Should dbsync run on a separate instance?

With the dbsync 4.0 being available. Should I run it on a separate instance and if so do I have to build the node out with cli and cabal before installing dbsync?

Hello again, Momo!

I personally don’t use dbsync, but I reached out to some of my operator friends that do.

Here are their responses:

“It’s probably best to run db sync along with a dedicated relay node on the same box. Performance should be slightly better for syncing the blocks between them. Unless memory is an issue, then separate machines.” @AndrewWestberg - BCSH - JorManager Creator

“db-sync needs a node to talk to. doesn’t matter how you get the node - by building or grab the binaries.” @papa_carp - LOVE - Creator

“You can use any cardano-node instance 1.18.0 I believe. I use a dedicated node because the db-sync chugs, especially at initial bootup” Adam Dean - SPIKE / BUFFY

Thanks and I hope this helps!


I’ll set it up on a separate instance but should I connect it to the core or relay or both and do I point to it just by changing the topology.json?

You just need a separate instance - connect it how you choose - I personally would not connect it to my core, but I don’t understand your particular use case either.

I personally would use an entirely separate topology architecture isolated from my pool operation.

I just assumed it would make my nodes have better connectivity and perform better.