Hi , We are currently facing problems with our server 2008 r2 where we have an application installed, the application service core listens to port 2910, during non peak times the application works ok and we only have problems once the peak begins... there are currently approx 1000 servers connecting to the application core on port 2910, this is the behavior we see after analysing with wireshark- this is just a sample of it.. so in between we also see good connections , We have checked the application logs and reviewed memory dumps where we are unable to see that the request is reaching the applicaton.. obvious.. so now we need to find out if windows 2008 R2 has any features or settings that may cause this behaviour we have considered that this may be the possible cause http://msdn.microsoft.com/en-us/library/aa560429%28BTS.10%29.aspx http://social.msdn.microsoft.com/Forums/en-US/winserver2008appcompatabilityandcertification/thread/36fd41b5-f57a-4401-af9b-328760ac5d38/ But we need someone from microsoft to shed some light if this behaviour is infact possible with the windows server 2008 r2 Thank you

Hi Hyder, Thanks for posting here. No sure the root cause form current symptom but have you got any warring or error form event viewer ? Please also take look the two registry keys that will control the connection number in the thread below: http://social.technet.microsoft.com/Forums/en-us/winservergen/thread/fb7a6ef7-5a70-43a4-b0fe-c0252877467b Thanks. Tiger LiTiger Li TechNet Community Support

Hi Hyder, Thanks for posting here. No sure the root cause form current symptom but have you got any warring or error form event viewer ? Please also take look the two registry keys that will control the connection number in the thread below: http://social.technet.microsoft.com/Forums/en-us/winservergen/thread/fb7a6ef7-5a70-43a4-b0fe-c0252877467b Thanks. Tiger Li Tiger Li TechNet Community Support Also during the peak time have you checked if the application is still listening on the required port by running netstat -ano. How many connections do you see in connected , time_wait and close_wait state.

Thank you both for the replies, we have already done the following TCPTimeWaitDelay is 30 MaxUserPort is 65534 MaxFreeTcbs is 65534 MaxHashTableSize is 65536 this did not make any difference... also we have been actively moniroting all the port usage.. the connection established can shoot up to 12000 ports this is where we see the problem (cannot telnet to 2910) once the connections are dropped to under 1000 the system becomes responsive(can telnet to 2910) Thanks

Hi Hyder, Are you still experiencing this issue. If yes then please let us know and we can look further. Also it would be really good you could let us know about basic server config e.g. CPU, RAM, etc.

Thank you, this is now resolved, i will post the resolution later

Thank you, this is now resolved, i will post the resolution later Thanks for letting us know. Please mark this issue as resolved and post the resolution as per your convenience.