Bo Liu

Hi Xun, Internally we use ldd to figure out all the .so files that are required by the counter service binary and package them together for deployment. Thanks, Bo On...

that sounds great to me. On Wed, Oct 9, 2019 at 6:48 PM Derek Argueta wrote: > Hi team 👋 > > Would you be open to building this repository...

It's a library, so you need to compile the library into your application process. You need to open rocksdb instances in your application code, and add opened rocksdb instance to...

Yes, opening db is still done by your application, but you need to add code to add db handlers to rocksplicator library. Also, you will need to write/read to/from db...

It's useful when you want to detect individual hot keys. For example, a single key is contributing to over X% of the total traffic.

Multiple column families within the same db instance share the same WAL, and the WriteBatch(BatchResult) read from GetUpdatesSince() has the column family info associated with each Update. So I guess...

I was talking more about the core replicating logic, which has no problem to replicate DBs with multiple column families. Changes are required in the admin side to support column...

Unfortunately, we haven't implemented it yet as of today.

Can you provide more context about how you setup the servers? Does your server use thrift for RPC? To give you a high-level idea about how it works. The library...

Hi, thanks for trying it out. Yes, you're gonna need a config which contains the shard mapping for your cluster. An config example can be found [here](https://github.com/pinterest/rocksplicator/blob/master/common/thrift_router.h#L515-L527). We are migrating...