1 Reply Latest reply on Oct 23, 2012 7:46 AM by dyjbc

    expiration, lifespan, and wakeupinterval

    dyjbc Newbie

      Hello, I've got infinispan 5.1.5 FINAL.

       

      I've got replication mode cluster.  I also set the lifespan for each cache entry during the put operation.  It seems to be working correctly.  After the expiration, in JMX I can see the numberofentries decrease appropriately.

       

      I did not configure eviction or the wakeupinterval so I'm wondering what is enabling the purging of the expired cache entries.  Can you tell me what is responsible for doing that?  I thought that entries would only be removed from cache if the wakeup thread was configured to run.  I tried an expirement and configured wakeupinterval to see if there would be a difference but there was not.  Also the "eviction" attribute always remained 0 in both test cases.  I would expect it to increase after an entry expired and the wakeup purge thread runs.  Thanks for any info.

        • 1. Re: expiration, lifespan, and wakeupinterval
          dyjbc Newbie

          After turning on trace, I'm starting to understand what is going on although it doesn't seem to be documented.  I noticed that a purge was happening every 60 seconds.  Note I do not have eviction or expiration enabled in my config.  However if I view the config via jmx, I can see this:

           

          <expiration wakeUpInterval="60000" maxIdle="-1" lifespan="-1" reaperEnabled="true"/>

           

          meaning it was autogenerated.  Can someone clarify this?  I'd like to see that added to the doc.  I know the doc on the subject of expiration/eviction is a bit messy right now due to recent changes (like moving wakeupinterval to expiration).  Also I'd like to get edit page access on the wiki so I can help.  Thanks