on 01-24-2014 02:26 PM
Hello,
I'm struggling with finding root causes of some intermittent error.noresource errors, also, actually gaining value from this exception. Having recently integrated Nuance 8.5 with our environment I've been struggling to decern what the error.noreasource event is telling me each time i see it (I can of course see the problem when i get to the Nuance logs but handling it effectively in call studio is a different story).
I see this come back when there is a bad compile on a grammar, bad fetch on grammar, when Nuance is out of licensing, and network issues seem to trigger it as well. Its frustrating that this event seems to be a catch-all for various types of errors encountered between gateway and Nuance ASR. My biggest issue is that we've moved from test into production where I'm now seeing about a .4% rate of error.noresources. I know we're not exceeding ports used on Nuance end, I know the grammars are there and can be fetched properly by ASR, there is nothing unique about certain calls which receive the noresource error, it sparatically appears about .4% of the time regardless of the application that is running.
My real question is, is there a way from our end (cisco side) to see further into this error? I am using hot events to handle the expception right now but there seems to be no way to know on our end whether it was a network error, licensing issue, bad grammar, missing grammar, etc.. they all throw the same exception. It would be really nice to handle these events disctinctly rather than lump them into the no.resource bucket where I need to drop into a DTMF only application.
Thank you,
Ryan
It looks like you can turn on mrcp error logging on the gateway and see
the errors coming back from the ASR server.
I don't know why they convert it into error.noresource, maybe it's a bug
- I also get that for missing audio files in CVP8.
Can you submit a TAC case on it?
Janine
Janine/Hemal, I currently have a TAC case open and upon referencing CSCto77602 (which is only kind-of related) I have been informed by my engineer that:
The essence of the defect seems to be that if the ASR doesn't have
available resources to handle the call (or I guess if it encounters a
failure in general), it throws an event, which should be correlated back
to the calling VXML application as
'error.com.cisco.media.resource.unavailable.asr', which can be handled
using the default catch statements to close the VXML application
gracefully and to pass the call back to ICM for further treatment. It
seems that somewhere around IOS 12.4(24)T05, that stopped happening
causing an abrupt ending of the call..
Later the engineer said:
"Yeah I've done some more reading since and I have the distinct
impression that any error from the Nuance end is mapped to
error.noresource, rather than proper VXML events as they were before. I suspect this needs to be
re-evaluated in IOS code as obviously you can't do proper handling like
this."
I have moved the case to the IOS team and am further troubleshooting...
Thought it was kind of interesting... hopefully some proper error handling will be the result of this.
It looks like you can turn on mrcp error logging on the gateway and see
the errors coming back from the ASR server.
I don't know why they convert it into error.noresource, maybe it's a bug
- I also get that for missing audio files in CVP8.
Can you submit a TAC case on it?
Janine
> --------------------------------------------------------------
> I do have access to those logs.. however they are off the domain and
> it is a painful process to get them. That has been my only way of
> troubleshooting this. I guess i'm just expressing my frustration with
> the error.noresource exception and wondering if anyone on the board
> has some suggestions to handling it nicely at an application level.
> Thanks for the reply Janine, glad i'm not the only one struggling with
> this one.
>
> -Ryan
> --
> To respond to this post, please click the following link:
>
> <http://developer.cisco.com/web/cvp/forums/-/message_boards/view_message/4246136>
>
> or simply reply to this email.
--
Janine Graves
Hey Heidi,
What I got from TAC was that any 15.1T release has a known bug "CSCto77602". This bug causes most events to come back as error.noresource. Even causes a disconnect in some situations such as ASR throwing SPEECH_TOO_EARLY. I can reproduce this by yelling into the phone as the menu application starts, sure enough there is an MRCP SPEECH_TOO_EARLY event sent and the gateway disconnects the call and I see a session timeout in the VXML server logs.
The solution given was to upgrade to 15.1(4)M2. Still working through the hoops to get that happening in my environment, so I cannot comment if this resolved the issues or not....
Please keep us updated if you find a resolution, I fear that the upgrade is only going to solve the unhandled events and still leave us with error.noresource coming back for every exception. I even get this instead of a badfetch for missing wav files now :/
-Ryan
Thanks Ryan for the update. We will let you know as well if we find a resolution.
Hey Heidi,
What I got from TAC was that any 15.1T release has a known bug "CSCto77602". This bug causes most events to come back as error.noresource. Even causes a disconnect in some situations such as ASR throwing SPEECH_TOO_EARLY. I can reproduce this by yelling into the phone as the menu application starts, sure enough there is an MRCP SPEECH_TOO_EARLY event sent and the gateway disconnects the call and I see a session timeout in the VXML server logs.
The solution given was to upgrade to 15.1(4)M2. Still working through the hoops to get that happening in my environment, so I cannot comment if this resolved the issues or not....
Please keep us updated if you find a resolution, I fear that the upgrade is only going to solve the unhandled events and still leave us with error.noresource coming back for every exception. I even get this instead of a badfetch for missing wav files now :/
-Ryan
Does anyone know if Cisco has fixed the IOS code so that Nuance ASR errors are not reported as 'error.noresource'?
It's been a long time since this issue began and I'd hope it's been addressed by now.
Thanks, Janine
Did anyone got the solution ?
Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: