I am running an application where in after receiving the response document from the other end i can see the service invoked by the trigger is stuck in the server.
Along with the subscribing service, an adapter service which is called inside it is stuck and there is no further processing happening.
This is a scheduled application and running in an clustered environment.
We have done several restarts of the server for the issue… But still this is happening very frequently at least twice or thrice a week.
Rafi --Is this a new environment ? which version of wM you are in ? Is the interface a fresh one ? Did you properly configure scheduler ? Did you configure the service to get kick-off once the previous service gets completed ? Check things properly.
to me it looks like server is waiting for adapter service to complete.
what kind of adapter service it is? JDBC/SAP/MQ etc.
If it is JDBC, can you see what value you have set for Query Timeout. if it is -1 that means infinite. can you change that value and see how server behaves.
before that can you see if there is any error in server logs? also is there any error at backend application server.
do reply with this additional information and we can try to debug issue.
Its wM 6.5 and its a JDBC Adapter which is stuck in the server… time out is not set for the Adapter Service and there was nothing in the server error logs.
Since this application is an old one and running since Quite a few years, this sudden behavior is lil bit confusing.
OK, I guess it must be a prod environment, just check the below items (ignore if you have already saw them)
1> Check the JDBC adapter fix levels and IS fix levels are same on both the node
2> Check the JDBC adpater settings from IS are same on both the node
3> Disable the JDBC adapter connection on the affected node or on both the node (make sure this should not impact your BAU)
3> Manually check the adapter services via developer on both the node to see if they are identical
4> Last option, to re-deploy the package containing the affected adapter service from the working node and check if its working. Make sure you do this without impacting the BAU.
At least the connection should have an inactivity timeout, meaning the service in doubt is no longer transferring data for a specified time (see connection pooling settings for details (can you share them?)), the connection will be closed by either database, database connection manager or transport layer (whatever occurs first).
If this happens the service should fail subsequently after next try to access the connection which no longer exists.
Make sure you have set min pool size to 0 which avoids keeping invalid connections in the pool.
If nothing helps try to shutdown and restart the IS.
Rafi – It will be good if you update us on this thread. I hope you fixed this. If not kindly share the current status, I am very interested this sort of issues.
