Open mcfriend99 opened 7 months ago
Yo, i'm using a nuc i5 12g, 64g ram, 2to disk... i ran the swap api in a docker and with the binary. i had to allow 48g of ram for the service to run properly... even tho if i stop it and restart it i'll get an oom too. So clean ubuntu 22.x restart with 8cores, 48g ram it runs ok... less than that == oom after few minutes
I suggest getting 50+ Gb. Couldn't start it on 48.
Really appreciate the recommendation. Please can anyone @Wouimbly and @paadoks can you guys answer three other questions for me?
Really appreciated your answers and will appreciate this more.
Guess the Jupiter team can do a better work with documenting the answer to the general questions of system requirements and question number 1.
Thanks
@mcfriend99 I've only tried to run jup API in non-recommended mode, without Yellowstone. I've heard, that with Yellowstone it won't consume so much RAM. Regarding your questions, I won't be much of a help :(
7.4Gb RAM using Yellowstone grpc colocated with the validator, but CPU usage is quite high
What causes such high CPU usage? This is running on same hardware used for RPC nodes:
To set up your own Jupiter-swap API, what are the requirements for the RPC? Does the RPC need to have account-index?
The app keeps getting on my ubuntu deployment killed due to system out of memory when running as CLI in foreground or background as well as when running as a daemon.
It will be nice to let us know the minimum system requirement.