Why does LMDB need to know how large our DB might be? about lmdbjava HOT 3 CLOSED

http://www.lmdb.tech/doc/group__mdb.html#gaa2506ec8dab3d969b0e609cd82e619e5

from lmdbjava.

I'm not the author of lmdb or lmdbjava but here are my thoughts.

It's a function of how lmdb works.

For many use cases this is not an issue at all. The size sets the memory map size and uses a sparse file so it doesn't take up as much space as you set, only that you use.

If it's an issue (it isn't for my use cases) you can detect when size goes above a certain threshold and copy/create a new larger file based db. Others use rolling databases for time series data or use for archival / immutable data.

from lmdbjava.

If it's an issue (it isn't for my use cases) you can detect when size goes above a certain threshold and copy/create a new larger file based db.

You just need to re-open the database with larger settings if required. I've never encountered it as an issue in practical situations. My larger workloads have a "data collection" phase where it's basically random writes, then an end of day write of the database so the keys are in perfect order. That reduces the size of the long-term file as well.

from lmdbjava.