Is this just BLF’ hints causing this? There are 16 phones online, each monitoring several of the other phones (with 3 of them having sidecars monitoring everyone). CPU usage goes to 100% when calls come in. Wondering if there’s a better way to do this, as in can the phones monitor each other on the LAN vs over the internet to the PBX? Or is there something else that could be causing this?
Periodically I also see these massive spikes in calls too, where it’ll look in the graph like 20-30 calls are concurrently running. It’s only every couple hours though, not every call. There are generally 1-3 people on the phone at once.
Are you using the sangomaconnect module by any chance? I’ve seen that on v16 systems using that module or something that depends on that module. I think it was because of previous HTTP sessions not shutting down for some reason, so it just kept maxing out the apache sessions allowed and loading down the server. I think fwconsole restart resolved it until it happened again a few days later. Seemed to happen on servers with fairly high call volume so that combination may of had something to do with it.
Similar to our situation. When we would turn on sms for a connect user it would cause the cpu to spike for long periods of time. We would turn off sms for the connect user, reboot the server and like magic every thing was fine.
Not sure who you are talking to. We are reporting what we went through, in the field, and our work around. You have a fix? Or you just telling us what your box shows. Thank you so much for all your help.
