Trung Dinh
Trung Dinh
I am encountering the same issue
We also need this feature as it's not currently possible to get back an already registered collector in a multi-threaded environment (i.e., multiple threads trying to register the same set...
> If you are referring to the parallal read I/O features (as in this [blogpost](https://rocksdb.org/blog/2022/10/07/asynchronous-io-in-rocksdb.html)), then it's as simple as adding a new flag to `ReadOptions`. We don't currently have...
Seems this is still a nice-to-have feature? I'd like to take a stab at this to understand more the datafusion ecosystem.
> @trungda I'm interested in this feature. Is this something you have solved? No I haven't yet. It's on my list but I don't have time to get to it...
Thank you! Could you approve the workflows just so I can merge?
> You don't mention how the tenant is sharded. The pageserver isn't meant to handle arbitrarily large shards -- large databases should be broken up into many shards, even if...
> If one thread is fast enough to keep up, then it's not really desirable to use more threads, because that introduces new issues of compaction competing for resources with...
Maybe related https://github.com/neondatabase/neon/issues/5331 @jcsp ?
> I think we usually see either: > - That image layer generation goes fast enough that it's done with a layer by the time it would be evicted. >...