Juju controller needs to be restarted
Bug #1729712 reported by
Thomi Richards
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
Triaged
|
Low
|
Unassigned |
Bug Description
Our production juju controller needed to be restarted today in order to fix an issue where `juju status` was taking in excess of one minute to respond, and almost all other juju commands were similarly slow.
What we found when logging into the controller was that jujud and mongod were each consuming about 50% of the available CPU time.
Restarting the controller fixed the issue, but this clearly is not a good solution long-term.
Attached are the machine-0 logs from the controller for that day. Note that the incident occured around 2017-11-02 23:35, and there's nothing particularly interesting in the controller log.
Changed in juju: | |
status: | New → Triaged |
importance: | Undecided → High |
tags: | added: controller-load restart |
To post a comment you must log in.
Let me know what other information I can provide. Attached is syslog for the same period.