Your problem is most probably the use/abuse of Local channels, where the call initially has a uniqueid but it changes to a different one do to a channel masquerading or similar.

It is not something that can be easily fixed, as the manager monitor application uses the monitor app (instead of mixmonitor), and I am not sure if we can pass arguments to be expanded when the call ends (and not when the recording starts/middle of call).

Mixmonitor has the option, but not sure if Monitor does. So, the script is ok, but the uniqueid does not match the cdr uniqueid and that is why the field is not populated. You could modify the script to do a fuzzy search of sorts, using src, dst, and aproximate uniqueid and timestamps, but it won't be too nice imho.

I am reviewing events and code, and I think the problem might be the channel you are monitoring. If you start recording on the channel that receives the call (leg B), the filename will have the uniqueid of that channel, but the cdr is saved with the uniqueid of the originating channel (leg A).

Try starting a recording for an extension that is doing an outgoind call and check uniqueids, then do it when you receive a call. Do you see the difference in the later case?

I am working on a fix, so fop2 sends the other leg uniqueid as a parameter to the post recording script. On my initial tests it seems to work, you should see it working on the next release. If you want to test it before release, catch me on the live help and I will provide a link to download a test version.

Regarding monitor_mix, it does not affect the script. If you set monitor_exec asterisk IGNORES the mix option, and records unmixed and passes 3 parameters to the script.

Best regards,

Yes, the issue is fixed on what it would be the next release. Just to be sure we are talking about this, I will describe the issue : using elastix and the update script provided for monitor_exec is that the recording appears depending on the call leg you start the recording in fop2. Suppose you have extension 100 talking to extension 200, and you start recording , it works when you record 100, but it does not work when you record 200. (actually , the recordings *are* made, but the uniqueid of the call does not match the cdr entry, and so the link is lost)

wait. wait. I smell a "I too have this problem" when you are actually not having the same problem.

One thing is not having the recordings populated in the recordings interface because you did not configure them, and a totally different thing is that you have SOME recordings on the *Elastix* recording interface (not fop2 interface) showing up when using the special purpose recordings_elastix.pl script supplied with the latest fop2 version.

If you do not see *anything* in the fop2 recordings interfaces then it is not the same problem, the most probable cause is what I said above, you did not configure the monitor_exec option in fop2.cfg.

Recordings, by default, are *not* enabled because they will most probably require some intervention from your part.

Look for this line in fop2.cfg,

; monitor_exec=/usr/local/fop2/recording_fop2.pl

remove the semicolon so it becomes active:

monitor_exec=/usr/local/fop2/recording_fop2.pl

Then restart fop2:

service fop2 restart

After that, you *might* start seeing recordings that you initiate via fop2 in the recordings tab. If you have "all" permissions you will see all recordings, if not you will see only your own.

Best regards,

Hi Gavin,

This issue is the one fixed on what will be the next release. Contact me in the live help tomorrow if you want to try out the patched version, as there is no official update yet.

http://download.fop2.com/recordings_elastix.pl.txt

That is the last incarnation of the monitor_exec script to use for elastix. Elastix rpm will use just a named file the way old ARI used (with file inspection on the hard disk). Elastix 2.3 uses now a field on the cdr table to link recordings (much more proper), so you need to use this script to set the field when the recording is done.

However, it works depending on the call leg you start the recording, not sure what direction, because the uniqueid stored on the cdr does not match the uniqueid of that leg. FOP 2.27 fixes that problem, but it is still in beta.

Best regards,

Hi,

download 2.27 beta that includes the fix and the modified script:

http://download.fop2.com/fop2-2.27-centos-i386.tgz
http://download.fop2.com/fop2-2.27-centos-x86_64.tgz

Do you use http or https? What browser?

One important change that might be an issue is that the new fop2 version will use secure web sockets instead of standard web sockets when you access it via https. You have to set the ssl certificate files in fop2.cfg, uncomment and set the appropiate certs, the same ones you use for your https web server:

; If you access fop2 via https, browsers will try to use wss (Secure 
; web sockets) and for that it requires a certificate file and key file,
; the same ones you have in your web server configuration:
;
;ssl_certificate_file=/etc/pki/tls/certs/localhost.crt
;ssl_certificate_key_file=/etc/pki/tls/private/localhost.key

Now, the problem lies in that some browsers require you to accept the certificate as you have to do when you browse an https site with self signed certificates, but the wss connection does not ask for it. I think firefox needs to be pointed to https://your.server:4445 or perhaps wss://your.server:4445 in the address bar so it ask for confirmation on the certificate. I did not have to do it with chrome. Another option if this is the case is to disable websockets in /var/www/html/fop2/js/presence.js