HP P6000 Continuous Access Implementation Guide (T3680-96431, August 2012)

also assumed that the packet loss and network latency jitter on the networks falls within HP’s defined
acceptable range for HP P6000 Continuous Access replication.
• HP currently supports three FCIP router families: B-series, C-series, and HP IP Distance Gateway.
The FCIP configuration can have either single or dual long-distance network links between the
local and remote gateways. Both the Brocade and the Cisco FC-IP routers support IP
acceleration, which optimizes the SCSI protocol across the IP network by reducing the number
of round trips. The HP IP Distance Gateway does not currently support IP acceleration capability.
Data compression remains supported on all of these products. For detailed information on
FCIP router support, see the HP SAN Design Reference Guide.
• In a single long-distance IP network link solution, consider enabling the FCIP router IP
acceleration capability to optimize the SCSI protocol on the IP network.
• If IP acceleration is not enabled in a dual network environment, the FCIP router load-balances
I/O equally across the two links. If IP acceleration is enabled in a dual network environment,
both the Brocade and Cisco FCIP routers will force all replication traffic to use a single network
path, even if redundant long-distance IP networks are available. This behavior is characteristic
of these routers and cannot be changed.
• In a redundant replication network solution, consider enabling IP acceleration only if both
links are capable of independently meeting the solution’s required RPO. IP acceleration should
not be enabled if either one of the networks is not capable of meeting the solution’s RPO.
• When using two equivalent (QOS) redundant long distance IP networks that are independently
capable of meeting the solution’s RPO, consider enabling IP acceleration.
• HP SCSI-FCP also allows a tunnel to use any or all of the available paths to the remote
controller, which may improve overall performance. To enable the simultaneous use of more
than one array host port for tunnel traffic, the ports used for data replication on both the source
and destination arrays should be set for an equivalent priority.
NOTE:
◦ These priority settings should only be applied to ports running the HP P6000 Continuous
Access protocol.
◦ If it is necessary to change the data replication protocol setting with existing DR groups,
data replication will stop until both arrays in the HP P6000 Continuous Access relationship
are set to a compatible protocol.
Tunnel thrash
Tunnel thrash is the frequent closing and opening of a tunnel while holding host I/O in the transition.
This occurs when peer controllers can see each other, but cannot sustain replication on any path.
Tunnel thrash can be caused by the following conditions:
• High volumes of packet loss
• Incorrectly configured routers
• Rerouted IP circuits
• Oversubscribed circuits
Although tunnel thrash is rare, if it occurs a critical event is placed in the controller event log and
displayed in HP Command View EVA. An event will be logged for each DR group that shares the
affected tunnel.
Planning the data replication protocol 41