User's Manual

The recommended setting is 16K. If this field is set to Default, you will receive the following error
message:
The format operation did not complete because the cluster count is
higher than expected.
Importing Windows dynamic disk volumes
If you create a snapshot, snapclone, or mirrorclone with a Windows 2003 RAID-spanned dynamic
volume on the source virtual disk, and then try to import the copy to a Windows 2003 x64 (64-bit)
system, it will import with Dynamic Foreign status. The following message displays in the DiskPart
utility:
The disk management services could not complete the operation.
This error occurs because the 64-bit version of DiskPart fails to import dynamic RAID sets on a new
server.
To avoid this issue, use the 32-bit version of DiskPart instead of the 64-bit version. Copy DiskPart
from a 32-bit x86 Windows system, located in C:\WINDOWS\system32. Place the DiskPart utility
in a temporary folder on the 64-bit x64 Windows system.
Losing a path to a dynamic disk
If you are using Windows 2003 with dynamic disks and a path to the EVA virtual disk is temporarily
lost, the Logical Disk Manager (LDM) will erroneously show a failed dynamic volume. For more
information, see the following issue on the Microsoft knowledge base website:
http://support.microsoft.com/kb/816307
To resolve the issue, reboot the Windows 2003 server to restore the dynamic volume.
Microsoft Windows 2003 MSCS cluster installation
The MSCS cluster installation wizard on Windows 2003 can fail to find the shared quorum device
and disk resources might not be auto-created by the cluster setup wizard. This is a known Windows
Cluster Setup issue that has existed since Windows 2003 was released.
There are two possible workarounds for this problem:
• Follow the workaround recommendation described in the Microsoft support article entitled
Shared disks are missing or are marked as "Failed" when you create a server cluster in
Windows Server 2003 (ID 886807), available for download on the Microsoft website:
http://support.microsoft.com/default.aspx?scid=KB;EN-US;886807
• Use the MPIO DSM CLI to set the load balancing policy for each LUN to NLB.
Microsoft is currently working on a resolution to address this issue.
Array I/O can halt after a power cycle with an HP-UX host
I/O halts after power cycling an HSV300-S enclosure connected to an HP-UX 11.23 host with
HP-UX driver versions earlier than B.11.23.08.02. To resolve this issue, either upgrade to version
B.11.23.08.02, or reset the HBAs on any server affected after an HSV300-S is power cycled.
Reboot can be required after EVA power cycle to resume access to LUNs
If an EVA4400 experiences a power cycle while the management GUI for Windows MPIO DSM
is open, the LUNs might not appear under Disk Management after they recover. This issue is seen
when the MPIO DSM Manager snap-in is open at the time of a power cycle. To reduce the
probability of this occurring, HP recommends that you close the GUI when you are not performing
management actions.
If this problem occurs, follow these steps:
28 EVA4400 operation