Server Recovery on the EX490 and EX495

From MediaSmartServerWiki
Jump to: navigation, search

This document describes the behavior of the Server Recovery and Factory Reset process used by the HP MediaSmart Server EX490, EX495, and X510 DataVault servers. This process is new for these server models, instead of using an on-board USB flash drive for booting the server, it is instead booted from an image that is loaded over the network from a Client PC. This has caused significant concern[1] and problems[2] amongst owners of these models of the server.


Overview

The MediaSmart Server is a headless device with no keyboard or display available to reload the operating system. Instead, a special DVD is included that enables the user to rebuild the server.

Server Recovery and Factory Reset are the mechanisms provided by Windows Home Server to allow the user to replace the operating system of their Home Server. Factory Reset is a destructive process that will erase all drives connected to the Home Server. Server Recovery will only replace the Windows Home Server operating system on the 20GB system partition of the system drive that resides in Bay 1 of the MediaSmart Server, and is used to recover from a failed system drive or a corrupt operating system. A Server Recovery should leave your data intact while a Factory Reset will erase all drives connected to the server.

Both the Server Recovery and Factory Reset process is performed in the same manner, with the only difference being the operation chosen by the user when they perform from the client PC. Both of these operations will hereafter be referred to as Server Recovery to simplify the documentation of the process.

The EX470/EX475/EX485/EX487 all contained a 256MB onboard USB flash memory that stored a special Windows PE boot image. This image allowed the server to boot into Server Recovery mode, where it was able to transfer the OS image from a client PC and apply it to the system drive of the server. See How_do_I_perform_a_Server_Recovery? and How_do_I_perform_a_Factory_Reset? for more details on this process.

With the EX490, EX495 and X510 DataVault HP has removed this onboard flash and instead uses a PXE boot process to boot the server over the network and requires the server and Client PC to be directly cabled together. This has some advantages in that it removes the possibility of DNS issues, but also introduces new possible failure points, specifically with the need to acquire an IP Address via DHCP from the Client PC that is performing the Server Recovery, and with the transfer and loading of the boot files.

Process

HP has documented the process in the User Guide[3] on pages 213 - 217 from the end user perspective. Unfortunately the details and troubleshooting tips often aren't adequate to resolve issues encountered during the Recovery process. Please review the instructions from HP for an understanding of how this process works.

  • Some users suggest it best to boot the client PC into "Safe Mode with Networking" in order to avoid many different problems with the recovery process.
  • Insert the Server Recovery DVD into the client PC. It should auto-run, if not then manually launch WHSRecovery.exe
  • Disable any wireless connections on the client PC.
  • Use an ethernet cable to directly connect the MediaSmart Server to your client PC. I have also performed a server recovery over a network switch, the important part is you want no other DHCP servers running on the "network" that the server and client are using.
  • Proceed through the Server Recovery process and put the Server into Server Recovery mode
  • You most likely will see an application flash on the screen and then disappear. This is the Tftpd application and will reside on the system tray of your client PC. It is the DHCP server that supplies an IP address to your client PC and to the Server. It also is an ftp server that transfers the Windows PE boot environment to the server. This is the most fragile and critical part of the Server Recovery process.
  • At this point, mysterious things happen until either you are presented with an error (Step 2 or Step 3 failed) or you are presented with a full green progress bar that states your server is ready for recovery.
  • Proceed through the remainder of the process, choosing whether you want to perform a Server Recovery or a Factory Reset.
  • When the Recovery completes, finish up the process, allowing the software to automatically reboot your server. Connect your Client PC and Server to your home network again, and use the PC Install CD to set up your server.

Troubleshooting

The following is an example of a Server Recovery, with the benefit of a display connected to the server and some troubleshooting tips and images that will hopefully help you work through any issues you may encounter. The images will follow the process described above. I performed the recovery on my netbook running Windows XP SP3, over a 100mbit connection wired into a gigabit switch, with the server being the only other device on the switch. This is not required, you can directly wire the server and client together, this was just more convenient for my setup. The server contained a single hard drive with all partitions removed, there was no data on this drive and the recovery process had no issues with the blank drive. I performed the Factory Reset process, but the steps are the same for a Server Recovery.

  • Server Recovery application is running, the Tftpd application is running (green icon in system tray) and I've double-clicked the icon on the system tray to bring the Tftpd application up.
Tftpd app.jpg


  • Navigate to the DHCP Server tab of the Tftpd application, you should see the client with the 169.37.37.37 IP address, and the server with the 169.37.37.38 address. This took approximately 30 seconds for the server to acquire the IP address. If the client or server cannot acquire an IP address, then the Server Recovery application will eventually fail with the "Failed at step 2 of 4" message.
Dhcpgood.jpg


Dhcpgood server.jpg


  • Navigate to the Tftp Server tab of the Tftpd application, you should see the Windows PE image file (BOOT.WIM) transferring to the server. This took just over 3 minutes to complete. If there is a failure in transferring the boot image to the server, then the Server Recovery application will eventually fail with the "Failed at step 3 of 4" message. This could be caused by a firewall on the client, a problem with the hard drive in the server, or possibly other reasons.
Boot.wimtransfer.jpg


Boot.wimtransfer server.jpg


  • When the boot image finishes transferring to the server, it will attempt to boot off of the image. This step also took about 3 minutes to complete.
Inrecoverymode.jpg


Inrecoverymod server.jpg


  • Choose Server Recovery or Factory Reset.
Choose.jpg


  • After starting the Factory Reset, it took about 20 minutes to transfer the full Windows Home Server image to the server at which point the client application reported success.
Done.jpg


References

  1. http://www.mediasmartserver.net/forums/viewtopic.php?f=25&t=6307
  2. http://www.mediasmartserver.net/forums/viewtopic.php?f=25&t=5662
  3. http://h10032.www1.hp.com/ctg/Manual/c01868379.pdf