Fix active_server for multi-cluster deployments
#205
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Found a few issues with the current approach. Reworking it and will update the PR as soon as possible. |
Don't show the summary only for the first configured cluster, but for each. Instead of utilizing `run_once` which limits the execution to once per play, just use the existing `when` condition to limit it to the 1st server of each cluster.
While `first_server.yml` would only be included for the 1st server of a cluster, `run_once` for setting the `active_server` fact also meant this would only happen once for all clusters instead of once per cluster. Use a loop across the members of each cluster and a fact delegation instead, to set the `active_server` fact instead on each cluster member to the `inventory_hostname` of the server, for which `first_server.yml` was included.
Instead of downloading only the kubeconf of the 1st cluster processed, download it for each cluster.
Don't limit the waiting for the remaining nodes to be ready to the 1st cluster with `run_once`. Instead, check whether `inventory_hostname` either matches `active_server` or the 1st member of the group of servers.
eliasp
force-pushed
the
multiple-clusters-active_server
branch
from
944e220 to
1994af1
Compare
|
Pushed now a new iteration which worked as expected in multiple local tests here so far. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
When addressing multiple clusters at once within a play, the
active_serverwould only be determined once - when processing the very first cluster.active_serverwould then be used to delegate the task to, to wait for all nodes and pods to be ready again.When all clusters are sized equally, this wouldn't cause any real troubles (yet still be incorrect in terms of actually verifying the state of the cluster's nodes), since the number of nodes in the first cluster would match the number of nodes in any other cluster checked for the nodes to be up.
But as soon as a cluster with a different number of nodes is present, this would cause failures/timeouts, since the check waiting for all nodes to be up and running again would never reach the expected number.
By not utilizing
run_onceand verifying for the include offirst_server.yml, whetheractive_serveris a member of the current cluster, this problem can be prevented.Type of change
How Has This Been Tested?
rke2successfully with those changesrke2role fail without those changes