HP AE370A - Brocade 4Gb SAN Switch 4/12, Руководство администратора

Страница: 336 / 447

336 Administering ISL Trunking
Monitoring traffic
To implement ISL Trunking effectively, you must monitor fabric traffic to identify congested paths or to
identify frequently dropped links. While monitoring changes in traffic patterns, you can adjust the fabric
design accordingly, such as by adding, removing, or reconfiguring ISLs and trunking groups in
problem areas.
There are three methods of monitoring fabric traffic:
• Advanced Performance Monitoring monitors traffic flow and allows you to view the impact of different
fabric configurations on performance. Refer to ” Administering Advanced Performance Monitoring
(APM) ” on page 327 for additional information.
• Fabric Watch allows you to monitor traffic flow through specified ports on the switch and send alerts
when the traffic exceeds or drops below configurable thresholds. Refer to the Fabric Watch
Administrator’s Guide for additional information.
• Use the
portPerfShow
command as described in the following procedure to record traffic volume for
each port in your fabric over time.
To use the
portPerfShow
command
1. Connect to the switch and log in as admin.
2. Enter the following command:
where
interval
is the number of seconds between each data-gathering sample (the default is one
sample every second).
3. Record the traffic flow for each port participating in an ISL.
4. Repeat step 1 through step 3 for each switch in the fabric until all ISL traffic flow is captured.
In a large fabric, it might be necessary to only identify and capture the key ISLs. However, you might
want to continue this process throughout the day (or an entire work cycle), to capture varying traffic
patterns under different conditions.
The following example shows a switch without trunking, and indicates that ports 0 through 2 are
underutilized and ports 4 and 5 are congested:
The following example shows traffic flowing through a trunking group (ports 5, 6, and 7). After port 6
fails, traffic is redistributed over the remaining two links in the group, ports 5 and 7:
portperfshow [interval]
switch:admin> portperfshow
0 1 2 3 4 5 6 7 Total
--------------------------------------------------------------------
0 0 0 145m 204m 202m 0 168m 719
0 0 0 145m 206m 208m 0 186m 745
switch:admin>
switch:admin> portperfshow
0 1 2 3 4 5 6 7 Total
--------------------------------------------------------------------
0 0 0 0 0 145m 144m 145m 434
0 0 0 0 0 144m 143m 144m 431
0 0 0 0 0 162m 0 162m 324
0 0 0 0 0 186m 0 186m 372
0 0 0 0 0 193m 0 192m 385
0 0 0 0 0 202m 0 202m 404
0 0 0 0 0 209m 0 209m 418
switch:admin>