Maintenance, Upgrades, and Troubleshooting
This domain tests planned change and fault handling: PowerStoreOS upgrades, expansion, drive and node replacement, and the evidence-gathering steps that precede a support case. The exam favours the safe sequence over the fast one.
Upgrades and expansion
- PowerStoreOS upgrades are non-disruptive: nodes are upgraded one at a time while the peer serves I/O. A pre-upgrade health check must pass, and the upgrade path may require an interim release before the target version.
- Drive firmware is updated as part of the upgrade process; confirm the health check and support-material collection complete first.
- Expansion enclosures can be added online. Adding an appliance to a cluster is also non-disruptive, but capacity does not automatically rebalance; volumes must be migrated deliberately.
- Drive replacement is hot-swap, but the correct slot must be identified from the alert and the enclosure layout before any drive is pulled.
Troubleshooting method
Confirm the symptom and scope, check active alerts and events for the affected object, then review performance and capacity for the same period. Collect support materials before making changes, and use SupportAssist to attach logs to a case. For path issues, compare host multipath output with array port and node status; for replication, check the remote system pair and session state.
Worked example
An upgrade is blocked at the health check because one drive is degraded. The correct sequence is: identify the drive from the alert, replace it in the indicated slot, wait for the rebuild and for health to return to normal, then rerun the health check and restart the upgrade. Forcing the upgrade or ignoring the check risks a second failure during the node reboots.
Traps that cost marks
- Starting an upgrade without the health check or without confirming the supported upgrade path.
- Assuming a node upgrade is disruptive to all I/O, or that both nodes upgrade simultaneously.
- Pulling a drive based on physical position without matching the alert to the slot.
- Changing configuration before collecting logs, destroying the evidence a case needs.