How do I troubleshoot LACP issues on a Direct Connect LAG?

Resolution

Note: If you receive errors when you run AWS Command Line Interface (AWS CLI) commands, then see Troubleshooting errors for the AWS CLI. Also, make sure that you're using the most recent AWS CLI version.

Check LAG and connection state

Use one of the following methods to check the status of your LAG and your connection.

Use the Direct Connect console

Complete the following steps:

1. Open the Direct Connect console.
2. In the navigation pane, choose LAGs.
3. Select your LAG and then check the status of the LAG and each member connection's state.
   For Minimum links, note the configured value.
   For Connections, confirm that the number of available connections meets the minimum links requirement.

Use the AWS CLI

To check your LAG status, run the following describe-lags AWS CLI command:

aws directconnect describe-lags --lag-id example-LAG-ID

Note: Replace example-LAG-ID with the ID of your LAG.

To check a specific connection's status, run the following describe-connections AWS CLI command:

aws directconnect describe-connections --connection-id example-connection-ID

Note: Replace example-connection-ID with the ID of your connection.

Check CloudWatch metrics

Check the ConnectionState Amazon CloudWatch metrics for your Direct Connect connection. If your ConnectionState value is set to 0, then your connection status is DOWN.

To check optical signal levels, use CloudWatch to check the ConnectionLightLevelTx and ConnectionLightLevelRx Direct Connect metrics.

The levels must be within the range of -14.4–2.50 decibel-milliwatts (dBm) for 1 gigabit per second (Gbps) and 10 Gbps Direct Connect connections. For 100 Gbps connections, keep the ConnectionLightLevelTx value between -4.3 and 4.5 dBm. For ConnectionLightLevelRx, keep the value between -10.6 and 4.5 dBm. There are four optical lanes for 100 and 400 Gbps connections. You must check the light signal for each optical lane.

Check the AWS Health Dashboard

Planned or emergency maintenance events cause your Direct Connect connections to go down. Check your Health Dashboard for notifications of planned maintenance or unplanned outages.

Check your LACP configuration

Make sure that you correctly configured your LACP for Direct Connect.

For the LACP mode, make sure that you configured the following settings:

• Make sure that the AWS side is set to Active and that the CGW side is set to Active or Passive (dynamic LACP).
  Note: If both sides are Passive, then the LAG doesn't come up.
• Make sure that static mode is off. AWS requires LACP protocol negotiation.
• For the LACP timer, confirm that the CGW side of your connection uses Fast mode. If the CGW side uses Slow mode, then you experience delayed failover detection. AWS uses Fast mode.
• To check that you correctly configured the member link requirements, make sure that all connections in a LAG have the same bandwidth. You can't mix 1 Gbps and 10 Gbps. Also, confirm that all connections terminate at the same Direct Connect endpoint.

For minimum links, the number of active links is greater than or equal to the minimumLinks value that you configured for your LAG.

Note: Check this setting when you add or remove connections from a LAG.

Check the CGW-side configuration

Run the following commands from your CGW to check its configuration.

Note: Command syntax for each vendor varies depending on the operating system (OS) version. Refer to the vendor's official documentation for exact syntax.

Interface and LAG operational status

To check the operational status of your interface and LAG, run one of the following commands based on your OS:

• For Junos OS:

show interfaces terse | match lag-name

Note: Replace lag-name with the name of your LAG.

• For Cisco IOS:

show etherchannel summary

LACP activity mode and partner state

To check your LACP activity mode and partner state, run one of the following commands based on your OS:

• For Junos OS:

show lacp interfaces aeinterface-number

Note: Replace interface-number with the Aggregated Ethernet number of your AE interface. For more information, see show lacp interfaces on the HPE Juniper Networking website.
In the following example, the number of the Aggregated Ethernet interface is 1:

show lacp interfaces ae1

• For Cisco IOS:

show lacp neighbor

In the output, take the following actions:

• Check that the Actor (CGW) mode is Active or Passive. AWS doesn't support static on modes for LACP.
• Confirm that the Partner (AWS) mode is Active.
• Check that the Partner System ID is consistent across all member links.

LACP timer configuration

To check the configuration of your LACP timer, run of the following show commands based on your OS:

• For Junos OS:

show lacp interfaces interface-name

Note: Replace interface-name with the name of your interface.

• For Cisco IOS, run the following command to check the LACP settings of the local device:

show lacp internal

Also, run the following command to check the LACP settings on the AWS side:

show lacp neighbor detail

Note: You must run both previous commands.

In the output, check that the periodic timer matches the AWS side.

Check optical signal levels (CGW side)

To check your optical signal levels, run one of the following show commands based on your OS:

• For Junos OS:

show interfaces diagnostics optics interface-name

Note: Replace interface-name with the name of your interface.

• For Cisco IOS:

show interfaces transceiver

Related information

Direct Connect link aggregation groups (LAGs)

Monitor with Amazon CloudWatch

AWS Direct Connect maintenance