Chassis Replacement

This section explains how to replace the chassis on a Qumulo node running on hardware you supply, transferring drives and components into the replacement chassis.

This procedure is generic. For mechanical specifics (rail release, lid removal, drive carrier handling, fault LEDs), follow the service manual provided by your hardware vendor. We strongly recommend having another person help you with this process.

Step 1: Prepare for Chassis Replacement

1. Shut down the node. You can use the power button at the front of the node, or connect to the node by using SSH and run the sudo shutdown -h now command.

2. Disconnect all cables from the back of the node.

3. Remove the node from the server rack. Follow your hardware vendor’s rail-release procedure.

4. Place the node on a flat, static-safe surface.

5. Open the chassis lid as described in your hardware vendor’s service manual.

Step 2: Transfer Drives and Components

1. Transfer the boot drive from the original chassis to the replacement chassis.

2. Transfer the data drives (NVMe SSDs and HDDs) from the original chassis to the replacement chassis, keeping each drive in the same physical bay position in the replacement.

   Important
   Use the physical drive bay (as labeled on the chassis and shown in your vendor's drive-bay diagram) to identify which drive to pull. Any other identifier — the Qumulo Core drive ID shown in qq cluster_slots, the Web UI, and failure alerts — is a logical software identifier and does not correspond to a physical chassis bay. Pulling the wrong drive can cause additional data loss or an unnecessary second drive failure.

3. Transfer any add-in cards (such as NICs and host bus adapters) into the same PCIe slots in the replacement chassis.

4. Install the replacement chassis in the server rack.

5. Reconnect the power and networking cables to the node.

6. Power on the node.

Step 3: Verify the Node Has Rejoined the Cluster

1. Log in to the Qumulo Web UI.

2. Verify that the node has rejoined the cluster and no red banner appears at the top of the page. The reprotect process might have started automatically.

3. If the node was previously recused from the cluster, reintroduce it by running:

   /opt/qumulo/sbin/reintroduce_node.sh
   

Step 4: Reconfigure Out-of-Band Management (IPMI)

If the current IPMI settings for your node are configured statically (rather than by using DHCP), you must reapply the static IPMI settings to the replacement node. For more information, see Configuring Out-of-Band Management.

For more information, see Default IPMI Usernames and Passwords 🔒 on Qumulo Care.