Updating Discourse at community.coops.tech

The latest Discourse update needs to be run on the command line rather than via the web admin interface, so I’m about to do a backup and then run the playbook to upgrade it and hopefully I’ll be able to post a comment in this thread in a moment to say it has completed without an issue…

The Discourse upgrade appears to have run without any issues :slight_smile:.

3 Likes

That’s great! Thanks, Chris :slight_smile:

There is a new version of Docker CE out (but the release notes haven’t been updated yet nor has the source document for them), but I have upgraded another server and it doesn’t appear to have caused any problems so I’ll upgrade this one in a moment, it will result in the site being unavailable for up to 5 minutes as this playbook is run.

The upgrade appears to have been completed without issues :slight_smile:. I’m now going to do the same for the Nextcloud server (Collabora Online Development Edition, which provides the WYSIWYG editing is running in a Docker container).

Thanks @chris - putting the rest of us to shame.

No problem, it turns out that it was a minor update, to Docker version 17.09.1-ce, build 19e2cf6 — I was assuming above that it was going to be 17.12.0-ce…

I’m about to update this Discourse site from v1.9.0.beta15 to 1.9.0.beta16, there will be a little downtime as a result of this.

The Docker release notes from the upgrade the other day are now available and a couple of issues fixed were security vulnerabilities.

I’m about to update this Discourse site from v1.9.0.beta16+1 to 1.9.0.beta17, there will be a little downtime as a result of this.

We are now at v1.9.0.beta17 +38 and everything appears to be working fine… have a great holiday everyone :slight_smile: :christmas_tree: :clinking_glasses:

1 Like

The last update, on 28th December appears to have resulted in a higher load and the odd 500 error, see:

https://meta.discourse.org/t/higher-load-and-500-internal-servers-errors-since-28th-dec-2017-upgrade/76970

I got a 500 error that disappeared eventually on reload.

Btw @chris, did you know you can set a banner for letting users know about downtime. Just search for “emergency” in the admin settings. :sunglasses:

That is because I restarted the container, see this thread for more details regarding the problems.

Thanks, I have added one.

So it appears that the issues with this server have been solved with a reboot (it had been up 100 days) — perhaps the latest version of Docker CE (Docker version 17.12.0-ce, build c97c6d6) for which the release notes have yet to be published have a fix related to a recent Kernel, but this is just guesswork — I don’t know the cause of the problems, only the symptoms and these have now gone.

Sorry for the downtime just now, I was upgrading Discourse — it wasn’t an accident :sunglasses:.