HP StorageWorks Fabric OS 5.X Diagnostics and System Error Messages Reference Guide (AA-RVHZB-TE, September 2005)

Fabric OS 5.x diagnostics and system error messages reference guide 83
• cfgTransAbort
• zoneObjectCopy
• zoneObjectExpunge
• zoneObjectRename
Dual-CP systems
For the Core Switch 2/64, SAN Director 2/128, and 4/256 SAN Director, each CP has a unique error
log, depending on which CP was active when that message was reported. To fully understand message
logging on these switches, you should enable the system logging daemon because the logs on the host
computer are maintained in a single merged file for both CPs and are in sequential order. Otherwise, you
must examine the error logs in both CPs, particularly for events such as firmwareDownload or
haFailover, for which the active CP changes.
For the security violations such as telnet, HTTP, and serial connection violations are not propagated
between CPs. Security violations on the active CP are not propagated to the standby CP counters in the
event of a failover, nor do security violations on the standby CP get propagated to the active CP counters.
System logging daemon
The system logging daemon (syslogd) is a process on UNIX, Linux, and some Windows systems that reads
and logs messages as specified by the system administrator.
Fabric OS can be configured to use a UNIX-style syslogd process to forward system events and error
messages to log files on a remote host system.
The host system can be running UNIX, Linux, or any other operating system that supports the standard
syslogd functionality.
Configuring for syslogd involves configuring the host, enabling syslogd on the SilkWorm model, and,
optionally, setting the facility level.
For information on configuring syslogd functionality, refer to the <Italic>Fabric OS Administrator’s Guide.
Port logs
The Fabric OS maintains an internal log of all port activity. Each switch or logical switch maintains a log
file for each port. Port logs are circular buffers that can save up to 8000 entries per logical switch. When
the log is full, the newest log entries overwrite the oldest log entries. Port logs capture switch-to-device,
device-to-switch, switch-to-switch, some device A-to-device B, and control information. Port logs are not
persistent and are lost over power cycles and reboots.
Run the portLogShow command to display the port logs for a particular port.
Run the portLogEventShow command to display the specific events reported for each port.
Refer to the <Italic>Fabric OS Administrator’s Guide for information on interpreting results of the
portLogDump command.
Port log functionality is completely separate from the system message log. Port logs are typically used to
troubleshoot device connections.
Panic dump and core dump files
The Fabric OS creates panic dump files and core files when there are problems in the Fabric OS kernel.
These files can build up in the kernel partition (typically because of failovers) and might need to be
periodically deleted or downloaded using the saveCore command. In case of a panic dump, the files
can be viewed with the pdShow command.
The software watchdog process (SWD) is responsible for monitoring daemons critical to the function of a
healthy switch. The SWD holds a list of critical daemons that ping the SWD periodically at a
predetermined interval defined for each daemon.
If a daemon fails to ping the SWD within the defined interval, or if the daemon terminates unexpectedly,
then the SWD dumps information to the panic dump files, which helps to diagnose the root cause of the
unexpected failure.