Does the problem known as slow train only apply to Nodejs mongodb driver? The problem description goes like this:
While Node.js is asynchronous, MongoDB is not. Currently, MongoDB uses a single execution thread per socket. This means that it will only execute a single operation on a socket at any given point in time. Any other operations sent to that socket will have to wait until the current operation is finished.
The description clearly states that this is a mongodb limitation, but I have only scene this problem mentioned in nodejs community and in mongodb driver for Nodejs, that is why I am asking this question, so maybe a better way to rephrase this question is that, why is the problem only seems to be a concern in nodejs? Do other programming languages like python or java have a way to bypass the limitation of which I am not aware?
The description clearly states that this is a mongodb limitation, but I have only scene this problem mentioned in nodejs community and in mongodb driver for Nodejs, that is why I am asking this question, so maybe a better way to rephrase this question is that, why is the problem only seems to be a concern in nodejs? Do other programming languages like python or java have a way to bypass the limitation of which I am not aware?
The link you shared is to a much older version of the Node.js driver, however the docs for recent versions of the driver contain similar information. Each driver contains an FAQ section that addresses common concerns, however as you’ve surmised this isn’t really a Node-specific challenge.
The MongoDB server doesn’t currently support socket-level multiplexing which results in operations blocking as a request is processed and a response is returned. For this reason drivers support connection pools to address this.
For example, setting a maxPoolSize=5 will ensure up to 5 connections can be checked out to service concurrent operations before subsequent operations begin to queue. Most operations complete extremely quickly and return their checked out connections to the pool so the size of 5 here is just meant to be illustrative - likely having a pool of 5 would only result in 1-2 connections ever being checked out at once.