Symptom: The switch reports incorrect CPU utilization in the output of the show system command.

Scenario: In certain conditions, the switch may fail to update the true CPU utilization information in the output of the show system command and keep reporting the last recorded value.

Workaround: Make a change to the polling interval of the system resources utilization using the system resource-utilization poll-interval command to get it out of the stuck state.

The default value for poll-interval is 10, hence toggle to some other value and then revert it back to 10. For example:

switch(config)# system resource-utilization poll-interval 15
switch(config)# system resource-utilization poll-interval 10

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: Port LED shows an incorrect link status.

Scenario: When 1G port is connected and negotiates a link with a peer device, the port status LED is amber instead of green.

Workaround: Use the show interface command to verify port link status.

Symptom: Port LED shows incorrect activity status for some of the 1G ports.

Scenario: When there is network activity on the 1G port, the port activity LED does not indicate the correct state. When the cable is removed while traffic is running through the port, the activity LED shows an inconsistent state.

Workaround: There is no functional impact on traffic.

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: Switch Interface remains stuck in "waiting for link state".

Symptom: The switch crashes and reboots when configuring MC-LAG.

Scenario: When adding a switch port to a multi-chassis interface, after removing the same port from the inter-switch-link (ISL) interface, the switch may crash and reboot.

Workaround: Add another port to the ISL interface, before configuring the previous port to a multi-chassis interface.

Symptom: The switch event logs show frequent additions and deletions of L3 host entries.

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: Switch fails to process some jumbo packets.

Scenario: Packets with a size greater than 9166 are not processed by the switch.

Workaround: Configure the switch MTU size to less than 9166.

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 switch is not able to apply certain ACLs.

Scenario: When an ACL in configured with VLAN rules (e.g.: "permit any any any vlan 40"), the switch does not apply the ACL and it returns an error message similar to

configuration does not match active configuration.
run 'access-list all reset' to reset all access-lists to match active configuration.

Workaround: Remove the ACL rule for VLAN. Please note, that packets will not be filtered on VLAN tags.

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: Switch LEDs incorrectly report the associated event or component state.

Workaround: Use the show environment fan | led | power-supply | temperature command for more details on the operation status of the respective component. The switch component alerts are also reported in the output of the show events command.

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: 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: The switch interface displays a lacp-block state.

Scenario: When using a 1000SX transceiver in an LAG interface, the switch interface may get stuck in lacp-block status in the output of the CLI command show lacp interfaces after a switch reboot. For example:

Actor details of all interfaces:
------------------------------------------------------------------------------
Intf   Aggr   Port    Port   State   System-id         System  Aggr Forwarding
       Name   Id      Pri                              Pri     Key  State     
------------------------------------------------------------------------------
1/1/1  lag1(mc)1029    1     ALFNCD  98:f2:b3:68:01:30 65534   1    up        
1/1/2  lag2(mc)1037    1     ALFOE   98:f2:b3:68:01:30 65534   2    lacp-block
1/1/3  lag3(mc)1012    1     ALFNCD  98:f2:b3:68:01:30 65534   3    up

Workaround: Turn the interface on the peer side where this issue is seen OFF and then ON.

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:A link fails to come up.

Scenario: After a hot-swap of one type of 40G transceiver with a different type (for example, DAC with Optical), the switch link may fail to come up.

Workaround: Remove the transceiver and wait at least 15-20 seconds before re-inserting it.

Symptom: The switch is not be able to identify 40G transceiver JH231A with part number 1990-4554.

Scenario: When 40G transceiver JH231A with part number 1990-4554 is inserted in the switch port, the transceiver is not properly identified in the output of the show interface transceiver or show interface brief commands and the interface is not linked up.

1/1/32    unknown     ??        ??              ??

Workaround: Use a transceiver with part number 1990-4557 instead of 1990-4554.

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.