Description:
We have just migrated our MongoDB cluster from Atlas to a Percona-based Cluster on AWS and are using PMM to monitor the cluster. When looking at the advisors, there are some warnings that don’t make sense - see the screenshot. Journaling is enabled as per the config & the log level is set to 0 (default). Regarding swap usage, since we are on AWS swap is turned off by default and we use instances with “enough” memory, we don’t think enabling it will make sense.
Version:
PMM: 2.40.0
Percona MongoDB distribution: 4.4.22-21
Does anyone know why these checks fail? And, is there any way to turn off / acknowledge these checks?
Thanks,
Alex
You should be able to disable an advisor by looking at the respective tab (configuration, performance, security) and disable it. Are you logged in as an editor or admin to PMM? I’m a version behind you on my main instance but I have an extra column after Details that allows me to silence the alert which I believe only silences the alert for the host in question but will continue to run the check against other hosts. You may have to be an admin to get that option.
I can’t answer why the journal setting isn’t being detected…looking at the logic of the advisor itself it’s looking at parsed docs and checking for storage.journal being “enabled”.
Thanks @steve.hoffman for your answer. Unfortunately, I am logged in as an admin and am still not able to see the column you are talking about. Maybe it got lost between your version and mine?
It would be extremely helpful to be able to acknowledge such warnings, because I don’t want to turn off the whole advisor but still be notified if the setting changes.
Regarding the journal, I saw at the documentation that it is enabled by default since 4.0+ and can not be disabled. Maybe this confuses the advisor to deliver false positives?
In recent versions (starting with versions 4.0 +), MongoDB enables journaling by default and doesn’t allow turning it off.
https://docs.percona.com/percona-platform/advisors/checks/mongodb-journal.html
Sounds like a bug…or two!
Would you mind submitting a ticket for each at https://jira.percona.com so we can get it looked at and resolved?
Thanks!
Steve