[BUG] StarRocks FE Component HScale Operation Fails While CN Operations Succeed

[iziang]

Description:

I'm encountering an issue while scaling a StarRocks cluster using KubeBlocks, where the horizontal scale (hscale) operation for the Frontend (FE) component fails. In contrast, both the hscale and vertical scale (vscale) operations for the Compute Node (CN) components, as well as the vscale operation for the FE, are successful.

Environment:

• Kubernetes: v1.21.5-eki.0
• KubeBlocks: 0.9.1-beta.12
• kbcli: 0.9.1-beta.16
• Note: There is a version difference warning between kbcli (0.9.1-beta.16) and kubeblocks (0.9.1-beta.12).

Steps to Reproduce:

1. Set up a StarRocks cluster with at least one FE and one CN.
2. Execute hscale and vscale operations for both the FE and CN components.
3. Observe that:
   • The hscale and vscale operations for the CN components succeed.
   • The vscale operation for the FE component succeeds.
   • The hscale operation for the FE component fails.

Expected Behavior:
The hscale operation should successfully increase the number of FE instances without any errors, similar to how the CN components and the FE vscale operation behave.

Actual Behavior:
The hscale operation for the FE component fails, indicating there may be an issue specific to the FE's hscale process or its interaction with KubeBlocks.

Additional Information:

• The problem has been identified as a known issue.
• It was suggested that upgrading the cloud and KubeBlocks versions might resolve the issue.

Logs/Errors:
[Please attach relevant logs or error messages here, if available. These could include operator logs, pod logs, or any other diagnostic information.]

Screenshots:
[If applicable, include screenshots that help illustrate the issue, such as the state of the pods or the output from the commands used.]

Attachments:

• [Any additional files, such as configuration files, that could be helpful for diagnosing the issue.]

Questions:

• Are there any known limitations or requirements for the FE hscale operation in the current environment?
• Is there a recommended approach or workaround to achieve the desired FE hscale?

[wangyelei]

fixed at #8541
upgrade to 0.9.2, and can specify opsrequest.spec.preConditionDeadlineSeconds to resolve this bug