Hi All,
For the past few days we are seeing a strange issue with WSUS Scan when clients
are trying to SCAN against the WSUS server, connections are dropping at WSUS
pool level and clients are resetting the connection . When we investigated the
issue, found that the connections are accumulating at worker process leave and
WSUS pool is unable to process them due to latency in processing the request.
To remediate this, as a workaround we have added the
MaxConcurrentRequestsPerCPU registry key and set the value to 5000 to increase
the Concurrent Requests. This is helping to process the requests but servers
are consuming the memory and server performance is degraded.
We have three servers in NLB and we are supporting approx 50K clients with
these three SUP servers.
We tried to remove one server from NLB thinking it was something to do with NLB
but that too did not help. We have tried to eliminate corruptions in IIS or
WSUS by uninstalling the IIS and WSUS on one server but still we are seeing
same issue (we have two servers behind NLB and both are behaving in same way).
Can someone give us some pointers where to look what might be the reason for
WSUS pool to take long time to process the request and how we can fix this
issue. Appreciate your help. Below is the connection drop error from http error
log . POST /SimpleAuthWebService/SimpleAuth.asmx Windows-Update-Agent - -
testlab.com - 0 880 59386 1 Connection_Dropped_List_Full WsusPool
Regards,
/SAM
