SPARC: Dynamic Reconfiguration Support
Oracle Solaris Cluster supports the dynamic reconfiguration software feature. This section describes concepts and considerations for Oracle Solaris Cluster support of the dynamic reconfiguration feature.
All the requirements, procedures, and restrictions that are documented for the Oracle Solaris dynamic reconfiguration feature also apply to Oracle Solaris Cluster dynamic reconfiguration support (except for the operating environment quiescence operation). Therefore, review the documentation for the Oracle Solaris dynamic reconfiguration feature before using the dynamic reconfiguration feature with Oracle Solaris Cluster software. You should review in particular the issues that affect non-network IO devices during a dynamic reconfiguration detach operation.
Dynamic reconfiguration implementation can be system dependent, and might be implemented differently as technology changes. For more information, see Oracle Solaris Cluster Data Service for Oracle VM Server for SPARC Guide .
SPARC: Dynamic Reconfiguration General Description
The dynamic reconfiguration feature enables operations, such as the removal of system hardware, in running systems. The dynamic reconfiguration processes are designed to ensure continuous system operation with no need to halt the system or interrupt cluster availability.
Dynamic reconfiguration operates at the board level. Therefore, a dynamic reconfiguration operation affects all the components on a board. Each board can contain multiple components, including CPUs, memory, and peripheral interfaces for disk drives, tape drives, and network connections.
Removing a board that contains active components would result in system errors. Before removing a board, the dynamic reconfiguration subsystem queries other subsystems, such as Oracle Solaris Cluster, to determine whether the components on the board are being used. If the dynamic reconfiguration subsystem finds that a board is in use, the dynamic reconfiguration remove-board operation is not done. Therefore, it is always safe to issue a dynamic reconfiguration remove-board operation because the dynamic reconfiguration subsystem rejects operations on boards that contain active components.
The dynamic reconfiguration add-board operation is also always safe. CPUs and memory on a newly added board are automatically brought into service by the system. However, the system administrator must manually configure the cluster to actively use components that are on the newly added board.
Note - The dynamic reconfiguration subsystem has several levels. If a lower level reports an error, the upper level also reports an error. However, when the lower level reports the specific error, the upper level reports Unknown error. You can safely ignore this error.
The following sections describe dynamic reconfiguration considerations for the different device types.
SPARC: Dynamic Reconfiguration Clustering Considerations for CPU Devices
Oracle Solaris Cluster software does not reject a dynamic reconfiguration remove-board operation because of the presence of CPU devices.
When a dynamic reconfiguration add-board operation succeeds, CPU devices on the added board are automatically incorporated in system operation.
SPARC: Dynamic Reconfiguration Clustering Considerations for Memory
Kernel memory cage
Non-kernel memory cage
For the purposes of dynamic reconfiguration, there are two types of memory:
These two types differ only in usage. The actual hardware is the same for both types. Kernel memory cage is the memory that is used by the Oracle Solaris Operating System. Careful consideration must be taken before performing a dynamic reconfiguration remove-board operation which will impact kernel memory cage. Oracle Solaris Cluster software does not reject the operation, but in most cases, such a dynamic reconfiguration operation will have a significant impact on the entire cluster. The tight coupling between cluster nodes, between multiple instances of scalable applications, and between the primary and secondary nodes of HA applications and services means that the quiescing of one node for repair can cause operations on non-quiesced nodes to be delayed until the repair operation is complete and the node is unquiesced.
In most cases, the preferred method of removing or replacing a system board with kernel cage memory is to bring the node requiring repair down. This allows the remainder of the cluster to cleanly take over the duties of the node being repaired. Only when circumstances prevent the node requiring repair from being brought out of the cluster, should dynamic reconfiguration be used to remove or replace a system board with kernel cage memory while the node is still part of the operating cluster. For suggestions on preparing the cluster for a dynamic reconfiguration kernel cage remove-board operation, see Preparing the Cluster for Kernel Cage Dynamic Reconfiguration in Oracle Solaris Cluster Hardware Administration Manual .
When a dynamic reconfiguration add-board operation that pertains to memory succeeds, memory on the added board is automatically incorporated in system operation.
If the node being repaired panics during the dynamic reconfiguration operation, or if the dynamic reconfiguration operation is otherwise interrupted, it may be necessary to manually re-enable heartbeat monitoring and reset the repaired node's quorum vote count. These two actions are normally done automatically at the completion of the dynamic reconfiguration operation to return the cluster to a stable state. For instructions on recovering in this case, see How to Recover From an Interrupted Kernel Cage Dynamic Reconfiguration Operation in Oracle Solaris Cluster Hardware Administration Manual .
SPARC: Dynamic Reconfiguration Clustering Considerations for Disk and Tape Drives
Oracle Solaris Cluster rejects dynamic reconfiguration remove-board operations on active drives on the primary node. You can perform dynamic reconfiguration remove-board operations on inactive drives on the primary node and on any drives in the secondary node. After the dynamic reconfiguration operation, cluster data access continues as before.
Note - Oracle Solaris Cluster rejects dynamic reconfiguration operations that impact the availability of quorum devices. For considerations about quorum devices and the procedure for performing dynamic reconfiguration operations on them, see Dynamic Reconfiguration Clustering Considerations for Quorum Devices.
See Dynamic Reconfiguration With Quorum Devices in Oracle Solaris Cluster 4.3 System Administration Guide for detailed instructions about how to perform these actions.
SPARC: Dynamic Reconfiguration Clustering Considerations for Quorum Devices
If the dynamic reconfiguration remove-board operation pertains to a board that contains an interface to a device configured for quorum, Oracle Solaris Cluster software rejects the operation. Oracle Solaris Cluster software also identifies the quorum device that would be affected by the operation. You must disable the device as a quorum device before you can perform a dynamic reconfiguration remove-board operation.
See Chapter 6, Administering Quorum, in Oracle Solaris Cluster 4.3 System Administration Guide for detailed instructions about how administer quorum.
SPARC: Dynamic Reconfiguration Clustering Considerations for Cluster Interconnect Interfaces
If the dynamic reconfiguration remove-board operation pertains to a board containing an active cluster interconnect interface, Oracle Solaris Cluster software rejects the operation. Oracle Solaris Cluster software also identifies the interface that would be affected by the operation. You must use an Oracle Solaris Cluster administrative tool to disable and remove the active interface before the dynamic reconfiguration operation can succeed.
 | Caution - Oracle Solaris Cluster software requires each cluster node to have at least one functioning path to every other cluster node. Do not disable a private interconnect interface that supports the last path to any node in the cluster. |
See Administering the Cluster Interconnects in Oracle Solaris Cluster 4.3 System Administration Guide for detailed instructions about how to perform these actions.
SPARC: Dynamic Reconfiguration Clustering Considerations for Public Network Interfaces
If the dynamic reconfiguration remove-board operation pertains to a board that contains an active public network interface, Oracle Solaris Cluster software rejects the operation. Oracle Solaris Cluster software also identifies the interface that would be affected by the operation. Before you remove a board with an active network interface present, switch over all traffic on that interface to another functional interface in the multipathing group by using the if_mpadm command.
| Caution - If the remaining network adapter fails while you are performing the dynamic reconfiguration remove operation on the disabled network adapter, availability is impacted. The remaining adapter has no place to fail over for the duration of the dynamic reconfiguration operation. |
See Administering the Public Network in Oracle Solaris Cluster 4.3 System Administration Guide for detailed instructions about how to perform a dynamic reconfiguration remove operation on a public network interface.