Configuration - Could not retrieve or read configuration (ConfigurationError) - Back again.

[Bag72]

Alex,

I'm back with my problem from a few days ago. Container has been running OK. I had problems before, Issue 355, but fixed by wiping and reinstalling if you remember.

At 0130GMT today the system stopped working again with the same error.
{"type":"log","timestamp":"2023-02-04T01:30:15Z","message":"Configuration - Could not retrieve or read configuration (ConfigurationError)","level":"error"}

Now nothing changed on the container and it has been using the same one for 4 days. Interestingly 4 days ago I had several hours of the same message at around 0700GMT which fixed itself and until I looked through the logs today I hadn't noticed.

At 0130GMT I was fast asleep and have not made any changes to the system. Last time I put it down to something I had done and there was a bit of a mix up with updates. I hadn't been doing them and I thought I was.

So I have a few questions and hopefully it can lead to a fix for me.

What error generates the above message? Unable to read, file doesn't exist, certain parameters not there?
Where is the configuration file stored? I have config/.env and that is w:rwe and still accessible.
When does it read the file? It looks like during every time speedtest is queued.
In the code after reading the config file does the error handling move on? If the next command errors will it use this error instead of one of its own?
Grasping at straws here and as it isn't just me it seems a particularly strange error.

What can I do to help diagnose this?

[alexjustesen]

You're getting this error in your results right?

[Bag72]

When the test is ran, either through the timer or by clicking the run now button it fails and the following is down as the error when you click on the red icon in the GUI.

{"type":"log","timestamp":"2023-02-04T01:30:15Z","message":"Configuration - Could not retrieve or read configuration (ConfigurationError)","level":"error"}

This is the same error I got last time and then I fixed it by deleting everything and reloading. This time that isn't working. I'm now convinced it is something on my system triggering an obscure bug, not something unique to me because others are getting it but definitely present on my system.

I've reloaded the brand new .env from Github, Loaded the new DC and started it up as new. Still no joy. See attachments.

It can clearly see and read the .env file as it filled in the App Key environment variable. So it is clearly in the right place and RW. It is clearly something simple but finding it will be an issue. How can I set it up to step through the code to find exactly where it goes wrong?

docker-compose.txt
env.txt

[cinderblockgames]

Having the same issue. Just started up the container for the first time.

[alexjustesen]

So this error occurs when the speedtest-cli can't reach the server it's trying to test against. Try setting another server under General settings (if you have one set) and if you don't have one set try specifying one.

[cinderblockgames]

I don't have any options in the General list at all.

[alexjustesen]

Thanks I'll look into this, might be a bug on the initial installs.

[Bag72]

I don't have any items in my drop down list to choose as a target. However it was getting some to target originally so they must have come from somewhere. Is there an environment setting you can use if there is none on the list and you can't type one in or alternatively a db setting? Where does it get these from?

[Bag72]

Perhaps a more relevant error would be better as well. I spent hours playing with the configuration files.

[MarcusSi2023]

Just wanted to confirm that I have this exact same issue. [

[kode54]

Same error here too. And the previous deprecated speed test tracker I was using, stopped passing any tests around 4am today.

Does it matter that my host machine is already using port 9000 for Portainer?

Edit: Running speedtest cli from the terminal inside the Docker container:

[2023-02-11 07:09:41.789] [error] Trying to get interface information on non-initialized socket.
[2023-02-11 07:09:46.795] [error] Configuration - Couldn't resolve host name (HostNotFoundException)
[2023-02-11 07:09:46.795] [error] Configuration - Cannot retrieve configuration document (0)
[2023-02-11 07:09:46.801] [error] ConfigurationError - Could not retrieve or read configuration (Configuration)
[2023-02-11 07:09:46.801] [error] ConfigurationError - Could not retrieve or read configuration (Configuration)
{"type":"log","timestamp":"2023-02-11T07:09:46Z","message":"Configuration - Could not retrieve or read configuration (ConfigurationError)","level":"error"}

And I see the problem now. DNS resolving wasn't working with host systemd-resolved. A workaround is to delete the symlink /etc/resolv.conf and replace it with a custom file pointing at the correct DNS server, which systemd-resolved will also use.

Edit 2: I see another possible problem, with my DNS services. My browser was unaffected because of course, Firefox has its own resolver.

[cinderblockgames]

@kode54 Thank you - you made me realize that, while I made a network for speedtest to hit the internet, I forgot to add it to the container. This is working for me now.

[Bag72]

@kode54. I tried editing the file /etc/resolv.conf but it wouldn't let me as it was in use. I stopped the container and again no joy. How did you replace yours? It was pointing to 127.0.0.11 for some reason and that isn't anything like my DNS servers so have no idea where that came from so I tried to edit by copying out and got this;.
So in container resolv.conf
nameserver 127.0.0.11
options ndots:0

docker cp speedtest:/etc/resolv.conf .
edited resolv.conf to point to my DNS server

docker cp resolv.conf speedtest:/etc/resolv.conf
Error response from daemon: Error processing tar file(exit status 1): unlinkat /etc/resolv.conf: device or resource busy

Any ideas?

[kode54]

I didn't replace the resolv.conf inside the container. I had originally changed the one on the host machine. Apparently my actual problem was with DNS services on my router.