Tuning Dgraph / Badger for HDDs instead of SSDs?

I am one of the unlucky few who isn’t running Dgraph on SSDs (although I do have lots of RAM).

Is there a way to tune Badger via Dgraph, or does that require a recompile? Besides setting GOMAXPROCS to something lower, I’m also interested in tuning some parameters at https://github.com/dgraph-io/badger/blob/1496af939d8533779f7b6b3f3a6cd251abf5c598/options.go#L32.

What are some recommendations? Generally, because of the HDDs, I believe I should “prefer a large number of sequential writes to reduce later random reads”.

P.S. I don’t mind sending a PR if pointed in the right direction. :smiley: