Upgrading the Software on all Active Nodes

You can centrally upgrade the software on all active nodes in a stack. To upgrade all nodes in the stack, all nodes must be running an ExtremeXOS release that supports stacking (ExtremeXOS release 12.0 or later).
  1. Download a new ExtremeXOS software release and install it on all nodes on the active topology using the command: download image [[hostname | ipaddress] filename {{vr} vrname}] {partition}


    ExtremeXOS offers the ability to synchronize a non-master slot from the master individually, but it does not support the upgrade of the entire stack with one image download to the master.
  2. If necessary, use the use image {partition} {primary | secondary} command to select the image partition (primary or secondary) into which the software was saved.
  3. Restart all nodes in the new release using reboot {[time mon day year hour min sec] | cancel}
    For example:
    download image [[hostname | ipaddress] filename {{vr} vrname}] {primary | secondary}
    use image {partition} [primary | secondary]
  4. Before you upgrade a stack, make sure that the active image partition is same across all nodes.

    To determine the active partition selected on all nodes and the ExtremeXOS versions installed in each partition, use the show slot detail command. You can install the image only on the alternate image partition and not on the active image partition.

  5. To run the upgraded software, you must reboot the stack after installation with the image partition that received the software being selected.
  6. If the active partition is different on some nodes, the action you take depends on what is stored in both partitions:
    If both primary and secondary partitions have the same ExtremeXOS release, you may use the following commands to cause a node to use the same active image as the rest of the stack:

    use image {primary | secondary} slot slot-number reboot slot slot-number

    If you are using the primary image on your master node and some other node primary image does not contain the same ExtremeXOS version as your master node's primary image, you may use the command: synchronize slot slotid to cause the node to contain the same ExtremeXOS versions in both partitions as it is on the master node, and to reboot the node into the primary partition.

    Hitless upgrade is not supported in a stack.