Hi there,

I have a production yaws server, that works perfectly - for a while, then over time slowly uses up ~70% of CPU time, with the rest spread between idle and system processes.

This happens with only a few simultaneous users (2-3!), and once it has reached that point, never returns, even once incoming traffic has been removed.

I don't seem to be able to replicate the conditions in testing, so I really need a way to see what is taking the time on the live server.  I had thought it was mnesia tables becoming excessively large, but now they are trimmed right down (to where they were before the problems started), the same thing happens.

etop for some reason doesn't work - I get badmatch,{error,nxdomain}} on startup.

Any ideas, anyone?!

Regards