HP 3PAR X9000 File Serving Software (IBRIX) Implementation Guide (QL226-96987, June 2013)

Persistent Cache is supported on all quad-node and larger HP 3PAR arrays, including the HP 3PAR

F400 Storage System.

Controller Node Redundancy

Controller nodes are configured in logical pairs in such a way that each controller node has a

partner. The two partner nodes have redundant physical connections to the subset of physical

drives owned by the node pair. Within the pair, nodes mirror their write cache to each other, and

each serves as the backup node for the logical disks owned by the partner node.

If a controller node were to fail, data availability would be unaffected. Upon the failure of a

controller node, the node failover recovery process automatically flushes the dirty write cache to

the physical drive, and transfers ownership for the logical disks owned by the failed node to its

partner node.

The Persistent Cache feature allows the surviving node to mirror write data to another node, enabling

systems with four or more nodes to survive a node failure with minimal performance impact. In a

system with two nodes, the node failover recovery process puts all logical disks owned by the

remaining partner node in write-through (non-cached) mode. Furthermore, under certain

circumstances, the system is capable of withstanding a second node failure (however rare) without

affecting data availability. After the node failover recovery process for the initial node failure is

complete, a second controller node from the remaining node pairs can fail without causing system

downtime.

Controller nodes are hot-pluggable and can be serviced or added to a system online and

non-disruptively. Similarly, the HP 3PAR OS and other associated node software can be upgraded

online and without disruption.

The following procedure describes how to execute the Persistent Cache feature using the HP 3PAR

OS CLI.

1. Create a mounted file system on the StoreAll 9300 hosts using the LUNs that are exported to

the StoreAll 9300 from HP 3PAR. See “Mounting a File System” (page 28).

2. When I/O is heavy on the file system, reboot one of the controller nodes as shown in the

following example:

inservvm cli%shutdownnode reboot 3

-----System Information----- System Name : inservvm System Model : HP_3PAR 7400 Serial Number : 9903993

--------------Partial List Of Hosts---------------

Id Name Persona -WWN/iSCSI_Name- Port IP_addr

0 eos172i1 Generic 50014380167CB13C 3:1: 2 n/a

1 eos172s2 Generic 50014380167CD3C 0 3: 1:2 n/a

2 eos172s3 Generic 50014380167CC59C 3: 1:2 n/a

3 eos172s4 Generic 500143800251BE3C 3:1:2 n/a

Shutting down this node will impact applications running on the hosts displayed above

Do you REALLY want to REBOOT this node? yes or no: yes

The node will REBOOT in 15 seconds