You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, we currently use pebble on large NAS ssd-like drives. Have been watching the development of the SharedStorage feature with interest, but I'm having trouble getting a big picture description of how it's used and the tradeoffs involved outside of just reading the code.
Is there anything you can share here about this new addition, it's purpose, production readiness, performance, etc?
Thanks! (pebble rocks btw)
The text was updated successfully, but these errors were encountered:
Thanks for the interest. This feature is currently in development and is not ready to use. The goal is to be able to store SSTs on systems like S3, either permanently or temporarily. This can allow faster restores from backup, or faster rebalancing of ranges between cockroach nodes.
There are a couple older issues regarding plans for an S3 backend (#640, #805). Does SharedStorage mean supporting object storage is now in the works?
I have the following use case. I'd like to be able to run a KV store API server that scales from 0 to 1. If Pebble can use an object store as a backend then this becomes very easy to implement.
We have the experimental ability to read and write sstables to S3, but the latencies are too high in practice to use directly and the per-IO pricing might also make it unfeasible. Unfortunately, for now, we don't have anything here that can be used in production.
Hi, we currently use pebble on large NAS ssd-like drives. Have been watching the development of the SharedStorage feature with interest, but I'm having trouble getting a big picture description of how it's used and the tradeoffs involved outside of just reading the code.
Is there anything you can share here about this new addition, it's purpose, production readiness, performance, etc?
Thanks! (pebble rocks btw)
The text was updated successfully, but these errors were encountered: