I have a system where the calls is not disconnected properly.
I get this error in syslog when receiving BYE:
Jun 24 08:23:54 scantronpbx /usr/local/sbin/kamailio[31588]: WARNING: uac [replace.c:645]: restore_uri(): invalid char found in the new uri at pos 30 (#012) [sip:004540294149@127.0.0.1:507#01257441;transport=TCP;ob] Jun 24 08:23:54 scantronpbx /usr/local/sbin/kamailio[31588]: WARNING: uac [replace.c:647]: restore_uri(): this can happen when URI values are altered by end points - skipping the update Jun 24 08:23:54 scantronpbx /usr/local/sbin/kamailio[31588]: WARNING: uac [replace.c:645]: restore_uri(): invalid char found in the new uri at pos 30 (#012) [sip:004540294149@127.0.0.1:507#01257441;transport=TCP;ob] Jun 24 08:23:54 scantronpbx /usr/local/sbin/kamailio[31588]: WARNING: uac [replace.c:647]: restore_uri(): this can happen when URI values are altered by end points - skipping the update Jun 24 08:24:11 scantronpbx /usr/local/sbin/kamailio[31588]: WARNING: uac [replace.c:645]: restore_uri(): invalid char found in the new uri at pos 30 (#012) [sip:004540294149@127.0.0.1:507#01257441;transport=TCP;ob] Jun 24 08:24:11 scantronpbx /usr/local/sbin/kamailio[31588]: WARNING: uac [replace.c:647]: restore_uri(): this can happen when URI values are altered by end points - skipping the update
Kamailio receives this From header: f: sip:004540294149@194.255.22.44;tag=e042cfe1-40f7-410d-95fd-e5bb4f8976d7;transport=TCP;ob
And passes this on, where it is obvious that '%' may give a problem f: sip:0001@192.168.2%:;tag=e042cfe1-40f7-410d-95fd-e5bb4f8976d7;transport=TCP;ob
The RR headers send forward with INVITE looks like this: Record-Route: sip:127.0.0.1:5069;r2=on;lr=on;ftag=MYEMZIUn-UWwH-R7g9CktpyvOn1ZFuZu;vsf=AAAAAEUAcAB5AxcDGAkYCgALLjEwMQ--;did=2f6.6cc;vst=AAAAAAAABAR0AQsLGgACAW4DHA4uMC4wLjE6NTA3MA-- Record-Route: sip:192.168.2.9:5069;r2=on;lr=on;ftag=MYEMZIUn-UWwH-R7g9CktpyvOn1ZFuZu;vsf=AAAAAEUAcAB5AxcDGAkYCgALLjEwMQ--;did=2f6.6cc;vst=AAAAAAAABAR0AQsLGgACAW4DHA4uMC4wLjE6NTA3MA--
And received on BYE Route: sip:127.0.0.1:5069;lr;r2=on;ftag=MYEMZIUn-UWwH-R7g9CktpyvOn1ZFuZu;vsf=AAAAAEUAcAB5AxcDGAkYCgALLjEwMQ--;did=2f6.6cc;vst=AAAAAAAABAR0AQsLGgACAW4DHA4uMC4wLjE6NTA3MA-- Route: sip:192.168.2.9:5069;lr;r2=on;ftag=MYEMZIUn-UWwH-R7g9CktpyvOn1ZFuZu;vsf=AAAAAEUAcAB5AxcDGAkYCgALLjEwMQ--;did=2f6.6cc;vst=AAAAAAAABAR0AQsLGgACAW4DHA4uMC4wLjE6NTA3MA--
RR does not seem to be the problem as mentioned here: https://opensips.org/pipermail/users/2010-October/014999.html
I have attached a pcap file with the full trace. The error appears in frame 101
The setup is a bit complex, as I have 8 Kamailio's running, where 6 is identical.
The first Kamailio is on port 5069, which receives a call from a door phone. It then branches the call to registrations or GSM targets, which is handled by the other Kamailio's Thus 5069 is only doing a lookup in a database and branching, plus it handles rtpengine.
I first saw the error on version 4.4, and now I replicated it on 5.1.4. My hardware is a Raspberry Pi.
/usr/local/sbin/kamailio -v version: kamailio 5.1.4 (arm6/linux) flags: STATS: Off, USE_TCP, USE_TLS, USE_SCTP, TLS_HOOKS, DISABLE_NAGLE, USE_MCAST, DNS_IP_HACK, SHM_MEM, SHM_MMAP, PKG_MALLOC, Q_MALLOC, F_MALLOC, TLSF_MALLOC, DBG_SR_MEMORY, USE_FUTEX, FAST_LOCK-ADAPTIVE_WAIT-NOSMP, USE_DNS_CACHE, USE_DNS_FAILOVER, USE_NAPTR, USE_DST_BLACKLIST, HAVE_RESOLV_RES ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE 262144, MAX_LISTEN 16, MAX_URI_SIZE 1024, BUF_SIZE 65535, DEFAULT PKG_SIZE 8MB poll method support: poll, epoll_lt, epoll_et, sigio_rt, select. id: unknown compiled on 23:45:36 Jun 19 2018 with gcc 4.9.2