Closed osowski closed 2 years ago
At the moment, I believe it is the inclusion of the FQDN sub-domain as Labels in the Pod specs for the StatefulSets. We will need to do more investigation on if these are required into the future and/or if the prior HA functionality will be carried through as-is into CC. For now, we can increase this limitation to allow for more robust deployments and circle back to ensure we are driving the need to remove the limitation altogether.
A manual Chart deployment can definitely still fail, but the feedback is much more direct to the Helm user than the bash deploy script user and is more easily rectified.
Signed-off-by: Rick Osowski rosowski@gmail.com
Description
As part of previous changes for multiple Data and Server Pods per Worker Node, Pod FQDNs became longer and certain cases were causing cluster deployment to break post-deployment or fail completely. As such, a previous limit on namespace length was implemented to limit the exposure to these scenarios. Additional work has been done to support longer FQDNs throughout CORTX Core components, so this maximum namespace length limit can be increased.
Type of change
Applicable issues
How was this tested?
Locally through many deploy/destroy loops with namespaces of length 8, 10, 15, 20, 25+ characters.
Additional information
Checklist
If this change requires newer CORTX or third party image versions:
image
fields in solution.example.yaml have been updated to use the required versions.appVersion
field of the Helm chart has been updated to use the new CORTX version.If this change addresses a CORTX Jira issue:
CORTX-XXXXX:
)View rendered README.md