Frequnet Warning message in IS logs

Hi Guys,

Offlate i happened to see the below warning message getting logged into IS server log frequently… I can see it’s a warning message and it successfully reconnect but my concern is whenver i see this log message I observed IS is losing cluster as well… which give me indication of some serious trouble… i saw the number of DB session’s its pretty much within the limits we have set in our enviornment… fyi… i can set 20 max connection for each pool and at DB end i can see the session are less than this number… so was wondring if you can point me to some direction…

thanks in advance!

I am running on IS 8.0.1 with core Fix 7 and jdbc pool fix 1 (SCG)

2010-11-17 15:24:59 EST [CommonLib.MWS.0002W] DB Server disconnected. Will try to reconnect
[SoftwareAG][SQLServer JDBC Driver]Object has been closed.java.sql.SQLException: [SoftwareAG][SQLServer JDBC Driver]Object has been closed.
at com.wm.dd.jdbc.sqlserverbase.BaseExceptions40.createAppropriateSQLExceptionInstance(Unknown Source)
at com.wm.dd.jdbc.sqlserverbase.BaseExceptions40.createSQLException(Unknown Source)
at com.wm.dd.jdbc.sqlserverbase.BaseExceptions.createException(Unknown Source)
at com.wm.dd.jdbc.sqlserverbase.BaseExceptions.getException(Unknown Source)
at com.wm.dd.jdbc.sqlserverbase.BaseConnection.createStatement(Unknown Source)
at com.wm.dd.jdbc.sqlserverbase.BaseConnection.createStatement(Unknown Source)
at com.webMethods.portal.service.sql.core.provider.pool.PoolableConnection.createStatement(PoolableConnection.java:57)
at com.webMethods.portal.service.sql.core.BaseSqlWorker.executeQuery(BaseSqlWorker.java:531)
at com.webMethods.portal.service.sql.runtime.base.CacheGetSysTime.executeQuery(CacheGetSysTime.java:32)
at com.webMethods.portal.service.sql.core.BaseSqlWorker.executeQueryProcOut(BaseSqlWorker.java:734)
at com.webMethods.portal.service.cache.impl.CacheInvalThread.getCheckTime(CacheInvalThread.java:141)
at com.webMethods.portal.service.cache.impl.CacheInvalThread.run(CacheInvalThread.java:71)

What is the Expire timeout setting? This should be slightly less than the time that the DB will wait before dropping an idle connection.

Also, what is the min connection setting? It most likely should be 0.

thanks for replying Rob!

yep the min connection is set to zero and expire timeout is 6000 ms and its in DB i think timeout is set as 10 s… so we are okay on that front as well…

That is a really, really short idle timeout. I’d be surprised if the idle timeout on the DB side is just 10 seconds. Usually it’s much longer–often as high as many hours. Short idle timeouts can result in session thrashing on the DB.

You’ll want to evaluate the block timeout as well–are threads being allowed enough time to wait for connections from the pool to become available?

When IS is moderately busy a DB connection might go idle (6 seconds isn’t very long) and get dropped. When a thread needs a connection at the 7th second it will need to wait for a connection to be established–which is usually not very speedy. By the time the connection is available, the thread might exceed its block timeout and give up.

I think you want to bump up your timeouts in coordination with the DBA.

Hello,

Actually my question is bit related to it…
I am too getting messages, which fills out the server log, those messages are of JDBCLocks and LockFactory, something like that…
How can I detect, from where they are arising, and what is the solution for it.

Regards.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.