The chunk migration operation takes too long to work with the Balancer window in the production environment.
So I’m thinking about always working through the “_waitForDelete setting”, but the manual says that the option is test option.
May I ask your opinion about enabling this option in production use?
I am using mongodb 4.2.
By default, this value is false. Do you want to enable it ? (meaning the balancer will wait for the delete phase)
I wonder if it is a setting that can be used in a production environment.
I want to do a stable chunk migration in an environment where steady input and output exist.
Having the same question.
Is enabling the “_waitForDelete” in production a valid use case?
My main concern about the regular async migration is:
I’m using self managed storage for the DB and that the default setting is not freeing up space (since the migration takes a long time).
oplog is filling up very fast because of the migration, which leads to loosing change-stream resume point.