sr-users November 2005

sr-users@lists.kamailio.org

245 participants
571 discussions

[Serusers] ACK before INVITE on re-INVITE causes Cisco AS's to return a 500 error
by Evan Borgstrom 16 Nov '05

16 Nov '05

Hey all, This has more to do with Asterisk & Cisco AS's IMO but I thought it would be interesting to get some responses from this group before I begin the battle with Cisco to get a bug report underway. We have an Asterisk box acting as a PBX sending calls to our SER instances. The asterisk box has "reinvite=yes" in the sip.conf file which causes Asterisk to send a re-INVITE for the call once it receives the 200 OK. This re-INVITE has it's SDP changed to that of the IP Phone behind the Asterisk box so that all media doesn't need to traverse it. This works fine except in certain instances where messages get out of order and it causes our Cisco GW's to error the call with a generic 500 message. When the messages are in the correct order the 500 error does not appear. So here's the situation, user behind the asterisk box picks up the phone and dials a PSTN destination. Asterisk fires off the initial invite and everything proceeds as normal and we eventually receive the 200 OK message to which the Asterisk box replies with ACK and a new INVITE that come in this order (X = Asterisk, Y = SER, Z = Cisco): U 2005/11/15 16:51:51.557322 X.X.X.X:5060 -> Y.Y.Y.Y:5060 ACK # U 2005/11/15 16:51:51.557948 X.X.X.X:5060 -> Y.Y.Y.Y:5060 INVITE SER accordingly processes the two messages and replies 100 Trying to the new INVITE message. Here's where the problem occurs, SER then sends these messages out to the PSTN gateway but the order has been reversed: # U 2005/11/15 16:51:51.558937 Y.Y.Y.Y:5060 -> X.X.X.X:5060 SIP/2.0 100 trying # U 2005/11/15 16:51:51.559000 Y.Y.Y.Y:5060 -> Z.Z.Z.Z:5060 INVITE # U 2005/11/15 16:51:51.559657 Y.Y.Y.Y:5060 -> X.X.X.X:5060 ACK When this happens the Cisco GW replies with a 500 error message and the following is sent to the Cisco's logs: # U 2005/11/15 16:51:51.711355 Z.Z.Z.Z:5060 -> Y.Y.Y.Y:5060 SIP/2.0 500 Server Internal Error. Nov 15 21:51:51.613: HandleUdpSocketReads :Msg enqueued for SPI with IPaddr: Y.Y.Y.Y:5060 Nov 15 21:51:51.613: *****CCB found in UAS Request table. ccb=0x653BCD0C Nov 15 21:51:51.613: CCSIP-SPI-CONTROL: act_sentsucc_new_message Nov 15 21:51:51.617: CCSIP-SPI-CONTROL: act_sentsucc_new_message_request Nov 15 21:51:51.617: Queued event from SIP SPI : SIPSPI_EV_SEND_MESSAGE Nov 15 21:51:51.617: sip_stats_status_code Nov 15 21:51:51.617: sipSPICheckRequest: CheckRequest fail on method 102 error code: 2 and status: 500 Is there any parameter that I can set to ensure that message leave SER in the order they enter? Should the Cisco be able to handle this condition (I check a number of different sections of the RFC but couldn't find anything specific for either side)? Thanks, Evan

3 4

[Serusers] avp_write/avp_pushto in database/memory?
by Ramin Nikaeen 16 Nov '05

16 Nov '05

Valued Colleagues, Do the avp_write and avp_pushto involve Database access operations?! For example: avp_write("sip:22@10.1.10.60:5060", "s:callfwd"); avp_pushto("$ruri", "s:callfwd"); Does avp_write store the string "sip:22@10.1.10.43:5060" for "s:callfwd" in database and avp_pushto looks up the string related to "s:callfwd" from the database or is the information stored in memory?! I want to avoid database access to ensure improved performance. Thanks ramin

2 1

[Serusers] call forwarding and *21
by Kyriakos Mavromichalis 16 Nov '05

16 Nov '05

Can someone point me to any resources on how to implement services like callforwarding by pressing *21 through the users' UA? Basically I need my users to be able to activate different services via the phone's keys. thank you -- Kyriakos Mavromichalis Otenet Telecom

3 3

[Serusers] Problem with running SER with mysql module
by Murděj Ukrutný 16 Nov '05

16 Nov '05

Hi, I am a beginner here. Could anyone please answer my question? I have downloaded the SER 0.9.4 and compile with mysql module it on SUSE Linux Enterprise Server 9. Mysql 4.1.15 compiled from source. I have problem whith running SER. Command 'serctl start' finished with message 'Starting SER : PID file /var/run/ser.pid does not exist -- SER start failed' Message in /var/logs/messages is 'Nov 16 05:34:34 suse1 ser: ERROR: load_module: could not open module </usr/local/lib/ser/modules/mysql.so>: libmysqlclient.so.14: cannot open shared object file: No such file or directory' 'libmysqlclient.so.14' is in '/usr/local/lib/mysql/' Any idea ? Thank. -- Murdej Ukrutny

3 2

[Serusers] Mediaproxy High Load
by Adrian Georgescu 16 Nov '05

16 Nov '05

The CPU will show in Linux high usage as soon as you get over 20 calls but you will be able to load the box up to 200 streams, the CPU will show the same high value while the proxy works fine. Use the rtpgenerator to test after what amount of calls your box will start loosing packets we reckon 70 streams per 1 GHz. Regards, Adrian <<<<<< Hi All, I tested the load of the mediaproxy (ver 1.4.2) and found that it is quite high. To give an idea, on a P4 2.4GHz machine, 20 sessions give a reading of about 35%, and it will go to almost 90% for 50 sessions. If we have a system that say can accommodate 10000 users (behind nat) or concurrent calls, that probably means we need about 10000/50=200 machines. That is way too many boxes. Does anybody has a solution for this? Thanks for any help available. Regards, TC Chan

1 0

[Serusers] difference between callsetup asterisk & ser and related security risks
by Arne Van Theemsche 16 Nov '05

16 Nov '05

Hi, I've been spending days on tcpdumping callsetups with ser OR asterisk as registar. This is my conclusion, can somebody confirm this? On asterisk: -asterisk maintains 2 independendly sipsessions between 2 clients, where asterisks stays the contact in the sipheaders. After picking up the call, asterisk sends an re-invite in the sip-body (address), and remains the contact for sipmessages. If someone hangs up, asterisk sends an re-invite to the other party, for redirecting stream back to asterisk, and then sending the bye to the other client (the one who didn't hang up). on ser: ser receives the invite, looking up the address of the callee, and sending the invite through to the callee, the contact header allways stays the caller's uri, but ser put's a record-route in the sipmessage, so that the bye first is send to the ser machine, and then to the other party. Am I seeing things correct here? What are the pro/contra's of both procedures? Imo gives the asterisk way a lot more overhead, but is more secure (2 clients keep communicating with asterisk, and are unable to fake things (or less easely), there were an not compliant client can bypass the record-route of ser, and so avoiding accounting. Again, am I correct here? thx Arne

3 2

[Serusers] ERROR: exec_str:
by davor jovanovic 16 Nov '05

16 Nov '05

Hi all, I'm using SER 0.9.3 and trying to do LDAP query but I keep getting this message: 0(17313) INFO: dont_fork turned on, living on 0(17313) exec_str: rtrim 0(17313) ERROR: exec_str: cmd /etc/ser/sipldap failed. exit_status=-1, errno=10: No child processes can someone explain to me why is that and how can I fix it? when I'm doing lookup with: (!lookup("location")) { if (!exec_dset("/etc/ser/sipldap")) { sl_send_reply("404", "Not Found"); break; } else { log(1," sipldap call"); }; }; I'm getting "Not Found", but when I'm doing (transferring to PSTN route without sl_send_reply) (!lookup("location")) { if (!exec_dset("/etc/ser/sipldap")) { #sl_send_reply("404", "Not Found"); route(5); break; } else { log(1," sipldap call"); }; }; route[5] { rewritehostport(xxxxxxxxxxx); I can extract LDAP telephone number and do get call transferred to PSTN GW but still getting "ERROR: exec_str:" thanx davor

1 0

[Serusers] Inaccurate Radius Accounting
by Lenir Santiago 16 Nov '05

16 Nov '05

Hey guys, I'm using Radius to do pretty much everything regarding my users, i.e. registrations, authorization, features and preferences. I implemented my own way of doing speeddial by using AVPs provided by radius and AVPOPs module, Here is the snippet: AVP-SIP = speedial:30:15615551212 avp_subst("s:caller_speedial/i:99", "/(.*):(.*)/\2/"); #15615551212 avp_printf("i:98","$rU:*"); #30 avp_printf("i:97","sip:$avp($temp2)@$td"); # sip:15615551212@xx.xx.xx.xx avp_printf("i:96","<sip:$avp($temp2)@$td>\r\n"); # <sip:15615551212@xx.xx.xx.xx>. xlog("L_N"," SPEEDIALING: $fU is speed-dialing $hdr(To)->$avp($temp2)\n"); if (avp_check("s:caller_speedial", "fm/$temp1/g")) { #avp_pushto("$To","i:96"); #avp_pushto("$ruri","i:97"); #avp_pushto("$duri","i:97"); #subst('/^To: (.*)(a)(.*)/To: "$avp($temp2)" <$avp($temp2)@\2/'); #append_hf("To: \"5615551212\" <15615551212(a)xx.xx.xx.xx>;user=phone"); #subst_uri('/^sip:(.*)@(.*)$/sip:$avp($temp2)@\2/i'); #subst_user('/(.*)/$avp($temp2)/'); #remove_hf("To"); rewriteuser("15615551212"); xlog("L_N"," SPEEDIALING: $fU is dialing $hdr(To) $avp($temp2)\n"); avp_delete("$temp1/g"); avp_delete("$temp2/g"); }; In this scenario, im forcing it instead of getting the value from AVPs just for ease of showing my point. The problem that I'm having is that no matter what I do to change the To header or change the R-URI, radius ALWAYS sends the accounting records with the original R-URI, in my case 30(a)xx.xx.xx.xx, how can I get acc.so to send the final resulting URI (15615551212(a)xx.xx.xx.xx) in the Called-Station-Id AVP? rad_recv: Accounting-Request packet from host xx.xx.xx.xx:39216, id=97, length=298 Acct-Status-Type = Start Service-Type = Sip-Session Acct-Terminate-Cause = 200 Error-Cause = Invite User-Name = "1000(a)xx.xx.xx.xx" Calling-Station-Id = "sip:1000@xx.xx.xx.xx" Called-Station-Id = "sip:30@xx.xx.xx.xx;user=phone" Sip-Translated-Request-URI = "sip:949449852150#15615551212@gateway;user=phone" Acct-Session-Id = "000d2890-d47f01e8-1dcc30bc-7a49c7a4(a)yy.yy.yy.yy" Sip-To-Tag = "230383F4-2AF" Sip-From-Tag = "000d2890d47f20f54576a906-290bdc33" Sip-CSeq = "102" NAS-IP-Address = xx.xx.xx.xx NAS-Port = 5060 Acct-Delay-Time = 0 rlm_sql (sql): Reserving sql socket id: 12 rlm_sql (sql): Released sql socket id: 12 Sending Accounting-Response of id 97 to xx.xx.xx.xx:39216 rad_recv: Accounting-Request packet from host xx.xx.xx.xx:39216, id=98, length=292 Acct-Status-Type = Stop Service-Type = Sip-Session Acct-Terminate-Cause = 200 Error-Cause = 8 User-Name = "1000(a)xx.xx.xx.xx" Calling-Station-Id = "sip:1000@xx.xx.xx.xx" Called-Station-Id = "sip:30@xx.xx.xx.xx;user=phone" Sip-Translated-Request-URI = "sip:949449852150#15615551212@gateway:5060" Acct-Session-Id = "000d2890-d47f01e8-1dcc30bc-7a49c7a4(a)yy.yy.yy.yy" Sip-To-Tag = "230383F4-2AF" Sip-From-Tag = "000d2890d47f20f54576a906-290bdc33" Sip-CSeq = "103" NAS-IP-Address = xx.xx.xx.xx NAS-Port = 5060 Acct-Delay-Time = 0 Thanks in advance! Lenir

3 3

[Users] no voice when using nathelper with rtpproxy
by messa＠innsof.com 16 Nov '05

16 Nov '05

Hi all I install rtpproxy in the same host (with the public IP) than openser. the installation of rtpproxy put rtpproxy.sock in /var/run directory.But when i make a call between 2 clients (each client in a different LAN), i have no voice after signalisation. And openser give me these errors: [root@Is-dev01 openser-1.0.0-tls]# /usr/local/sbin/openser Listening on udp: 62.128.173.101 [62.128.173.101]:5060 tcp: 62.128.173.101 [62.128.173.101]:5060 Aliases: tcp: 62-128-173-101.iwayafrica.com:5060 udp: 62-128-173-101.iwayafrica.com:5060 WARNING: no fork mode stateless - initializing 0(0) Maxfwd module- initializing textops - initializing 0(0) INFO: udp_init: SO_RCVBUF is initially 110592 0(0) INFO: udp_init: SO_RCVBUF is finally 221184 2(0) INFO: fifo process starting: 20621 2(0) SER: open_uac_fifo: fifo server up at /tmp/ser_fifo... 2(0) INFO: no fifo_db_url given - fifo DB commands disabled! 0(20619) ERROR: force_rtp_proxy2: no available proxies 0(20619) ERROR: force_rtp_proxy2: no available proxies 0(20619) ERROR: unforce_rtp_proxy: no available proxies What's the matter? Best regards Serge

3 2

[Serusers] Mediaproxy High Load
by tcchan 16 Nov '05

16 Nov '05

Hi All, I tested the load of the mediaproxy (ver 1.4.2) and found that it is quite high. To give an idea, on a P4 2.4GHz machine, 20 sessions give a reading of about 35%, and it will go to almost 90% for 50 sessions. If we have a system that say can accommodate 10000 users (behind nat) or concurrent calls, that probably means we need about 10000/50=200 machines. That is way too many boxes. Does anybody has a solution for this? Thanks for any help available. Regards, TC Chan

1 0

← Newer
1
...
30
31
32
33
34
35
36
...
58
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

sr-users November 2005