Time server seems not reachable via address only via IP

[VincentVerweij]

The Problem

I have two ESP32 devices which I updated today from 10.5.2 to 12.0.1, it all works fine. But as of that 10.5.2 version one of these two devices has had the Unix start datetime as a value. I did enter the Expert mode for the configuration and entered two different timeservers: time.google.com and nl.pool.ntp.org. After that did not work, I did enter undefined and unchecked the box for TimeServer, but no results.

Today after updating to the latest release, I still saw the issue. In issue #758 I did see that putting in an IP could help, so I tracert nl.pool.ntp.org which gave me 93.94.224.67. And that does seem to help, so it must be something with the DNS not properly resolving.

I don't understand why one device works without touching the TimeServer and the other doesn't. I am able to connect to the device via a static IP and it sends out MQTT which arrives at my Home Assistant, so it able to communicate with the outside world (albeit my local network).

I did see in the release note of v10.6.0 and v10.6.1 that there was a bullet with: Bug Fix: no time with fixed IP, Postprocessing, MQTT, could this be related?

The thing is, if they change the IP for whatever reason then I would have to go in and change it, so it works now but it's not ideal.

Any suggestions on how to troubleshoot or fix? Thanks in advance!

Version

v12.0.1 (Rev: dfe2c46)

Expected Behavior

Proper time syncing with the timeserver when defined as address and not IP as my other ESP seems to do well.

Screenshots

No response

Additional Context

No response

[caco3]

if it works by IP but not by hostname, then it also could be that the device does not find a DNS server or does not get a response from it.

[VincentVerweij]

Do you perhaps have an approach on how to verify which of the two it is that is causing the issue? I probably need to SSH into the ESP, right?

[caco3]

It does not have an SSH server. But it provides a log through USB. I have seen on my systems that it needs more than the 10 retries and thus et its time to 1. January 1970. I added an issue for this: https://github.com/jomjol/AI-on-the-edge-device/issues/1292 Maybe that was also the issue on your side? A restart would have fixed this.

[VincentVerweij]

Nice that you already noticed the retries that might cause an issue 😄 And with a restart I assume that it's a hard one (as in cut the power), not the reboot button that is available in the software?

[caco3]

And with a restart I assume that it's a hard one (as in cut the power), not the reboot button that is available in the software?

No, a soft reset through the REST API (resp. UI) should be fine.

[VincentVerweij]

Ok, then I can confirm that a soft reset does not work. I rebooted multiple times while upgrading from 10.5.2 to 12.0.1, since I installed all intermediate releases too.

[caco3]

can you show the USB logs of the bad or both systems? Only the startup until the ======================== Main Started ============================ line is relevant

[VincentVerweij]

Sorry for not getting back to you yet. Is that the logs that can be found under System > Logs? Or do I need to attach it to my computer to get the so-called USB logs?

[caco3]

The ones from System > Logs are ok. Make sure it contains the startup sequence as that window only shows the newest 80kB of the log

[VincentVerweij]

Hi @caco3, sorry to let you wait. I will try to post the logs in the coming days. Just to let you know that I did not forget about this.

[caco3]

the next release will contain an improved NTP handling.

You can already play with it with the rolling build.

[github-actions[bot]]

You can use the latest Automatic Build of the the rolling branch. It might already contain a fix for your issue. Pick the most top passing entry (it has a green circle with a tick in it), then scoll down to the Artifacts and download the file named update_*.