hi,
We are experiencing a peculiar problem, where TN goes unresponsive. Our partners are unable to send us documents, cant log on to TN console, can log on to IS UI though, cant reload TN.
After going thru activity logs we found that many threads were stuck on calls to TN services like tn.recognize, tn.log and so on. Has anybody encountered any problems with TN database connections hanging/starvation ? We are using Sequelink Server 5.1 and SQL Server database.
-rg
Does activity resume after a few minutes? Say about 10 minutes?
Contact wM Tech Support. We encountered this issue. wM was unable to recreate the problem. We switched JDBC drivers to CJDBC 3.2 and tweaked some pool and timeout settings. This seems to have eliminated the stall.
No the activity doesnt resume, it just remains hung.
CJDBC, u mean connectjdbc. Also for TN in its config/properties there are no pool/timeout settings. This pool is internal to TN, so is it configurable via some properties file.
We were having a problem with SQL Server and TN several months ago.
In our case attempts to insert docs into the TN DB were failing. A look at the SQL Server event log indicated that SQL Server was getting “Insufficient memory” errors. Those errors had something to do with how “SQL Profiler” was configured. Anyway, what the SQLServer DBA observed was that there was a very large number connections between wM and SQL Server and, apparently, no more connections could be obtained. It seemed like TN was not reacting nicely to the sql failures and connections were being kept open.
Once the SQL Server people made a configuration change to how SQL Profiler was running all problems disappeared.
Please review the SQL Server event logs and make sure that it is not having a problem.
Yes, CJDBC = Connect for JDBC.
In our case, the DB was Oracle. Activity would resume after just under 10 minutes of stall time, which smelled suspiciously like a timeout.
TN has a properties.cnf file that contains various settings. Check with wM support about the settings you can add/tweak. We increased the connection pool (30 to 45) and decreased the timeout (15 mins to 3 mins).