This article explores how to properly execute the opatchauto command in , ensuring a successful and controlled application of patches across all nodes in your Oracle Grid Infrastructure (GI) and Database home, specifically addressing scenarios involving opatchauto session errors. 1. What is opatchauto Non-Rolling Mode?
Patching an Oracle Real Application Clusters (RAC) environment requires a strategic approach to balance security updates with system availability. While rolling upgrades are preferred to keep the database online, certain patches (such as major Grid Infrastructure patches or specific critical security updates) require a approach.
$GRID_HOME/OPatch/opatchauto apply /path/to/patch -nonrolling Use code with caution.
The nonrolling mode took over. Instead of the delicate dance of one node at a time, opatchauto laid waste to the entire stack at once, updating the binaries across all nodes in parallel. By 1:30 AM, the terminal finally blinked a green success message.
# opatchauto apply /path/to/patch/directory -oh -nonrolling Use code with caution. Best Practices to Prevent Patching Failures opatchauto72030 execute in nonrolling mode
Verify that the patch requires a non-rolling, full-system shutdown.
Verify no lingering Oracle processes remain active: ps -ef | grep ora_ or ps -ef | grep d.bin Phase 4: Validating Central Inventory and SSH
Download the latest OPatch patch (Patch 6880880) matching your major Oracle Database release from My Oracle Support (MOS). Back up the existing OPatch directory in both homes:
: Change to the directory where patch 72030 is located. This article explores how to properly execute the
Ensure your patching utilities are capable of handling the payload.
mv $GRID_HOME/OPatch $GRID_HOME/OPatch_old mv $ORACLE_HOME/OPatch $ORACLE_HOME/OPatch_old Use code with caution.
Utilize the binary pre-check capability to find configuration conflicts before committing to downtime: # opatchauto apply /path/to/patch/directory -analyze Use code with caution.
Nodes are patched one by one. The cluster remains online, ensuring zero downtime for applications. The nonrolling mode took over
Failing storage, network, or OS-level system checks on one or more nodes in the cluster.
Sometimes, opatchauto may fail, or a session may be improperly locked as "rolling" when it should be "non-rolling". If you encounter issues during patching, especially if the analyze phase fails or indicates wrong mode, you might need to modify the session info. Fixing "Invalid Non-Rolling Session" (e.g., Session F7ET)
As the root user, run the apply command. The critical switch here is -nonrolling .
crsctl stat res -t
Discrepancies between the node configurations listed in the oraInventory and the actual physical nodes of the cluster.