CloudLink: What happens when a CloudLink node in a cluster fails with PowerFlex

Summary: This is a tested scenario for Disaster Recovery (DR) purposes. What happens when a CloudLink node in a CloudLink cluster fails in a PowerFlex environment?

This article applies to This article does not apply to This article is not tied to any specific product. Not all product versions are identified in this article.
Check out other resources

Instructions

Scenario:
  • There is a two node CloudLink cluster in full sync. 
  • There are four SDS agent machines connected, most likely to the first node.
  • All machines are showing connected in the CloudLink Machines UI status. 
  • The command Svm status co shows both nodes, the first as connected, the other as accessible.

Steps Taken:
  1. The first CloudLink cluster node is shut down. The machine agents should switch to the other node once they timeout on the first node connection. A practical observation is that sometimes the agents may be slow to switch. If they are given time to switch and connect to the other node, the svm status shows connected. The svm status co would show inaccessible and connected respectively.
  2. The first node is dropped or removed from the cluster. If the agents performed the switch to the second node, they receive the updated list of nodes in the cluster (only the second node). It takes a few seconds for the agents to be given an updated list of the cluster. If the agents did not perform the switch, they may keep trying to connect to the deleted node.
  3. A new node is deployed and added to the cluster on the IP of the first deleted node. The second node is shut down. The agents that performed the switch should now see the updated cluster list. Other nodes (that keep trying to connect to the deleted node) now find a new node at the same IP while the other node is inaccessible. 
  4. In this lab environment, some nodes showed svm status as connected and svm status co as connected and inaccessible. The others showed svm status as connected, while svm status co as available and inaccessible. The running CloudLink node shows the first machines as connected, the others as disconnected.
  5. To fix the "disconnected" machines, use the command svm -S -G. 
  6. This test was performed too fast, The agents did not follow up with the updates to environment changes. The agents are prepared for a CloudLink node crash to switch to the other node.
  7. Administrative activities, such as deleting or adding a node while the node is shut down may lead to an unclear state.

Affected Products

CloudLink SecureVM, CloudLink
Article Properties
Article Number: 000222483
Article Type: How To
Last Modified: 05 Sept 2025
Version:  3
Find answers to your questions from other Dell users
Support Services
Check if your device is covered by Support Services.