Dell SmartFabric OS10 User Guide Release 10.5.3

Configuring delay-restore orphan port in VLT domain

Perform the following steps to configure delay-restore orphan ports on ethernet1/1/1:

Enter CONFIGURATION mode.
```
OS10# configure terminal
```
Enters INTERFACE CONFIGURATION mode.
```
OS10(config)# interface ethernet 1/1/1
```

Enable delay-restore orphan port.

OS10(conf-if-eth1/1/1)# vlt delay-restore orphan-port enable

Exit INTERFACE CONFIGURATION mode and enter CONFIGURATION mode.
```
OS10(conf-if-eth1/1/1)# exit
```
Enter VLT domain mode.
```
OS10(conf)# vlt-domain 1
```
Configure VLT delay-restore timer to 150 seconds. This command enables the VLT delay-restore timer to control the orphan ports.
```
OS10(conf-vlt-1)# delay-restore 150
```

After you configure delay-restore orphan ports on both the VLT peers, the behavior of the port in different scenarios are explained below:

Fresh VLT configuration

If VLT is not configured earlier, the ethernet1/1/1 interface is up on both the nodes.
When VLT is configured on both peers, VLT election occurs.
In the primary VLT peer, ethernet1/1/1 remains up.
In the secondary VLT peer, ethernet1/1/1 is brought down and delay-restore is started. A log indicating that the VLT delay-restore has started is thrown on console.
The ports are brought up after delay-restore timer expires. A syslog indicating that the VLT delay-restore timer has stopped is thrown on the console.

NOTE: A fresh VLT configuration is treated the same way as a reload case. All configured orphan ports irrespective of the ignore vlti-failure configuration are brought down.

NOTE: If VLT and DROP are configured in a system for the first time (or being converted from DRP), DROP is configured after the VLT election and after the initial delay-restore timer expires. This configuration must be applied if you do not want the orphan ports to be brought down.

VLT peer reloads and joins as secondary (single VLT peer save and reload):

During boot up, ethernet1/1/1 is kept down.
After the VLT domain is created locally and the VLT peer joins the VLT fabric as the secondary node, the VLT delay-restore timer is started. A syslog indicating that the VLT delay-restore has started is thrown on console.
After expiry, the ports are brought up. A syslog indicating that the VLT delay-restore timer has stopped is thrown on the console.

Reloading both Peers:

During boot up, ethernet1/1/1 in both the nodes is kept down.
VLT domain is created and election occurs.
Ethernet1/1/1 is brought up immediately in the primary VLT peer.
In the secondary VLT peer, ethernet1/1/1 remains down and delay-restore timer is started. A syslog indicating that the delay-restore timer has started is thrown on the console.
After the delay-restore timer expires, a syslog is thrown on the console and the port is brought up.

VLTi fail with heart-beat configured and up:

If VLTi fails and the VLT heart-beat is up, ethernet1/1/1 is brought down (since ignore vlti-failure configuration is disabled) immediately in the secondary VLT peer.
When VLTi recovers, the delay-restore timer starts and a syslog is thrown indicating that the timer has started is thrown on the console.
After the delay-restore timer expires, ethernet1/1/1 is brought up and a syslog is thrown on the console.
Ethernet1/1/1 is up in the primary VLT peer.

VLTi fail with heart-beat not configured or down:

When VLTi fails and the VLT heart-beat is down, both the VLT peers become primary (split brain).
Ethernet1/1/1 in both the VLT peers are kept up.
When VLTi recovers, election occurs.
The port remains up in the peer elected as the primary node.
In the secondary VLT peer, ethernet1/1/1 is brought down (since ignore vlti-failure configuration is disabled) and the delay-restore timer is started. A syslog indicating that the delay-restore timer has started is thrown on the console.
After the delay-restore timer expires, ethernet1/1/1 is brought up and a syslog is thrown on the console.

On VLTi failure, DROP-enabled ports are treated the same way as VLT port-channels (when ignore vlti-failure configuration is disabled) in all scenarios; except in a scenario where VLT MAC (common MAC) is not configured and VLT heart-beat is down or not configured.

If VLT MAC is not configured on the VLT peers, the primary VLT peer's MAC is used as the system ID in the LACP BPDUs. When VLTi fails, the VLT peers become split-brain, as there is no VLT heart beat. Each VLT peer starts sending its own system MACs as the system ID in the LACP BPDUs. As a result of this mismatch, the VLT port-channel arm of the old secondary peer becomes operationally down (since it changed its system ID from the primary's MAC to its own MAC) due to LACP protocol behavior. In this scenario, the DROP configured orphan ports will not be brought down. Any incoming traffic from the orphan ports is black-holed as the VLT port-channel arm and VLTi are operationally down.

After VLTi recovers, both DROP-enabled ports and VLT port-channels are brought down in the peer elected as the secondary node for running delay-restore timer. After the timer expires, DROP-enabled ports and port-channels are brought up.

NOTE: It is recommended to always have VLT MAC and heart-beat configured.

Perform the following steps to configure orphan ports to ignore VLTi failures:

Enter CONFIGURATION mode.
```
OS10# configure terminal
```
Enter INTERFACE CONFIGURATION mode.
```
OS10(config)# interface ethernet 1/1/1
```

Enable orphan port to ignore VLTi fail scenario.

OS10(conf-if-eth1/1/1)# vlt delay-restore orphan-port ignore-vlti-failure

Display the current status of the timer and delay-restore orphan-port configurations.

OS10# show vlt 1 delay-restore-orphan-port
VLT Delay-Restore timer                                       : 90 seconds
Remaining time                                                : 60 seconds
Delay-Restore Orphan-Port enabled interfaces                  : Eth1/1/1
Delay-Restore Orphan-Port Ignore VLTi Fail enabled interfaces : Eth1/1/1

Perform the following steps to disable the delay-restore orphan ports:

Enter CONFIGURATION mode.
```
OS10# configure terminal
```
Enter INTERFACE CONFIGURATION mode.
```
OS10(config)# interface ethernet 1/1/1
```

Disable orphan port to ignore VLTi failures.

OS10(conf-if-eth1/1/1)# no vlt delay-restore orphan-port ignore-vlti-failure

Disable delay-restore orphan port.

OS10(conf-if-eth1/1/1)# no vlt delay-restore orphan-port enable

The following table provides the behavior of orphan ports with different DROP configurations and events:

Table 1. DROP Configurations and EventsDROP Configurations and Events
	Reload and join primary	Reload and join as secondary	Primary node VLTi down or recover	Secondary node VLTi down with HB	Secondary node VLTi recovery with HB	Secondary node VLTi down without HB	Secondary node VLTi recovery without HB
None	No action	No action	No action	No action	No action	No action	No action
Ignore VLTi failure only (no action since DROP is not configured)	No action	No action	No action	No action	No action	No action	No action
DROP only	Up	Down. Brought up after delay-restore timer expires.	Up	Down	Brought up after delay-restore timer expires (Orphan ports were already brought down when VLTi failed).	Up	Down. Brought up after delay-restore expires.
DROP and ignore VLTi failure	Up	Down. Brought up after delay-restore timer expires.	Up	Up	Up	Up	Up

Changing configurations from delay-restore orphan port to delay-restore port and vice-versa

If you want to change configuration from delay-restore orphan port (VLT) to delay-restore port (non-VLT), the system would throw error even if the delay-restore port configuration is done on an interface where no delay-restore orphan port configurations are present. You must first remove all the delay-restore orphan port configurations from all interfaces and then configure delay-restore port. For changing configuration from delay-restore port to delay-restore orphan port, the same steps should be taken.

The configurations and behavior of this feature for VLT and non-VLT are different and mutually exclusive. Meaning, only one of the delay-restore port configurations or delay-restore orphan port configurations can be present in a system at a time. If delay-restore port is configured on few interfaces earlier and now delay-restore orphan port needs to be configured, delay-restore port configurations must be removed from all ports first. Otherwise, delay-restore orphan port configuration will be rejected with error message.

If delay-restore orphan ports are configured in the system and if VLT domain is removed, delay-restore orphan ports commands is considered to be inactive; the delay-restore timer is not applied for orphan ports. After the VLT domain is configured back, the command becomes active again on the configured interfaces.

Configuring delay restore port or delay restore orphan port when delay restore timer is running

When delay-restore port or delay-restore orphan port is enabled on an interface and the respective delay-restore timer is running, the port is immediately brought down. This is done to comply with the behavior of VLT ports (When a normal port-channel is converted into a VLT port-channel when delay-restore timer is running, the port-channel is immediately brought down).

NOTE: If you want to enable delay-restore port or delay-restore orphan port on an interface, but do not want it to be brought down, you must ensure that delay-restore timer is not running or wait for the timer to expire if it has started already.

Welcome

Welcome to Dell

Dell SmartFabric OS10 User Guide Release 10.5.3

Configuring delay-restore orphan port in VLT domain

Rate this content