Getting the volunteers back together! Time for another database migration on Unicorn!

Hi friends. Continuing the discussion from Keeping this Forum Updated:

Morning @bfb @the @fnord if you all have a little time this weekend let’s jump into tmux and do #unicorn maintenance! It was both fun and fruitful on our last database migration. We could meet on https://jitsi.noisebridge.io/unicorn

What are we addressing?

Upgrade to PostgreSQL 13 from 12. We did this migration last time together, too!!

  • v12 database backup will need to be stored externally since we are at 92% disk capacity.
  • Perhaps we can reclaim some disk space on the server.

Discourse rebuild will fix Follow plugin, which is causing errors. Those bugs were fixed upstream.
Rebuild will remove Babble chat plugin as it is broken, causing errors, and unmaintained. :frowning: Too bad, it was cool.
As a bonus, we could setup some simple gallery tool for photos as Tyler requested over on Good photo album software for Noisebridge?

This data migration should address the majority of our paper cut forum errors impacting performance. :+1:

So, I just rebuilt the forum like a complete lunatic. All seems to be okay… please do check my work if you feel inclined. I’m confused because it seems we are already on PostgreSQL 13… which is confusing me.

1 Like

Awesome I see things are up. I’m seeing what looks like some weird loads on the server. There a bunch of random threads firing under “rando” account and mine as well. I’m not running any persistent services that I’m aware of.

Time for a general system update and reboot?

Also hanging on unicorn jitsi if anymore support is needed, @fnord is catching up as well.

1 Like

P.S. didn’t see any active tmux, so have one spin up split with a nano text and perf. “tmux a” on noisebridge account.

What are we doing? This is a wiki post; feel free to edit, expand, and improve it!

In Space

“bunker” server, currently cia.noise (Windows 10), running virtualized Ubuntu 20 jitsi.noise. 32GB Ram TB+ SDD, shit load of thread ripper cores.

“hackitorium” server(s) temple.noise (Windows 10) looking to migrate to Arch, currently has backup of previous Windows version of devil.noise, that server is now running Arch and hosting live streaming to YouTube of Flaschen Taschen.

Upstairs there’s faa.noise running Capp St. Driveway jitsi webcam, unkown.noise is jitsi enabled workstation in the music room.

Trying to build and maintain specs here https://www.noisebridge.net/wiki/Computers

nsa host extravm running jitsi in Utah on 200gb

Installed VBox on temple.noise, setting up NextCloud ova (VM).

On ssdnodes

Unicorn - 24gb ram, 120gb ssd, 12tb traffic in Dallas
u0 - 32gb ram, 480gb ssd, 15tb traffic in Los Angeles

What would help James?

  • Need multisite yml file forum so we can avoid future downtime when rebuilding the forum.

Backups for our infrastructure

  • All of our servers need backups in multiple locations.
  • Operation Backup - Let’s Try some experiments

Where do we keep these backups?

  • at Noisebridge on some machine.
  • James could host at least some backups on personal hardware.

SSBC is wondering if we would host a Pub channel for them. A nice prospect and project for the space if we want to.

Weirdness, what’s running in the background (Unicorn)ˇ
this gem of a thread:
x> /bin/sh -c node -e “console.log(JSON.stringify(’$EXTERNAL_SCRIPTS’.split(’,’)))” > external-scripts.json && ?nmp install $(node-red blah blah blah…

npm stuff is possibly related to our Secure Scuttlebutt Pub channel, which is seeing quite a bit of use.