Upgrading an Active/Passive Cluster

You can upgrade an operational Active/Passive cluster if required.

The following points are important during any upgrade process:

•During all tasks, if output text is presented in red text, read the instructions closely. Do not issue any further commands (such as "cluster promote") or create new configurations (such as adding appliances) until the issue is resolved.

•Don't attempt the use of destructive commands (such as "system destroy") without the involvement of Pulse Support.

•If an upgrade fails, don’t proceed further. Contact Pulse Support before attempting to upgrade again.

•Do not import a backup from an older Pulse One version into a newer Pulse One version (or vice versa). The import and export of configurations is only supported from within the same version of Pulse One.

The specific process will depend on whether your cluster is configured to use auto-failover.

If your cluster is configured for auto-failover, the FQDN will resolve to the IP address of the EXT interface on both Active and Passive nodes, see Working with Automatic Failover. If this is the case, first perform the following steps:

1.Run the following CLI command on both nodes:

cluster status

2.Ping the FQDN of the cluster to confirm that it resolves to the IP address of the Active node’s EXT interface.

Possible outcomes:

•If the Active node is working, remove the Passive node DNS entry that resolves to the cluster FQDN.

•If the Active node is down, and the Passive node has been demoted to standalone and is receiving requests, remove the original Active node DNS entry that resolves to the cluster FQDN. Also, treat the Passive node as Active for the purposes of the upgrade procedure.

•If neither node is working, do not proceed with the upgrade.

It is important to allow time for the DNS entries to reach all managed appliances. The propagation delay depends on your DNS infrastructure. Five minutes is typically sufficient.

Then, to upgrade any cluster, perform the following procedure:

You must ensure that each step is complete before proceeding to the next step.

1.On the Passive node, demote the node to standalone and then confirm the demotion:

cluster demote

cluster status

2.On the Active node, demote the node to standalone and then confirm the demotion:

cluster demote

cluster status

3.On the (formerly-Active) standalone node, assess the size of a backup of the node’s configuration:

system storage show

If this is too large, you can exclude the syslog files from the backup (see below).

4.Perform a backup of the node’s configuration using one of the following commands:

system backup export

system backup export -e syslog

After the appliance backup file is created, download the backup from the displayed URL. This backup can be used to perform a restore if required. See Working with Backup and Restore for full details of the backup/restore processes.

The restore process for a configuration backup must be performed on the same Pulse One software version from which it was backed up. That is, if a backup is created from 1901.1, it should be restored to 1901.1.

5.Perform a system snapshot. This may be required by Pulse Support:

system snapshot

6.On the (formerly Active) standalone node, perform a Pulse One upgrade, as described in Upgrading a Pulse One Appliance.

7.Restart services on the (formerly Active) appliance:

services restart

Wait until all services have started on the Active appliance.

8.On the (formerly Passive) standalone node, perform a standalone Pulse One upgrade, as described in Upgrading a Pulse One Appliance.

9.Restart services on the (formerly Passive) appliance:

services restart

10.Perform the following actions on both nodes:

•Ensure that the node is in standalone mode:

cluster status

•Configure the internal interface:

network interface internal --ip <ip> --netmask <mask> --gateway <ip>

•Confirm the new configuration of the internal interface:

network show

11.After both standalone nodes are upgraded and restarted successfully, re-cluster the nodes. This process is described in Setting Up an Active/Passive Cluster.

12.(Optional) If you intend to use auto-failover, add the Passive node entry back into the DNS.

This addition to the DNS will take some time to propagate to managed appliances. The length of time will depend on your network.