interesting. thanks for the pointers.. I'm not ready to take this project on, but at 85% disk fullness, I'm close to needing to accomplish this.

itsrealfake

Raw block storage is on the HDD, they are sequential and don't need the benefit of the SSDs random I/O writing speed. The chainstate/ utxo set are on the SSD and this is where the speed is needed for validation.

No pruning needed. You can store on HDD and SSD < 1TB will handle everything else. Can even fit electrum server on the SSD as well in this scenario.

Even a 1TB SSD and a 1 TB HDD will allow you to work with what you have for longer and then by the time we need to get new hardware the hardware situation should look different.

I use proxmox and an ubuntu VM to achieve this. Claude helped a lot in working through the pain points.

bitcoin_beginners

Upgrading a full node in 2026

Raw block storage is on the HDD, they are sequential and don't need the benefit of the SSDs random I/O writing speed. The chainstate/ utxo set are on the SSD and this is where the speed is needed for validation. 

No pruning needed. You can store on HDD and SSD < 1TB will handle everything else. Can even fit electrum server on the SSD as well in this scenario. 

Even a 1TB SSD and a 1 TB HDD will allow you to work with what you have for longer and then by the time we need to get new hardware the hardware situation should look different.

I use proxmox and an ubuntu VM to achieve this. Claude helped a lot in working through the pain points.