• Bug
  • Status: Closed
  • Resolution: Fixed
  • drb
  • Reporter: sourceforgetracker
  • September 21, 2009
  • 0
  • Watchers: 0
  • September 22, 2009
  • September 22, 2009


In RMIBootstrapCacheLoader the list keys of all keys is fetched from a random CachePeer. Later the Elements belonging to those keys are fetched in chunks. These elements could meanwhile have expired at the CachePeer. If this is the case a null element is in the received chunk. Since it is not allowed to put null elements in the cache these elements must be ignored.

See attached patch which also helps to generate a backtrace in case of an exception during bootstrap. The discovered cache peers are also logged in debug level.

Shouldn’t the log level of errors during bottstrap be increaed?

Sourceforge Ticket ID: 1728950 - Opened By: fleiter - 31 May 2007 14:02 UTC


Sourceforge Tracker 2009-09-21

Logged In: YES user_id=693320 Originator: NO


Thanks for the patch. It is in trunk and will be in ehcache-1.3.0.

I have increased the bootstrap error to warn. The reason it is not higher is that I have been sort of thinking of bootstrapping as a performance optimisation. If it fails there is no harm done. It is also a relatively new and somewhat experimental feature which is not widely used as of yet compared with the other parts of replication.

Thanks for both of your patches.

Comment by: gregluck - 2 Jun 2007 01:08 UTC

Fiona OShea 2009-09-22

Re-opening so that I can properly close out these issues and have correct Resolution status in Jira