Cisco MDS 9000 Family NX-OS System Management Configuration Guide

Saving Cores

You can save cores (from the active supervisor module, the standby supervisor module, or any switching module) to an external CompactFlash (slot 0) or to a TFTP server in one of two ways:

• On demand—Copies a single file based on the provided process ID.
• Periodically—Copies core files periodically as configured by the user.

A new scheme overwrites any previously issued scheme. For example, if you perform another core log copy task, the cores are periodically saved to the new location or file.

Saving the Last Core to Bootflash

This last core dump is automatically saved to bootflash in the /mnt/pss/ partition before the switchover or reboot occurs. Three minutes after the supervisor module reboots, the saved last core is restored from the flash partition (/mnt/pss) back to its original RAM location. This restoration is a background process and is not visible to the user.

Tip The timestamp on the restored last core file displays the time when the supervisor booted up not when the last core was actually dumped. To obtain the exact time of the last core dump, check the corresponding log file with the same PID.

To view the last core information, enter the show cores command in EXEC mode.

To view the time of the actual last core dump, enter the show process log command in EXEC mode.

First and Last Core

The first and last core feature uses the limited system resource and retains the most important core files. Generally, the first core and the most recently generated core have the information for debugging and, the first and last core feature tries to retain the first and the last core information.

If the core files are generated from an active supervisor module, the number of core files for the service is defined in the service.conf file. There is no upper limit on the total number of core files in the active supervisor module.

To display the core files saved in the system, use the show cores command.

Online System Health Management

The Online Health Management System (OHMS) (system health) is a hardware fault detection and recovery feature. It ensures the general health of switching, services, and supervisor modules in any switch in the Cisco MDS 9000 Family.

The OHMS monitors system hardware in the following ways:

• The OHMS component running on the active supervisor maintains control over all other OHMS components running on the other modules in the switch.
• The system health application running in the standby supervisor module only monitors the standby supervisor module, if that module is available in the HA standby mode.

The OHMS application launches a daemon process in all modules and runs multiple tests on each module to test individual module components. The tests run at preconfigured intervals, cover all major fault points, and isolate any failing component in the MDS switch. The OHMS running on the active supervisor maintains control over all other OHMS components running on all other modules in the switch.

On detecting a fault, the system health application attempts the following recovery actions:

• Performs additional testing to isolate the faulty component.
• Attempts to reconfigure the component by retrieving its configuration information from persistent storage.
• If unable to recover, sends Call Home notifications, system messages and exception logs; and shuts down and discontinues testing the failed module or component (such as an interface).
• Sends Call Home and system messages and exception logs as soon as it detects a failure.
• Shuts down the failing module or component (such as an interface).
• Isolates failed ports from further testing.
• Reports the failure to the appropriate software component.
• Switches to the standby supervisor module, if an error is detected on the active supervisor module and a standby supervisor module exists in the Cisco MDS switch. After the switchover, the new active supervisor module restarts the active supervisor tests.
• Reloads the switch if a standby supervisor module does not exist in the switch.
• Provides CLI support to view, test, and obtain test run statistics or change the system health test configuration on the switch.
• Performs tests to focus on the problem area.

Each module is configured to run the test relevant to that module. You can change the default parameters of the test in each module as required.

Loopback Test Configuration Frequency

Loopback tests are designed to identify hardware errors in the data path in the module(s) and the control path in the supervisors. One loopback frame is sent to each module at a preconfigured frequency—it passes through each configured interface and returns to the supervisor module.

The loopback tests can be run at frequencies ranging from 5 seconds (default) to 255 seconds. If you do not configure the loopback frequency value, the default frequency of 5 seconds is used for all modules in the switch. Loopback test frequencies can be altered for each module.

Loopback Test Configuration Frame Length

Loopback tests are designed to identify hardware errors in the data path in the module(s) and the control path in the supervisors. One loopback frame is sent to each module at a preconfigured size—it passes through each configured interface and returns to the supervisor module.

The loopback tests can be run with frame sizes ranging from 0 bytes to 128 bytes. If you do not configure the loopback frame length value, the switch generates random frame lengths for all modules in the switch (auto mode). Loopback test frame lengths can be altered for each module.

Hardware Failure Action

The failure-action command controls the Cisco NX-OS software from taking any action if a hardware failure is determined while running the tests.

By default, this feature is enabled in all switches in the Cisco MDS 9000 Family—action is taken if a failure is determined and the failed component is isolated from further testing.

Failure action is controlled at individual test levels (per module), at the module level (for all tests), or for the entire switch.

Performing Test Run Requirements

Enabling a test does not guarantee that the test will run.

Tests on a specific interface or module only run if you enable system health for all of the following items:

• The entire switch
• The required module
• The required interface

Tip The test will not run if system health is disabled in any combination. If system health is disabled to run tests, the test status shows up as disabled.

Tip If the specific module or interface is enabled to run tests, but is not running the tests due to system health being disabled, then tests show up as enabled (not running).

Tests for a Specified Module

The system health feature in the NX-OS software performs tests in the following areas:

• Active supervisor’s in-band connectivity to the fabric.
• Standby supervisor’s arbiter availability.
• Bootflash connectivity and accessibility on all modules.
• EOBC connectivity and accessibility on all modules.
• Data path integrity for each interface on all modules.
• Management port’s connectivity.
• User-driven test for external connectivity verification, port is shut down during the test (Fibre Channel ports only).
• User-driven test for internal connectivity verification Fibre Channel.

Clearing Previous Error Reports

You can clear the error history for Fibre Channel interfaces, iSCSI interfaces, an entire module, or one particular test for an entire module. By clearing the history, you are directing the software to retest all failed components that were previously excluded from tests.

If you previously enabled the failure-action option for a period of time (for example, one week) to prevent OHMS from taking any action when a failure is encountered and after that week you are now ready to start receiving these errors again, then you must clear the system health error status for each test.

Tip The management port test cannot be run on a standby supervisor module.

Interpreting the Current Status

The status of each module or test depends on the current configured state of the OHMS test in that particular module (see Table 6-1 ).

The status of each test in each module is visible when you display any of the show system health commands. See the “Displaying System Health” section.

On-Board Failure Logging

The Generation 2 Fibre Channel switching modules provide the facility to log failure data to persistent storage, which can be retrieved and displayed for analysis. This on-board failure logging (OBFL) feature stores failure and environmental information in nonvolatile memory on the module. The information will help in post-mortem analysis of failed cards.

OBFL data is stored in the existing CompactFlash on the module. OBFL uses the persistent logging (PLOG) facility available in the module firmware to store data in the CompactFlash. It also provides the mechanism to retrieve the stored data.

The data stored by the OBFL facility includes the following:

• Time of initial power-on
• Slot number of the card in the chassis
• Initial temperature of the card
• Firmware, BIOS, FPGA, and ASIC versions
• Serial number of the card
• Stack trace for crashes
• CPU hog information
• Memory leak information
• Software error messages
• Hardware exception logs
• Environmental history
• OBFL specific history information
• ASIC interrupt and error statistics history
• ASIC register dumps

Saving Cores

Prerequisites

• Be sure to create any required directory before performing this task. If the directory specified by this task does not exist, the switch software logs a system message each time a copy cores is attempted.

Detailed Steps

To copy the core and log files on demand, follow this step:

To copy the core and log files periodically, follow these steps:

Examples

If the core file for the specified process ID (PID) is not available, you see the following response:

To copy the same PID with different instance number, do as follows:.

Clearing the Core Directory

Use the clear cores command to clean out the core directory. The software clears all the core files and other cores present on the active supervisor module.

Task Flow for Configuring System Health

Follow these steps to configure system health:

Step 1 Enable System Health Initiation.

Step 2 Configure Loopback Test Configuration Frequency.

Step 3 Configure Loopback Test Configuration Frame Length.

Step 4 Configure Hardware Failure Action.

Step 5 Perform Test Run Requirements.

Step 6 Clear Previous Error Reports.

Step 7 Perform Internal Loopback Tests.

Step 8 Perform External Loopback Tests.

Step 9 Perform Serdes Loopbacks.

Enabling System Health Initiation

By default, the system health feature is enabled in each switch in the Cisco MDS 9000 Family.

Detailed Steps

To disable or enable this feature in any switch in the Cisco MDS 9000 Family, follow these steps:

Configuring Loopback Test Configuration Frequency

Detailed Steps

To configure the frequency of loopback tests for all modules on a switch, follow these steps:

Configuring Loopback Test Configuration Frame Length

Detailed Steps

To configure the frame length for loopback tests for all modules on a switch, follow these steps:

Configuring Hardware Failure Action

Detailed Steps

To configure failure action in a switch, follow these steps:

Performing Test Run Requirements

Detailed Steps

To perform the required test on a specific module, follow these steps:

Clearing Previous Error Reports

Use the EXEC-level system health clear-errors command at the interface or module level to erase any previous error conditions logged by the system health application. The bootflash, the eobc, the inband, the loopback, and the mgmt test options can be individually specified for a given module.

The following example clears the error history for the specified Fibre Channel interface:

The following example clears the error history for the specified module:

The following example clears the management test error history for the specified module:

Performing Internal Loopback Tests

You can run manual loopback tests to identify hardware errors in the data path in the switching or services modules, and the control path in the supervisor modules. Internal loopback tests send and receive FC2 frames to and from the same ports and provide the round-trip time taken in microseconds. These tests are available for Fibre Channel, IPS, and iSCSI interfaces.

Use the EXEC-level system health internal-loopback command to explicitly run this test on demand (when requested by the user) within ports for the entire module.

Use the EXEC-level system health internal-loopback command to explicitly run this test on demand (when requested by the user) within ports for the entire module and override the frame count configured on the switch.

Use the EXEC-level system health internal-loopback command to explicitly run this test on demand (when requested by the user) within ports for the entire module and override the frame length configured on the switch.

Note If the test fails to complete successfully, the software analyzes the failure and prints the following error:

External loopback test on interface fc 7/2 failed. Failure reason: Failed to loopback, analysis complete Failed device ID 3 on module 1

Performing External Loopback Tests

You can run manual loopback tests to identify hardware errors in the data path in the switching or services modules, and the control path in the supervisor modules. External loopback tests send and receive FC2 frames to and from the same port or between two ports.

You need to connect a cable (or a plug) to loop the Rx port to the Tx port before running the test. If you are testing to and from the same port, you need a special loop cable. If you are testing to and from different ports, you can use a regular cable. This test is only available for Fibre Channel interfaces.

Use the EXEC-level system health external-loopback interface interface command to run this test on demand for external devices connected to a switch that is part of a long-haul network.

Use the EXEC-level system health external-loopback source interface destination interface interface command to run this test on demand between two ports on the switch.

Use the EXEC-level system health external-loopback interface frame-count command to run this test on demand for external devices connected to a switch that is part of a long-haul network and override the frame count configured on the switch.

Use the EXEC-level system health external-loopback interface frame-length command to run this test on demand for external devices connected to a switch that is part of a long-haul network and override the frame length configured on the switch.

Use the system health external-loopback interface force command to shut down the required interface directly without a back out confirmation.

Note If the test fails to complete successfully, the software analyzes the failure and prints the following error:

External loopback test on interface fc 7/2 failed. Failure reason: Failed to loopback, analysis complete Failed device ID 3 on module 1

Performing Serdes Loopbacks

Serializer/Deserializer (serdes) loopback tests the hardware for a port. These tests are available for Fibre Channel interfaces.

Use the EXEC-level system health serdes-loopback command to explicitly run this test on demand (when requested by the user) within ports for the entire module.

Use the EXEC-level system health serdes-loopback command to explicitly run this test on demand (when requested by the user) within ports for the entire module and override the frame count configured on the switch.

Use the EXEC-level system health serdes-loopback command to explicitly run this test on demand (when requested by the user) within ports for the entire module and override the frame length configured on the switch.

Note If the test fails to complete successfully, the software analyzes the failure and prints the following error:

External loopback test on interface fc 3/1 failed. Failure reason: Failed to loopback, analysis complete Failed device ID 3 on module 3.

Configuring OBFL for the Switch

Detailed Steps

To configure OBFL for all the modules on the switch, follow these steps:

	Command	Purpose
Step 1	switch# config terminal switch(config)#	Enters configuration mode.
Step 2	switch(config)# hw-module logging onboard	Enables all OBFL features. Note This CLI only enable OBFL features that are disabled by no hw-module logging onboard command. For OBFL features that were individually disabled, please enable those using hw-module logging onboard obfl-feature command.
	switch(config)# hw-module logging onboard cpu-hog	Enables the OBFL CPU hog events.
	switch(config)# hw-module logging onboard environmental-history	Enables the OBFL environmental history.
	switch(config)# hw-module logging onboard error-stats	Enables the OBFL error statistics.
	switch(config)# hw-module logging onboard interrupt-stats	Enables the OBFL interrupt statistics.
	switch(config)# hw-module logging onboard mem-leak	Enables the OBFL memory leak events.
	switch(config)# hw-module logging onboard miscellaneous-error	Enables the OBFL miscellaneous information.
	switch(config)# hw-module logging onboard obfl-log	Enables the boot uptime, device version, and OBFL history.
	switch(config)# no hw-module logging onboard	Disables all OBFL features.

Configuring OBFL for a Module

Detailed Steps

To configure OBFL for specific modules on the switch, follow these steps:

Clearing the Module Counters

Restrictions

• The module counters cannot be cleared using Device Manager or DCNM-SAN.

Detailed Steps

To reset the module counters, follow these steps:

To reset the counters for all the modules, follow these steps:

Displaying System Processes

Use the show processes command to obtain general information about all processes (see Example 6-1 to Example 6-6).

Example 6-1 Displays System Processes

Where:

• ProcessId = Process ID
• State = process state.

– D = uninterruptible sleep (usually I/O).

– R = runnable (on run queue).

– S = sleeping.

– T = traced or stopped.

– Z = defunct (“zombie”) process.

• NR = not running.
• ER = should be running but currently not-running.
• PC = current program counter in hex format.
• Start_cnt = number of times a process has been started (or restarted).
• TTY = terminal that controls the process. A hyphen usually means a daemon not running on any particular TTY.
• Process Name = name Name of the process.

Example 6-2 Displays CPU Utilization Information

Where:

• MemAllocated = Sum of all the dynamically allocated memory that this process has received from the system, including memory that may have been returned
• Runtime CPU Time (ms) = CPU time the process has used, expressed in milliseconds.microseconds
• Invoked = number of times the process has been invoked.
• uSecs = microseconds of CPU time on average for each process invocation.
• 1Sec = CPU utilization in percentage for the last one second.

Example 6-3 Displays Process Log Information

Where:

• Normal-exit = whether or not the process exited normally.
• Stack-trace = whether or not there is a stack trace in the log.
• Core = whether or not there exists a core file.
• Log-create-time = when the log file got generated.

Example 6-4 Displays Detail Log Information About a Process

Example 6-5 Displays All Process Log Details

Example 6-6 Displays Memory Information About Processes

Where:

• MemAlloc = total memory allocated by the process.
• StackBase/Ptr = process stack base and current stack pointer in hex format.

Displaying System Status

Use the show system command to display system-related status information (see Example 6-7 to Example 6-10).

Example 6-7 Displays Default Switch Port States

Example 6-8 Displays Error Information for a Specified ID

Example 6-9 Displays the System Reset Information

The show system reset-reason command displays the following information:

• In a Cisco MDS 9513 Director, the last four reset-reason codes for the supervisor module in slot 7 and slot 8 are displayed. If either supervisor module is absent, the reset-reason codes for that supervisor module are not displayed.
• In a Cisco MDS 9506 or Cisco MDS 9509 switch, the last four reset-reason codes for the supervisor module in slot 5 and slot 6 are displayed. If either supervisor module is absent, the reset-reason codes for that supervisor module are not displayed.
• In a Cisco MDS 9200 Series switch, the last four reset-reason codes for the supervisor module in slot 1 are displayed.
• The show system reset-reason module number command displays the last four reset-reason codes for a specific module in a given slot. If a module is absent, then the reset-reason codes for that module are not displayed.

Use the clear system reset-reason command to clear the reset-reason information stored in NVRAM and volatile persistent storage.

• In a Cisco MDS 9500 Series switch, this command clears the reset-reason information stored in NVRAM in the active and standby supervisor modules.
• In a Cisco MDS 9200 Series switch, this command clears the reset-reason information stored in NVRAM in the active supervisor module.

Example 6-10 Displays System Uptime

Use the show system resources command to display system-related CPU and memory statistics (see Example 6-11).

Example 6-11 Displays System-Related CPU and Memory Information

Where:

• Load average—Displays the number of running processes. The average reflects the system load over the past 1, 5, and 15 minutes.
• Processes—Displays the number of processes in the system, and how many are actually running when the command is issued.
• CPU states—Displays the CPU usage percentage in user mode, kernel mode, and idle time in the last one second.
• Memory usage—Displays the total memory, used memory, free memory, memory used for buffers, and memory used for cache in KB. Buffers and cache are also included in the used memory statistics.

Displaying Core Status

Use the show system cores command to display the currently configured scheme for copying cores. See Examples 6-12 to 6-15 .

Example 6-12 Displays the Message when Cores are Transferred to TFTP

Example 6-13 Displays the Message when Cores are Transferred to the External CF

Example 6-14 Displays All Cores Available for Upload from the Active Supervisor Module

Example 6-15 Displays Logs on the Local System

Figure 6-1 Show Cores Dialog Box

Verifying First and Last Core Status

You can view specific information about the saved core files. Example 6-16 provides further details on saved core files.

Example 6-16 Regular Service on vdc 2 on Active Supervisor Module

There are five radius core files from vdc2 on the active supervisor module. The second and third oldest files are deleted to comply with the number of core files defined in the service.conf file.

Displaying System Health

Use the show system health command to display system-related status information (see Example 6-17 to Example 6-22).

Example 6-17 Displays the Current Health of All Modules in the Switch

Example 6-18 Displays the Current Health of a Specified Module

Example 6-19 Displays Health Statistics for All Modules

Example 6-20 Displays Statistics for a Specified Module

Example 6-21 Displays Loopback Test Statistics for the Entire Switch

Example 6-22 Displays Loopback Test Statistics for a Specified Interface

Note Interface-specific counters will remain at zero unless the module-specific loopback test reports errors or failures.

Example 6-23 Displays the Loopback Test Time Log for All Modules

Example 6-24 Displays the Loopback Test Time Log for a Specified Module

Verifying Loopback Test Configuration Frame Length

To verify the loopback frequency configuration, use the show system health loopback frame-length command.

Displaying OBFL for the Switch

Use the show logging onboard status command to display the configuration status of OBFL.

Displaying the OBFL for a Module

Use the show logging onboard status command to display the configuration status of OBFL.

Displaying OBFL Logs

To display OBFL information stored in CompactFlash on a module, use the following commands:

Displaying the Module Counters Information

This example shows the device IDs of all the devices in a module:

Monitoring the CPU Utilization

To display the system CPU utilization, use the show processes cpu command.

This example shows how to display processes and CPU usage in the current VDC:

Obtaining RAM Usage Information

You can obtain the processor RAM usage by using this SNMP variable: ceExtProcessorRam.

Monitoring Rx and Tx Traffic Counters

When monitoring Rx and Tx traffic counters, you should include the Rx counter OID:

Monitoring Status of Interfaces

To monitor status of interfaces, use the IETF extended-linkDown trap, which has ifAlias (this trap can set interface description) and ifDescr, which shows port name in the ascii format as shown below:

The following is an example of the trap:

Monitoring Transceiver Thresholds

Use the cisco-xcvr-mon-status-chg trap way to monitor digital diagnostics statistics for thresholds as shown below:

The trap MIB is as show below:

This example shows how to display transceiver details:

Configuring Supervisor Switchover Notification

The supervisor switchover notification can be monitored by listening for the ciscoRFSwactNotif trap:

Configuring a Counter to Include CRC and FCS Errors

You can include CRC and FCS errors of interfaces by polling dot3StatsFCSErrors counter as shown in the example:

dot3StatsFCSErrors Counter32

Configuring CallHome for Alerts

The call home feature enables receive call home e-mail when exceptions occurs in the system. Use the CLI or SNMP to setup call home configurationss and to enable all alert-groups as shown below:

Monitoring User Authentication Failures

You can monitor any user authentication failures by listening the authenticationFailure trap:

MIBs