I have been performing zfs replication on several remote pools and since I have migrated my system pool to the new native ZFS encryption, I have been monitoring my replication closely.
Over the past few months, I have noticed my iocage jails root directory are growing steadily, especially on one particular jail running Nextcloud and causing replication to slow down significantly.
I thought this was specific to the type of data being handled within the jail, but I couldn’t understand why this would happen knowing that my data are stored outside the jail and only exposed to the jail via a mount point. Upon further investigation, I found that the issue is related to the clientmqueue folder accumulating a large amount of files over time, found here:
iocage\jailname\root\var\spool\clientmqueue
Digging deeper, I noticed each and every running jails are exhibiting the same behavior.
Only a handful of comments have been made on the subject in the old forum:
https://www.truenas.com/community/threads/memory-and-swap-usage.64536/
https://www.truenas.com/community/threads/jail-sendmail-cannot-chdir.17001/page-2
One of the main reason this is happening is that every cron job being run need to report the result by email.
From what I can see, I am not convinced having those emails sent outside of my network is a wise decision as it can expose HW and other configuration settings to the outside world.
What should be the proper course of action to deal with this particular issue/feature?