VoIP monitor Files

version 3.0.1
=============

Upgrade from 2.2:
-----------------

cat cdrtable.sql.2.2-3.0 | mysql voipmonitor # this will take many time
depending on number of cdr records. 


Buf fixes:
----------
- fix memory leak and double CDR during wav recording 
- fix crashes caused by bad fclose
- fix memory corruptions and race condition causing crashes
- Statistics from caller (a_*) was swapped to caller in some cases. 
- Fix ab_received packet. It didnt correspond to what wireshark sees. It was
always two packet less.
- Despite the documentation, the only the long form --pid-file was parsed. Add
missing -P option.
- currently voipmonitor <= 2.2 insists on IPv4 addresses on the sniffing 
interface. That's not necessary, often it's error prone. Furthermore the original 
code provided the IP address to the pcap library, but they expect the netmask. So
the code was wrong and might be a cause for sporadic errors with filter compilation.
- ignore packets where datalen < 0 (corrupted packets? just in case?)
- check if RTP packet is really RTP packet by checking RTP version bit which
has to be equel 2. Without this check, some packets like STUN was confusing 
voipmonitor and graph file and raw file was created for each non RTP UDP packet 
which has the same SRC/DST port.
- calculation of packetization was wrong for G723 in case of 60ms which causes
drops on WAV recording. Simplify and fix packetization.
- fix out of sync WAV recording in case where both RTP streams do not start at
the same time (probably the most cases are for those who sends 200 OK
immediately after 180 RINGING to simulate progress inband).
- Fix recording WAV and RAW rtp for calls with more than one reINVITEs. Name of
files were not unique and some raw files overwrited previouse file so final
file was not in sync and shuffled.
- Fix sync WAV issue for cases where RTP stream is not sending for a while
before reINVITE.  
- fix handling malloc and open files failures
- rise MAX_IP_PER_CALL to 30 

New features:
-------------

- implement --ring-buffer and set it in MB (feature of newer >= 2.6.31 kernels). 
If you see voipmonitor dropping packets in syslog upgrade to newer kernel and 
increase --ring-buffer to higher MB. It is buffer between pcap library and 
voipmonitor. The most reason why voipmonitor drops packets is waiting for I/O 
operations (switching to ext4 from ext3 also helps.
- Implement configuration file. See config/voipmonitor.org. Config is read
only if --config-file /etc/voipmonitor.conf specified on command line
- Implement generic init script. See config/init.d/voipmonitor
(thanks Telephonic http://telephonic.ca for sponsoring this work)
- do not allocate 30 RTP classes for each call, allocate it only whan needed. It
saves RAM and CPU.
- Implement new cdr.whohanged column which represents who hanged up or canceled
call.
- Implement lastSIPresponse and lastSIPresponseNum (cat cdrtable.sql.2.3-2.4
|mysql voipmonitor) you can now filter calls by bad response - example 
SELECT * FROM cdr WHERE lastSIPresponseNum >= 400 (403 Wrong password or domain)
- support compact headers according to rfc3261.txt section 20. Compact headers
means that Call-ID: is replaced by i:. This resulted that voipmonitor ignored
all calls with compact headers on.
- add --rtp-oneleg option which is important option if voipmonitor is sniffing on 
SIP proxy and see both RTP leg of CALL. in that case use this option. It will 
analyze RTP only for the first LEG and not each 4 RTP streams which will confuse 
voipmonitor. Drawback of this switch is that voipmonitor will analyze SDP only 
for SIP packets which have the same IP and port of the first INVITE source IP 
and port. It means it will not work in case where phone sends INVITE from 
a.b.c.d:1024 and SIP proxy replies to a.b.c.d:5060. If you have better idea 
how to solve this problem better please contact support@voipmonitor.org
- Implement SIP register messages storing to PCAP and MySQL new table register. 
This feature is optional with --sip-register new switch and configuration option 
in voipmonitor.conf.
- implement native conversion to OGG vorbis audio, now you can choose if you
want to record to wav or ogg. Ogg is about 25kbps. It can be configured in 
voipmonitor.conf or --audio-format = <wav|ogg> on command line
- raise jitterbuffer for recording stream from 200 to 500 which improves
dejittering audio
- Put write buffers for RAW and WAV files which solves IO throughput. Now it
can convert wav files in realtime for 200 concurent calls on usual sata 2GB
disk and ext4 fs. Before this change queue was filling by calls until it
reached memory limit and voipmonitor crashed. But I'm not recommending writing
WAV file in realtime as it uses a lot of RAM and can cause instability. 
- Change core dump to unlimited
- change default maximum opened files from system default (usually 1024) to
65535 to not block writes for very large network traffic
- make core dump if voipmonitor crashes to voipmonitor directory