Replace a Controller

Use this procedure to replace a controller in your HPE storage array.

Time required:	45 minutes per controller
Part(s) required:	Replacement controller(s) ordered from Nimble Storage
Information required:	RMA (Return Materials Authorization) work order Case number Array name or serial number The array serial number can be found on the back of Controller A, as shown below.
Tools required:	Phillips screwdriver ESD (Electrostatic Discharge) wrist strap
Support contact:	See Support.

NOTE:

HPE recommends using the Maintenance Window functionality within the HPE InfoSight portal to inform HPE Support of upcoming maintenance windows.

Refer to https://infosight.hpe.com/dashboards/nimble/wellness/maintenance-windows for more information.

You can disable automated support case generation during planned maintenance from the Manage Maintenance Windows page. After a maintenance window is set, no further communication with HPE Support is required.

NOTE:

When replacing a controller, reuse the USB device from the old controller. If the old USB device cannot be used, contact Nimble Storage Support.
If one controller fails, the other controller takes over. The array can function with one healthy controller while the replacement is installed.
Replace the array controller if any of its components, such as the NVDIMM, have failed.
Replacement controllers are shipped without the PCIe cage installed. Before installing the replacement controller, you must remove the PCIe cage from the failed controller and install it into the replacement controller.

Caution:

Before you attempt to replace the controller:

Do not remove the failed controller until you are ready to install the replacement.
Verify that the replacement controller matches the part shown on the RMA work order.
Determine the NimbleOS version running on the array, and the state of the controllers.
See Check Array Model Number, NimbleOS Version, and Operating State Using the GUI.

Prepare the failed controller for replacement.
1. Log in to the array using the CLI.
2. Verify that the failed controller is in Standby mode.
  ctrlr --list --array array_name
3. If the failed controller is in Active mode, perform a manual failover.
  failover --array array_name
4. Verify that the manual failover placed the failed controller in Standby mode.
  ctrlr --list --array array_name
5. Halt the standby controller.
  halt --array array_name --controller {A | B}
  Example
```
NimbleOS $ halt --array array123 --controller A
```
Remove the failed controller.
1. Label all the cables (to ensure proper recabling), then remove them from the controller.
2. Loosen the keeper screw to release the controller handle.
  IMPORTANT: Use caution while loosening the keeper screw to avoid the screwdriver tip slipping through the fan grill and hitting the fan blade.
  1. Keeper screw
  2. Handle
  3. Thumbscrew
  Caution:
  Do not remove the power cord above the controller. If retention clips are installed, it is recommended to secure the clips around the power cord to avoid accidental power cord removal/displacement.
3. Grasp the failed controller by the handle, pull down, and use a firm, steady pressure to pull the controller out of the chassis bay. Use two hands to support the controller during removal.
  
  IMPORTANT:
  To avoid electrostatic damage, place array components on a grounded anti-static mat. Use an ESD wrist strap during removal and replacement.
IMPORTANT: During controller replacement, you must remove the USB stick from the failed controller and install it into the replacement controller. Failure to perform the next step will prevent the replacement controller from coming online.
Install the USB stick from the failed controller.
All components are fully visible when the controller is removed from the chassis bay.
1. Locate the USB stick in the failed controller.
  1. USB port
  2. Fans
  3. Controller handle
2. Gently but firmly pull the USB stick up and out of its socket.
3. Mark or label the USB stick from the failed controller.
4. Remove the USB stick from the replacement controller and set that USB stick aside.
5. Insert the USB stick from the failed controller into the new controller. You should feel it seat into the USB slot.
Move the PCIe cage from the failed controller to the replacement controller.
1. Loosen the green thumbscrew to unsecure the PCIe cage from the failed controller.
2. Gently pull the PCIe cage up from the failed controller to unseat the PCIe cage riser card from the connector.
  1. PCIe cage
  2. Controller panel
  3. Thumbscrew
3. Carefully align the PCIe cage riser card with the connector on the replacement controller, then gently seat the PCIe cage into place.
4. Fasten the green thumbscrew to secure the PCIe cage to the controller.
Install the replacement controller.
1. Slide the replacement controller into the chassis bay. Do not force the controller. If it does not slide in easily, remove it and retry.
2. After the new controller is in place, return the controller handle to its original closed position and tighten the keeper screw.
3. Reconnect all the cables to the new controller (as labeled from step 2.)
Verify that the new controller is powered on.

Once installed, the replacement controller powers on automatically. The controller LEDs on both the left and right side of the array remain on. If the controller does not power on automatically, contact Nimble Storage Support.

NOTE: If firmware updates are required, the new controller can take from 15 to 45 minutes to appear in the UI. If the controller takes longer than 45 minutes to report an Active/Standby state, contact Nimble Storage Support for assistance.
Return the failed controller as directed on the RMA work order.