We're having issues updating the driver from version 1.4.40 to 2.2.9.

The main issue is that all the queries are freezed or 10s every 30 seconds or so (it’s random). Please find the details bellow.

Context

• mongo cluster in version 2.6.11 ;
• node is in version 4.5.0 ;
• Please find attached a very simple code sample to reproduce.

Old driver behaviour

• 1 db connection is opened and never close ;
• 5 TCP connections are opened and never close ;
• Queries are running without any issue.

New driver behaviour

• 1 db connection is opened and never close ;
• At first only 5 TCP connections are opened but then the number of connections starts to increase progressively until it reaches 15 ;
• Looking at a tcpdump only one connection seems to be in use at a time ;
• Sometimes one of those TCP connections is closed (without going in TIME_WAIT state) and all queries are freezed for 10s ;

Further testing

• We've tested to decrease and increase the pool size. The result is the same but the number of tcp connections differs. For a pool size of 1, we've 3 TCP connections (with the old driver it opens only 1 tcp connection) at the same time and it's working fine. As soon as the 4th one gets created it freezes all the queries.
• We've tested to do only one query at a time, behaviour is the same but it get's frozen faster.

Thanks for your help!