lxd high load with many models
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
Triaged
|
Low
|
Unassigned |
Bug Description
I was load testing by creating a single controller, and then creating 200 models inside that controller, all of which don't have anything in them.
When I 'juju upgrade-juju' while the machine is coming back up, I see:
27235 root 20 0 4023008 160752 17120 S 200.0 2.0 49:29.35 /usr/bin/lxd --group lxd --logfile=
I reported https:/
My best guess is that we are doing something like running a Provisioner per model which ends up connecting to LXD per model, and requires a lot of CPU overhead.
It might be that we're creating 100s of profiles (we create one per model), and this action is causing us to ensure that all 200 still exist, and LXD doesn't handle it particularly efficiently.
It takes several minutes before LXD recovers and stops thrashing.
This bug has not been updated in 2 years, so we're marking it Low importance. If you believe this is incorrect, please update the importance.