TTS Outbound Survey Readling Wrong List ID

[mstopkey]

Well, I've been researching and testing my head off on this one. Hope someone can help. Here it is. We have one server in our cluster dedicated to Cepstral TTS outbound surveys. It has always performed well. It is relative light load and uses a single HDD. We always clone with CloneZilla after any updates to it. 2 weeks ago, the drive went bad. No problem, replace the bad drive with identical spare, restore from clone image and fire it up. Not quite. I ran my test list with my cell phone number in the list to verify correct function. It called me, no audio. No such tts wav file.

[Apr 26 10:07:20] WARNING[10507][C-00000007]: file.c:701 ast_openstream_full: File /var/lib/asterisk/sounds/tts/d3/tts-d3a773a9a11032ad4384e050898f3f79 does not exist in any format

Created new TTS entry and got audio to work but, it calls my number in the assigned camp list and Allison is reading the fname lname from a list that is in a different campaign!!!! It does this every time, the same lead_id name from the wrong list. First thought, database server load. Nope, it's fine. Is the Cepstral install toast?

[mflorell]

Might want to try clearing out your /var/lib/asterisk/sounds/tts/ directory, it will use cached audio files if they are present and that may be what's happening.

[mstopkey]

Gotcha. Will try

[mstopkey]

Did that and tts not playing . CLI says:
WARNING[7823][C-00000016]: file.c:701 ast_openstream_full: File /var/lib/asterisk/sounds/tts/d3/tts-d3a773a9a11032ad4384e050898f3f79 does not exist in any format

That is the same file it was looking for earlier.

[mflorell]

Well, it's not generating for some reason, try looking in the Asterisk CLI output to see if there are any messages about it. Also, if you have AGI output set to BOTH for the server, it should show you the actual command that is being run to generate the audio file, then you can test that command.

[mstopkey]

I set AGI Ouput to BOTH. Tested again and I do not see the command happening to generate. You are correct, not generating the file but it does generate the directory for it in /var/lib/asterisk/sounds/tts such as d3 or c9 etc. I have posted to the Cepstral forums for help

[mstopkey]

[Apr 26 16:43:16] -- <SIP/voipinnovationsoutbound2-00000018>AGI Script agi://127.0.0.1:4577/call_log completed, returning 0
[Apr 26 16:43:16] -- Executing [8374@default:3] AGI("SIP/voipinnovationsoutbound2-00000018", "agi-VDAD_ALL_outbound.agi,SURVEYCAMPCEP-----LB") in new stack
[Apr 26 16:43:16] -- Launched AGI Script /usr/share/asterisk/agi-bin/agi-VDAD_ALL_outbound.agi
[Apr 26 16:43:16] -- <SIP/voipinnovationsoutbound2-00000018> Playing 'sip-silence.gsm' (escape_digits=) (sample_offset 0) (language 'en')
[Apr 26 16:43:16] WARNING[16509][C-00000019]: file.c:701 ast_openstream_full: File /var/lib/asterisk/sounds/tts/d3/tts-d3a773a9a11032ad4384e050898f3f79 does not exist in any format
[Apr 26 16:43:17] -- <Local/912107222479@default-00000019;2>AGI Script agi://127.0.0.1:4577/call_log--HVcauses ... --11-----0 completed, returning 0
[Apr 26 16:43:17] == Spawn extension (default, 912107222479, 3) exited non-zero on 'Local/912107222479@default-00000019;2'
[Apr 26 16:43:27] -- <SIP/voipinnovationsoutbound2-00000018>AGI Script agi-VDAD_ALL_outbound.agi completed, returning -1

This is what I am seeing, looks like running wrong script.

[mflorell]

I forgot that I threw those in a separate log file, take a look at "Cepstral.2018-04-26"

[mstopkey]

I ran swift -V and get this on the CLI:
swift -V
/opt/swift/bin/swift.bin: Symbol `swift_param_descriptors' has different size in shared object, consider re-linking

[mflorell]

No clue on that one, haven't installed cepstral in a while though either, but I've never seen that error.

[mstopkey]

Fixed it! If anyone has similar issue, uninstall Swift by doing the following.
As root, "rm -rf /opt/swift/"
This un-installs all swift applications, voices, voice licenses, port licenses and write-to-file licenses.
Reboot, then run "updatedb".
Re-install your purchased voices by extracting your voice files, cd into the extracted directory and run "./install.sh"
Re-active your voice license(s). cd /opt/swift/bin and run "swift -reg-voice" and follow prompts.
Re-active your port license(s). cd /opt/swift/bin and run "swift -reg-ports" and follow prompts.
Re-activate your write-to-file license. cd /opt/swift/bin and run "swift -ref-filewrite" and follow prompts.

Worked for me.

[mflorell]

Thanks for posting your solution!

[mstopkey]

Like to help others where I can.