Re: System-Type retrying on failure

+1, commited to SVN.

Thanks,
Alexander Malysh

Am 21.09.2010 um 13:24 schrieb Rene Kluwen:

> Patch by Roy Walker. See below.
>  
> Improvement or upgrade breaker?
>  
> == Rene
>  
>  
> From: Roy Walker [mailto:[email protected]] 
> Sent: Tuesday, 21 September, 2010 01:02
> To: Rene Kluwen
> Cc: [email protected]
> Subject: RE: System-Type retrying on failure
>  
> Here is a patch that disables retries on an invalid system type aka 0x53 error.  Please let me know if you have any issues.
>  
> Roy
>  
> From: Rene Kluwen [mailto:[email protected]] 
> Sent: Monday, September 20, 2010 10:27 AM
> To: Roy Walker
> Cc: [email protected]
> Subject: RE: System-Type retrying on failure
>  
> I have not gotten input yet.
> Also I am a little bit busy at the moment. Any patch is welcome.
>  
> == Rene
>  
> From: Roy Walker [mailto:[email protected]] 
> Sent: Monday, 20 September, 2010 17:10
> To: Rene Kluwen
> Cc: [email protected]
> Subject: RE: System-Type retrying on failure
>  
> Rene,
>  
> Did you get any input from anyone else on added this error code?
>  
> Thanks,
> Roy
>  
> From: Rene Kluwen [mailto:[email protected]] 
> Sent: Thursday, September 09, 2010 10:48 AM
> To: Roy Walker; 'Alejandro Guerrieri'
> Cc: [email protected]
> Subject: RE: System-Type retrying on failure
>  
> In that case: Code 0x53 has to be added and handled the same way as “invalid login”.
>  
> If the other developers agree with this, I don’t mind to make a patch. It should be straightforward.
>  
> == Rene
>  
> From: Roy Walker [mailto:[email protected]] 
> Sent: Thursday, 09 September, 2010 17:43
> To: Rene Kluwen; 'Alejandro Guerrieri'
> Cc: [email protected]
> Subject: RE: System-Type retrying on failure
>  
> Rene,
>  
> I wish it was that simple… we are trying to get certified by At&t and they require us to go through a bunch of tests before they will give us a production SMPP bind.  One of these tests are to have an incorrect system-type… it does not return an error 13 though… see below, but does recognize it as an “Invalid system_type field”:
>  
> 2010-09-09 10:39:41 [12178] [11] ERROR: SMPP[Att]: SMSC rejected login to transmit, code 0x00000053 (Invalid system_type field).
> 2010-09-09 10:39:41 [12178] [11] ERROR: SMPP[Att]: Couldn't connect to SMS center (retrying in 45 seconds).
>  
> Thanks,
> Roy
>  
> From: Rene Kluwen [mailto:[email protected]] 
> Sent: Wednesday, September 08, 2010 4:01 PM
> To: Roy Walker; 'Alejandro Guerrieri'
> Cc: [email protected]
> Subject: RE: System-Type retrying on failure
>  
> Then what error is returned upon entering a wrong system-type?
> If it is not 0x0d (13) then it is an error in your smsc.
>  
> But then again: Is it that hard to just enter the correct system-type? Once you do that, the problem will be solved. Or do I see things wrong?
>  
> == Rene
>  
> From: Roy Walker [mailto:[email protected]] 
> Sent: Wednesday, 08 September, 2010 22:22
> To: Rene Kluwen; 'Alejandro Guerrieri'
> Cc: [email protected]
> Subject: RE: System-Type retrying on failure
>  
> Well I had hoped that my carrier would let me slide on this, but it’s not an option.  It seems to me that this should be a bug… if the system-type is wrong, then kannel should fail the same way it does a bad password.  As it is now the example Rene outlined below is not what it does.  It retries…
>  
> I created Bug #559 to track this.  If you disagree or think this should not be the case please let me know.
>  
> Thanks,
> Roy
>  
> From: Rene Kluwen [mailto:[email protected]] 
> Sent: Monday, August 23, 2010 11:04 AM
> To: Roy Walker; 'Alejandro Guerrieri'
> Cc: [email protected]
> Subject: RE: System-Type retrying on failure
>  
> I think it depends on what the smsc returns as an error code.
> If system-type is wrong and smsc returns 0x0d (13) then it is considered “wrong credentials” and afaik, Kannel doesn’t retry.
>  
> == Rene
>  
>  
> From: [email protected] [mailto:[email protected]] On Behalf Of Roy Walker
> Sent: Monday, 23 August, 2010 17:45
> To: Alejandro Guerrieri
> Cc: [email protected]
> Subject: RE: System-Type retrying on failure
>  
> Should this be opened as a feature request or bug report…?  Seems like a decision one way or the other… change it or make it configurable.
>  
> From: Alejandro Guerrieri [mailto:[email protected]] 
> Sent: Friday, August 20, 2010 4:30 PM
> To: Roy Walker
> Cc: [email protected]
> Subject: Re: System-Type retrying on failure
>  
> That depends on which carrier do you ask, but yes I agree, many of them require system-type to be treated as a "fatal" error and not retried.
>  
> Imho should be a configurable option or a compile switch at least.
>  
> Regards,
>  
> Alex
> 
> On Fri, Aug 20, 2010 at 10:27 PM, Roy Walker <[email protected]> wrote:
> Found what some might consider a bug, but when an invalid system-type is passed on an SMPP bind, it will retry based on the reconnect-delay setting.  This should be a stop failure and should work the same as an invalid smsc-username/system-id or smsc-password.  Where it does not retry… right?
>  
> Roy
>  
> <system-type-no-retry.patch>