We are running IFS 10.
This is related another topic , since issues more cleared I’d like to create a new topic to get some helps.
After tested with several scenarios, we found when processing is executed for 4 minutes or more, client (IEE) will lost response.
It is only happens on clustered application server and database server environment. We use Azure Application Gateway to load balance accesses on Application Server, and use Azure Load Balancer to provide a stand-by database service.
Environment and idle time setting:
- There is a timeout setting on the Azure Application Gateway (Application Server side), and currently it is set to 2 hours on the AGW. Since 7200> IFS setting value 5400, it is recognized that there is no problem.
- Azure Load Balancer (DB server side) TCP idle timeout is 10 minutes.
Details of Issue:
1st: Application server lost response from database when using Azure Load Balancer when processing is executed for 4 minutes or more. Even if we have set Load Balancer TCP Reset and Idle Timeout as 10 minutes
→ Here we use oracle keep-alive (The DB server (Oracle) sets the cycle for sending alive status use SQLNET.EXPIRE_TIME)
2nd: After 1st work around applied to environment, the Application Server get response, and post to AGW. But Client lost response from AGW.
We are looking for
- The reasons for the 4 minutes timeout
- Work around solutions: Is there any parameter setting on IFS IEE or Azure Application Gateway to keep the communication between IFS IEE and AP Server?
We would appreciate it if someone could give us some advices!
