My AWS free tier is about to run out, and I’ve noticed some issues with my current setup, so I’m looking for tips on how I should fix it. Any help is appreciated!
My main issues right now are:
-
Cloud sync tasks consistently fail due to hash mismatches. I assume this is because I’m not using the snapshot option due to nested datasets. Seems to only be a few files each time, but still should probably fix this.
-
My AWS costs seem to be higher than I’m prepared to actually pay. I’m using Glacier Flexible Retrieval, and most of the cost seems to be from 450,000 PutObject requests. I’m thinking of switching to another provider, but I’m having some trubble understanding what the costs would be with my usage,
About my setup:
-
The TrueNAS server (Scale 25.10.3) is mostly used for photo storage through Immich and various file backups plus various smaller services.
-
I have a pool of 2 8TB HDDs in a mirror configuration.
-
I also keep a Linux VM running that I can remote into as well as occasionally running Minecraft servers. Both of these happen on a separate pool of 1 SSD, which is replicated to the main pool once per day.
-
Availability isn’t super important to me; it’s fine if it takes a few days to restore cloud backups, as it’ll anyway take time to fix whatever hardware issue I had.
-
Right now I have about 1 TB of data stored. I’d like to keep cloud backups of at least a few versions.
-
I have one or more datasets per service, in total about 15 of them.
-
I have the cloud sync task configured as PUSH, SYNC, Snapshot off (because nested datasets) and storage class Glacier.
My main questions:
-
Should is split up my sync task into one per dataset? This seems like a lot of work, and it feels like there should be something easier to snapshot everything. Or am I just misusing nested dataset in general?
-
Would Cloudflare R2 or Backblaze B2 (or somthing else?) help with costs? Could I configure something better to reduce AWS costs?