We run a Wildfire 3.1.1 server and connect to it using Psi 0.10 clients. We have around 400 users in total, spread between three different servers. Under normal conditions everything works mostly as expected. When an incident occurs, the population of the main chatroom increases from 100 users to around 150-170 as everyone tries to get the information in order to react to it. It can take more than five minutes between hitting enter and a message actually appearing in the room.
Restarting the wildfire service usually alleviates the problem, but this takes up to fifteen minutes. When we try and stop the service, it fails to respond in a timely fashion and then displays “Stopping”. We have to wait until it finally stops before we can actually restart the server.
What could be causing our problem? Our users are beginning to get exasperated, and we’re running out of ideas!