We have a system configured as a vertical cluster (2 jboss instances on a single server using gemfire). In late December node 1 stopped processing files and seemed to interfere with node 2 processing: if node 2 was processing files and node 1 was started, the processing on node 2 would stop, and then when node 1 was stopped the processing on node 2 would resume. No node 2 is dead as well. Just emits messages in the log like:
2013-01-25 14:41:42,582 info [com.adobe.idp.dsc.workmanager.workload.statistic.MemoryStatistic] Memory statistic is 73.0
No amount of restarting, clearing out the stage folders, clearing the LiveCycle temp area, redeploying jboss components seems to unjam the system. Adobe support has been unhelpful to date. Willing to entertain suggestions from the field.
LiveCycle version is 18.104.22.168GM, Patch Version 9.0.0.COR-2.234, Service Pack Version 22.214.171.124.20120813-1-320288
PDF Generator ES2 v 126.96.36.199 patch 9.0.0-BCF-2.203
Barcoded Forms ES2 v 188.8.131.52 patch 9.0.0-BCF-2.203
Reader extensions ES2 v 184.108.40.206 SP2
Output ES2 v 220.127.116.11 patch 9.0.0-OPT-2.205
Windows serv er 2008 sp2 enterprise edition 64 bit for sqlserver db server, 32 bit for application server.
run.bat sets the following java heap args: -XX:PermSize=512m -XX:MaxPermSize=512m -Xms4096 -Xmx4096 -XX:+UseParallelGC -XX:+UseParallelOldGC -XX:ParallelGCThreads=4 -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=D:/logfiles/node1 -XX:-PrintGCDetails -XX:-PrintGCTimeStamps -verbose:gc -Xloggc:D:/logfiles/node1/gc.log -XX:NewRatio=4 -XX:MaxTenuringThreadshold=25
Thanks in advance,
For what it's worth, and if I recall, this was resolved by stopping JBoss, and clearing out the Adobe_LiveCycle_Temp area, and the old cruft left by processes that died under LiveCycleStorage (assorted folders with names beginning "docm"), removing the svc*, tmp and work folders under opt/jboss_node/server/lc_sqlserver_cl/, and then restarting the JBoss node. Your mileage may vary.