In case you missed it, I have an update in this post. TL;DR we are losing our host and will need to move to a VPS.

It seems pretty clear that once we have to start paying for storage, it will be very expensive to keep everything on hard drive storage, so we will need to move to storing images on object storage.

Unless anyone has any objections, instead of doing everything in one go, I would like to do the migration to object storage for images in advance of the main move to a VPS in order to better assess what we will need for storage once that part has been offloaded. The plan is to use Backblaze B2 storage, because it's cheap, reliable, and already used for our backups.

This means we will need two patches of downtime, most likely this weekend and next as we need to be off the current host by the end of the month.

This post is to propose an outage from this Friday after 8pm, through to whenever it is finished on Saturday. I have been doing some testing, and it seems it will be a bit flaky if I try to do the migration too fast. My plan is to set it going the night before, run it overnight, and start Lemmy back up in the morning assuming the migration has finished by then. But there is every chance that it just stops while I'm sleeping and needs to be restarted in the morning, taking most of Saturday.

My current plan is to keep Lemmy itself running during this time, meaning there will be no images but federation will continue so we shouldn't need to catch up after it's done.

TL;DR

• Outage Friday after 8pm through to when it's done on Saturday
• Most likely Lemmy text content will remain available but no guarantees
• Moving image storage to Backblaze B2 - please tell me ASAP if there are objections (American company, but only real alternatives are Cloudflare or big tech and American anyway)
• A further outage will be required the following weekend to move the actual Lemmy service