LtU is now running in a new, more stable environment

LtU has experienced a long period of downtime recently. Its software infrastructure was outdated enough that it became difficult to maintain when problems arose. It has now been migrated to a brand new environment. It should be much more stable from now on.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Thank you

Thank you for keeping LtU healthy and fun!

Can you reopen comments?

Because we've been down so long, it's no longer possible to comment on the posts that were made last month even though they are among the more recent posts.

Example?

Can you give an example of an affected post? I'm not aware of a limitation like that, but there could be some data integrity issue that's preventing comments.

Your efforts are appreciated, Anton

Thanks for your work on bringing LTU back from the undead. :-)

Your efforts are appreciated, Anton

Thanks for your work on bringing LTU back from the undead. :-)

Ah! The double post...

... seems to be alive and well. Sorry, all.

How about switching to HTTPS?

How about switching to HTTPS?

HTTPS

Coming soon.

Curious about the new environment

I'm curious what was unstable about the old environment, indeed what it was, and something about the new environment and why it is better.

about the double post

Sorry about the double post. The first time I posted, it came back with a server error (502), so I tried again, this time going through Preview first, then Post. When my post appeared, I saw there were two. It looks like I can't delete the extra one.

Why it’s better

This pretty much captures it: End of the Line.

I’ll post a more serious response over the weekend.

What I did on my summer vacation

The LtU app - a slightly customized old version of Drupal - and its database were containerized during the previous site move, which was nearly 5 years ago. The containers were deployed on a Kubernetes cluster, along with a number of other sites.

Back in those ancient times, there were no major vendors with a managed Kubernetes offering, so the cluster was unmanaged, created using the kOps utility, and it wasn't kept up to date with newer versions.

One issue that occurs semi-regularly for LtU is that bot attacks - often just from overly aggressive crawlers - can cause the database container to run out of resources and crash. In that case, Kubernetes usually restarts the container and everything recovers automatically within a few minutes.

This process worked well for a few years, but at some point, it stopped working seamlessly. Even after manual attempts to recover, there would still be an intermittent but frequent symptom where it seemed like the app was still trying to reach the old crashed container, or something like that. I spent a bit of time looking at it, but soon decided it would be easier to switch to a new cluster than debug that years-old setup.

The new cluster is a fully managed Google Kubernetes Engine cluster running the latest versions of everything, and upgrades of the infrastructure in future will be automatic.

Who is promoting topics?

Anton: Who is deciding what topics go to the front page, and how often are they looking?

"contributing editors"

LtU was traditionally a "group blog" (back when that sort of meant anything to anyone). Stories are posted to the home page, as opposed to the forum, by folks known as contributing editors who specifically have that privilege. The forum is a separate thing from posts on the home page, and I don't remember whether there's any way to promote forum threads directly to the home page, even for editors. Here is a very old discussion of related topics which might shed some light on the mindset operative in those days. Whether any of this is still relevant is anyone's guess.

Short answer: they aren't looking. 😂

(dupe)

(dupe)

(dupe)

(dupe)