Google Cloud VM Upgrade for Mining? (e2-medium (2 vCPUs, 4 GB memory))

Hello everyone. I just started getting into NKN mining, and I have deployed several nodes. Still no rewards. I have read through the tutorials and threads about mining, and i’m a bit confused about just how important the hardware specs are when it comes to mining.

I’m currently running all g1-small 1.7 GB, 1 shared vCPU vm’s. Are these not effective enough? Would it be wiser to reduce the total # of nodes, but upgrade the vm to something more powerful like e2-medium (2 vCPUs, 4 GB memory)?

Please have mercy, I’m a noob. :slight_smile: Any help would be appreciated–I’m in love with this project. Thanks everyone!


g1-small is enough for NKN full node to run. Actually in normal mining mode, nknd should consume about 20-25% of CPU and less than 15% of memory. Why I know this? because I also run a nkn node on g1-small.

Occasionally, the CPU usage will exceed the average, and the google cloud monitor will complain that CPU load is too high and suggest you to upgrade to bigger node type. But bigger node type will cost you more money.

  1. When the node is first created, the GenerateID process will consume a lot of CPU, for an hour or even more. This happens only once in the entire life of a NKN mining node.
  2. When the node is restarted (e.g. to upgrade to new version), and the prune database switch is ON (default is OFF), then the node will try to compress the chain database, delete unused data and to save disk space. This process can take hours or even days, depending on when the node did the last pruning. But after pruning is done, the node will get into regular mining mode and CPU use will drop back down to 20-25%.

So I would recommend to stay with g1-small to keep your mining cost down. It will be fine.

1 Like

Thank you so much :slight_smile: – I will keep my g1-smalls!

You are welcome. And also to make sure your node is in proper mining mode, you can use community tools like these to check your node:

Is it normal for nodes to take 3 days to sync? It’s weird, my nodes deployed on Digital Ocean were ready by the end of the night, but my g1-smalls have taken two days and are only about 75% sync complete. I didn’t alter the SyncBatchWindow in config because I figured the hardware was enough to handle the default window size…


Google g1-small seems to be a CPU shared by other users, and not as powerful as the DO vCPU. And it does take much longer than DO nodes to do heavy lifting stuff like initial sync, generateID, and prune database. I don’t know why they can charge more money for less service. :joy: So in the end, I only keep one node on Google (just to make sure it works) and all other nodes are either on Digital Ocean or Linode.

Thanks for the clarification, I’ll definitely look into moving them over to Digital Ocean… :slight_smile: