After running multiple nodes for a while I noticed that my all my nodes are keeping a live state of the mempool of tx’s while some do not even produce blocks. This seems unnecessary and from a resource view I would like for the nodes that just sync the chain locally for application to be optimized. Is there a possibilty to turn of this behavior? In the mainnet-config.json there is a setting for the mempool but I think the question is more broad. What are the optimal cardano-node settings for running just a syncing node that needs to sync the chain (possibly with cardano-db-sync running in parallel) and can send transactions?
There is pretty massive memory usage by cardano-node. I believe many SPO perform rolling restarts to update topology and cycle their relays without interrupting the block producer.
Thank you for the reply. I saw that post and I am using the tricks mentioned there. But I wanna go even further. I am not running a pool that constructs blocks or tries to validate. My nodes only need to add minted blocks that peers add to the chain. A lot of the functions of my node are in that sense unnecessary and waste a lot of resources. I’m not sure but I would guess that a lot could be striped out, like the Mempool. Unfortunately I’m busy with other projects at the moment, will come back to this one.