Google Distributed Cloud air-gapped 1.14.6 hotfixes

Hotfix 2

Backup and restore:

• Backup pods constantly crashloop, primarily affecting the Managed Kubernetes service environment.

DNS:

• A race condition occurs during IP address allocation for GDC services, caused by DNS components sharing the same broad subnet parent as other services.

Documentation:

• IO documentation is missing from the documentation hosted on the well-known server.

Identity and access management:

• Role assignment propagation is slow.
• A 500 Internal Server Error occurs when fetching the login config.
• Custom roles created with resourceNames specified might not take effect.

Key and secret management:

• The KMS controller experiences crashlooping due to out-of-memory (OOM) issues, specifically the root key reconciler consumes excessive memory.

Managed Kubernetes Service:

• Obsolete node pools are not cleaned up.

Networking:

• There are intermittent DNS resolution failures related to border leaf switch configuration.

Platform authentication:

• A certificate issuance using the ACME protocol fails due to DNS-related problems when a certificate is requested for both wildcard and non-wildcard versions of a DNS name.

Security:

• Restarting the strongswan service causes storage volume access to fail.

Virtual machines:

• When creating disks, a race condition occurs when a PersistentVolumeClaim (PVC) is not quickly garbage collected and causes lingering PersistentVolumes (PV).

Hotfix 1

Vertex AI:

• Users can't provision open search pods.