Performing a rolling upgrade

Prerequisites

Ensure that you have a scheduled maintenance window. Do not change network configurations while the primary and secondary switches are at different versions. Make sure that both switches are updated to the same version within two hours by completing the steps in this topic.

This procedure is NOT hitless, but it does aim to minimize the amount of traffic loss. It is advisable to have a support engineer from HPE available during this procedure.
Back up the current software to the secondary image bank. Run the show version command. If the Active Image is primary, copy the primary image to the secondary image bank:
```
switch# copy primary secondary
```
If the Active Image is the secondary image, you do not need to run the copy primary secondary command because the current software is already in the secondary image bank.
Save a copy of the running configuration on both devices. One way to save a copy of the running configuration is to enter the following command:
```
switch# copy running-config sftp://<username>@<IP>/filename vrf mgmt
```
Check the status of the show vsx brief command and validate the ISL is in-sync and the keepalive is established. Also, verify which switch is the secondary switch. This procedure will have you upgrade the secondary switch first.
Check the output of the show lacp interfaces multi-chassis command and note which LACP interfaces are in a forwarding state of up.

Procedure

Ensure that you are in a maintenance window. There will be some amount of network disruption even with the upgrade of a single switch.
Copy the software from this release onto both switches by using TFTP, SFTP, or USB:
```
switch# copy sftp://<username>@<ip>/XL_10_02_0001.swi primary vrf mgmt
```
If the upstream link is a point-to-point link, disable the Route Only Port (ROP) or change the routing cost on the upstream nodes to offload traffic of the link connected to the secondary switch.
On the VSX LAG secondary switch, boot the switch to the 10.02 software release:
```
switch# boot system primary
```
While the secondary switch is being upgraded, the primary switch is still actively forwarding traffic with VSX LAGs from version 10.01 of the software. When you are prompted, save the configuration.
Enter:
```
switch# show vsx status 
```
A software version mismatch is shown because only one switch has been upgraded.

CAUTION:
VSX synchronization has stopped on the secondary switch you updated. Do not change network configurations while the primary and secondary switches are at different versions. It is advised not to change the VSX configuration at this step. For example, if you create a VLAN at this time, VSX synchronization will not work for the new VLAN. Make sure that both switches are updated to the same version within two hours by completing the steps in this topic.
Verify that the ARP, MAC, and LACP states are fully synced:
```
switch# show vsx status linkup-delay
```

Verify that the VSX LAGs are up:

switch# show lacp interfaces multi-chassis

Enable any ROPs, SVIs, or layer 3 LAGs that were previously disabled to the uplinks. The VSX configuration is applied, the upgrade is complete, and the secondary switch is ready to forward network traffic.
On the designated primary switch, boot the switch to the 10.02 software release:
```
switch# boot system primary
```
Save the configuration from the running configuration to the startup configuration by entering the write memory command.
Run the show vsx brief command on both switches.
Verify that the ISL state is In-Sync and the keepalive state is Keepalive-Established.
Check the output of the show lacp interfaces command. Validate that the VSX LAGs have a forwarding state of up.
Configure VSX-related features as needed. Consult this guide for additional information about the newest VSX features.
Save the running configuration to the startup configuration:
```
switch# write memory
Success
```