i hope everyone had a good start into the new year. :>
Those are my questions:
- Does couchbase server (3.x) internally always compress (via snappy?) values for each key/value pair? So that in memory AND disk only the compressed value reside, which are then transparently decompressed, before being sent back to any client?
- If it is compressed internally, does XDCR use those compressed bytes for replication as well?
I ask this, because i could compress JSON on the client side via my own transcoder. And I wonder, if that is actually necessary? It is just important to me, that the data is compacted on the couchbase server side. And that this compacted data is transferred between the data centers.
There are some google pages, that mention, that couchbase uses snappy internally. But i find no definite proof.
Thank you for any help or pointer.
Have a nice day!