The cardano-node relay stops running after about 15min, and can no longer SSH into the VM (the serial port log on the VM in GCP says ‘not enough disk space’ or no space left)
My machine is a e2-medium (2 vCPU, 4GB memory) w/ 10GB SSD VM hosted as a compute engine in GCP. I’m following instructions from the Stake Pool Course.
cardano-node run \
--topology testnet-topology.json \
--database-path db \
--socket-path db/node.socket \
--host-addr 0.0.0.0 \
--port 3001 \
--config testnet-config.json
When I do various ls and du commands to check the size of my ~/cardano-node/db directory, its contents don’t look that large (the .dat files), does this imply that maybe it’s a file that’s not in ~/cardano-node/db that is consuming too much disk space? Log file hanging around somewhere?
Here is the output from running htop, why is cardano-node listed so many times, that normal? 1 instance of it running but using or sharing more more CPU’s?
I run the commands tree and df to observe memory and disk consumption, any key metrics to look for when running those? When I ran ‘tree’ I noticed the “available memory” was slowly dropping over time.
I captured this log right after a fresh cardano-node install but before running a node.
Any params I can tweak? Flags passed to cardano-node run. I’m checking them now.
Need to update the testnet-config.json at all?
I have fresh install backup up via machine image and snapshot so can iterate on any suggestions! Thanks
First and foremost, thanks @Alexd1985 for your answer.
This is at least the third time I’m seeing something like this, and it is really frustrating.
I get that minimum requirements are just that – minimum, but you need “at least 40gb [of storage] to be safe”. You also need 16GB of memory for the node to actually try to function.
I think “Realistic Hardware Requirements” would really help noobs like me, rather than having to fail at setting up a node several times then search the internet for error messages and cobble together useful bits of information like this from disparate sites.
Again, I really appreciate this answer because I was using 30GB of storage, thinking that was above the requirements and might be sufficient.
I understand that things are relatively new and under intense development, but as someone trying to set up and run a node, I am having serious trouble finding the information I need.
I worry that other people trying to do the same thing are not finding what they need, getting frustrated, and giving up without speaking up. So, we end up with people turned off by Cardano, rather than more stake pool operators. Maybe that’s by design–so only people who are truly dedicated get to run a node?
Wow. OK. Thanks for the heads up.
BTW, I just found your How to set up a pool in a few minutes. Thank you for that!
What are the real world requirements? Is an updated list maintained somewhere?
So far, here is what I’ve found:
CPU – compute power is not that important and the minimums do seem sufficient.
MEMORY – 16 GB seems to be what you actually need
STORAGE – 100 GB is a realistic amount
NETWORK – needs to be able to handle ~1GB / hour
STATIC IP ADDRESS – kind of obvious, but necessary