![background image](http://html1.mh-extra.com/html/ibm/totalstorage-ds300/totalstorage-ds300_best-practices-manual_3462750060.webp)
42
IBM TotalStorage DS300 and DS400 Best Practices Guide
Logical drives defined on a DS300 controller are visible only on the iSCSI ports
of the controller that owns the array. The alternate controller in a dual controller
DS300 does not see arrays that are defined on the other controller. Nor does the
alternate controller respond with information such as passive paths as is done in
the DS400.
Each DS300 controller has an Eth2 and Eth3 port for iSCSI traffic (see
Figure 3-5 on page 34). Each of these ports must be assigned to a separate
network segment. The Eth2 ports for both controller A and B must be assigned a
different IP address in the same network segment, while the Eth3 ports for both
controller A and B must be assigned a different IP address in another network
segment.
The DS300 supports four modes for IP failover:
remote
the IP interfaces will failover between Eth2
→
Eth2 and
Eth3
→
Eth3 between the controllers
none
the IP interfaces will not failover between controllers
local
the IP interfaces will failover from Eth2
→
Eth3 or Eth3
→
Eth2
within the same controller only
both
the IP interfaces will first failover between the two ports on the
same controller and then to the remote interface for that port
on the alternate controller. For example, Eth2[A]
→
Eth3[A]
→
Eth3[B]
→
Eth2[B]
For example, assume the DS300 is configured to support remote failover mode.
The controller failover is initiated when a controller goes down or when a link to a
controller goes down (a host bus adapter link failure does not cause a failover).
The IP addresses of both Eth2 and Eth3 interfaces, and the arrays owned by the
failing controller, are failed over to the alternate controller. If Eth2 port on
controller A fails, sessions are failed over to Eth2 port on Controller B. The IP
address on Eth2 port on controller A is migrated to Eth2 port on controller B.
After failover occurs, the sessions are reconnected to the same IP address (but
on Eth2 port on controller B, with a different MAC address) and I/O continues.
Once the interface has failed over, and the IP address is online on Eth2 port on
controller B, a Gratuitous ARP is sent to invalidate the current MAC address. A
second Gratuitous ARP is sent to the switch to update the MAC address
forwarding table for the new MAC bound to that IP address.
It is important to allow enough time for all the failover processing to occur. After
the IP address has been moved to the new port (about 15 seconds), failover of
the logical drives is sequential. It takes about 45-60 seconds for a failover of a
single logical drive.
Summary of Contents for TotalStorage DS300
Page 2: ......
Page 20: ...2 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 28: ...10 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 226: ...208 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 228: ...210 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 240: ...222 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 450: ...432 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 581: ...Copyright IBM Corp 2006 All rights reserved 563 Part 4 Appendixes Part 4...
Page 582: ...564 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 638: ...620 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 640: ...622 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 648: ...630 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 664: ...646 IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 677: ...1 0 spine 0 875 1 498 460 788 pages IBM TotalStorage DS300 and DS400 Best Practices Guide...
Page 678: ......
Page 679: ......