Does Trident works with settings Non-Graceful Node Shutdown (NGNS) without configuring on Kubernetes
Applies to
Trident
Kubernetes
Answer
- Enabling Trident's enableForceDetach feature without configuring Kubernetes for Non-Graceful Node Shutdown (NGNS) can lead to unintended consequences.
- The enableForceDetach feature in Trident is designed to work in conjunction with Kubernetes' NGNS to safely manage storage resources during unexpected node failures.
- Without NGNS enabled, force detaching volumes may result in data inconsistencies or corruption. Therefore, it's crucial to ensure that NGNS is properly configured in your Kubernetes cluster before enabling enableForceDetach in Trident.
- Refer to Trident Document and Kubernetes Document :
- If modified the CustomResourceDefinition (CRD) YAML to set enableForceDetach: True and applied it, Trident will accept the configuration change. However, this bypasses the safeguard mechanisms in place and can lead to the issues mentioned above.
Potential Risks:
- Unexpected Volume Behavior*: If a node fails ungracefully, Trident might attempt to force-detach volumes even though Kubernetes is unaware of NGNS, potentially leading to data integrity issues or conflicts.
- Unsupported Configuration: Since NGNS is meant to work with **enableForceDetach, running Trident with **enableForceDetach* enabled while NGNS is disabled might result in undefined behavior.
Conclusion:
- If your goal is to *enableForceDetach*, it's best to enable NGNS in Kubernetes to ensure a supported and stable configuration.
- We recommended to enable NGNS in Kubernetes before enabling Trident's "enableForceDetach" option.
Additional Information
additionalInformation_text
