Hi all. I've tried to update to Opensips 1.6.1, but encountered the following problem. Opensips starts successfully, but soon almost all it's processes die one by one and only two processes remain. For example, if right after start we have:
# ps ax | grep opens 26182 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26183 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26184 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26185 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26186 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26187 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26188 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26189 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26190 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26191 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26192 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26193 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26194 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26195 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26196 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26197 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26198 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26199 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26200 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26201 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26202 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26203 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26204 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26205 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26206 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26207 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26208 ? S 0:00 ./opensips -k 0x11110204 -u opensips When processes die, we have only: #ps ax | grep opens 26182 ? S 0:00 ./opensips -k 0x11110204 -u opensips 26184 ? S 0:00 ./opensips -k 0x11110204 -u opensips If I set debug=6, the following is written to /var/log/messages: Dec 22 11:02:03 srv rtpproxy[17011]: INFO:rxmit_packets: caller's address filled in: 195.182.195.206:1024 (RTP) Dec 22 11:02:03 srv opensips[26184]: Route 5 - NOTIFY Dec 22 11:02:05 srv opensips[26185]: Route 5 - PUBLISH Dec 22 11:02:06 srv opensips[26183]: Route 5 - NOTIFY Dec 22 11:02:06 srv opensips[26185]: Route 5 - NOTIFY Dec 22 11:02:06 srv opensips[26185]: Route 5 - NOTIFY Dec 22 11:02:06 srv opensips[26186]: Route 5 - NOTIFY Dec 22 11:02:06 srv opensips[26186]: Route 5 - NOTIFY Dec 22 11:02:08 srv rtpproxy[17011]: INFO:handle_command: lookup on ports 36664/35096, session timer restarted Dec 22 11:02:08 srv rtpproxy[17011]: INFO:handle_command: pre-filling callee's address with 87.251.142.50:5006 Dec 22 11:02:08 srv opensips[26208]: CRITICAL:core:receive_fd: EOF on 13 Dec 22 11:02:08 srv opensips[26182]: INFO:core:handle_sigs: child process 26186 exited by a signal 11 Dec 22 11:02:08 srv opensips[26182]: INFO:core:handle_sigs: core was not generated Dec 22 11:02:08 srv opensips[26182]: INFO:core:handle_sigs: terminating due to SIGCHLD As I see, the last message received by process with PID 26186 is NOTIFY, and then it crashes. "Route 5 - NOTIFY" is in this block of configuration file: # SUBSCRIBE and PUBLISH Message Handling # -------------------------------------- route[5] { if (!t_newtran()) { xlog("L_INFO", "Failed to create transaction\n"); sl_reply_error(); exit; } if (is_method("PUBLISH")) { xlog("L_INFO", "Route 5 - PUBLISH \n"); handle_publish(); } else if (is_method("SUBSCRIBE")) { xlog("L_INFO", "Route 5 - SUBSCRIBE\n"); handle_subscribe(); } else if (is_method("NOTIFY")) { xlog("L_INFO", "Route 5 - NOTIFY\n"); t_reply("200", "OK"); exit; } exit; } In main routing logic: if (method == "SUBSCRIBE" || method == "PUBLISH" || method == "NOTIFY") { route(4); return(0); } As I see, Opensips sets core dump limit, if it's turned off, but no core is produced (OS is CentOS 5.3). What can be wrong? Version 1.6.0 did not crash like this.
_______________________________________________ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users