cloud_connect failed: -111

[Terje K Wahl]

Hi, the thread at https://devzone.nordicsemi.com/f/nordic-q-a/48859/nrf9160-dk-connecting-problems-with-nrf-cloud/211627 has been very useful for me to get past several problems getting the asset tracker example to work. I connect successfully to a NB-IoT network here in Norway now, thanks to (as explained in the linked thread) setting modem mode to NB-IoT, using legacy PCO mode and using a static IP address since DNS is not supported.

Finally, the problem I am left with is "cloud_connect failed: -111". I had high hopes that a certificate update (devzone.nordicsemi.com/.../nrf-cloud-certificate-update) would solve this, but after successfully updating my certificates the same problem still persists. 

I have a nRF9160 DK v 0.8.5, with modem FW updated. I can successfully connect to the network using the at_client sample. Since the SIM included with the dev. kit only supports LTE-M1 (no coverage where I am located), I am using a SIM card from com4.no to connect to NB-IoT.

Here is my output from the terminal:

***** Booting Zephyr OS build v1.14.99-ncs3-snapshot2-2647-gd6e67554cfeb *****

[00:00:00.378,173] 

[0m<dbg> nrf9160_gps.init: MAGPIO set: AT%XMAGPIO=1,0,0,1,1,1574,1577[0m

Asset tracker started

[00:00:00.389,282] 

[0m<dbg> nrf_cloud_transport.nct_client_id_get: client_id = nrf-352656100223414[0m

[00:00:00.398,498] 

[0m<dbg> nrf_cloud_transport.nct_topics_populate: shadow_base_topic: $aws/things/nrf-352656100223414/shadow[0m

[00:00:00.410,156] 

[0m<dbg> nrf_cloud_transport.nct_topics_populate: accepted_topic: nrf-352656100223414/shadow/get/accepted[0m

[00:00:00.421,661] 

[0m<dbg> nrf_cloud_transport.nct_topics_populate: rejected_topic: $aws/things/nrf-352656100223414/shadow/get/rejected[0m

[00:00:00.434,204] 

[0m<dbg> nrf_cloud_transport.nct_topics_populate: update_delta_topic: $aws/things/nrf-352656100223414/shadow/update/delta[0m

[00:00:00.447,082] 

[0m<dbg> nrf_cloud_transport.nct_topics_populate: update_topic: $aws/things/nrf-352656100223414/shadow/update[0m

[00:00:00.458,923] 

[0m<dbg> nrf_cloud_transport.nct_topics_populate: shadow_get_topic: $aws/things/nrf-352656100223414/shadow/get[0m

Connecting to LTE network. This may take several minutes.

[00:00:00.477,172] 

[0m<inf> lte_lc: Using legacy LTE PCO mode...[0m

[00:00:00.483,428] 

[0m<dbg> lte_lc.w_lte_lc_connect: Network mode: AT%XSYSTEMMODE=0,1,1,0[0m

[00:00:04.653,717] 

[0m<dbg> lte_lc.at_handler: recv: +CEREG: 2,"0CEB","0202791C",9,0,0,"11100000","11100000"[0m

[00:00:04.664,123] 

[0m<dbg> lte_lc.parse_nw_reg_status: Network registration status: 2[0m

[00:00:05.574,218] 

[0m<dbg> lte_lc.at_handler: recv: +CEREG: 1,"0CEB","0202791C",9,,,"11100000","11100000"[0m

[00:00:05.584,472] 

[0m<dbg> lte_lc.parse_nw_reg_status: Network registration status: 1[0m

Connected to LTE network

[00:00:05.595,123] 

[0m<dbg> nrf_cloud_transport.nct_connect: IPv4 Address 52.54.149.10[0m

cloud_connect failed: -111

LTE link disconnect

Shutdown modem

***** Booting Zephyr OS build v1.14.99-ncs3-snapshot2-2647-gd6e67554cfeb *****

...

Any other suggestions would be greatly appreciated.

Kind regards,

Terje

[Didrik Rokhaug]

Hi.

The mqtt_socket library merges all errors on connecting to -111 (connection refused).

However, the reason for your failure to connect might be other things than just some issue with the certificates.

Can you add a print inside the "if block" ncs/nrf/subsys/net/lib/mqtt_socket/mqtt.c:322 that prints the value of err_code?

AWS (which nRF Cloud) is built upon uses load balancing, so the IP address of the endpoint you are looking for might change without warning.

Best regards,

Didrik

[Terje K Wahl]

Hi, thank you for the feedback.

Added a print of the error code (and renewed the IP address to connect to).

In mqtt_connect(): client_connect() returned error code -116

This seems to indicate a connection timeout, which might make sense if connecting via a higher latency NB-IoT network (and not LTE-M1).

I'm now trying to figure out how to increase timeout values, but there is a lot of code to read... Do you know how to increase timeout values across the board (socket timeout, AWS timeout and possibly other timeouts)?

[Didrik Rokhaug]

I suspect that you are not getting a connection to the AWS server. Due to load balancing, the IP address you use might not be the correct one anymore.

When it comes to finding configuration options, menuconfig or the corresponding tool in SES (Project -> Configure nRF Connect SDK Project) can be quite handy.

However, I do not believe there is a single option to extend all kinds of timeouts.

[Terje K Wahl]

Hi,

The IP address was refreshed (using https://whatismyipaddress.com/hostname-ip) just before retrying.

I searched the settings for "timeout" and increased NET_CONFIG_INIT_TIMEOUT to 60 seconds, and increased NET_SOCKETS_CONNECT_TIMEOUT to 60000 ms.

The results are still the same. Timeout after around 15-20 seconds after calling nrf_connect() from within nrf91_socket_offload_connect() inside the file nrf91_sockets.c.

Do you know how I can increase the timeout for TLS handshake/connection?