When maintaining Oracle Grid Infrastructure (GI) or RAC environments, you may encounter the requirement to use the -nonrolling opatchauto

. This is often triggered by specific patch constraints or the OPATCHAUTO-72030

error, which serves as a directive from the utility that the current environment state or patch content does not support a rolling application. Understanding the Non-Rolling Mode By default, opatchauto operates in rolling mode

, patching one node at a time while keeping the cluster operational. In contrast, non-rolling mode

requires a coordinated shutdown of services across the cluster to apply patches that might otherwise cause version conflicts during a phased rollout. Downtime Requirement:

All remote nodes must be down to start a non-rolling session. Execution Order:

You patch a single node (the local node) first, followed by others. Version 12c+ Behavior:

Starting with 12c, non-rolling patching requires the GI stack to be up on the local node being patched, while all other nodes are shut down. Resolving OPATCHAUTO-72030 OPATCHAUTO-72030

error typically occurs when the tool detects that a rolling update is not feasible—often because the cluster is already in an inconsistent state or the patch itself is "non-rollable". Stop Remote Nodes:

Ensure all remote nodes in the cluster are completely shut down before execution. Verify Local Node:

The local GI stack must be active on the node where you are running the command. Execute with Flag: Use the explicit -nonrolling parameter as shown in this 12c GI PSU Guide # opatchauto apply -nonrolling Use code with caution. Copied to clipboard High-Quality Execution Steps

To ensure a successful execution and avoid common pitfalls, follow this structured workflow: 1. Preparation and Health Checks Download Latest OPatch: Ensure you have the latest version of the OPatch utility for your specific Oracle release. Analyze First: Run the command with the

flag to simulate the process and identify prerequisite failures without modifying the system. Verify Permissions: opatchauto from a directory other than to avoid write permission errors, as noted in troubleshooting logs 2. The Patching Workflow Concepts of Multi-Node Patch Orchestration Using OPatchAuto

OPATCHAUTO-72030 typically emerges when a Database Administrator attempts to apply an Oracle Grid Infrastructure (GI) patch in rolling mode on a shared CRS home. Because the Grid Infrastructure home is shared across multiple nodes, the standard "one-at-a-time" rolling method is logically impossible; the files on disk cannot be simultaneously patched and running for different nodes. The Patching Story: A Shift to Non-Rolling

The journey of resolving OPATCHAUTO-72030 often follows a predictable path for DBAs: The Default Attempt : Most administrators start with the standard opatchauto apply command, which defaults to rolling mode to avoid downtime. The Roadblock : The system triggers the error:

“Cannot execute in rolling mode, as CRS home is shared. Execute in non-rolling mode”

. This signifies that a cluster-wide outage is mandatory to update the shared binaries. Transition to Non-Rolling : To proceed, the administrator must stop the GI stack on Executing the Solution : The patch is applied using the -nonrolling # As root user

/u01/app/grid/OPatch/opatchauto apply -nonrolling Use code with caution. Copied to clipboard

This command instructs OPatchAuto to handle the shared home correctly by assuming all services are down across the cluster. Key Technical Rules for Non-Rolling Mode Complete Outage

: Unlike rolling updates, a non-rolling update requires all databases and services using that home to be offline. Explicit Flag : You must manually add -nonrolling

to the command line; the utility will not automatically switch modes after a failure. Sequential vs. Parallel

: In a multi-node environment with a shared home, you typically run the non-rolling session from the local node, which may handle remote nodes if configured, but the stack must remain down globally. Shared Home Verification : This error is often a safeguard. If your environment is

supposed to be a shared home, you may need to investigate "Bug 31639069" or check your inventory configuration. exact command syntax for your specific Oracle version, or help reviewing the session logs to confirm why the rolling mode was rejected? Doc ID 2957442.1 OPATCHAUTO-72030 During Opatchauto

The error OPATCHAUTO-72030 typically occurs when you attempt to patch a Shared Grid Infrastructure (GI) home in rolling mode. Because the home is shared across multiple nodes, it cannot be updated incrementally while other nodes are still using it; it must be updated in non-rolling mode. Understanding Error OPATCHAUTO-72030

This error is an instruction from the OPatchAuto utility signaling that the current execution mode is invalid for your configuration.

Cause: You initiated opatchauto apply (which defaults to rolling mode) on a shared CRS home.

Requirement: All nodes sharing the Oracle home must be down during the patching process to ensure binary consistency. How to Execute in Non-Rolling Mode

To resolve this, you must explicitly tell the utility to use non-rolling mode using the -nonrolling flag.

Stop All Stacks: Manually shut down the Oracle Grid Infrastructure and Database stacks on all nodes that share the home.

Run the Command: Execute the following as the root user from the patch directory:

# /OPatch/opatchauto apply -nonrolling Use code with caution. Copied to clipboard

(Example: /u01/app/19.0.0/grid/OPatch/opatchauto apply /soft/35319490 -nonrolling)

Verify Permissions: Ensure you are not running the command from a restricted directory like /root. Move to a directory where the home owner has write permissions.

Restart Stacks: Once patching completes, bring the GI stack and databases back up on all nodes. Best Practices for Success

Update OPatch First: Always ensure you are using the latest OPatch utility version (e.g., 12.2.0.1.30+ for Oracle 19c) before starting.

Analyze First: Run the command with the -analyze flag first to identify potential conflicts or space issues without making changes.

Check Logs: If the session fails again, review the system configuration log found in /cfgtoollogs/opatchautodb/ for specific step failures.

3 Concepts of Multi-Node Patch Orchestration Using OPatchAuto

5. Staging

  • Copy patch to each host OR use central shared patch location reachable by opatchauto.
  • Verify patch integrity (GPG/sha256 checksums).
  • Ensure opatch utility is present and has correct permissions.

3. Environment Settings

  • Set ORACLE_HOME (Grid home), PATH, and ORACLE_SID correctly.
  • Ensure SSH equivalence is configured between all nodes.
  • Disable any monitoring or cron jobs that might interfere.

Practical command pattern

Use opatchauto72030 with explicit nonrolling mode and quality settings. Example pattern (adapt paths and flags to your environment):

/path/to/opatchauto72030 -nonrolling -quality HIGH -oh /u01/app/oracle/product/19.0.0/dbhome_1 -oh_grid /u01/app/19.0.0/grid -patch 12345678 -apply -invPtrLoc /etc/oraInst.loc -skip_noncritical_checks false

Notes:

  • Replace paths, patch IDs, and flags with your environment-specific values.
  • Include full inventory pointer location and necessary privileges (run as oracle/grid OS users as required).
  • Use -dryrun or equivalent pre-check flags if available before actual apply.

1. Environment Validation

# Verify cluster status (if RAC)
crsctl stat res -t