This section provides the following repair procedures to follow if changes were inadvertently made to the storage configuration during upgrade:
Any changes to the storage topology, including running Sun Cluster commands, should be completed before you upgrade the cluster to Solaris 9 software. If, however, changes were made to the storage topology during the upgrade, perform the following procedure. This procedure ensures that the new storage configuration is correct and that existing storage that was not reconfigured is not mistakenly altered.
Ensure that the storage topology is correct. Check whether the devices that were flagged as possibly being replaced map to devices that actually were replaced. If the devices were not replaced, check for and correct possible accidental configuration changes, such as incorrect cabling.
Become superuser on a node that is attached to the unverified device.
Manually update the unverified device.
# scdidadm -R device |
Performs repair procedures on the specified device
See the scdidadm(1M) man page for more information.
Update the DID driver.
# scdidadm -ui # scdidadm -r |
Loads the device-ID configuration table into the kernel
Initializes the DID driver
Reconfigures the database
Repeat Step 2 through Step 3 on all other nodes that are attached to the unverified device.
Return to the remaining upgrade tasks.
For a nonrolling upgrade, go to Step 3 in How to Perform a Nonrolling Upgrade of Sun Cluster 3.1 8/05 Software.
For a rolling upgrade, go to Step 4 in How to Perform a Rolling Upgrade of Sun Cluster 3.1 8/05 Software.
If accidental changes are made to the storage cabling during the upgrade, perform the following procedure to return the storage configuration to the correct state.
This procedure assumes that no physical storage was actually changed. If physical or logical storage devices were changed or replaced, instead follow the procedures in How to Handle Storage Reconfiguration During an Upgrade.
Return the storage topology to its original configuration. Check the configuration of the devices that were flagged as possibly being replaced, including the cabling.
As superuser, update the DID driver on each node of the cluster.
# scdidadm -ui # scdidadm -r |
Loads the device–ID configuration table into the kernel
Initializes the DID driver
Reconfigures the database
See the scdidadm(1M) man page for more information.
If the scdidadm command returned any error messages in Step 1, make further modifications as needed to correct the storage configuration, then repeat Step 1.
Return to the remaining upgrade tasks.
For a nonrolling upgrade, go to Step 3 in How to Perform a Nonrolling Upgrade of Sun Cluster 3.1 8/05 Software.
For a rolling upgrade, go to Step 4 in How to Perform a Rolling Upgrade of Sun Cluster 3.1 8/05 Software.