Ken Cox
2007-05-07 16:51:15 UTC
I am attempting to attain a load of 1000 calls, split between 4 instances of
sipp on the same host running 250 calls each. Somewhere around 600-700
calls, I start seeing timeouts, e.g.
Aborting call on UDP retransmission timeout for Call-ID '
103-***@10.1.1.242'.
The funny thing is, I can't find this Call-ID in my SIP
logs. I capture the SIP traffic using 'ngrep' on both the UAC host and the
UAS host.
I know UDP is unreliable, but it would be surprise me if sipp was actually
sending multiple retransmissions with the same Call-ID and the UAC dropped
all of them. My UAC and UAS are both on dual-Xeon dual-core Linux
boxes, on an isolated Gig-E switch. My scenario file was created using
'sipp -sd uac_pcap', and includes retrans="500" on INVITE.
Any suggestions would be most welcome,
sipp on the same host running 250 calls each. Somewhere around 600-700
calls, I start seeing timeouts, e.g.
Aborting call on UDP retransmission timeout for Call-ID '
103-***@10.1.1.242'.
The funny thing is, I can't find this Call-ID in my SIP
logs. I capture the SIP traffic using 'ngrep' on both the UAC host and the
UAS host.
I know UDP is unreliable, but it would be surprise me if sipp was actually
sending multiple retransmissions with the same Call-ID and the UAC dropped
all of them. My UAC and UAS are both on dual-Xeon dual-core Linux
boxes, on an isolated Gig-E switch. My scenario file was created using
'sipp -sd uac_pcap', and includes retrans="500" on INVITE.
Any suggestions would be most welcome,
--
-Ken
-Ken