So — the notifications account for this blog was on mstdn.party and the movie account was on mstdn.plus, of course, for maximum fun. So I’ve moved those to other servers (that no doubt will also have problems soon). I guess the only way to cut down on this type of fun is to run your own instance? But I’m so fed up with sysadmining…
You can follow updates to this blog on @email@example.com and the movie blog on @MovPic23@toot.community. I chose these servers the same way I chose the previous two: They were listed on joinmastodon.org as supposedly well-run servers that aren’t going to go away soon.
Perhaps I can just take this opportunity to do some unstructured kvetching about the whole Fediverse/Mastodon thing? Sure? Sure.
It’s like they’ve managed to replicate all the drawbacks of having a distributed system without any of the advantages. If a server just goes away like *snap*, all your account’s data is gone: People you’re following, who’s following you, etc, but most importantly: All your toots are basically gone. That is, they may be cached somewhere until the cache is gone, but basically gone, long term.
So: It’s not a distributed system. It’s just a system that maximises the number of single points of failure.
If your server hasn’t *snapped*, you can export some of the data. Look at this howto:
This is the official way to do this; it’s not the natterings of some know-nothing on the web somewhere, like this blog is. How is it possible for somebody to type that howto without going “hang on, this is insane. We’ve got computers. Computers are good at doing stuff like shovelling data from one computer to another. Can’t the computers do this?”
You have to do one thing on one account, then wait five minutes, then another thing on another account, then wait a bit, then download five CSV files, then upload five CSV files.
And… after doing all this, my old, so so valuable toots from the old account aren’t copied over to the new account, of course, because… because… I can’t even come up with a joke to explain this, because it’s just unfathomable to me how somebody could even come up with a system where you can’t replicate the fucking posts.
Perhaps it was just difficult? They couldn’t come up with a flag like “this is an old post; not need to re-replicate, just copy” so that it doesn’t get re-federated?
It’s just… it’s just…
And another thing! (I swear I haven’t had a drop of alcohol!) If I now go to an account on the old server, I can’t read anything there. I’m just redirected to the account settings:
To read any toots on mstdn.party, I have to open an incognito window.
It’s just… it’s just…
I conclusion: Elno has nothing to fear from Mastodon.
Oh, and speaking of Elno! Like I predicted in the previous post he’s now changed his mind about API access for posting, anyway. The fun thing about this isn’t that this reveals that he’s an asshat (we already knew that), but that he now has absolutely nobody around him that can give push-back on even the most trivial thing (“well, Mr. Musk, sir, Twitter has many fun automatic accounts — should we do a carve-out for those? Sir? Mr. Musk? Why are you calling security? Why did my access card suddenly explode? Sir? SIR?!”). They’re totally schtumm.
But perhaps it’s because they enjoy seeing him making an ass of himself, and not because they’re afraid to get fired. Sounds more likely.
New Mastodon accounts.
Now I’m getting a drink.
One thought on “Mastodon Is Going Great”
That’s why, at the time, I chose the main Mastodon developer instance. If that is gone, everything is gone. 😅
Anyway, followers migration should be automatic now , as long as the origin instance remains active), but for posts migration, well, there is a GiHub issue on it opened long time ago. From what I see, it is finally taking some steam. But who know? I think it is really crazy that moving your data between instances was not a feature from day one. 🤷♂️