Skip to main content
Migration Notice
We're migrating documentation from the old portal into this one. Some things may look a little different or out of place in the meantime — we know, and we're working to get it right. If something's unclear or doesn't look right, let us know.
Version: 2.15.0

PowerScale OneFS Upgrade Procedure with Eyeglass

Introduction​

This page covers how to upgrade PowerScale OneFS on clusters that are managed by Eyeglass, without losing Eyeglass configuration replication, failover readiness, or Data Security functionality during the process. It is separate from the Upgrade Guide, which covers upgrading the Eyeglass appliance itself.

Determine Which Procedure to Follow​

Before you begin, check the Feature Release Compatibility table in the Release Notes for your Eyeglass release to confirm the source and target cluster OneFS versions are supported together.

Then choose your procedure based on the upgrade scope:

  • Minor release upgrade — the upgrade does not change the first two version digits (for example, 8.2.1.1 to 8.2.1.2, or 9.x to 9.x) and the source and target clusters remain in a supported configuration throughout. Follow Minor Release Upgrade.
  • Major release upgrade — the upgrade changes the first two version digits (for example, OneFS 8.x to 9.x). Follow Major Release Upgrade.
  • Unsupported intermediate state — your upgrade process will put the source and target clusters into an unsupported configuration at any point. Follow Functions Impacted During an Unsupported Upgrade State in addition to the applicable procedure above.
note

As of OneFS 8.2 and later, the cluster no longer supports API or SSH connections to the SmartConnect Service IP (SSIP). Before upgrading a cluster to 8.2 or later, change the Eyeglass-managed IP for that cluster to an IP from a dynamic pool with the System Access Zone assigned (for example, an IP from the Management Pool): in the Eyeglass WebUI, go to Inventory, right-click the cluster, select Edit, and replace the SmartConnect Service IP field with the new IP. Re-enter the Eyeglass service account password and submit.

Major Release Upgrade​

Major release upgrades are performed one site at a time: the target cluster is upgraded first, then the source cluster, using a planned failover to move production to the already-upgraded cluster in between.

Step 1: Upgrade the Target Cluster​

  1. Confirm Eyeglass is configured correctly and all Eyeglass Jobs show OK status with no errors.
  2. Prepare the target cluster for upgrade per Dell PowerScale documentation: run pre-upgrade checks, confirm the supported upgrade path from the current release to the target release (an intermediate release may be required), and stage the installer on the cluster (for example, under /ifs/data).
  3. Temporarily disable Eyeglass Jobs for the upgrade window.
  4. Perform the OneFS upgrade on the target cluster. Eyeglass will report the cluster as unreachable during this step — this is expected.
  5. Once the upgrade completes, bring the target cluster back under Eyeglass management:
    • Update the Eyeglass admin role privileges on the target cluster with any additional permissions required by the new OneFS release. See Assign Privileges to the Admin Role in the Installation Guide.
    • Update the eyeglass service account's sudo privileges on the target cluster to match.
    • Restart the Eyeglass sca service: SSH to the appliance as the admin user, run sudo -s and enter the admin password, then systemctl stop sca.service followed by systemctl start sca.service. Confirm with systemctl status sca.service.
    • Re-enable the Eyeglass Jobs, either with Run Now from the Jobs window or by waiting for the next scheduled replication job.
    • Once jobs complete, confirm Eyeglass status is OK with no errors, and confirm the DR Dashboard shows healthy audit and readiness status for all zones.
Mixed mode

At this point the environment is in mixed mode: the OneFS API uses the lower cluster version's API and takes priority when syncing configuration from source to target. See What You Need to Know in Mixed Mode below.

Step 2: Upgrade the Source Cluster​

  1. Confirm Eyeglass Jobs show OK status with no errors before proceeding.
  2. Prepare the source cluster for upgrade per Dell PowerScale documentation.
  3. Perform a planned Eyeglass failover to the (already-upgraded) target cluster. Follow the Failover Planning checklist and use the failover mode appropriate to your environment.
  4. Confirm the failover completed successfully with no errors in the failover log.
  5. Eyeglass Jobs for Mirror Policies are User Disabled by default after a failover. If they are not (for example, if this environment has already been failed over and back before), manually disable all Eyeglass Jobs for the upgrade window.
  6. Perform the OneFS upgrade on the source cluster. Eyeglass will report the cluster as unreachable — this is expected.
  7. Once the upgrade completes, bring the source cluster back under Eyeglass management using the same steps as Step 1.5 above (admin role privileges, sudo privileges, restart sca.service, re-enable jobs, verify status).

What You Need to Know in Mixed Mode​

While clusters in an Eyeglass-managed pair are running different OneFS major releases (any combination including 8.x, 8.2, or 9.x):

  • Configuration replication is supported from a lower-version cluster to a higher-version cluster. Confirm the exact combination is supported in your release notes.
  • Failover is supported.
  • Failback is supported.
  • Runbook Robot is supported.
  • RPO reporting is supported.
important

Mixed mode is intended to be a temporary state — typically weeks, not months — while you complete the upgrade of the second cluster. It is not a supported long-term operating mode. Mixed-mode API behavior only clears once all clusters managed by Eyeglass are on the same major release.

Minor Release Upgrade​

For minor release upgrades where the source and target clusters remain in a supported configuration throughout:

  1. Log in to the Eyeglass WebUI before starting the upgrade and confirm configuration replication is in a healthy state.
  2. Begin the OneFS upgrade. Eyeglass remains running throughout — no Jobs need to be disabled.
  3. If either cluster becomes unreachable during the upgrade, you may see failed Eyeglass configuration replication jobs and related alarms. These typically clear automatically once both clusters are reachable again and the next scheduled job runs successfully.

Expected Alarms During a Minor Release Upgrade​

While a cluster is unreachable during a minor release upgrade, expect:

  • Replication jobs failing while the cluster is unreachable.
  • DR Dashboard showing an error status, with readiness alarms generated.
  • Multiple alarms raised, one per scheduled replication job — this is expected and not itself a sign of a problem.

After the Upgrade​

Leave Eyeglass running for at least 30 minutes after the upgrade completes, then confirm Eyeglass Jobs have returned to their pre-upgrade healthy state. Review the Alarms History for any alarms raised during the upgrade window.

Functions Impacted During an Unsupported Upgrade State​

If your upgrade process will place the source and target clusters into an unsupported configuration at any point, expect the following Eyeglass, Ransomware Defender, Easy Auditor, and Performance Auditor functions to be impacted for the duration of that unsupported state only:

  • Configuration replication will show errors — these can be ignored until the cluster upgrade completes.
  • Runbook Robot will not execute without errors.
  • Ransomware Defender and Easy Auditor will not be able to collect events.
  • Planned failover execution is not supported while in this state.
  • The DR Readiness Dashboard will not be accurate.
  • Reports will not execute correctly.

Post-Upgrade Recovery Steps​

Eyeglass DR:

  1. SSH to the appliance as the admin user, run sudo -s and enter the admin password.
  2. Restart the SCA service: systemctl restart sca.
  3. Wait 30 minutes, then log in to the Eyeglass WebUI and confirm there are no errors in Jobs or the DR Dashboard.

Ransomware Defender, Easy Auditor, and Performance Auditor (ECA cluster):

  1. Log in to node 1 of the ECA cluster.
  2. Run ecactl cluster down and wait for shutdown to complete.
  3. Run ecactl cluster up and wait for startup to complete.
  4. In the Eyeglass WebUI, open the Manage Services icon and confirm all nodes show green (this can take up to 5 minutes).
  5. For Ransomware Defender, run Security Guard and confirm it completes successfully.
  6. For Easy Auditor, run Robo Audit and confirm it completes successfully.
  7. For Performance Auditor, open the tool icon and confirm performance data is displayed.

See Also​

  • Upgrade Guide – Upgrading the Eyeglass appliance itself: release/platform compatibility, disk space requirements, and known upgrade issues.
  • Failover Planning – Pre-failover planning checklist referenced during a major release upgrade's target-cluster failover step.
  • Failover Process and Customer Role – Support entitlement policy for planned failovers, including this upgrade procedure as an acknowledgment item.
  • Release Notes – Feature Release Compatibility table for confirming supported OneFS version combinations.