Users get errors: "Sorry the server has been restarted..." or "Session Refused"

Document ID : KB000044684
Last Modified Date : 14/02/2018
Show Technical Document Details

Symptoms: 

While using CA Service Desk Manager, users get the following messages:

"Sorry the server has been restarted.  Press OK to close all CA Service Desk windows and return to the login form"

OR

"Session Refused"

 

Environment:  

CA Service Desk Manager 12.9, 14.1 and up

Conventional or Advanced Availability architecture with Load Balancer

 

Cause: 

These error messages are often caused by the same conditions.  There are several conditions which can cause these messages as follows:

Scenario 1:   If a load balancer is in use to distribute the user load among secondary servers or application servers, and session persistence (also known as "sticky sessions") is not enabled on the load balancer, when a user clicks on something in the CA Service Desk Manager user interface, the load balancer sends the request to a different secondary or application server, however that user's session is only valid on the server in which it was initiated, thus when the session hits the different server, its no longer valid and one of the error messages referenced above appears.

Scenario 2:  The webengine process which the user's session was connected to has failed and restarted for some reason, and as such, the user's session is no longer valid on that webengine, so newly started webengine will display one of the above messages.

 

Resolution:

For Scenario 1: Work with your netowork team to ensure that "Session Persistence" or "Sticky Sessions" is enabled and enforced.  This will keep all user sessions on the secondary server or application server in which those sessions were initiated on and will not attempt to bounce the user sessions between servers.

For Scenario 2:  If you run a pdm_status command on the server for which the user was connected to at the time (or all servers if you are unsure of which server the user was connected to at the time),  focus on the start date/time for the webengine (webengine.exe) process in the list.  If it appears that the date/time in which any webengine started is more recent than all of the other processes, then you can take a look at the stdlogs on that server to see if any messages appear during that time frame related to the webengine dying and restarting.  Hint: Search for the word "died" in the log file which covers that time frame.  If you see that the webengine died and restarted during that time frame, please review the following document and follow any necessary steps:  http://www.ca.com/us/support/ca-support-online/product-content/knowledgebase-articles/tec595780.aspx

 

Additional Information:

 

See Configuring the Load Balancer in the documentation wiki site.