Hello,
We have some log.nsf entries that are getting logged at an incredible rate. The agent message we are getting can fire off more than 100 times in the same timestamp and they are firing off constantly. Our server CPU is also spiking at between 50-100% constantly. The HTTP service (nhttp.exe) is what is spiking.
The message is:
HTTP Server: Agent Message: xx
where xx is a number… always one of 2 numbers.
Does anyone know what I can do to try and track down what specific agent that is throwing this message? More verbose logging of some sort? Any logs on the file system that would give more info than what the log.nsf db is showing?
I am not a Notes developer or admin at all, but the server has been affected for several days now and it needs a resolution. The site is up and running… but users have been complaining obviously about things being slow and at some times not working.
Thanks in advance for any advice!
Subject: A way to trace
Since it is an HTTP server message the agents must be getting trigged by the web. So, around the same time, check the urls being accessed on the server in the web server logs (if they are setup, if not, set them up in the server document or web config). See what document, form, agent, etc. is being accessed at the time. If a form/document, than check the webquerysave and webqueryopen setting on the form to see what agent is set to run. The code will have a print statement which in a web server type agent will print out to the console.
Howard
Subject: agent managers
It seems we have some scheduled agents that run every 20 minutes that may have something to do with what is going on. Our errors in the log.nsf db appear after these run.
The server is set to be able to run 5 agents concurrently. We have 6 namgr.exe processes running on the server. Should the concurrent agent setting (5 in our case) on the server match up with the number of namgr.exe processes?
The agents that are scheduled use two of these processes… 2 agents a piece. I can see that the namgr.exe processes being used have what appears to be a memory leak. Each run, the memory usage on those 2 processes grows.