BROCADE 10 Gbps Fibre Channel over Ethernet PCIe Converged Network Adapter and 8 Gbps & 4 Gbps Fibre Channel PCIe Host Bus Adapter Release Notes - Software Release 2.3.0.3 (March 2011)

Brocade Adapter v2.3.0.3 Release Notes Page 13 of 19
10. In this release the WinPE image creation is limited to Windows 2008 x64 and
IA64. As most of the systems are x64, it is recommended to use platform other
than x86 e.g. x64 or IA64.
11. Windows Crashdump Configuration. On a system that is configured with Boot
From SAN on a Brocade HBA, the following steps should be followed:
BIOS must be enabled on all HBA port instances that can see the boot
LUN.
Same discovery mechanism (i.e., First LUN/Auto/Flash under "Boot
Device Settings" in BIOS config menu or bcu CLI) should be used for all
HBA port instances for which the boot LUN is exposed.
If multiple storage ports (having unique PWWNs) are configured to access
the same boot LUN in the storage array and all PWWNs are zoned to a
specific HBA port instance, then all such PWWNs must be listed under
“Boot Device Settings” in BIOS config menu or bcu CLI.
Note: If CLI or HCM GUI is used to configure a boot LUN, a reboot is required
for the change to be effective.
12. Windows 2008 x86 boot over SAN installation may suspend on some servers
(e.g. Dell R900). To minimize this issue, install the Brocade HBA on any one of
the PCIE slots, while leaving the other slots empty. Then, install Windows 2008
onto a boot over SAN LUN connected to one of the HBA ports. Once the
installation is complete, shutdown the system to populate the rest of the empty
slots.
13. In boot over SAN setup with Windows 2003 R2 SP2 x86 and a HDS AMS 1000
Array, the system may halt during the OS installation (after the first reboot).
Reboot the system and continue with the installation.
14. For Windows 2003, the Windows device manager may require a manual re-scan
to show the current devices after a zone change.
Linux:
1. When Network Manager is enabled, network traffic may not pass through multiple
CNA network interfaces. The workaround is to disable Network Manager or set up
Network Manager to use the appropriate NIC interface.
2. Always use KMP binary RPMs for SLES11.x and avoid using NOARCH RPMs. If a
non-KMP RPM of driver is installed over this, any YaST2-based install will revert
back to the inbox version of the driver.
3. On SLES11 SP1, after FC-Director reboot, not all paths to the target device
might recover. The workaround is to delete the offline device and do a rescan.
4. On Linux RH5.4, server may get stuck or reboots in rare case on the Linux
RedHat screen during or after typing "linux dd" to inject Network driver DUD. This
issue happens very rarely and retrying the same steps normally makes it work.