Accusys ExaRAID GUI - Manual (Page 188)

Advanced Functions
5-15
5.1.6 MAC Multi-Path Solution
Mac OS X provides multi-path support on a basis since OS X 10.3.5,
providing both path redundancy and load balancing. Mac OS X
implements miltipathing at the target device level, and it requires that the
RAID controller presents the same World Wide Node Name (WWNN) to all
the host interfaces connected to the MAC systems. Please refer to
• Setting FC Worldwide Node Name on page 2-61 to select identical
WWNN. After restarting the RAID system to make this change effective,
then follow 5.1.3 Configuring MPIO Hosts and RAID Controller on page 5-2
to complete the LUN mapping configurations.
The Mac OS multipath I/O driver by default supports only round-robin I/O
policy. With Apple Xsan software, you may set the policy to be either
round-robin (rotate) or static (failover-only). For more information please
visit Apple web site and read “Apple Xsan Administrator’s Guide”.
5.1.7 VMware ESX Server Multi-Path Solution
VMware ESX Server 2.5 or the later by default is loaded with hardware-
independent multi-path drivers and management interface. After
completing the RAID system configurations and attaching the RAID
system to the host computer, you may use VMware Management
Interface or the command, vmkmultipath or esxcfg-mpth, at Service
Console to manage the paths.
There are three multi-path policy options supported:
1. fixed -- using user-predefined path.
2. mru -- using most recently used path.
3. rr -- using round-robin algorithm, which is available only after ESX Server
The first two polices are use only one active path, and move to a standby
path only when the active path is down. The third policy can use all paths
at the same time to deliver the best performance. You may use esxcfg-
advcfg to set path performance parameters. The single-controller RAID
system supports all the options.
When a cable is pulled, I/O freezes for approximately 30-60 seconds, until
the SAN driver determines that the link is down, and failover occurs.
During that time, the virtual machines (with their virtual disks installed on a
SAN) might appear unresponsive, and any operations on the /vmfs
directory might appear to hang. After the failover occurs, I/O should
resume normally.
After changing the configuration of LUN or paths, please use esxcfg-
rescan command to do rescan and esxcfg-vmhbadevs to know the
mapping between device names and the LUNs.
For more information, please refer to VMware ESX Server Administration
Guide and VMware ESX Server SAN Configuration Guide, or go to
http://www.vmware.com/.