in-memory

hz-stop --help
No Hazelcast server found to stop

Currently, aggregate functions have CALL and FINALIZE steps, and their accumulator is initialized to NULL at the beginning. It is possible that adding a BEGIN step instead of the current initialization will allow us to simplify some of the CALL/FINALIZE steps and improve the performance of aggregate functions.

Currently we don't have any mechanism to limit the maximum number of clients that could be handled simultaneously.

This feature should be designed properly. Here is some clue: https://redis.io/topics/clients#maximum-number-of-clients

There are no external core dependencies other than LZ4 and Snappy which are not used unless compression is configured. So they should be installable.

libraryDependencies += "io.swaydb" %% "lz4" % "version"
//OR 
libraryDependencies += "io.swaydb" %% "snappy" % "version"

For now we have to exclude

Use of the $geoNear pipeline stage results in:

[Error 40324] Unrecognized pipeline stage name: '$geoNear'

It's clear it is not currently supported in de.bwaldvogel.mongo.backend.aggregation.Aggregation.java

Example Mongo query usage:

db.places.aggregate([
   {
     $geoNear: {
        near: { type: "Point", coordinates: [ -73.99279 , 40.719296 ] },
        distance

Tracking issue

Currently, the Go client does not support Hazelcast's partition predicate.