Volumes are effectively extra hard disks that you can attach to a droplet. In DO admin console, click the droplet to go to the primary droplet management page, then on the left menu click Volumes and you can see if the volume is attached to the droplet. If it’s not, then AFAIK it’s not really doing anything.
Spaces are a little different, that’s more like AWS S3 file storage. Discourse might be configured to use that for file uploads or for backups?
but why would 250GB not be sufficient for a backup? Is it because the backup tar file has to be created locally before it can be moved to the S3 location, and there’s not enough room for that locally even if there would be sufficient room to keep it on S3?
So it looks like the primary disk has 160GB, of which 109GB is currently used. The mounted volume is at /dev/sda, and shows 36GB of 100GB being used. It looks like virtually all of that 36GB is a single backup file from 2022. I wonder if I should be pointing the backups there, but I’m not sure how to do that. In the Discourse admin console, it offers the Backup location choice to be either S3 (current choice) or Local. But it doesn’t seem to give an option for where that Local location would be. I think this conversation describes what I want to do, but I don’t fully get the solution:
I dont know how formal it should be but i would like to have some sort of “meet here if lost” idea. I went to check the UP twitter but ofc id forgotten elon went nazi
I don’t know anything about the software that runs the forum, but isn’t there a way to stop backups from growing forever, for example by deleting older backups? There is no real need to have more than one “snapshot”, no?
The current setting is for backups to run every 7 days, with a maximum of 5 (so older ones are automatically deleted). The problem is that the backups are not being successfully run (or at least I can’t find them anywhere), presumably for the same reason that I can’t manually run a backup - the file size is too large when including uploaded images.
I’m going to disable the “Backup with uploads” option to confirm that’s the issue.
i’m imagining bluesky is so small that our site going down would make a huge impact on their traffic and we would all recognize each other instantly. but you probably meant just like let’s get eachother’s names and follow eachother now, which is smart, it’s a good idea
honestly might be overkill. I had little idea what happened and was kind of worried the site had been pwned, but a site like this having an outage isnt the end of the world, as long as its recoverable. as long as the database is intact and not corrupted, you can always rebuild the site (although it sounds like we may lose stuff like uploaded images which would create a minor tragedy that existed on 2p2 when all the image hyperlinks in old threads disappeared)
I’ve had a not as lucrative as it should be side business for a little while consulting and pitching self hosted discourse solutions and the one thing I am always impressed by is how idiot proof the self hosting is, even if you dont know what you’re doing. the fact that this site has had less than 2 total days downtime in 6 years is an impressive stat.
Maybe something like https://uptimerobot.com? Sounds like it can email folks if a website goes down. At least that way a larger group would hear about it sooner than trying to log in? Could probably set up a free email list someplace for people who wanted to be notified of such a thing.