Multiple same MM submits over MM1

Multiple same MM submits over MM1 SearchSearch
Author Message
Dziugas Baltrunas
New member
Username: Menulis

Post Number: 3
Registered: 02-2006
Posted on Thursday, March 09, 2006 - 03:20 pm:   

Hi,

we're facing multiple MM submit over MM1 problem here for quite a long time. The situation is not easily reproducable but basically what we see from logs (and also user complains) is that user sends one MMS for one recipient, but phone keeps sending the _same_ MMS for several times. This number varies from 2 to 100 or so. It means that the user might be charged 50 times.

The proof of the situation is obvious - we see that during a short period of time (say 15 minutes) there are N attempts to send MMS with the same message size to the same recipient and according to NowSMS logs, all of them are sucessfull (accounting callbacks are possitive and a MMSIN record in mmsc log appears).

Have others been facing the same problem? Initially we thought it might be issue with buggy phones, but looks like it is not limited with one particular phone manufacturer.

I was thinking wheather this might be related to M-send.conf parameters (like X-Mms-Transaction-ID or X-Mms-Response-Status) which might be interpreted differently by different phones.

Thanks,
Dziugas
Bryce Norwood - NowSMS Support
Board Administrator
Username: Bryce

Post Number: 5617
Registered: 10-2002
Posted on Thursday, March 09, 2006 - 07:42 pm:   

Hi Dziugas,

I'll be curious to hear if anyone else has observed anything similar.

I fyou're getting the positive accounting callback, then that should indicate that the MMSC is returning a postive response.

That response should be an m_send_conf with a transaction-id header (a copy of the same header from the message submission), MMS version header (always 1.0 in v5.x), X-Mms-Response-Status: Ok, and an X-Mms-Message-Id header.

I can't think of any other headers that might be required by some devices.

But here's the strange part ... if sometimes the devices can send ok, but other times they encounter this repeating error ... then this would suggest to me that it is not a problem of a missing header, but rather something about the message content that is triggering an unexpected problem.

I'd look for any common denominators between the devices that are having occasional (or consistent) problems. Perhaps there is a problem occurring at the proxy? Maybe there is a common denominator of all affected devices using HTTP/WAP2 vs. WSP/WAP? I'm not sure what that would tell us, but basically any common denominators would help.

The fact that the accounting callbacks are indicating at the task was completed tells me that it was. But apparently the original submitter does not know this.

If you're using NowWAP, try updating to the v2.56 version at www.nowwap.com. I do recall some HTTP/WAP2 proxy issues involving MMS submissions and unusual chunking boundaries. But I thought in those instances the end result was truncated MMS messages.

-bn