Short answer: no.
Long answer: The way to automatically switch failures in MongoDB is that for a set of replicas, a qualified majority is required to successfully select a new primary. Detained members have votes in the election. Therefore, if any of your nodes fails, the set of replicas discovers that it does not have this majority, and the current main steps down, even if it does not work. So what you basically do is double , the chance that your replica set will fail. Arbitrator is a very cheap process in terms of using RAM, processor and even disk space when starting with --smallfiles --no-journal --noprealloc or equivalent parameters specified in the configuration file. Please note that these options are safe to use, since the arbiter essentially only checks the heartbeats of nodes that carry data. For example, you can put an arbiter on an application server.
Disclaimer: The following procedure is highly recommended. Act at your own risk.
You can set the votes of the pending server to 0. Thus, the undelayed node will cause an election in case of an unsuccessful participant, concludes that it is the only node of the online replica and that it has the majority of votes (1/1) and will continue to work as expected. This course of action requires some attention, since you will again have an even number of votes if you add a member to the replica set later and require reconfiguration of the replica set. This also has serious implications for network fragmentation issues. Again: use at your own risk
Markus W Mahlberg
source share