Removal and replacement procedures 42
5.
Disconnect the cable from the cache module or PCIe controller.
6. Remove the FBWC capacitor pack.
The server may look different than the one shown in the following illustration, depending on the
server model.
To replace the component, reverse the removal procedure.
Recovering data from the flash-backed write cache
If the server fails, use the following procedure to recover data temporarily stored in the FBWC.
CAUTION: Before starting this procedure, read the information about protecting against
electrostatic discharge ("Preventing electrostatic discharge" on page 26).
1. Perform one of the following:
o Set up a recovery server using an identical server model. Do not install any internal drives or
FBWC in this server. (Hewlett Packard Enterprise recommends this option.)
o Find a server that has enough empty drive bays to accommodate all the drives from the failed
server and that meets all the other requirements for drive and array migration.
2. Power down the failed server ("Power down the server" on page 25).
3. Transfer the drives from the failed server to the recovery server.
4. Perform one of the following:
o If the array controller has failed, remove the cache module and capacitor pack from the failed
array controller, and install the cache module and capacitor pack on an identical array controller
model in the recovery server.
o If the server has failed, remove the controller, cache module, and capacitor pack from the failed
server, and install the controller, cache module, and capacitor pack in the recovery server.
5. Power up the recovery server. If there was data in the cache at the time of the controller or server
failure, a 1792 POST message appears, stating that valid data was flushed from the cache. This
data is now stored on the drives in the recovery server. You can now transfer the drives (and
controller, if one is used) to another server.
To Next Page
Loading...
