Ice Rock @IceRock@glaceon.social

Hey, uh, we're back? I guess?

bundle decided to eat all the goddamn memory and basically DOSed the whole instance.

Threw a reboot at it and am currently praying that it doesn't happen again?

Slightly more detail on my backup account, where you'll find me whenever g.s is down: https://chitter.xyz/@monorail/101115748636089866

Everything seems to have gone well! Just a couple more seconds of downtime and we'll be fully updated! Hang tight!

Hey, everyone! Expect some brief downtime in the next half hour or so as I reboot the server in a desperate attempt to free up enough memory to precompile assets.

If that doesn't work you can expect some longer downtime while I stop the instance entirely so that the assets can precompile.

lb: someone remind me when i get home

You may be asking yourself: Why the downtime? Well, here's the deets: It was fun! I just felt like it.

...nah, I'm kidding.

I wrote a patch that slightly modifies the behaviour when you reply to a toot with a content warning. Recently, an update made it so that this would stick "re:" onto the beginning of the default CW (unless it was already there). This is still mostly true. However, it's now omitted if you're replying to your /own/ toot.

The logic here is that if you reply to your own toot, you're very likely writing a thread, and whatever content warning you applied is very likely to still be relevant. It also makes the behaviour consistent with thread mode on the web UI.

That's it! Small change but I think it'll be nice to have.

We are back! Thanks for holding tight.

Downtime incoming! Weeeee

Anyone else on this instance active right now? *checks local timeline* No? Great!

In about 20 minutes (that's 11:45 EST for those of you playing along at home) we'll have some SCHEDULED DOWNTIME! 🎈 🎉

I expect the downtime to last maybe 5 seconds. On the off chance that I broke something, it may take up to... let's say half an hour while I furiously google git commands.

Wish me luck!

Hey everyone! The character limit for toots on this instance has been bumped up to 1000! Having it at 471 was a pretty good joke but I don't think it was worth it.

I'm planning on making anything longer than X characters get automatically CW'd with "long" but haven't done it yet. Don't abuse this. :P

hey heads up

5 seconds or so of downtime incoming

Hey everyone! I have good news!

A while ago, @Siphonay gave me a copy of all the donphan.social emojis (other than the Mutant Standard ones) and a script to import them into this instance. Unfortunately, through no fault of his, it didn't work.

For a little while now, Mastodon itself has had a tool for mass-importing custom emojis. I could never get it working, either. Until today!

This is a longwinded way of saying: We have tons of pokemon emojis now! Have fun!

That's all for now. Ice Rock, /out/. ​

The second is that in, oh, let's say T minus four seconds, we'll be joining @Sir_Boops's relay, yiff.social! If your federated timeline speeds up significantly, that'll be why. This isn't necessarily permanent, but it sounds like it could be fun!

I'm actively seeking feedback with this. I don't know how it's going to work out. If it's too much for you, or you're not a fan of the content you're seeing, /please/ let me know.

Hello, everyone! Two quick announcements.

The first is that the instance blocklist has been updated. We're up to date with the current version of dzuk's list, and raki.social has been suspended as well.

We're officially back, and running glitch-soc! If you still see an error when you visit https://glaceon.social there's nothing to be worried about (I'm actually in the same boat, I'm using data on my phone just to make it work). It just means the name server you use to look up websites hasn't updated yet, give it some time and everything will work itself out.

Guess who got glitch-soc working!!!

All that's left now is to move over the database and point the domain name at the other server. We'll be going down soon, and unless I manage to seriously screw it up we'll be back as soon as the DNS changes propagate.

This still doesn't mean anything will be lost! I've ensured that a database backup was loaded successfully, so that process does work. The last thing that I'll do before changing the IP address that the domain name points to is make a new backup shut down the instance so no one accidentally toots, restore the backup under glitch-soc and get everything running over there. Other than the brief downtime, it should be totally seamless.

Quick update about this. You may have noticed that we're not running glitch-soc.

Well, surprise! It turns out computers are hard. Spent hours tearing my fur out over it and nearly got it working, but I'm running into a roadblock that I just don't understand.

Don't worry, I'll keep at it. But I've spun up a seperate server to tinker with, so glaceon.social can stay up in the meantime. When the "new" instance is up and running, I'll move everything over at once.

glaceon.social is temporarily shutting down now! If all goes well, we'll be back very soon.

See you on the other side.