Overview

This article outlines some specific steps that must be taken when upgrading a ScaleArc HA (High Availability) deployment to version 3.12. If upgrading a standalone deployment without HA, please refer to Upgrading ScaleArc to the latest version.

Prerequisites

ScaleArc version 3.12 came with a HA technology upgrade from the Linux HA-based heartbeat system to a pacemaker + corosync + CMAN based HA system. For this reason, there is a slight variation in the way upgrades are done to version 3.12 GA when existing systems are configured in HA. 

Prechecks

The upgrade package checks for the correctness of the VIPs (Virtual IP address) configured. A comparison is made between the network the VIP belongs to with that of the network the interface over which the VIP is configured. If they do not match, the upgrade fails at the pre-check stage and must be corrected before proceeding with the upgrade.

Prior actions

ScaleArc has determined that there is a minor problem involving the use of /dev/random during the HA configuration phase of the upgrade and has developed a workaround. Prior to applying the upgrade, complete the following procedure to ensure that your upgrade will not appear to hang during this phase.

1. Install the rng-tools package with YUM: sudo yum install rng-tools Note that this requires Internet access for YUM to access the repositories online. If you are unsure how to proceed with this, the YUM install fails, or you need assistance in evaluating how to configure your firewall to permit this operation, please stop at this point and contact ScaleArc Support for assistance.
2. Automatically edit the rngd configuration file: sudo perl -p -i -e 's/EXTRAOPTIONS=""/EXTRAOPTIONS="-r \/dev\/urandom"/' /etc/sysconfig/rngd. This should create the /etc/sysconfig/rngd file. You may verify this with the ls command.
3. Enable the rngddaemon for startup in the appropriate init states: sudo chkconfig rngd on This should add rngd to init states 2, 3, 4, and 5. This can be checked with sudo chkconfig --list | grep rngd which should show rngd being on in the appropriate init states as stated.
4. Start the rngd daemon: sudo service rngd start This can be verified with ps
5. sudo service rngd stop and disable it: sudo chkconfig rngd off as the service has not been fully verified to be compatible with ScaleArc. This could change with future upgrades but the recommendation at the moment is to disable the service.

Information

Upgrading pre-version 3.11 configured in HA

Follow these steps carefully to upgrade a ScaleArc system (pre-version 3.11) configured in HA to version 3.12:

1. Upgrade the current secondary system using the upgrade processes described here for the standalone Instance.  Once the upgrade is complete on the Secondary ScaleArc, the machine will reboot.
2. The secondary machine will now become the primary and start processing traffic. The original primary machine at this point will appear in a non-operational state where HA services are disabled. As a result, the machine will not be capable of handling traffic.
   • Note: At this point, HA operations such as "Become Secondary, Force to be Primary" will not be available.
3. Confirm the upgrade on the secondary machine was successful, then repeat the same upgrade steps on the old primary machine.
4. After a reboot, the primary will join as a secondary machine into the HA pair.

Upgrading post-version 3.11 configured in HA

Follow these steps carefully to upgrade a ScaleArc system (version 3.11+) configured in HA to version 3.12:

1. Upgrade the current secondary system using the upgrade processes described here for the standalone Instance.  Once the upgrade is complete on the Secondary ScaleArc, the machine will reboot.
2. The secondary machine will come up as secondary in a non-operation state. Perform HA Switch to make this machine the Primary and in Operation state to handle traffic.
3. Upgrade the earlier Primary System (Current secondary) using the upgrade processes described here for the standalone Instance. Once the upgrade is complete, the machine will reboot and come up as the secondary ScaleArc machine.
4. If you prefer the earlier Primary machine to handle traffic, perform another HA Switch to make it primary. This step is optional.