Hi, Good news about this leak of memory on my server.The path send by Hans G Knudsen (JIRA SANDESHA2-93) apply to the sandesha 1.2 TAG version and used with Axis2 1.2 seems corrected my problem.
Memory on my server is stable after 220 GC executions
Thanks Jérôme Jerome Camilleri wrote:
Hi,My application use Axis2 1.2 to send WebService with MTOM attachement (MEP IN-OUT with synchronous call). I had tested my application without fiability and the memory used by catalina during 12 hours seems constant.But when I engaged Sandesha2 1.2, the memory used by tomcat grows regulary. If I use some profiler I see that the class org.apache.sandesha2.storage.inmemory.InMemoryStorageManager.getTransaction [106] used an HashMap to save the transaction without (it seems) release it... (number of byte[] instances growth) ?Because I find into the Sandesha code that the traitement to remove this transaction is different if we used 1.0 or 1.1 RM SPEC I try with this two client configurations but without improvement ;(Someone have an idea about this problem ? (see my garbage collector activity on the catalina server memory : memory used initially 21748K; after 237 sequences of 3 messages : 74395K)Regards. Jérôme ------------------------------------------------------------------------ --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
<<inline: GC.jpg>>
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
