submitted 11 months ago by [email protected] to c/[email protected]

7 comments fedilink hide all child comments

As you might have noticed, we've been experiencing repeated crashes and downtimes for the last 12 hours or so. This was because, after four long and tireless months, our storage drive had finally filled its 50GB capacity.

I noticed this today, at around 9:00 UTC and quickly took action. Fortunately we had already prepared a different partition with an additional 15GB ready for situations of this kind. It took me some additional 40 minutes of tinkering to move the database to said partition, but after that the instance was back online without any majour issues.

This is our current storage breakdown:

Secondary partition (currently: 10 GB; max capacity: 15GB)
- Database (postgres): 10GB
Main partition (currently: 38GB; max capacity: 50GB)
- Image hosting (pict-rs): 23GB
- Lemmy executables, other services (Kaleidoscope, AutoMod, Flair) and operating system: 15GB

Of these, the only ones expected to grow are the database and the image hosting, with the latter being by far the fastest growing.

In the future we are considering moving pict-rs to a separate, more cost effective storage, however for the time being this should hold. We apologize for the disservice in the last hours.

Cheers and stay Based!

top 7 comments

sorted by: hot top controversial new old

[-] [email protected] 3 points 10 months ago

Ew. Cringe and <99% uptime pilled

[-] [email protected] 2 points 10 months ago

I bet you feel real stupid right now

Source: https://uptime.lemmings.world/lemmy.basedcount.com.json

[-] [email protected] 3 points 10 months ago

Oh, that's pretty cool! Thanks for reminding me how great a webmaster I am 😜

[-] [email protected] 2 points 10 months ago

a different comment but it seems like text-only federation scales easier (not very helpful to think of for this instance which is often based on meme sharing I'd guess)

[-] [email protected] 1 points 10 months ago

Not too sure about that. A good chunk of all the images that end up filling our image hosting are cached images from other instances that reach us through federation. All images from all comunities on other instances that have at least one of our users subscribed to them get cached by the image host. Month after month, it fills up quite rapidly.

So the only solution to that would be either not federating at all, or doing so only with text-only communities. Of course none of these are really feasible.

[-] [email protected] 2 points 10 months ago

only with text-only communities

if it wasn't clear this is basically what I meant, no images at all, even from other instances; it would require a kind of change of software perhaps (I am suggesting a totally different "low tech" social medium setup). Text is much more sustainable than images or sound, which is more sustainable than video

[-] [email protected] 1 points 10 months ago

Ah got it. No, I didn't undestand your point.

Yeah it would be an interesting medium but radically different from what we currently have. And it simply wouldn't be a good social media, compared to the "competition", text-only means very little user engagement. Pretty sure the average user would look at maybe two or three posts, then get bored and move to Instagram. Or worse, TikTok.

this post was submitted on 27 Oct 2023

5 points (100.0% liked)

Announcements and Meta

100 readers

1 users here now

A community reserved for communications from the Based Count admin team to its users.

Occasionally we will also hold community polling in here. Make sure to subscribe to not miss any updates.

Everyone is welcome to join the discussion! We accept suggestions on how to run the instance from everybody, including users from other instances.

Posting to this community is restricted to the admin team to avoid spam, however feel free to speak your mind about the status of the instance and how you'd improve it in our [email protected] community.

founded 1 year ago

MODERATORS

[email protected]