Here's a picture of one of my XT clones that I recently got working.

It's a sharp pc-7100, and it has a dead HDD, but it is beautiful.

I'm joking, of course, but in all seriousness: If any other want to help me out with this, I'd be really appreciative (and willing to compensate financially.)

This week especially, I'm just not going to have the time to get things cleaned up and repaired the way that I want. (I'll explain why in another post, under a couple of CWs.)

Fixing this is pretty simple. I could just blow out everything I have, take a fresh database dump, and do it all again, but install it the right way this time.

I'm just not sure when I'm going to have the time to do it.

death; sad; travel; family; work; stress; Show more

death; sad; travel; family; work; stress; Show more

death; sad; travel; family; work; stress; Show more

First I tried to go straight to 2.8 (this doesn't work, because the DB can't migrate over that many versions.) I did this following the current install guide.

Then, I decided I'd just install the next version after the one we're on, and migrate the DB then (the right way, right?)

Except I decided to do it in place, and check out the version we're going to.

This doesn't work, because the install guide installs a bunch of pre-reqs for later versions, so I've got pre-req mismatches

Details of our situation:
We're running in Docker. Well, specifically, we're running via Nanobox, and that's at least part of the problem.

Sidkiq is not working.

I have a database backup, and our secrets saved.

We're a couple of versions behind, which is what is causing the biggest portion of our problems.

I have another server, and I was going to migrate everything over. Here's where I fucked up:

I'm joking, of course, but in all seriousness: If any other want to help me out with this, I'd be really appreciative (and willing to compensate financially.)

This week especially, I'm just not going to have the time to get things cleaned up and repaired the way that I want. (I'll explain why in another post, under a couple of CWs.)

(if anyone is curious about how/why I fucked up today, I'll be happy to explain it all to you, on the condition that you help me fix it. :-P )

Alright, things are still broken, but I mostly know why and I have a plan to fix them.

It's just 2AM again, so Retro is going to be dead at least until tomorrow.

One of the bits that I'm excited about, if I can get everything else working tonight, is a new database backup strategy.

If everything goes according to plan, the server will be much easier to maintain in the future.

Okay, retro.social is still down, but this time it's mostly intentional!

I just finished dumping the database, and I'm copying it over to the new server.

Alright, I'm installing mastodon on bare hardware, rather than in a VM in a container.

It'll take me a while to get everything set up, and then I have to migrate all our data over, and then update, upgrade, update, upgrade.

Then swap DNS and everything will be fine.

The thing that is confusing to me about the errors that I'm receiving on retro.social is that I'm getting 503/service unavailable errors for servers that are clearly up.

So something about the way my container is trying to connect to them is resulting not in a 404, but in a 503?

I can't think of any reason that would happen, and I don't know redis/sidekiq well enough to comment further.

Outbound federation on Retro Social still appears to be screwy.

I'm trying to find a short term solution.

There will be some down time tomorrow while I enact the long term solution (which is moving us out of docker, and on to hardware.)

Inbound posts are still showing up, and showing up at a reasonable time (aside from the fact that the sidekiq process keeps crashing, leaving gaps of a few hours between when it crashes and when I can fix it.)

I think the solution here is to do a fresh install of 2.8 and migrate our data over to that, but if anyone has any faster suggestions, I'd love to hear them.

Alright, so Retro.Social is still having major issues. This is my backup/alt account.

I'm not going to be home most of today, so I'm not going to be able to try again to fix it until late tonight or early tomorrow.

The gist, posts do not appear to be federating outbound.

I'm seeing thousands of errors like this in the sidekiq queue Mastodon::UnexpectedResponseError: Delivery failed for https://[someserver]/inbox: https://[someserver].cat/inbox returned code 503

Like, 15k of those last night.

Anybody in the Reston, VA, area or nearby (on public transit lines) got a couch or other crash space I might be able to use until I get a paycheck and can get a space of my own? I'm up for a job there, find out Mon or Tues if I got it, but am dead broke right now. I'm quiet, nonsmoking (allergic to cigarettes actually), hygienic, and well practiced at staying out of people's way in situations like this.

Show more
Cybrespace

cybrespace: the social hub of the information superhighway

jack in to the mastodon fediverse today and surf the dataflow through our cybrepunk, slightly glitchy web portal