[Freeswitch-users] Call quality issues
Eric Beard
eric at loopfx.com
Fri Apr 29 21:10:19 MSD 2011
It seems to be a CPU issue.
I started from scratch today with a new machine, same exact hardware, and I am running 20 calls at less than 1% CPU, with perfect audio quality (same as the bad machine before it went bad).
On the bad machine, 20 calls causes the CPU to be at 20%, and quality is shaky. 30 calls and it's terrible. Only the freeswitch process is eating CPU.
I have diffed the conf folders and everything is the same. This is driving me nuts. I have no idea what I did to that machine to ruin the performance.
-----------------------
Eric Z. Beard, CTO
Loop LLC
w (877) 850-2010 x9249
m (727) 776-2768
eric at loopfx.com<mailto:eric at loopfx.com>
From: freeswitch-users-bounces at lists.freeswitch.org [mailto:freeswitch-users-bounces at lists.freeswitch.org] On Behalf Of Michael Collins
Sent: Friday, April 29, 2011 12:57 PM
To: FreeSWITCH Users Help
Subject: Re: [Freeswitch-users] Call quality issues
On Thu, Apr 28, 2011 at 2:52 PM, Eric Beard <eric at loopfx.com<mailto:eric at loopfx.com>> wrote:
Hello,
I ran into some major call quality issues this week, and I'm trying to figure out how to troubleshoot things. I've been running FreeSwitch for a few weeks, and suddenly a few days ago my call quality dropped drastically. I had been running more than 100 concurrent calls, with the CPU at less than 20%, but now at 20 concurrent calls, the CPU is still at a little less than 20%, and the call quality is bad - any higher and calls go almost completely silent. There is a direct correlation between the number of simultaneous calls and call quality.
I have tested against multiple gateways, same results against each, so it's not an issue with the gateway.
I have captured packets on the machine and analyzed them with Wireshark. It seems like the inbound packets are all fine, no jitter or loss. But the packets being sent by FreeSwitch are degraded.
One sample call showed:
Drop by Jitter Buff:158(14.1%) Out of Seq 0 (0.0%) Wrong Timestamp 96(8.6%)
What is the topology of the network path for the above call? Also, where on the LAN/WAN did you capture these packets? Wrong timestamps and 14% dropped packets suggests that something on the network is interfering with the delivery of these packets in a timely manner.
-MC
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freeswitch.org/pipermail/freeswitch-users/attachments/20110429/c303e8fc/attachment-0001.html
More information about the FreeSWITCH-users
mailing list