You still need multiple servers to make up a full node, but with this change, you are able to use 64GB instances. (Minimum of 3 servers like this to make up a full node).
You are viewing a single comment's thread from:
You still need multiple servers to make up a full node, but with this change, you are able to use 64GB instances. (Minimum of 3 servers like this to make up a full node).
To mimic our setup on a smaller scale, it would be 2 x 61GB servers and 1 x 32GB for Hivemind's postgres. Not 3 x 64GB.
So does this mean I could run the 2 x 61Gb on a 128Gb HEDT machine and the 1 x 32 Gb on another machine on the same local gigabit Ethernet? Because I have this physical hardware already and would like to run a full node. I also have very fast internet.
Posted using Partiko iOS
With a configuration like this I would run a single node that includes all of the plugins except for tags/follows in one on the machine with 128GB. The downside would be a longer replay/sync time, but it would likely be more performant than running two separate nodes on one machine (due to disk paging). Fast disk would be required, such as nVME SSD. The two machines talking over local gigabit ethernet would work just fine yes.
Thanks. Would it be better to move 64Gb to another computer and run 3 separate machines. How fast does the processor need to be on the 32Gb machine? Is recent Pentium 4560 dual core 4 thread enough for the third machine?
Posted using Partiko iOS