Hello everyone, We are building a new * server based on a Supermicro motherboard with a 2.8 Xeon processor and a TE220B card. We're using the PBX In a Flash distribution. What we've found is that with a 4 user MeetMe conference, the CPU usage is consistently around 16%. This in comparison to our existing PSTN gateway * box running 1.09 (it hosts our conferences and terminates our T1s). With 23 users and processing all PSTN phone calls, CPU usage averaged from 3-8%. This is an older Supermicro, with a 2.4 Xeon processor. In both cases, the connections are via IAX trunks from our main PBX here, and in two remote locations. We use g711 u-law only - no other codecs are used. If we connect the same number of users through a PRI connection directly to the new server, the CPU is 1% or less, so obviously we've pooched something.
We saw this same behavior when we split off the users to a 1.4x based PBX, and we thought it was the server hardware in the new machine, which was an older Dell 2650. But now we're not so sure. I know this is kind of vague, but can anyone suggest what might be happening? New Server CentOS 5, Kernel version 2.6.18-53.1.14.el5 Asterisk 1.4.19.1, and the SVN Zaptel drivers for the TE220B problems posted recently 2.8 Xeon, Hyperthreading disabled, 4GB RAM, 3Ware 9550SX RAID Old Server Fedora, Kernel version 2.4.22-1.2199.nptl Asterisk 1.0.9 2.4 Xeon, Hyperthreading off, 1GB RAM Thanks for the help, Kevin
_______________________________________________ -- Bandwidth and Colocation Provided by http://www.api-digital.com -- asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
