This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Procedure for breaking HA to test new firmware releases.

I have two appliances running with HA (autoconfigure mode).   Since appliances do not have the rollback ease of a virtual machine, I want to break HA to provide a safety net.   Is this script the best approach?

Assume:

  • Unit1 is currently master,
  • Unit2 is currently slave.

Preparation:

  • Have CDs for both the new and old firmware versions.
  • Have a configuration backup from both machines.

Begin Testing:

  • Disconnect Unit2 from the production network and shut it down.
  • Upgrade Unit1 to new firmware.
  • Operate on Unit1 until a decision is made to keep the new version or roll back.

If new version is accepted, upgrade Unit2:

  • Connect Unit2 to an isolated network or laptop, then reboot.
  • If up2date kits are available on Unit2, use these to upgrade.  Otherwise, do a factory reset, install new firmware from CD, and reload configuration backup.
  • Shut down Unit2, then reconnect to production network.
  • Boot into HA to begin slave sync.
  • Both units will be running new firmware with Unit1 as master.

If new version is rejected, roll back to the slave:

  • Shut down Unit1 and disconnect from production network.
  • Reconnect Unit2 to production network, and boot it up.
  • Connect Unit1 to an isolated network or laptop, factory reset, install old firmware from CD, restore original configuration backup.
  • Shut down Unit1, then reconnect Unit1 to production network.
  • Boot Unit1 and verify that it begins slave sync.
  • Both units will be running old firmware with Unit2 as master.
  • Optionally, force an HA failover to make Unit1 master again.


This thread was automatically locked due to age.
Parents
  • mostly we use HA-feature "Keep node(s) reserved during Up2Date"

    this updates only one node and the second stops syncing.

    If all is OK you trigger next upgrade from HA menu.

    If new FW is not ok, you powerdown the master and the slave takes over.

    now you restore the other device and rejoin this device to HA.

    • Shut down Unit1 and disconnect from production network.
    • Unit2 is active now
    • Connect Unit1 to an isolated network or laptop, factory reset, install old firmware from CD, configure HA-auto-join
    • Shut down Unit1, then reconnect Unit1 to production network.
    • Boot Unit1 and verify that it begins slave sync.
    • Both units will be running old firmware with Unit2 as master.
    • Optionally, force an HA failover to make Unit1 master again.


    Dirk

    Systema Gesellschaft für angewandte Datentechnik mbH  // Sophos Platinum Partner
    Sophos Solution Partner since 2003
    If a post solves your question, click the 'Verify Answer' link at this post.

Reply
  • mostly we use HA-feature "Keep node(s) reserved during Up2Date"

    this updates only one node and the second stops syncing.

    If all is OK you trigger next upgrade from HA menu.

    If new FW is not ok, you powerdown the master and the slave takes over.

    now you restore the other device and rejoin this device to HA.

    • Shut down Unit1 and disconnect from production network.
    • Unit2 is active now
    • Connect Unit1 to an isolated network or laptop, factory reset, install old firmware from CD, configure HA-auto-join
    • Shut down Unit1, then reconnect Unit1 to production network.
    • Boot Unit1 and verify that it begins slave sync.
    • Both units will be running old firmware with Unit2 as master.
    • Optionally, force an HA failover to make Unit1 master again.


    Dirk

    Systema Gesellschaft für angewandte Datentechnik mbH  // Sophos Platinum Partner
    Sophos Solution Partner since 2003
    If a post solves your question, click the 'Verify Answer' link at this post.

Children