failure on one does not mean the other will
+	fail or fail to complete. Fixes components showing up as lost
+	occasionally when a worker is control-C'd.
+
 	* flumotion/manager/manager.py
 	(Vishnu._getComponentState.verifyExistingComponentState): If a
 	component logs in with stale configuration, update our

Modified: flumotion/trunk/flumotion/job/job.py
==============================================================================
--- flumotion/trunk/flumotion/job/job.py	(original)
+++ flumotion/trunk/flumotion/job/job.py	Thu Jan 11 13:52:44 2007
@@ -154,7 +154,7 @@
                 dlist.append(medium.callRemote("cleanShutdown"))
 
         # might call back immediately if we aren't connected to anything
-        return defer.DeferredList(dlist, fireOnOneErrback=True)
+        return defer.DeferredList(dlist, fireOnOneErrback=False)
 
     ### our methods
     def shutdown(self):