I am using jgroups in a clustered application for distributed caching of frequently used objects. The application is deployed on JBoss, websphere and weblogic. I have noticed that the system occassionally runs out of memory and has full gc's after running for a couple of days. On the websphere cluster deployment I inspected the heap dumps generated. These dumps show that there are about 200,000 objects in org/groups/util/Queue$Element holding up about 1GB memory. It seems that jgroups are holding on to these objects and are not getting gc'ed.
Does anybody know if this is a known issue?
I suspect there is some message that is being sent repeatedly. Can we limit the number of retries with a configuration setting?
I am running this on
websphere ND 5.1.1.6 ( on Redhat Linux )
IBM jdk 1.4.2.
jgroup 2.2.9.1
min and max memory setting in the app server is 256m and 1536m
jgroup config properties
UDP(mcast_addr=237.0.0.101;mcast_port=7500;ip_ttl=32;mcast_send_buf_size=150000;mcast_recv_buf_size=80000):PING(timeout=2000;num_initial_members=3):MERGE2(min_interval=5000;max_interval=10000):FD_SOCK:VERIFY_SUSPECT(timeout=1500):pbcast.NAKACK(gc_lag=50;retransmit_timeout=300,600,1200,2400,4800):UNICAST(timeout=600,1200,2400):pbcast.STABLE(desired_avg_gossip=20000):FRAG(frag_size=4096;down_thread=false;up_thread=false):pbcast.GMS(join_timeout=5000;join_retry_timeout=2000;shun=false;print_local_addr=true)
Any help appreciated.
Thanks,
Srinivas
