<tt><font size=2>asterisk-users-bounces@lists.digium.com wrote on 01/16/2014
08:55:31 AM:<br>
<br>
> From: Gareth Blades <mailinglist+asterisk@dns99.co.uk></font></tt>
<br><tt><font size=2>> To: Asterisk Users Mailing List - Non-Commercial
Discussion <br>
> <asterisk-users@lists.digium.com>, </font></tt>
<br><tt><font size=2>> Date: 01/16/2014 08:55 AM</font></tt>
<br><tt><font size=2>> Subject: Re: [asterisk-users] Weird issue with
Set(CALLERID(name)=string);</font></tt>
<br><tt><font size=2>> Sent by: asterisk-users-bounces@lists.digium.com</font></tt>
<br><tt><font size=2>> <br>
> Very little as the amount of data being captured is quite small. We
<br>
> have it running on our production servers which routinely handle a
<br>
> couple of hundred concurrent calls.<br>
> <br>
> This is the script we use to start off the capture. It uses rolling
<br>
> capture files so we will always have the last X number of capture
<br>
> logs. It works very well and we have a custom system which enables
<br>
> us to search for calls and request traces for them for when we have
<br>
> to diagnose problems.<br>
> <br>
> #!/bin/bash<br>
> cd /var/lib/asterisk/siptraces<br>
> DATE=`date +%Y%m%d%H%M%S`<br>
> TRACEFILE=/var/lib/asterisk/siptraces/$DATE-siptrace.pcap<br>
> nohup /usr/sbin/tcpdump -p -i eth0 -s 0 port 5060 -w $TRACEFILE -C
10 -W 500 &<br>
> <br>
> <br>
</font></tt>
<br><tt><font size=2>Not to derail the conversation, Gareth, but do you
leave this running full time on your asterisk boxes or just turn it on
when you are trying to track problems?</font></tt>
<br>
<br><tt><font size=2>On average, how far back can you go for looking at
problems?</font></tt>