How to remotely upgrade Protector client nodes

Ops Center Protector Quick Start Guide

Version
7.9.x
Audience
anonymous
Part Number
MK-99PRT001-10
Note:
  • The Manage Software Updates RBAC activity must be assigned to users who perform upgrades. It is recommended that this activity is restricted to administrative users only.
  • Upgrade of a Microsoft failover cluster node needs to be done locally, following a similar procedure to that described in How to install a Protector master or client on a Windows cluster. It's not possible to upgrade remotely because the standby nodes in the cluster are not available. Attempting a remote upgrade will place the clustered Protector service in a failed state, taking it offline.
  • Before upgrading unmount any mounted snapshots.
  • DO NOT upgrade while replications are in the process of pairing. Any active replications must be in the paired state before upgrade is carried out.

An upgrade can be performed if the existing installation has become corrupted or a newer version of Ops Center Protector is available.

If upgrading, obtain the upgrade installer files from your Hitachi Vantara support representative.

When an upgrade starts, the Ops Center Protector services are shutdown, causing the upgrading OS Host node and any nodes for which it serves as a proxy, to go offline in the Nodes Inventory. Any active data flows using those nodes will be temporarily interrupted. These nodes will come back online again when the services are automatically restarted on the OS Host node, after the upgrade process is completed, and the affected data flows will resume operation.

Note: We recommend upgrading Ops Center Protector only after current backups have been completed.
Note:
When upgrading, it is highly recommended (and often necessary) to upgrade all nodes to the same version. The recommended order in which to upgraded is:
  1. Internet Connected Nodes. If ICNs are not upgraded first, they may not be able to be 'push upgraded' through the UI. ICNs may appear to go offline until the master is upgraded.
  2. Master Node(s).
  3. Clients acting as ISMs, controlling Hitachi Block and File storage devices.
  4. Clients acting as Data Destinations, such as Repositories.
  5. Clients acting as Data Sources, such as application servers.
  1. Locate the installers and accompanying configuration files appropriate for the target OSs and hardware architectures in your Protector environment.
    The installer's executable and configuration filenames have the following format:

    Protector-Rm.n-m.n.n.nnnnn-ostype-proctype

    where:
    • m.n-m.n.n.nnnnn - is the version and build number
    • ostype - is the target operating system type:
      • Win
      • Linux
      • AIX
    • proctype - is the target processor type:
      • x32
      • x64
      • PPC
  2. Copy both the installer and configuration files to the C:\Programs Files\Hitachi\Protector\runtime\updater folder on the Master node.

    Copy both the installer and configuration files to the “/opt/hitachi/protector/runtime/updater” folder on the Master node.

    This folder will need to be created manually if this is the first time an update has been applied.
  3. Click Nodes on the Navigation Sidebar to open the Nodes Inventory
  4. Select the nodes to be upgraded (Master first, then Clients), then click Upgrade Clients to start the upgrade process.
    Note:
    • Only OS Host Client node types can be upgraded remotely
    • It is recommended to push out the upgrade to clients in batches of up to 20 nodes at a time.
    When upgrading the Master node, the UI will logout when the node's services are stopped by the installer. Wait a few minutes, then log back in again and complete the upgrade of the remaining nodes.

    When upgrading Client nodes, each one (and any nodes for which that Client acts as a Proxy) will go offline temporarily while the upgrade is applied.

    When the nodes come back online, the Version shown on the respective tile will be updated accordingly.

  5. All active data flows should now be reactivated and all destination nodes should be resynchronized with their sources.