Updating all nodes except the configuration node

When you are updating nodes individually, before you update the configuration node, you must update all of the non-configuration nodes in the clustered system.

To update a non-configuration node canister, follow these steps:

1. Ensure that all hosts have all paths available to volumes that are presented to them by the system.
   If there are any unavailable paths, wait up to 30 minutes, and check again. If any path is still not available, investigate and resolve the connection problem before you start the code update. Ensure that the multipathing driver is fully redundant, with every path available, and online. During the update, you might see multi-pathing driver errors that are related to paths that are going away, and to the increased multi-pathing driver error count.
2. In the management GUI, check that no incomplete volume synchronization tasks are in progress.
   In the status bars that are at the bottom of the page, expand Running Tasks to display the progress of actions. Ensure that all synchronization tasks are complete before you remove the node.
3. In the management GUI, use the dynamic image of the system to view the canisters. Right-click the canister that you want to remove and select Remove.
4. Open a web browser and type https://service_ip in the address field, where service IP is the service IP address for the node that you deleted.
   The service assistant login page displays.
5. Verify that the node is no longer a member of the system by checking that the node status, as shown in the display, is service.
   The node has an error code of 690. The removed node is no longer visible to the system. If the node status is active, you are probably connected to the wrong node.
6. On the service assistant home page, select Update manually from the left menu.
   Attention: Each node must run the same code version; nodes with different versions are not compatible.
7. Select the correct update package, and click Update.
   The node restarts and updates. Access to the service assistant is lost while the node restarts, but you can still access the service assistant from a different node.
8. Eventually the node canister that you removed and updated automatically rejoins the system.
   When the canister is online, go to step 9.
9. If you have any remaining nodes to update that are not configuration nodes, repeat this task for the next non-configuration node that is yet updated, starting at step 1.