We live in the days of 20+TB drives and I’m looking to build more bigger nodes.
I’m wondering:
The Storj docs recommends maximum 24TB nodes, but since the 30TB CMR nodes are around the corner, and bigger drives are already anounced, we all know sky is the limit. 
Currently, my biggest filed node is 21TB of data on Exos X22 22TB drive, formated in 4Kn, ext4 fs.
Synogy DS220+, 18GB RAM, db-es on USB SSD flash drive, 2 nodes running in Docker: 9TB/16TB and 21TB/22TB.
ext4 fs, no RAID, noatime, write cache enabled.
Startup piece scan and lazzy mode disabled.
The nodes are running good, with no lockups. The trash is cleaning, the OS reported space is the same on the dashboard. Will see when the TTL of the synthetic SaltLake data hits, because there are like 19TB of it on the big boy.
The small node has no SL data (GE last year).
It’s not a limit, it’s a recommendation. However, if you would get a single 30TB HDD, nobody will stop you except equilibrium.
Yeah, I don’t want to be the lab rat. This is one of the reasons why I made this thread.
We’d never call you a lab rat! You’re more of a … canary.
Yes… a beautiful beautiful healthy canary! 
snorkel that drive seems to be some sort of unicorn miracle.
a) how are you not having reported space and garbage collection problems like so many others?
b) are you running the filewalkers at all? if so how long do they take?
No filewalker! but they run stable, no lockups, databases on SSD flash. The majority of my nodes are on Synology 18GB RAM, 2 nodes per machine, Exos drives, ext4, noatime. No VPN. No RAID. Nothing fancy. I have no old trash directories, and all trash and retain jobs are finishing.
I’m guessing you left the 2GB stick in there. Why not replace it too? Or maybe your model is the new fancy one with soldered on ram?
Fancy model. 
Synology likes you to remember you are buying the “cheap” version, Apple style.
10TB, at the moment, on single drive. But the node is sharing traffic with the second one, and both have 15.8TB in total… after 4.5 years.
The second drive is 22TB, with only 5.8TB occupied.
The biggest drive I have is 24TB, Seagate Exos, but is the newest node, with only 4TB of data.
I believe that taking into account the deletions of old data/backups/etc, the retention period seems like around 2 years. So bigger drives don’t realy make sense. They won’t get filled ever, or at leat in the first 5 years. I think, now, the smart thing to do is to use drives under 20TB. 16-20TB is the sweet spot, depending on the prices you find.
Amazon in europe is plenty of seagate v7 12TB, they are the cheapest. i can find plenty of then for 160€ (around 180$). I think they are best for nodes because 16 or 18TB seems impossible to fill. 16 or 18TB are like 17€/TB and 12TB are like 14€/TB
I crossed the 16TB threshold a few times, on a single IP, in less than 5 years. So I will go with 16+ TB, just to spin less drives in 10 years.
I’ve got about 13TB at one IP address but it’s spread across several drives.
