Highlighted
Absent Member.
Absent Member.
5583 views

HP Service Manager - SOAP Fault Occured Session no longer valid

Dear Friends,

 

I am getting SOAP Fault Occured Session no longer Valid - Even when I am using Service Manager and it kicks me out of it.

 

Just trying to explain the setup.

 

Service Manager 7.11

Vertical Servlet design on Unix Solaris 10 Server.

Total 11 Servlets and each has 50 thread per process.

 

For load balancing we are using Cisco Ace Hardware. Here the vip accepts connection on port 80 and routes the traffic to 11 different ports to the server.

 

in Service Manager we have 45 Minutes inactive timeout value. In Cisco Load balancer we have 4 hours persistent value which remembers the user connection for each servlet.

 

I dont find any details in sm.log file as it just says session is no longer valid ..sending soap fault.

 

After 45 minutes when user gets disconnectted as per inactive timeout, he is not able to login again. He has to close the window and reopen it again.

Sometimes gets this error in the middle of using it.

0 Likes
9 Replies
Highlighted
Absent Member.
Absent Member.

Request someone to provide some idea's here to start toubleshooting the issue.

I could not find any possible config error in network equipment or application.

0 Likes
Highlighted
Absent Member.
Absent Member.

Here are some options:

 - test the Windows client and compare its stability to Web clients

 - if Apache and Tomcat are used, try connecting to Tomcat directly

 - bypass/disable the Cisco Load Balancer

 - replace the Cisco Load Balancer with the Service Manager Load Balancer

 - setup a debug serlvet with tracing enabled to capture the issue

 - apply the latest RTE patches to SM 7.11

 - reduce the number of threads per process

 - increase the physical RAM on the server

 - disable the inactivity process and compare the error frequency

 

Highlighted
Absent Member.
Absent Member.

Hi,

 

Thanks a lot for this response. Putting my thoughts again. May be I have to put points properly across to my stake holders to get the approval

 

- test the Windows client and compare its stability to Web clients.

   For web client the number of users are very less. It is also customized to log and check the status of 

   the ticket. No one can work on the tickets or update it. Web client does not have this issue.

 - if Apache and Tomcat are used, try connecting to Tomcat directly.

   Not sure how to do this, Can you provide some more details please.

 - bypass/disable the Cisco Load Balancer

   This would be complete architecture change and I have to justify the reason of change to stake holders. We

   are getting  SM9 saas model by end of 2012 which makes the approval process difficult.

- replace the Cisco Load Balancer with the Service Manager Load Balancer

   Bypassing the cisco load balancer by giving the direct IP and servlet port numbers, users are not facing any

   problem at all.But this will increase the load on server and users will start getting 50 - Max 

   user error if already 50 users are connected to one servlet.

- setup a debug serlvet with tracing enabled to capture the issue.

  I have done this but message just shows.

  2011/11/17 11:29:25 EST [DEBUG] SOAP.transact started
2011/11/17 11:29:25 EST [DEBUG] |request:
<stop />
2011/11/17 11:29:25 EST [ERROR] SOAP Failure - Session no longer valid
2011/11/17 11:29:25 EST [DEBUG] |response:
<Fault>
 <faultcode>SOAP-ENV:Server</faultcode>
 <faultstring>Session no longer valid</faultstring>
 <faultactor>Server</faultactor>
</Fault>
2011/11/17 11:29:25 EST [DEBUG] |SOAP.transact finished in 32 ms

 - apply the latest RTE patches to SM 7.11

   Will try this.

- reduce the number of threads per process

   Will try this.

 - increase the physical RAM on the server

   Will try this.

 - disable the inactivity process and compare the error frequency.

   We have 525 floating user license and 11 servlets(50 per thread). If we dont use this it will not clear the

   inactive users and same 50 max error will at one point of time. Here People always forget to signout from the Application.

 

 

0 Likes
Highlighted
Absent Member.
Absent Member.

Just to add to the above note.

Web client users are not facing any problem. Number of Connections are less and very few activities.
I think Apache and Tomcat suggested above are related that.
0 Likes
Highlighted
Absent Member.
Absent Member.

Yes, the suggestions were based on the assumption that Web clients were impacted (not Windows clients).

 

Please note the suggestions were intended to be tested against a DEV or TEST system (not PROD). Is it not possible to bypass the hardware load balancer just for a few users? For example, what if the Windows client was running on the server?

 

If you can replicate the issue on demand, then consider engaging HP Support as well.

0 Likes
Highlighted
Absent Member.
Absent Member.

Hi,

 

I have given 13 users direct access to server bypassing the load balancer in DEV and PROD.

The users are not getting any error at all in PROD. in DEV anyways we were not getting any error before.

 

Web clients are not impacted at all.

 

I have already tried to log a ticket with HP and provided all the logs. But did not get any solid info still now.

0 Likes
Highlighted
Absent Member.
Absent Member.

Replacing the Cisco Load Balancer (hardware) with the Service Manager Load Balancer (software) does not require providing direct access to every servlet. Users that connect to the SM loadBalancer are automatically redirected to available servlets.

 

If users are not getting the error when bypassing the hardware load balancer, then consider engaging the vendor (Cisco) and your network team. It might be useful to research how the traffic/connections are different for users that pass-thru the Cisco ACE versus users that bypass it.

0 Likes
Highlighted
Absent Member.
Absent Member.

If it is of any help, I have had exactly the same error on my standalone PowerEdge server running 7.11 in Windows Client with just sysadmin logged in. The error occurs whenever I am in a small CI Relationship Graph and open up a boxed container of 30 servers connected to a switch , the above SOAP error is generated and I ultimately have to log back in again as connectivity to the session is lost..

0 Likes
Highlighted
Absent Member.
Absent Member.

I am facing this problem too. Everytime i tried to connect the user  ii should connect twice. The first time i will met this prolem and the next time i can usually connect it after restart the services.

 

I still do not what is  wrong with it. Hope someone can come and helps to solve the problem.

 

Thanks.

0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.