Symptom: Traffic mirroring does not function properly.

Scenario: If a redundancy failover is performed while the line modules are still initializing, after the failover, traffic mirroring for enabled mirroring sessions may not work correctly.

Workaround: Disable and re-enable the mirroring session after the failover is completed and all line modules are fully initialized.

Symptom: VLAN interfaces are always up, even when they are not active on any physical port.

Scenario: In a split-brain condition, the VLAN interfaces are not brought down when all the ports associated with them are also down.

Symptom: The gateway IP address randomly becomes unreachable.

Scenario: In an Active-Gateway setup, the gateway IP address may randomly become unreachable.

Workaround: Disable the ISL link between the switches, remove and re-add the active-gateway configuration.

Symptom/Scenario: The switch experiences CPU utilization spikes in a relatively steady environment.

Workaround: These CPU utilization spikes are transient.

Symptom: The switch experiences intermittent DHCP forwarding failures.

Scenario:The switch incorrectly sets source port 68 for DHCP Relay packets.

Symptom: The switch event log is incorrectly classified as WARN.

Scenario: The switch is incorrectly classifying the debug event log A transaction has failed with error code 5 as a warning (WARN) event.

Workaround: Disregard the WARN classification, as this is really a DEBUG message.

Symptom: The switch incorrectly forwards ARP requests through the inter-switch link instead of the correct switch LAG interface.

Scenario: After disabling a switch interface followed by removal and addition to a LAG interface, the switch may incorrectly forward the traffic through the ISL link instead of the correct LAG interface.

Workaround: Disable one link on either of the LAG interfaces.

Symptom: In certain cases, spanning tree fails to converge.

Scenario: When an interface participating in the spanning tree is changed from L2 to L3 and back to L2 interface, spanning tree may fail to converge.

Workaround: Disable and re-enable spanning tree whenever converting an L2 interface to an L3 interface.

Symptom: The network experiences spanning-tree instability issues.

Scenario: In a mixed spanning-tree topology with an ArubaOS-CX switch running RPVST on VLAN 1 and interoperating with a peer device running RSTP or RPVST, the spanning-tree may experience instability issues and frequent topology changes (TCN).

Workaround: Disable and re-enable the extended system-id on the ArubaOS-CX switch.

Symptom: The switch event log reports a crash for the 'rsyslogd' process.

Scenario: In certain conditions, the switch may report in event logs a crash for "rsyslogd" process in a message similar to:

rsyslogd crashed due to signal:6

Workaround: The process will automatically restart after the crash and generate a core dump file listed in the show core-dump all command.

Symptom/Scenario: When deleting an interface with a PIM-RP configuration, the RP-Candidate configuration remains.

Workaround: Remove the RP configuration prior to deleting the interface.

Symptom: The switch enters a hung state and fails to reboot or failover to the second management module (if running on a chassis switch with dual management modules).

Scenario: When multiple IGMP reports for well known multicast group addresses are received, over time the switch may enter into a hung state and fail to reboot the switch to the second management module (if running on a chassis switch with dual management modules).

Workaround: Monitor switch memory utilization and if it is observed to increase over time, manually reboot the entire switch or switch over to the second management module to prevent entering the hung state. If the switch is already in a hung state, reboot the switch to clear the hung state.

Symptom: The switch fails to fallback to an alternate route.

Scenario: When using the default static route and specific static routes, the switch fails to fallback to an alternate route if the primary path is disabled.

Workaround: Disable and re-enable the default route.

Symptom: The SNMP process randomly crashes.

Scenario: When there are simultaneous SNMP queries processed by the switch, such as SNMP walks and show tech collection, the SNMP process may crash and generate a core file listed in the show core-dump all command.

Workaround: The SNMP process will restart immediately after the crash.

Symptom: Spanning tree enters an inconsistent state.

Scenario: After a switch reboot, the switch interfaces participating in the spanning tree path may be incorrectly initialized causing the spanning tree topology to enter into an inconsistent state and potentially cause network loops.

Symptom: The DHCP-relay debug message displays with an incorrect severity level.

Workaround: This is a debug message indicating the valid reason for the packet discard.

Symptom: The switch experiences traffic loss on non-ECMP next-hop routes.

Scenario: After one of the MCLAG nodes is rebooted, the switch may fail to redirect non-ECMP next-hop routes over the MCLAG and the switch may experience traffic loss on these links.

Workaround: Disable the affected non-ECMP links or reboot both MCLAG switches to clear the state.

Symptom: The SSH daemon crashes with an error similar to signal - 11.

Workaround: Use RADIUS authentication for switch SSH access or configure the "priv-lvl" TACACS+ attribute in lower case.

Symptom: Unexpected messages are displayed on the switch console.

Scenario: After issuing the boot system command, the switch may randomly display unexpected messages on the switch console, similar to [FAILED] Failed unmounting /run/netns.

Workaround: None. These messages do not have any impact to the switch or protocol functionality.

Symptom: Network clients are not able to obtain an IPv6 address from some DHCPv6 servers.

Scenario: When the switch is configured as a DHCPv6 relay agent, network clients may not be able to obtain an IPv6 address from some DHCPv6 servers.

Workaround: Do not allow the DHCP server to use the UDP source port from the packet forwarded by the agent.

Symptom/Scenario: When running multiple and repetitive SNMP queries, the switch memory utilization may increase over time.

Workaround: If observed switch memory utilization increasing over time, disable the SNMP agent on the switch using the no snmp-service vrf mgmt | default command and then re-enable the agent using the snmp-service vrf mgmt | default command.

Symptom: The switch incorrectly places ports in "blocking" state.

Scenario: In an MSTP configuration, if an event (such as disabling or disconnecting a port) is causing a topology change, switch ports may be incorrectly placed in "blocking" state, potentially causing two switches to become root and preventing the spanning-tree topology from properly converging. When this condition happens, the received and sent BPDU counters do not match in the output of the show spanning-tree detail command.

Workaround: Rebooting the switch will clear the incorrect port status and allow the spanning tree topology to properly converge.

Symptom: In certain conditions, the switch experiences traffic loss.

Scenario: In a switch configured in an MCLAG topology with VRRP, when there is a MAC or ARP aging event or when the events are cleared using the clear mac-address [ port | vlan ] <PORTNAME | VLAN-ID> or clear arp commands, the switch may experience traffic drops.

Workaround: Reboot the switch.

Symptom: Unable to remove the password for a BGP neighbor.

Scenario: When attempting to remove the BGP neighbor password using the no neighbor <ip-address> password <password-string> command, the configured password is not removed.

Workaround: Remove the <password-string> from the command, using just no neighbor <ip-address> password.

Symptom: A route that should be denied per the AS path filter list is being permitted.

Scenario: A denied route in the AS path list is selected as best route in the BGP table when the routemap has a permit rule.

Workaround: Add a deny rule to the route map to filter out the route.

Symptom: The switch does not properly honor the configured "estabilished" flag on TCP matching criteria.

Scenario: The switch incorrectly applies the "estabilished" flag to TCP synchronizing (SYN) packets matching the criteria.

Workaround: Do not use established flag in TCP matching criteria.

Symptom/Sceanario: The switch fails to display the active-gateway configuration in the output of the show running-config interface <IFNAME> command.

Workaround: Use the show running-config command to display the active-gateway configuration for the VLAN interfaces.

Symptom: The switch incorrectly allows the user to configure a policy action for PCP remarking.

Scenario: Although the switch does not support policy action for PCP remarking, the CLI configuration is not correctly rejected.

Workaround: Do not use PCP remarking with a policy action.

Symptom: The switch terminates all active sessions to the switch.

Scenario: When using show commands, the switch may abruptly terminate the active session to the switch.

Workaround: Use the specific interface ID show command, show interface <interface_id>.

Symptom/Scenario: Switch fails to display the configured description for the LAG interface in the output of the show running-config command.

Workaround: Reconfigure the description on the interface.

Symptom/Scenario: Under certain conditions, the switch ports appear in a "waiting for link" state.

Symptom: A switch port does not successfully come up.

Scenario: When a switch port which is a member of a LAG interface with an ACL applied is moved to another LAG interface, the port fails to come up in the new LAG interface.

Workaround: Reboot the entire chassis.

Symptom: The switch sends the incorrect event severity level for loop detection messages.

Scenario: The switch sends the syslog messages for loop detection as informational instead of warning messages.

Workaround: Consider the loop detection message as having a warning level, rather than an info level.

Symptom: The switch does not correctly display the loop detection.

Scenario: When loop protection action is configured for do-not-disable, the switch fails to update the loop detection status in the output of the show loop-protect command. For example:

Interface 1/1/1 
 Loop-protect enabled        : Yes 
 Action on loop detection    : Do not disable 
 Loop detected count         : 0 
 Loop detected               : No 
 Interface status            : up 
	 

Workaround: Verify the loop detection in the switch event logs:

|hpe-lpd|2803|LOG_INFO|AMM|1/5|Loop detected on port 1/1/1

Symptom: The switch line cards crash after execution of the show ntp status command.

Scenario: When the switch is configured with an NTP server and the switch time is subsequently manually changed using CLI commands, it may cause a software abort error in the dune-agent and the switch line modules entering an endless reboot cycle. The switch will generate event logs similar to

systemd-coredump|1201|LOG_CRIT|LC|1/1|dune_agent_0 crashed due to signal:6
hpe-cardd|3207|LOG_ERR|AMM|1/5|Line module 1/1 has failed: ASIC Error
hpe-cardd|3207|LOG_ERR|AMM|1/5|Line module 1/1 has failed: ASIC Error

Workaround: Perform the following:

1. Remove NTP configuration from the switch.

    no ntp server <ip_address>

2. Configure the switch with manual date and time

   clock time <time>
   clock date <date>
   

3. Save the switch configuration and reboot the switch.

Symptom/Scenario: After a period of runtime, the switch fails to display the web Interface.

Workaround: Reboot the switch or preform a redundancy switchover to the standby management module, if available.

Symptom/Scenario: When configuring peer-group route-reflector-client, the BGP sessions belonging to a peer-group do not get automatically reset.

Workaround: Use the clear ip bgp peer-group <peer-group> command to clear the BGP sessions.

Symptom: The AS number is displayed as zero.

Scenario: The output of the show ip bgp summary command is incorrect when the neighbor is assigned to the peer-group without configuring a remote AS neighbor.

Workaround: Configure the neighbor's remote AS before assigning to the peer-group.

Symptom: A switch with a configuration file out of the box may fail to correctly apply the added configuration.

Workaround: Zeroize the switch management module before additional configuration is added to the switch using the erase all zerioze command in the config context.

Symptom: SNMP traps are not working.

Scenario: If the trap server is running on the management VRF, SNMP traps do not work.

Workaround: Run the SNMP server on the default VRF.

Symptom: Switch cannot enable loop protection on some interfaces.

Scenario: Switch cannot enable loop protection on a VLAN when all VLANs are allowed on a trunk interface. For example:

interface 1/1/24
   no routing
   vlan trunk allowed all
   loop-protect vlan 1000 

interface lag1
    vlan allowed all
    loop-protect  

Workaround: Have specific VLANs in the allowed list and then enable loop protection.

Symptom: In certain conditions, the switch may not properly detect network loops.

Scenario: When loop protection is enabled on an interface, the switch does not properly detect the network loops generated in the network segment connected on that switch interface.

Workaround: This issue is specific to switches running ArubaOS-Switch as the loop pair. Identify the redundant link and disable the alternate path.

Symptom: OSPF interface statistics are not accurate with overlapping IP addresses across VRFs.

Scenario: If the same IP addresses are used on more than one VRF, then the OSPFv2 statistics are updated only on one VRF.

Workaround: Use only non-overlapping IP addresses under different VRFs.

Symptom: Virtual link is not functioning as expected.

Scenario: Virtual link may not function as expected if the remote end is initially configured to be area 0 and later changed to a different area.

Workaround: Avoid area ID changes "on the fly". You can reboot the switch to solve the issue, if needed.

Symptom: sFlow sampling will not work when line modules are in non-consecutive slots.

Scenario: If there is a gap between line modules, sFlow sampling will only work on the lowest numbered module before the gap. For example, if slots 1, 3, and 4 have line modules installed, but slot 2 does not, sFlow sampling will only work on the ports on line module 1.

Workaround: Install line modules in contiguous order starting at slot 1.

Symptom/Scenario: When using the Firefox browser, a dialog box reporting an error on attempt to upload new firmware is displayed.

Workaround: When seeing this dialog, user needs to log out and clear the cache in the browser, then try to log in again and reattempt the upload. To clear cache, enter URL: about:preferences, then type cache in the search. Select Clear Now button to clear cache.

Symptom: Switch does not display the proper error message in the output of show class <...> CLI commands

Scenario: When there is no class configured, the switch does not display the expected No Class found message in the output of show class <...> CLI commands.

Workaround: Use CLI command show running-config to verify classifier configuration.

Symptom: Loop protection may not be enabled on all VLANs.

Scenario: Switch fails to enable loop protection on all VLANs using the CLI command loop-protect vlan all in the interface context.

The switch also returns an error message similar to VLAN <VLAN-ID> is not configured on the interface when enabling loop-protect on an interface where all VLANs are allowed.

interface <INTF-NAME>
   no routing
   vlan trunk allowed all
   loop-protect vlan <VLAN-ID>

Workaround: Use a VLAN range to enable loop-protect on all VLANs, such as loop-protect vlan <1-4094>.

Symptom: An error string is shown in the event log about invalid command no shutdown for the loopback interface.

Scenario: The loopback interface displays no shutdown in the running configuration; however, the shutdown and no shutdown commands are not available for the loopback interface. This results in error messages when importing the configuration to the switch. This error can be ignored and will not impede importing the configuration file.

Workaround: This is a display issue only, with no functional impact.

Symptom/Scenario: When running snmpwalk, the entPhysicalVendorType is not correctly updated.

Workaround: Use the following CLI commands to find out the vendor information.

show system
show module

Symptom/Scenario: Out of the box, OOBM and SSH are enabled. After factory reset, the OOBM port and the SSH server are disabled.

Workaround: After the factory reset, manually enable the OOBM port and start the SSH server.

Symptom: PSU insert shows input fault state and blinking green LED.

Scenario: When a PSU is hot inserted or re-inserted into slot, PSU shows blinking green LED and the show environment power-supply command displays input fault.

Workaround: Remove the PSU and wait 5-10 seconds before re-inserting.

Symptom: Line modules are rebooted causing momentary traffic loss until the modules are back in a "Ready" state.

Scenario: When a VLAN which has IGMP snooping enabled gets deleted without removing the IGMP snooping configuration, all the line modules see a fatal error and get rebooted.

Workaround: Before deleting the VLAN, un-configure IGMP snooping on the VLAN.

Symptom/Scenario: Incorrect serial numbers for switch components may be displayed in the System Info widget of the Overview web page of the Web UI.

Workaround: Use the show module command to display the correct serial numbers.

Symptom: The switch fails to update the switch image version.

Scenario: After uploading a new switch image version using the Web UI Firmware Update page, the switch fails to update the information on the Web UI page.

Workaround: Use the browser Refresh button to refresh the information displayed on the Firmware Update page.

Symptom: Switch may fail to honor a "drop" policy action.

Scenario: When a rate-limit and drop action are combined in a single policy, the switch fails to honor the drop action.

Workaround: Configure separate policies for drop and rate-limit actions.

Symptom: Switch may fail to display the diagnostic information for some line modules.

Scenario: Switch may fail to display the diagnostic information such as memory or CPU for some line modules. An error message similar to System resource utilization data not available displays when using the show system resource-utilization module <MODULE-ID> command.

Workaround: Use the show system resource-utilization command to check memory, CPU, and open FDs.

Symptom: A LAG IP address may not be correctly programmed.

Scenario: If the IP address is assigned to a LAG interface before assigning port members to the LAG interface, the LAG IP address may not be correctly programmed when the line module is reloaded or after a system reboot.

Workaround: Remove and re-assign the IP address to the LAG interface.

Symptom: A LAG interface may enter a blocked state.

Scenario: When all line modules are removed and reinserted, a LAG interface status may show blocked in the output of the show lacp interfaces command.

Workaround: Reload the module with the affected LAG interface using the boot line-module <MODULE-ID> command.

Symptom: LLDP MIB does not display Port Description values.

Scenario: Network Management Systems fail to retrieve port description information from LLDP MIB.

Workaround: Use the show lldp neighbor-info command to check neighbor details..

Symptom: The switch may fail to re-enable an interface disabled by loop protection.

Scenario: When the loop protection timer is configured on an interface, the switch fails to re-enable the interface after consecutive disable/enable triggers on the interface due to loop detections.

Workaround: Once the loop is removed, manually enable the interface using the no shutdown command from the interface context.

Symptom: Network Analytics Engine (NAE) allows for empty parameter values to be present. The NAE user interface does not correctly display this particular case properly, and you may see red colored error text when editing a parameter that indicates that the value is missing or invalid; however, you can ignore it and continue to save edited parameters. The UI also does not display default values specified in the Python script while editing an agent. Empty parameters display red error text, indicating that the value is missing or invalid when it is in fact valid.

Scenario: When a script with empty parameter values is created and instantiated via REST and then the agent is edited in the UI, NAE displays red error text, indicating a value is missing or invalid, when it really is valid.

Workaround: The issue is visual only, you can save the agent without changing any parameter values, or cancel updating the agent.

Symptom: Switch may fail to honor OSPF active/passive mode configuration.

Scenario: When OSPF passive mode is configured globally for all interfaces, OSPF interfaces configured individually for active mode may enter passive mode after a switch reboot. Similarly, when OSPF global passive-interface configuration is enabled on another VRF, the OSPF interfaces on the default VRF are set to passive mode after the switch reboot.

Workaround: Avoid creating a global "passive-interface default" configuration and configure the interface individually.

Symptom: SNMP traps are not working.

Scenario: If the trap server is running on the management VRF, SNMP traps do not work.

Workaround: Run the SNMP server on the default VRF.

Symptom: The web UI displays a timeout error if the request takes more than 60 seconds to complete.

Scenario: When using the web UI to perform various operations, a timeout error may be displayed if the request takes more than 60 seconds to complete. One example of this is trying to copy one configuration to another (i.e. copy a configuration to startup-config).