Il mio webhost è inaffidabile o Discourse in Docker è veramente instabile?

I run a Discourse instance on something that can be described as a semi-managed docker container. Every few months the container doesn’t resolve any DNS names, causing external images to fail to upload and e-mails fail to send. After a reboot it’s fine again.

Now the webhost says that this is a known problem with Discourse in Docker. I have a hard time believing it since there are hardly any topics about this kind of problems here or on Github.

I’m looking for experiences of others running the Discourse Docker container. Do you recognize this problem? Do you have to reboot every few months? Basically double checking if the webhost is talking nonsense, or is this a known problem with Discourse in Docker?

2 Mi Piace

If you are using official Discourse docker installation - there is no such problem as far as I know.

It is the only best way to keep things stable and get deployment help here.

4 Mi Piace

And please - check
your-domain/logs/
to see if there are any related errors

2 Mi Piace

Thanks @4ong. The installation was done using the official Discourse docker installation.

2 Mi Piace

If there will be no related errors in /logs that I’ve proposed to check - one additional question. Do you use any kind of CDN? CloudFlare f.e.

2 Mi Piace

Yes there are errors. They are all:

Job exception: getaddrinfo: Temporary failure in name resolution

That’s why I know the DNS lookups are failing. This is a returning problem that disappears after a reboot of the server that runs the container. I suspect an intermittent network problem at our webhost. It’s a pretty plain setup, nothing behind Cloudflare or a CDN.

4 Mi Piace

Indeed, it looks that way. At least I have no additional ideas.

2 Mi Piace

That’s what it sounds like to me, though it’s odd that it would get fixed with a reboot.

I edited the title to make it a bit more polite. :wink:

4 Mi Piace