vADC Docs

Upgrading Stingray Traffic Manager on Amazon EC2

by on ‎02-22-2013 10:43 AM (2,338 Views)

These instructions describe how to upgrade Stingray Traffic Manager AMIs on Amazon EC2. For instructions on upgrading on other platforms, please refer to Upgrading Stingray Traffic Manager.


Upgrade strategy

Unlike physical machines or regular virtual appliances which tend to be very long-lived, EC2 instances are intended to be transient. If an EC2 instance develops a fault, it is easier to terminate it and replace it with a new one than to try to repair it.

Upgrades can be handled in a similar way: instead of upgrading a running instance when a new version of the software is released, it is easier to start an instance of the newer software, migrate the configuration over from the old one, and then terminate the old instance. Using clustering and fault tolerant Traffic IP addresses it is possible to upgrade a cluster in place, replacing each traffic manager with one running a newer version of the software, while continuing to serve application traffic.

Upgrade howto

Important: When the cluster is in a mixed state (i.e. the traffic managers are using different software versions) do not make any configuration changes. This should remain the case until all traffic manager instances in the cluster are running the upgraded version.

For each traffic manager in your cluster, perform the following steps:

  1. Start an instance of the new AMI.
  2. Using the Admin Server, or the user-data pre-configuration parameters, join the new instance to your cluster.

If using user-data pre-configuration, you can set the new instance to join the Traffic IP Groups by setting  join_tips=y but do not use this option if there are multiple Traffic IP groups configured in the cluster. Also for Amazon VPC instances, it will be ignored if the instance doesn't have a secondary IP address assigned whilst launching.

Note that per-node hostname mappings (configured the in System > Networking page) will not be migrated automatically - you must set these manually on each new instance.

  3. Terminate one of the old instances in your cluster.

Repeat these steps until all the traffic managers in your cluster have been replaced. Replace instances one by one - do not terminate an old instance until its replacement has successfully joined the cluster.

Upgrading to new product functionality

The instructions described in the previous section can also be used to change the product version you are running. For example, you can use this method to upgrade from one Stingray Traffic Manager to another that offers a higher bandwidth capacity or more functionality.  You can also use this method to upgrade from one Amazon instance size to another.

Note that you cannot use this method to downgrade the product you are currently running to one with fewer features, because your current configuration will not be applicable to a lower-featured instance. In this case, you must create a new cluster of simpler Stingray instances and migrate the relevant configuration from the old cluster manually.