Revert the system_id generation
What does this MR do?
In !417 (merged) we introduced the automatic generation of the system_id to prevent system_id cluttering in the UI. The change made didn't take in account the use of replicas when deploying GitLab Runner with Helm Chart.
As consequence, when Values.replicas > 1
, all the Pods registered use the same system_id
.
A fix is not straightforward because of the nature of k8s deployments/replicasets
. In order to unblock users, it has been decided to revert the changes while we think about a more robust solution.
Why was this MR needed?
To allow a system_id
per replica
What's the best way to test this MR?
- Set
Values.replicas
to2
- Install GitLab Runner
- SSH each GitLab Runner Pod and validate that the
system_id
are all differents
What are the relevant issue numbers?
close #556 (closed)
Edited by Romuald Atchadé