Our Glassfish Application Server has suddenly started crashing this week.
The Glassfish java process suddenly dies (although the wrapper service is still running). The system and application logs are not very informative. The crashes keep re-occurring even after appserver restarts or system reboots.
Environment:
Glassfish, Windows 2003 Server, Microsoft SqlServer 2005.
- This system is under strict change control and no known code changes or IT related system patches have been made to the system lately.
- One of our database tables is beginning to grow relatively large (50million rows). Users in past weeks have reported significant slowdown with some database queries.
We can't tinker with the production system, but some ideas we are thinking of pursuing are.
1) figure out way to increase size of database connection pool. (explore coyote/grizzly alternatives)?
2) We don't have a DBA, so the SQL-Server database does not go through any type of regular maintenance. Worthwhile to re-index, recalc stats or something of that sort? Would a significant slowdown in queries be sufficient to crash glassfish?
3) Figure out way to increase debugging/logging. What would we change? Change one of the logging.properties to FINE? which one?
4) Explore the ideas about a possible Windows Server TCP leak presented here:
http://www.ryandelaplante.com/rdelaplante/entry/glassfish_lockups_were_microsoft_s
Any recommendations on suggested next steps to debug would be greatly, greatly appreciated.
[Message sent by forum member 'dlca' (dlca)]
http://forums.java.net/jive/thread.jspa?messageID=340388