investigate why ohwr hangs randomly
ohwr had a hang today.
The number of processes, memory consumption and cpu usage skyrocketed today and the server had to be restarted. Our server is usually more than able to handle the workload, but not when it goes "asymptotically up".
I'm investigating why this happened.
So far I've been able to detect the sudden increase in resource comsumption, and I'm investigating how to "auto-detect" and kill those processes in Apache.
Later I'll do some stress tests on the repository views; we suspect they might be involved on these rogue processes.