AVMS over 3G - unexpected behaviour



#define POLICY “now”
swi_av_asset_PushString(asset, “industrialMonitorJsonData”, POLICY, SWI_AV_TSTAMP_AUTO, data);

This works great with ethernet bearer

But often fails over 3g

the above returns OK but the data is not sent

sometimes I see

Jun 10 15:14:13 swi-mdm9x15 user.err Legato: =ERR= | AirVantage[793]/AirVantage T=main | M3DA-SESSION(default.lua) () 85 | Closed session #70 with error connection_closed
Jun 10 15:14:13 swi-mdm9x15 user.err Legato: =ERR= | AirVantage[793]/AirVantage T=main | SRVCON(srvcon.lua) () 115 | Error while sending data to server: connection_closed

Any ideas?


Hey John,

When it returns OK, it does not mean that the data has been sent successfully. It just means that the API call was successful.

Is your device in a ‘DEPLOYED’ state in AirVantage? When you mentioned that it works great with ethernet bearer, were you able to see the data in AirVantage that you just pushed?

Hi Enoch

I am using policyPtr “now”

Yes the system is communicating with the AVMS server - I can see live data arriving in the following bearer cases:

Over Ethernet
Over 3G roaming at the SW developer day

But it fails in my office - it isn’t a hard fail - the data gets lumped together and occasionally gets uploaded by the AVMS agent all at once

I think that the “now” fails and subsequent "now"s don’t cause the agent to call again - the timed agent calls sometimes trigger all the avms local data to be uploaded clearing the backlog



Hi John,

The issue you are describing has been discovered very recently, and I think you have a good understanding of it.
The issue is located in the module managing the communicatiion with the AVMS server, the fix is already done, and we are working on integrating it in Legato.
We hope to do that as soon as possible.

I just have one question:
what do you mean by “the timed agent calls”?



Hi Laurent

I can’t remember the proper event name - but they are avms timed server callbacks configured from the AVMS backend UI - I think one is called “heartbeat” and the other “status report” - on my system I set the rate to 15 minutes

It could be that the behaviour was just a coincidence and that something else was unjamming the data I didn’t have time to check the target log in more detail - hard to tell as there is a lot going on under the hood :smiley:

Great that a fix is on the way - maybe a bug list thread on the forum could be opened so that others can see open / close / in progress stuff like this - or maybe it should be in git hub?



Hey John,

Currently we have a list of known issues located under the forum thread ‘New Legato Release!’ (legato.io/legato-af/14_04/le … a14_04.htm). But it’s not the easiest information to find, so I will create a new thread on the forum with this information + any additional bugs discovered by users (maybe also with workarounds where applicable).

Thanks for the suggestion!
Enoch :smiley:

Hi Enoch

Maybe git hub:

Could be a better place for bugs / issues?

That way the guys outside the forum would also be able to see / contribute?



Agreed. Github issues is maybe a better place for that.
Although the forum is now open to everyone!