[OpenSIPS-Users] [dispatcher] long delay in failed gw detection

Pascal Maugeri pascal.maugeri at gmail.com
Tue Oct 28 10:23:03 CET 2008


We have a configuration with one dispatcher that load-balances traffic to
two SIP registrars.

When I shutdown one registrar server, the failure is detected in 30 seconds
approximately. Is that a normal response time ? Or should we expect shorter
response times ? (I expect yes :-) ).

More precisely, having two target gateways named REG1 and REG2, if Is
shutdown REG2 at t0:
- traffic (SIP REGISTER messages) will be still load-balanced to REG1 and
REG2 during the first 7 seconds (t0+7seconds),
- then untill t0+30seconds, traffic will be sent exclusively to REG2 (the
failed gw),
- from t0+30s to t0+35, traffic will be sent to both gateways REG1 and REG2
- from t0+35s traffic will be sent to REG1 exclusively.

During the test we see the SIP OPTIONS messages sent to both REG1 and REG2.

For the dispatcher we are using the following configuration:

modparam("dispatcher", "list_file",
modparam("dispatcher", "flags", 3)
modparam("dispatcher", "dst_avp", "$avp(i:271)")
modparam("dispatcher", "grp_avp", "$avp(i:272)")
modparam("dispatcher", "cnt_avp", "$avp(i:273)")
modparam("dispatcher", "ds_ping_method", "OPTIONS")
modparam("dispatcher", "ds_ping_interval", 3)
modparam("dispatcher", "ds_ping_from",
"sip:fend1 at<sip%3Afend1 at>
modparam("dispatcher", "ds_probing_threshhold", 1)
modparam("dispatcher", "ds_probing_mode", 1)
modparam("dispatcher", "force_dst", 1)


        ds_select_dst("1", "0");
        xlog("L_INFO","---> $avp(i:271) $avp(i:272) $avp(i:273)");


failure_route[1] {

        xlog("L_INFO", "Marking GW as failed...\n");


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.opensips.org/pipermail/users/attachments/20081028/bf91176e/attachment.htm 

More information about the Users mailing list