nRF9160 thingy91 - AWS IOT (sample): Connect errors -11 and -10

[johnwest]

Hello All,

This is from an app I had working 1 1/2 months ago - when I've tried recently to get it connected again, I haven't been successful.

The latest asset tracker appears to work find and posts data to the nRF Cloud.

I've seen previous posts - DNS is working - or appears to be ok.

I'm getting the following:

===

+CSCON: 0

RRC mode: Idle

Next connection retry in 30 seconds

AWS_IOT_EVT_CONNECTING

+CSCON: 1

RRC mode: Connected

E: getaddrinfo, error -11

E: client_broker_init, error: -10

AWS_IOT_EVT_CONNECTING

+CSCON: 0

RRC mode: Idle

Next connection retry in 30 seconds

AWS_IOT_EVT_CONNECTING

+CSCON: 1

RRC mode: Connected

E: getaddrinfo, error -11

E: client_broker_init, error: -10

AWS_IOT_EVT_CONNECTING

+CSCON: 0

RRC mode: Idle

====

I can reload the asset tracker and it'll work fine.  I'm running the AWS_IOT sample code.

Thanks in advance as to what -11 and -10 mean - it would appear getting the address fails - I tried to find error codes that make sense but not sure and wanted to ask in case something has changed in the last 1 1/2 months.  This app used to run without issue.

Thanks!
jwest

[johnwest]

Didrik,

I guess the root cause here is the prj.conf doesn't appear to get updated after it's changed.

I got the demo running again - but posting to an S3 location that was just for testing.  I changed it - and the tools got the settings stuck again.
I got in the debugger and found the settings 'stuck' in autoconf.h (a name from the unix world, obviously); it kicked off cmake - then the v5.60 Embedded Studio environment promptly crashed.  I've seen that before.  But, it made merged.hex before croaking and I'm able to post now to S3:

SPM: NS image at 0x1c200

SPM: NS MSP at 0x200238d8

SPM: NS reset vector at 0x2231d

SPM: prepare to jump to Non-Secure image.

*** Booting Zephyr OS build v2.6.0-rc1-ncs1  ***

I: Modem library is not yet initialized, AT commands not sent

I: Configuration of MAGPIO and COEX0 is left to drivers

The AWS IoT sample started, version: v1.0.0

Board: thingy91_nrf9160

Host: ABCDEFG-ats.iot.us-west-1.amazonaws.com

Port: 8883

Client ID: my-thing

Security Tag=12345678

Command MaxLen=2700

LTE cell changed: Cell ID: 170924863, Tracking area: 35667

RRC mode: Connected

Network registration status: Connected - home network

Next connection retry in 30 seconds

PSM parameter update: TAU: -1, Active time: -1

AWS_IOT_EVT_CONNECTING

DATE_TIME_OBTAINED_NTP

AWS_IOT_EVT_CONNECTED

Publishing: {	

"state":

{

"reported":

{		

"app_version":	

"v1.0.0",	

"batv":

4394,	

"ts":	

1636008654905

}	

}

} to AWS IoT broker

PSM parameter update: TAU: -1, Active time: 60

AWS_IOT_EVT_READY

AWS_IOT_EVT_DATA_RECEIVED

Data received from AWS IoT console:

Topic: $aws/things/my-thing/shadow/get/accepted

...

So, is Segger supporting Embedded Studio or is that waning at the moment?  The stability of the tools appears tenuous at best.

So, getaddrinfo() was getting an incorrect endpoint because the process to update the endpoints in prj.conf is apparently broke.

This seems to be a rather well-known issue at Nordic - I'm at a loss as to why that wasn't pointed out early on as -11 is a 'fingerprint'

if you will for this error.

Regards,
johnwest

[Didrik Rokhaug]

It's good to hear you got it working.

johnwest said:

So, is Segger supporting Embedded Studio or is that waning at the moment?  The stability of the tools appears tenuous at best.

We recently released a plugin for VS Code, which can be used instead of Segger Embedded Studio.

https://www.nordicsemi.com/Products/Development-tools/nRF-Connect-for-VS-Code/

We will continue to support Segger Embedded Studio, but I recommend you look at the VS Code plugin as it better fits the nRF Connect SDK workflow.

Amongst other things, the VS Code plugin will automatically detect changes to the prj.conf file, which SES isn't able to.