Upgrading a version 9.3 double HA configuration with DR to version 9.4

Where you have a configuration that has two high availability groups that can provide disaster recovery for each other, you can upgrade high availability (HA) configuration while maintaining the availability of your HA queue managers.

Before you begin

Before upgrading, you should back up your queue managers as described in Backing up a queue manager. You should also ensure that none of your queue managers are in the partitioned state, resolving the state before upgrade if necessary.

About this task

This procedure updates the type of configuration described in Configuring disaster recovery fail over to another high availability group.

You must use the command line interface to upgrade your HA queue managers from version 9.3 to version 9.4.

Figure 1 shows an example configuration of two HA groups on different sites that also provide disaster recovery for each other. When a disaster occurs on one appliance, an instance with up-to-date data can be added to the high availability group on the other appliance. The description of the upgrade procedure uses this example configuration to explain the sequence of actions.

Using the example procedure, there is a staged upgrade of the appliances at the Rome site, followed by a staged upgrade of the Stockholm site. During the upgrade DR queue managers run on the recovery site, while HA queue managers run on the other appliance in the HA pair. The HA and DR queue managers are running throughout the upgrade procedure (while the standalone queue manager must be stopped).

Procedure

Upgrade the appliances at the Rome site:
1. Set queue manager terentia1 to be DR primary on appliance Hugin at the Stockholm site (this makes terentia1 secondary on Castor and Pollux and it runs at Stockholm). Also ensure that freya1 is DR primary at Stockholm.
```
makedrprimary -m terentia1

makedrprimary -m freya1
```
2. Suspend the Castor appliance from the HA group:
```
sethagrp -s
```
  Queue managers cicero1 and cicero2 fail over to the Pollux appliance.
3. Upgrade the Castor appliance to the new firmware level:
```
boot image accept-license firmware_file
```
  Where firmware_file is the file that contains the appliance firmware upgrade (see Overview of installing a new level of firmware by using the command line for details on obtaining the file).
4. The Castor appliance restarts. Resume it into the HA group:
```
sethagrp -r
```
5. Restart the stand alone queue manager, tullia2.
6. Suspend the Pollux appliance from the HA group:
```
sethagrp -s
```
7. Upgrade the Pollux appliance to the new firmware level:
```
boot image accept-license firmware_file
```
  Where firmware_file is the file that contains the appliance firmware upgrade (see Overview of installing a new level of firmware by using the command line for details on obtaining the file).
8. The Pollux appliance restarts. Now it has been upgraded, queue managers can run on Pollux again. Resume Pollux into the HA group:
```
sethagrp -r
```
  Note: terentia1 should become HA primary on Pollux, due to Pollux being its preferred location, but terentia1 is still DR secondary, so continues to run on the Hugin appliance at the Stockholm site.
Upgrade the appliances at the Stockholm site:
1. Switch the freya1 and terentia1 queue managers to be DR primary on Rome appliances.
```
makedrprimary -m terentia1

makedrprimary -m freya1
```
2. Suspend the Hugin appliance from the HA group:
```
sethagrp -s
```
  Queue manager freya2 fails over to the Mugin appliance.
3. Upgrade the Hugin appliance to the new firmware level:
```
boot image accept-license firmware_file
```
  Where firmware_file is the file that contains the appliance firmware upgrade (see Overview of installing a new level of firmware by using the command line for details on obtaining the file).
4. The Hugin appliance restarts. Resume it into the HA group:
```
sethagrp -r
```
5. Suspend the Mugin appliance from the HA group:
```
sethagrp -s
```
  Queue manager freya2 fails back to the Hugin appliance.
6. Upgrade the Mugin appliance to the new firmware level:
```
boot image accept-license firmware_file
```
  Where firmware_file is the file that contains the appliance firmware upgrade (see Overview of installing a new level of firmware by using the command line for details on obtaining the file).
7. The Mugin appliance restarts. Resume it into the HA group:
```
sethagrp -r
```
8. Switch freya1 back to be a DR primary on the Stockholm appliances.