What are indices, are those Vertices ?

node ids, from`Return id(v)`

Ok, so those are actually VertexID-s

so something simple like

MATCH (v:Post) return v(id) LIMIT 5000000

takes 5 mins for me on SNB SF100

Did you tried using Nebula native query language ? Did you tried creating index for TAG which is used for v:Post ?

i haven't, will try these, thanks!

Yes, as Goran suggested, after creating an index on TAG: Post, you could do

LOOKUP ON Post YIELD id(vertex) | LIMIT 5000000