System information

12-120
Cisco SN 5428-2 Storage Router Software Configuration Guide
78-15471-01
Chapter 12 Command Line Interface Reference
fcip destination config
Command History
Usage Guidelines Each FCIP instance requires three active elements:
• The networkif element assigns an interface and IP address for use by the FCIP peer.
• The destination element assigns the peer’s protocol and IP address.
• The device interface element associates the FCIP instance with an internal Fibre Channel interface
and is automatically assigned and enabled.
An FCIP instance runs with a point-to-point connection to an FCIP instance on a peer SN 5428-2 Storage
Router. Each FCIP instance requires one and only one peer. Both FCIP instances must be configured to
use the same connection protocol, TCP/IP or raw IP.
TCP/IP connection protocol uses the FCIP standard and TCP flow control and error recovery algorithms.
FCIP, using TCP connections, allows you to configure TCP receive and transmit window sizes. Raw IP
uses a proprietary protocol, but allows you to configure a wider variety of operational settings, providing
more granular control over flow control and error recovery. All changes to raw IP operational parameters
are applied immediately. Changes to TCP operational parameters are not applied until the FCIP instance
is stopped and restarted, or the storage router is rebooted.
Raw IP connections have unique flow control settings for network transmissions and FC transmissions.
The FC transmission counter (the frinhiwater keyword) limits the number of frames given to the FC
interface for transmission. The FC frames are only acknowledged once they have been transmitted out
of the SN 5428-2. The network transmissions are controlled by both a byte counter (the burstouthiwater
keyword) and a frame counter (the frouthiwater keyword). FCIP will not transmit data when either
counter has reached its high water mark. The data is queued until the counters have receded from their
high water mark.
Raw IP uses four error recovery algorithms. These algorithms use five different operational settings,
which control which error recovery algorithm is used.
• rexmitcount—The maximum number of times a packet can be retransmitted, before it is discarded.
• maxtimeout—The maximum amount of time, in ticks, that can be used for any one retransmission,
before the packet is discarded.
• timeoutincrement—The amount of time, in ticks, to add to a packet's time out value before
retransmitting the packet.
• initialtimeout—The initial amount of time, in ticks, to delay before retransmitting a packet.
• totaltimeout—The maximum amount of time, in ticks, that a packet is kept alive, before it is
discarded.
The following are the available error recovery algorithms:
1. For error recovery using maxtimeout and rexmitcount:
– timeout = maxtimeout / rexmitcount--
For example, using a maxtimeout value of 48 and rexmitcount value of 4 would result in
retransmissions at 12, 16, 24 and 48 ticks. This is the default error recovery algorithm.
2. For error recovery using timeoutincrement and rexmitcount:
– timeout += timeoutincrement
For example, using a timeoutincrement value of 8 and a rexmitcount value of 4 would result in
retransmissions at 8, 16, 24 and 32 ticks.
Release Modification
3.3.1 This command was introduced.