CCE Terraform Node Pool Disk Update Not Working as Expected

[Wolfslicht]

Terraform provider version - 1.5

Affected Resource(s)

Hey there,
I have to open a ticket.

We are experiencing an issue with updating the disk space for a node pool using Terraform on Swiss Open Telekom Cloud (OTC). The expected behavior is to add or extend data storage for nodes, for example, increasing the disk size from 100 GB to 300 GB for a node pool VM. However, the current behavior results in the destruction of the disk pool when attempting to extend the disk.
Even with the lifecycle setting create_before_destroy, the old node pool is destroyed as soon as the new node pool is created.
This issue leads to data loss, especially with STS or other persistence applications, as there is no tainting or evacuation possible.

Expectation:
• Add more disk space to a node pool.
• Extend data storage for nodes if corrected/adjusted through Terraform.
• Example: Increase disk size from 100 GB to 300 GB for a node pool VM.
Current Behavior:
• Extending disk results in the destruction of the disk pool.
• Lifecycle setting create_before_destroy does not prevent the old node pool from being destroyed, before persistence workload can be transferred to new node pool vms.
• No tainting or evacuation possible, leading to data loss with STS or other persistence applications.

Steps to Reproduce:

1. Attempt to extend the disk size for a node pool using Terraform.
2. Observe that the disk pool is destroyed and recreated, leading to data loss.

Impact:
• Data loss with STS Workload like longhorn or other persistence applications.
• Inability to safely extend disk space for node pools.

[anton-sidelnikov]

Hi @Wolfslicht, the issue is that our API supports updating only a limited set of node pool attributes (see API documentation). From my perspective, the best approach would be to request this feature through our customer support at [email protected].

[Wolfslicht]

Hey Anton,

I already did, they told me to raise the issue here.

Hello Mr. Schröder,

first note that disk changes to node pools will only be applied to new nodes.
https://docs.sc.otc.t-systems.com/cloud-container-engine/umn/node_pools/managing_a_node_pool/updating_a_node_pool.html

Persistent data should also be stored on a PVC so the data is not affected by changes to the worker nodes.

And this seems to be an issue about the exact operations performed by the OTC terraform provider for changes in these values.
So please create an github issue about your problems. [https://github.com/opentelekomcloud/terraform-provider-opentelekomcloud/issues](https://deu01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fopentelekomcloud%2Fterraform-provider-opentelekomcloud%2Fissues&data=05%7C02%7CWSchroeder%40t-systems.com%7C099c9f3b445b4153efbf08dd450de809%7Cbde4dffc4b604cf68b04a5eeb25f5c4f%7C0%7C0%7C638742648199178903%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=Vpg795p1C%2F2DISpy7byzvmQYsJTYiV6wKqx5XCeAFK0%3D&reserved=0)

[oeotvos]

We are aware of similar problem updating labels and tags. I'm going to update our internal demand with this request also.
[#OTCPR-21476]

[anton-sidelnikov]

HI @oeotvos, thank you!