View the original community article here
Last tested: May 1, 2020
It's an identifier for an instance. Instance slug is the combination of
- stable_root, which is the path to the directory where Looker's JAR file is running
- mac_address of the server on which Looker is running
- ip_address of the server on which Looker is running
- The license_key
Though it is supposed to be stable, it is not unusual (for self-hosted instances) for the instance slug to change as server/hardware changes are made, so over time, a license can accumulate multiple instance slugs.
How many instance slug should a cluster have?
Q: Is the instance slug the same among all instances within a cluster?
A: Yes, an "instance" can be clustered or single. A "node" is always single.
Q: If i have two clusters with the same license key, is the instance slug unique for each cluster?
A: Yes, it's supposed to be.
In relationship to PDTs:
With our recent changes to PDT naming / stable PDTs - is it still a bad idea to share a scratch dataset among multiple instances?
The best practice is one per instance, but we do support multiple instances using the same scratch schema. As long as you take care to have a different instance slug per instance, you will be fine. If you wind up with the same instance_slug though, you’ll run into enemy reaper problems.