As of the time of this writing, Rancher Logging is broken when the Docker root is configured to something other than
This issue is tracked in GitHub issue #21112.
- Rancher 2.x managed/imported cluster with logging enabled.
- Docker root configured to something other than
/var/lib/dockeron the nodes (confirmed with
docker info | grep Root).
These steps will assume you have the Docker data root set to
/other-docker-root to whatever your custom path is:
Rancher UI -> Cluster -> System Project -> Workloads -> cattle-logging Namespace
Find workload rancher-logging-fluentd-linux
Edit volume dockerroot
Change "Path on the Node" from
Add volume (with the following details):
Volume Name: dockerrootcustom Type: bind-mount Path on the Node: /other-docker-root Mount Point: /other-docker-root
- Click Save
At this point logging should be working with your non-default Docker root directory. You should be able to verify this on your logging target. Keep in mind it may take a few minutes for logs to show up there as fluentd is configured to clear its buffer every 60 seconds by default.