Turns out I was getting 500 errors because my instance didn't have permission to upload to the S3 bucket, due to it being a cross-origin request. This was accumulating in the background along with some pre-existing issues caused by relays. Both have now been fixed so my is finally using an appropriate amount of RAM again! 🔧 🐜

My instance is back online and shouldn't go offline again for a long time now that I've moved all the resources to object storage :kirby_happy:

I'm gonna try to move my instance's media to object storage today... this is the only newish writeup I've found about it: github.com/cybrespace/cybrespa It looks thorough and simple but Mastodon is never as easy as it seems 🙄😬😛🤷‍♀️ I'll stop the Mastodon services in 15 minutes so wish me luck while you can 🤞 ...Bye! :tassmuffin:

I haven't been using Mastodon much lately because I'm trying to avoid fixing my technical debt 😒 I have to transfer to object storage because it's cheaper (and update by quite a few versions too) 😛

*looks at federated* guess I have some more gabs to block after work this evening 😑 where's the much-needed regex domain block for mastodon?

Sidekiq is suddenly using way less memory?? that's cool I guess ?_? I hope it stays this way because I dislike using 1.5gb all the time... now it's back to 1.

I'd heard about Mastodon and GNU Social quite a bit by the time I actually bothered checking out the fediverse in August 2018... I regret not looking into it sooner. I wrote it off as a "ghost town" but it turns out this ghost town is perfect for me because it's full of like-minded ghosts who care about free software and the evolution of social media 😁👍 I love this place!

My is on a $10/month Digital Ocean VPS which has 2gb of RAM and 50gb of storage... It's one year old now and is using about 38gb. I try to have a lively federated timeline but I also do extreme pruning of remote media so I can probably get most of another year out of the remaining 12gb... My instance typically uses 1.2gb of RAM

I had followers from dead instances which my Sidekiq was continuously retrying to push my statuses to, and there's also a "development and testing" instance which is in the mastodon.host relay which I've suspended for giving weird errors repeatedly. My RAM usage is down significantly so I'm satisfied now. Some errors are probably inevitable with this many relays 🤷‍♀️

My has been using a bit more RAM than usual lately. The thing using most memory according to top is , which flickers between having 1 busy job and 0. I see there are a bunch of jobs that were retried 20-30 times each (some caused by a relay I think?)... Should I delete these? Is this even a problem? The Sidekiq dashboard is not very informative to someone who barely knows what it's doing 😅

cron uses tootctl but I have the 'unofficial' Masto-cleaning commands in a shell script called 'do_sometimes' which I do sometimes, because I'm afraid to automate_this 🙃

I started sending the output of my daily Mastodon-cleaning cronjob to a log file. So far it looks like tootctl deletes about 1337 remote media per night :mario:

Updated my non-Docker of Mastodon from v2.8.2 to v2.9.2. Custom changes preserved. Born in v2.4.5. All seems well 😀

I definitely would have tried to run my own multiuser Mastodon instance if Mastodon existed 14 years ago… I ran a lot of web forums back then 😛 I also ran a MediaWiki site about card games once, for about a year… No one ever edited it but me although I put lots of work into it for no reason 🤔🙂 Anyway... is all I can responsibly commit to nowadays. But I like the idea of small federated social communities :tassmuffin:

LB is useful to me as a trying to lower my disk usage :)

oops my masto was offline for a day because it ran out of disk space 🙃 I'm too cheap to buy more right now so it may happen again, but I'll eventually fix it when I have time and can justify spending. I'm already aggressively deleting remote media every day. This stability isn't a high priority for me but it's definitely not going to disappear entirely! I like the fediverse and self-hosting too much for that to happen 🙂

I keep fixing my tootctl cronjob for removing remote media and it's somehow broken AGAIN. Ugh. Why does this simple task keep breaking? 😠

I've helped other Mastodon admins fix broken cronjobs too, so I know it's a common issue. The Mastodon project needs more people to help with communication because it's frustrating to keep track compared to some other open source projects. The official docs seem deliberately sparse. Maybe I should make my own posts about troubleshooting?

Easily updated from Mastodon v2.8.0 to v2.8.2 😌

Show more
tassaron dot com

This is Brianna's federated microblog homepage! To follow my posts, find another Mastodon instance and join the fediverse.