Same here. Old nodes having very little uploads.
I don’t know how node selection works in detail but it would make a lot of sense to spread data more evenly across nodes and therefore preferring new (small) nodes.
Same here. Usualy I run 2 nodes per IP, older and newer, older have around 9TB and younger around 5-6TB. So around 15TB here, too, per IP. I have only one old node solo with it’s IP, but looks similar to ones that are in 2 node config. So my conclusion is that 10TB per node is the limit, for now, no matter how many nodes are on that IP.
I imagine the clients don’t keep data older than 2 years, the time it takes a node to stabilise itself at 10TB.
Would be interesting to get some stats about the age of stored pieces.
With old filestore I did this just by using file creation date from piece files. Unfortunately there seems to be no way to do such statistics with hashstore. Or am I missing something?
1 Like