fix(CSI-296): node registration fails after switch transport from NFS to Wekafs due to label conflict

[sergeyberezansky]

TL;DR

Added a node labeling step to the nodeserver daemonset initialization process.

What changed?

Modified the init container's command in the nodeserver daemonset to include a step that removes the topology.csi.weka.io/transport label from the node before fetching and storing the node metadata.

How to test?

1. Deploy the updated chart with allowNfsFailback = true (without weka client installed on node)
2. Verify that nodes are labeled with "topology.csi.weka.io/transport" set to "nfs"
3. Ensure that the node has a running weka client and restart the node. The label should set to "wekafs"

Why make this change?

This change prepares nodes for potential topology-change operations in the CSI WekaFS plugin. Since CSI cannot modify existing labels upon node registration, having to remove the existing label prior to setting up a new one

[sergeyberezansky]

• fix(CSI-296): node registration fails after switch transport from NFS to Wekafs due to label conflict #375 👈
• fix(CSI-297): nfsTargetIps override is handled incorreclty when empty #374
• fix(CSI-292): parse NFS version 3.0 to correctly pass it to mountoption #372
• fix(CSI-294): caCertificate, NfsTargetIps, localContainerName are not hashed in API client #371
• docs: fix the example of static provisioning of directory-backed volume #370
• dev

This stack of pull requests is managed by Graphite. Learn more about stacking.

Join @sergeyberezansky and the rest of your teammates on Graphite

[graphite-app]

Graphite Automations

"Request reviewers once CI passes" took an action on this PR • (10/23/24)

1 reviewer was added to this PR based on Sergey Berezansky's automation.