0

I have a 2 node IIS cluster using Microsoft Network Load Balancing under Windows Server 2019 hosting multiple websites/application pools. The cluster ran without problems for a year. However, beginning a few months ago, around once a month, one of the two web servers will stop responding to all http/https requests. The NLB doesn't detect the server as down so half the requests to the website are failing during this time. When this happens, IIS manager locks up and an iisreset at the command line responds with "restart attempt failed". During this time, the HTTP Error logs are filled with 503 response codes/QueueFill messages. There are no specific events in the Windows System/Application logs indicating an issue. Performance on the server(s) is not at all an issue. A reboot of the troubled server makes the issue go away.

What can I do to track down the cause of this issue?

kittyhawk
  • 171
  • 3
  • Two issues are there, 1) Network Load Balancing didn't detect the unhealthy server 2) HTTP Error log filled with 503 response codes/QueueFill messages. Both require further data to be collected from your environment, so you'd better open a support case via https://support.microsoft.com – Lex Li Aug 05 '22 at 16:48
  • From what I understand, Microsoft NLB does not detect 503 errors. https://serverfault.com/questions/167281/windows-nlb-detect-dead-iis. Additionally, I have seen you make multiple recommendations to open a support case with Microsoft which is fine but that seems odd to do on StackExchange unless you feel there is a bug in the OS or in the Microsoft environment. I was hoping to get assistance from people on here and not jumping to Microsoft right away. – kittyhawk Aug 05 '22 at 17:47
  • While you have seen me recommending Microsoft support, you probably also see that I answered tons of IIS related questions on StackExchange sites when clearly those can be discussed without touching the machines. Anyway it's always your call to decide what to do next. – Lex Li Aug 06 '22 at 00:29

0 Answers0