Upgrade requirements for MetroCluster configurations
If you have to upgrade a MetroCluster configuration, you should be aware of some important requirements.
Required methods for performing major and minor upgrades of MetroCluster configurations
Patch upgrades to MetroCluster configurations can be performed with automatic non-disruptive upgrade (NDU) procedure.
Starting with ONTAP 9.3, major upgrades to MetroCluster configurations can be performed with automatic non-disruptive upgrade (NDU) procedure. On systems running ONTAP 9.2 or earlier, major upgrades to MetroCluster configurations must be performed with the NDU procedure that is specific to MetroCluster configurations.
Both clusters must be running the same version of ONTAP.
You can verify the ONTAP version by using the version command.
The MetroCluster configuration must be in either normal or switchover mode.
Upgrade in switchover mode is only supported in minor patch upgrades.
For all configurations except two-node clusters, you can nondisruptively upgrade both clusters at the same time.
For nondisruptive upgrade in two-node clusters, the clusters must be upgraded one node at a time.
The aggregates in both clusters must not be in resyncing RAID status.
During MetroCluster healing, the mirrored aggregates are resynchronized. You can verify if the MetroCluster configuration is in this state by using the
storage aggregate plex show -in-progress truecommand. If any aggregates are being synchronized, you should not perform an upgrade until the resynchronization is complete.
Negotiated switchover operations will fail while the upgrade is in progress.
To avoid issues with upgrade or revert operations, do not attempt an unplanned switchover during an upgrade or revert operation unless all nodes on both clusters are running the same version of ONTAP.
Configuration requirements for normal operation
The source SVM LIFs must be up and located on their home nodes.
Data LIFs for the destination SVMs are not required to be up or to be on their home nodes.
All aggregates at the local site must be online.
All root and data volumes owned by the local cluster’s SVMs must be online.
Configuration requirements for switchover
All LIFs must be up and located on their home nodes.
All aggregates must be online, except for the root aggregates at the DR site.
Root aggregates at the DR site are offline during certain phases of switchover.
All volumes must be online.