A successful PowerStore Operating Environment (OE) upgrade mandates two key practices. First, running the upgrade pre-check (also known as a System Health Check) is a mandatory step enforced by the system. This check verifies the cluster's health and readiness, and the upgrade cannot proceed if it fails. Second, scheduling the upgrade during a planned maintenance window is a critical best practice. Although the upgrade is non-disruptive, performing it during periods of low I/O activity minimizes potential performance impact and mitigates business risk should any unforeseen issues arise. These two practices are essential for a smooth and successful upgrade process.
Why Incorrect Options are Wrong
C. Disconnect all hosts: This is incorrect because PowerStore OE upgrades are designed to be Non-Disruptive Upgrades (NDU), allowing hosts to remain connected and I/O to continue throughout the process.
D. Shut down one node manually: This is incorrect because the upgrade is an automated, rolling process managed by the system. Manually shutting down a node would disrupt this procedure and likely cause a failure.
References
1. Dell PowerStore: Software Upgrade Guide (Version 3.6)
Reference for A: In the "Upgrade the PowerStore software" section, Step 4 states, "The system automatically runs a pre-upgrade health check... If the pre-upgrade health check fails, the upgrade process stops." This confirms the pre-check is a mandatory, system-enforced step. (Page 14)
Reference for B: In the "Before you begin" section, a key planning step advises, "Dell Technologies recommends performing system software upgrades during a maintenance window when I/O is low." This is presented as a foundational best practice for a successful upgrade. (Page 9)
2. Dell PowerStore: Administration Guide (Version 3.6)
Reference for C & D: The "Software upgrades" chapter describes the Non-Disruptive Upgrade (NDU) process. It states, "During the NDU process, the system remains online and I/O services continue to run." This confirms that disconnecting hosts is not required. The description of the automated, node-by-node upgrade process also confirms that manual node shutdown is not part of the procedure. (Page 179)
