Query on MID Server

[Mathews1]

Hi All,

I have an issue with the MID Server. When I check the MID Server from the Service Now GUI, sometimes it shows as Down. We are having the MID Server in Windows Environment and I was able to RDP the server without any issues and server seems to be up.

Checked for the below and all seems to be running.

1) snc_agent service

2)       java.exe process

3)       wrapper.exe process

Can you advise me why the MID Server status is displayed as Down in Service Now even though the server is up and running fine ?

Do we have any other process related to discovery apart from the ones mentioned above.

Thanks,

Mathews

[Josh Cooper]

Just throwing a possibility out there - whenever the MID server shows as down, are you able to access the internet through your RDP Session?   Even if the server is up, it will appear down to ServiceNow if it's not able to check in every so often.

A more remote possibility, check to see which user is running the MID Server service - I have seen an issue before with Windows services (not the MID Server) where the service user was forced to logout when the RDP session ended, particularly if running as system and logging in on terminal 0.

[Mathews1]

Hi Joshua,

Checked in ServiceNow ECC queue and found that there is no "Input" Queue existing for HeartbeatProbe for that particular MIDServer whereas both Input and Output queues are existing for other MID Servers.

Do you have any idea on why Input queue is missing for HeartbeatProbe

Thanks,

Mathews

[StephenHey]

Has it ever connected before?   If it's been connected at one time, but not now, I would try the following things for troubleshooting:

• Look at the related list "Threads" in the MID Server form.   There should be around 20 something.   If there are hundreds then the MID server is too busy processing threads or you have some stuck threads and cannot send the heartbeat in time.
• Look in the wrapper log and the agent logs in the log directory for any mention of socket issues (indicative of a loss in network connectivity).
• Bounce the service on the Windows Server.   Does it reconnect?   If it reconnects, then it could be one of a few things.
  • Network connectivity was lost, but regained from bouncing
  • Stuck threads stopped heartbeat thread from being processed, but threads are cleared upon bounce
• Validate the user account running the service has the correct permissions to execute commands on the Windows box

Depending on how your routes are set up and how many NICs you have on your Windows server, you may lose connectivity on the NIC that is allowed to go to service-now.com, but your other NIC takes over and tries to route, but might get denied by your outbound firewall/proxy.   This can only be resolved by bouncing the server, correcting your routes, or allowing your other NIC access to the internet.

[garyopela]

Min mid-server service process crashes frequently. The windows service shows as up and everything, the mid-server in SN shows down, and the discovery statuses are not running. I restart the service and it usually resumes. Additionally, sometimes I have to reboot the server itself. Whenever it is hosed up like this, even though nothing is working, looking through task manager, you can see the processor on the server is pegged.

I attribute this to my running an old version for my discovery instance. I know under aspen this happened a lot. I upgraded it to berlin and it is doing much better, but i really need to upgrade to calgary or dublin, i'm guessing it would probably be even more stable.

What version are you running?