AB291A Fabric Clustering System Support Guide (12-port Switch), April 2004

Table Of Contents
Chapter 6
Monitoring and Troubleshooting
Monitoring and Troubleshooting the Switch
122
0x10: This flow shows the Related Multipack Protocol (RMPP) related processing.
0x20: This flow shows the partition management related processing.
0x1000: This flow shows the internal state machine (may create more flow for more granularity for path,
mcast, …).
0x2000: This flow shows the user configuration management related processing.
0x4000: This flow triggers the fabric switch Subnet Manager to take a snap shot of the current logging
information for the trace.
For example, use:
0x4001 to take a snapshot of the SM logging information.
0x4002 to take a snapshot of the SA information.
0x8000: This flow triggers the MAD packet to be displayed.
For example, use:
0x8002 to display MAD packets and SA related processing.
0x8001 to display SM MAD packets and SM related processing.
Troubleshooting the HP Fabric Clustering System Network
Ping
As on any IP network, you can use the UNIX ping command to test connectivity between IPoIB
fabric-attached hosts. For more information, see “Verifying Connectivity” in Chapter 3.
Setting Trace Levels
Enabling Tracing Tracing is disabled by default. To set a trace level for the purpose of debugging or
troubleshooting, perform the following steps:
Step 1. Enter global-configuration mode.
HB-IB> enable
HB-IB configure
HB-IB(config)
Step 2. View the trace help output, as the number of applications and modules may have changed.
HP-IB(config)# trace app ?
app <1-25>
app numbers:
APP_ID_CLI = 1
APP_ID_OSPF = 2
APP_ID_RIP = 3
APP_ID_IP_AGENT = 22
APP_ID_FIB_AGENT = 23
APP_ID_KERNEL = 24
APP_ID_CARD_AGENT = 25
APP_ID_SM = 26