I have setup a windows service and I am using the comet style notifications to receive messages. I have noticed over the last week and a half, that each long lived get request is now not waiting until the 2 minute timeout but comes back with an error message in the web response - "Unable to connect to the remote server". This behaviour is intermittent and it means that service may take take a long time to respond and you don't get the desired instant responses.
Something is wrong with the comet style notification strategy, can somebody please advise?
I've noticed this same problem happening again since 18:00 GMT yesterday and it's happening at the moment
Let me try and explain more clearly -
I have a windows service which creates a comet style get request to listen for messages from yahoo buddies. It sits there until the default 2 minutes is over and if no messages have been sent and then I raise another comet style get request, as advised in the yahoo documentation.
I have not needed to set the idle parameter. I simply wait for the default timeout to occur and when this happens and I create a new comet style get request.
I have set the timeout for the request to the yahoo api server to be 130 seconds, so that I give the yahoo server a chance to wait in idle state for 120 seconds or 2 minutes.
Now the same problem is ocurring as before. when I raise a comet style get request, I am getting back responses like "Unable to connect to the remote server" or "The underlying connection was closed, Unexpected error occurred on a receive". These responses seem to be coming back every 20 seconds.
The solution does work most of the time for many days without problems, but there have been sustained periods where I see this behaviour and it's happening now.
If you need more information please let me know. Any ideas?
[Fiddler] Connection to rproxy2.messenger.yahooapis.com failed.<BR>Exception Text: A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond 188.8.131.52:80
What is VIP/LB, could you give a little more detail about this? Also are you saying the issue is fixed, has more capacity been added? Is this related to the load on the notification server? I don't think using port 5050 will be an option for us. Any information including dates when the capacity issue was fixed, will be greatly appreciated?
The capacity issue was fixed about 2 weeks back. VIP/LB is a hardware load balancer sits before the real commet server. The issue we found is not on notification server, it is purely capacity configuration at load balancer.