http worker threads periodically hang causing MII to become unavailable.

[Former Member]

We see periodic instances where our MII instance begins to quckly degrade in performance and within the hour causes a "503 Service Unavailable" error. When this begins to occur we can see one (or several) http worker threads that have been running for a while (10-15 min). This seems to propagate to the other threads and eventually they all hang, which is when the error occurs.

In looking at the log, there are no errors showing for the user that has the hung thread.

Also, there are no deadlocks occuring in on our SQL database.

Lately, this is occuring once a week.

Does anyone know what is causing this and if we may be missing a parameter setting in our NW instance?

MII 12.1.9 (Build 116)

NW CE 7.11 SP10.

Thanks,

Bob Koshalek

[Parthis]

Bob, Did you figure out exact solution for this issue? Thanks!

[former_member204240]

Hi Bob,

Please check below threads may help you

http://scn.sap.com/thread/1407733

Regards,

Padma Rao

[former_member185280]

When we moved to 14/7.3 we had some issues with zombie threads that seemed to occur when a node was reaching its memory limit and there was some CG panic. Giving the nodes more memory seems to have alleviated the issue so far.