This chapter describes the procedures about how to maintain SCSI JBOD storage devices in a SunTM Cluster environment.
The procedures in this chapter apply to the following SCSI JBOD storage devices:
SunTM StorEdge 3120 storage array
Sun StorEdge 3310 SCSI storage array
Sun StorEdge 3320 SCSI storage array
SPARC: Sun StorEdge D1000 storage array
SPARC: Sun StorEdge Multipack storage array
SPARC: Sun StorEdge D130/S1 storage array
SPARC: Sun StorEdge D2 storage array
This chapter contains the following procedures:
How to Replace a Host Adapter When Using Failover and Scalable Data Services Only
SPARC: How to Replace a Host Adapter When Using Oracle Real Application Clusters Only
How to Replace a Disk Drive Without Oracle Real Application Clusters
How to Replace a Disk Drive With Oracle Real Application Clusters
The maintenance procedures in FRUs That Do Not Require Sun Cluster Maintenance Procedures are performed the same as in a noncluster environment. Table 2–1 lists the procedures that require cluster-specific steps.
Table 2–1 Task Map: Maintaining a Storage Array
Task |
Information |
---|---|
Remove a storage array | |
Replace a storage array To replace a storage array, remove the storage array. Add a new storage array to the configuration. |
SPARC: How to Add a Storage Array to an Existing SPARC Based Cluster |
Add a JBOD as an expansion unit. Follow the same procedure that you use in a noncluster environment. |
Sun StorEdge 3000 Family Installation, Operation, and Service Manual |
Remove a JBOD expansion unit. Follow the same procedure that you use in a noncluster environment. |
Sun StorEdge 3000 Family Installation, Operation, and Service Manual |
Replace a chassis | |
Replace a SCSI cable | |
Replace a host adapter | |
Add a disk drive | |
Remove a disk drive | |
Replace a disk drive |
How to Replace a Disk Drive Without Oracle Real Application Clusters How to Replace a Disk Drive With Oracle Real Application Clusters |
Upgrade disk drive firmware | |
Upgrade host adapter firmware |
Each storage device has a different set of FRUs that do not require cluster-specific procedures. Choose among the following storage devices:
The following is a list of administrative tasks that require no cluster-specific procedures. See the Sun StorEdge 3000 Family FRU Installation Guide for the procedures for the following FRUs. For a URL to this storage documentation, see Related Documentation.
The following is a list of administrative tasks that require no cluster-specific procedures. See the Sun StorEdge 3000 Family FRU Installation Guide for instructions on replacing the following FRUs. http://www.tokensbeads.com/education.htm
The following is a list of administrative tasks that require no cluster-specific procedures. See the Sun StorEdge A1000 and D1000 Installation, Operations, and Service Manual and the Sun StorEdge A1000 and D1000 Installation, Operations, and Service Manual for the procedures for the following FRUs. For a URL to this storage documentation, see Related Documentation.
The following is a list of administrative tasks that require no cluster-specific procedures. See the Netra st D130 Installation and Maintenance Manual for the procedures for the following FRUs. For a URL to this storage documentation, see Related Documentation.
The following is a list of administrative tasks that require no cluster-specific procedures. See the Sun StorEdge D2 Array Installation, Operation, and Service Manual for the procedures for the following FRUs. For a URL to this storage documentation, see Related Documentation.
The following is a list of administrative tasks that require no cluster-specific procedures. See the SPARCstorage MultiPack Service Manual for the procedures for the following FRUs. For a URL to this storage documentation, see Related Documentation.
Removing a storage array enables you to downsize or reallocate your existing storage pool.
This procedure relies on the following prerequisites and assumptions.
You plan to remove the references to the disk drives in the array.
You do not need to replace the storage array's chassis.
If you need to replace your storage array's chassis, see How to Replace the Chassis.
Your nodes are not configured with dynamic reconfiguration functionality.
If your nodes are configured for dynamic reconfiguration, see Sun Cluster system administration documentation, and skip steps that instruct you to shut down the node.
Does the storage array that you want to remove contain a quorum device?
# scstat -q |
If no, proceed to Step 2.
If yes, add a new quorum device that will not be impacted by this procedure. Remove the old quorum device.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
If necessary, back up the metadevice or volume.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
On each node that is connected to the storage array, perform volume management administration to remove the storage array from the configuration.
If a volume manager does manage the disk drives, run the appropriate volume manager commands to remove the disk drives from any diskset or disk group. For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation. See the following paragraph for additional VERITAS Volume Manager commands that are required.
Disk drives that were managed by VERITAS Volume Manager must be completely removed from VERITAS Volume Manager control before you can remove the disk drives from the Sun Cluster environment. After you delete the disk drives from any disk group, use the following commands on both nodes to remove the disk drives from VERITAS Volume Manager control.
# vxdisk offline cNtXdY # vxdisk rm cNtXdY |
Identify the disk drives that you plan to remove.
# cfgadm -al |
On all nodes, remove references to the disk drives in the storage array that you plan to remove.
# cfgadm -c unconfigure cN::dsk/cNtXdY |
Disconnect the SCSI cables from the storage array.
On all nodes, update device namespaces.
# devfsadm -C |
On all nodes, remove all obsolete device IDs.
# scdidadm -C |
Power off the storage array. Disconnect the storage array from the power source.
For the procedure about how to power off a storage array, see your storage documentation. For a list of storage documentation, see Related Documentation.
Remove the storage array.
For the procedure about how to remove a storage array, see your storage documentation. For a list of storage documentation, see Related Documentation.
If you plan to remove a host adapter that has an entry in the nvramrc script, delete the references to the host adapters in the nvramrc script.
If other parallel SCSI devices are connected to the nodes, you can delete the contents of the nvramrc script. Then, at the OpenBoot PROM, set setenv use-nvramrc? false. Afterward, reset the scsi-initiator-id to 7 as outlined in Installing a Storage Array.
If necessary, remove any unused host adapters from the nodes.
For the procedure about how to remove a host adapter, see your host adapter and server documentation.
From any node, verify that the configuration is correct.
# scdidadm -L |
You might need to replace a chassis if the chassis fails. With this procedure, you are able to replace the chassis and retain the storage array's disk drives and the references to those disk drives. By retaining the storage array's disk drives, you save time because you no longer need to resynchronize your mirrors or restore your data.
If you need to replace the entire storage array, see How to Remove a Storage Array.
This procedure relies on the following assumptions.
You plan to retain the disk drives in the storage array that you need to replace.
You plan to retain the references to the disk drives.
Your replacement chassis is unpacked and ready to be placed into the cluster.
Your nodes are not configured with dynamic reconfiguration functionality.
If your nodes are configured for dynamic reconfiguration, see the Sun Cluster system administration documentation, and skip steps that instruct you to shut down the node.
You are not using a StorEdge D2, a StorEdge 3310 SCSI, or a StorEdge 3320 SCSI storage array in split-bus mode.
If you are running your storage array in split-bus mode, you must shut down the entire cluster to replace the chassis of the storage array.
Does the chassis contain a disk drive that is a quorum device?
# scstat -q |
If no, proceed to Step 2.
If yes, add a new quorum device that will not be impacted by this procedure. Remove the old quorum device.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
If possible, back up the metadevices or volumes that reside in the storage array.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
On each node that is connected to the storage array, perform volume management administration to remove the storage array from the configuration.
Disk drives that were managed by VERITAS Volume Manager must be completely removed from VERITAS Volume Manager control before you can remove the disk drives from the Sun Cluster environment. After you delete the disk drives from any disk group, use the following commands on both nodes to remove the disk drives from VERITAS Volume Manager control.
If a volume manager manages the disk drives, run the appropriate volume manager commands to remove the disk drives from any diskset or disk group. For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
# vxdisk offline cNtXdY# vxdisk rm cNtXdY |
Disconnect the SCSI cables from the storage array.
You can disconnect the cables in any order.
Power off and disconnect the storage array from the power source.
For more information, see your storage documentation. For a list of storage documentation, see Related Documentation.
Connect the new storage array to the power sources.
Connect the SCSI cables to the new storage array.
You can connect the cables in any order.
Ensure the cable does not exceed bus-length limitations. For more information on bus-length limitations, see your hardware documentation.
One disk drive at a time, remove the disk drives from the old storage array. Insert the disk drives into the same slots in the new storage array.
Move your other components to the new chassis as well. For the procedures about how to replace your storage array's components, see your storage documentation. For a list of storage documentation, see Related Documentation.
Power on the storage array.
For the procedure about how to shut down and power off a node, see the Sun Cluster system administration documentation.
On each node that is attached to the storage array, run the devfsadm(1M) command.
# devfsadm |
From one node, attach the new storage array to the global device namespace.
# scgdevs |
One at a time, shut down and reboot the nodes that are connected to the storage array.
For the procedure about how to shut down and power off a node, see the Sun Cluster system administration documentation.
Perform volume management administration to add the storage array back into the configuration.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Use this procedure to temporarily disconnect a node from shared storage. You need to temporarily disconnect a node from shared storage if you intend to replace an HBA.
This procedure relies on the following assumptions.
You intend to reconnect the shared storage to the same host adapter that the storage array was connected before you disconnected the SCSI cable.
This procedure defines Node A as the node you want to disconnect from the shared storage. Node B is the remaining node.
This procedure assumes that you might have more than one storage array connected to Node A.
You must maintain proper SCSI-bus termination during this procedure. The process by which you disconnect the node from shared storage depends on whether you have host adapters available on Node B. If you do not have host adapters available on Node B and your storage device does not have auto-termination, you must use terminators.
To determine the specific terminator that your storage array supports, see your storage documentation. For a list of storage documentation, see Related Documentation.
Determine the resource groups and device groups that are running on Node A.
Record this information because you use this information in Step 6 to return resource groups and device groups to these nodes.
# scstat |
Identify the submirrors on the storage array that is connected to Node A.
Move all resource groups and device groups off Node A.
# scswitch -S -h nodename |
(Optional) If necessary, detach the submirrors on the storage array that is connected to Node A.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Shut down Node A.
For the procedure about how to shut down and power off a node, see your Sun Cluster system administration documentation.
To maintain proper SCSI-bus termination during this procedure, perform one of the following steps. The approach you choose depends on whether you have an available host adapter on Node B.
Disconnect the SCSI cable between Node A and the storage array. Attach this SCSI cable to Node B on the storage array. For an illustration, see Figure 2–1.
Disconnect the SCSI cable between Node A and the storage array. Install an appropriate SCSI terminator to this SCSI connector on the storage array. For an illustration, see Figure 2–2.
Boot Node A into cluster mode.
If you detached the submirrors in Step 4, reattach the submirrors. Wait for the submirrors to resynchronize.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Repeat Step 4 through Step 8 for each remaining storage array that is connected to Node A.
Reconnect the node to the shared storage as outlined in Reconnect the Node to Shared Storage.
(Optional) If necessary, detach the submirrors on the storage array.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Disconnect the SCSI terminator. Reattach the SCSI cable between the storage array and Node A.
Connect this storage array to the same host adapter that the storage array was connected before you disconnected the SCSI cable.
Boot Node A into cluster mode.
If you detached submirrors in Step 1, reattach the submirrors on the storage array. Wait for the submirrors to resynchronize.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Repeat Step 1 through Step 4 for the remaining storage array that you want to reconnect to Node A.
(Optional) Return the resource groups and device groups that you identified in Step 1 to Node A.
# scswitch -z -g resource-group -h nodename # scswitch -z -D device-group-name -h nodename |
Use this procedure to replace a failed SCSI cable in a running cluster.
Is one of the disks in the storage array a quorum device? This array is the array that is attached to the cable you are replacing.
# scstat -q |
If no, proceed to Step 2.
If yes, relocate that quorum device to another suitable storage array
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
Detach the submirror or submirrors on the storage array.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Replace the SCSI cable.
Reattach the submirror or submirrors on the storage array.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
(Optional) Did you relocate the quorum device in Step 1?
If no, this step completes the procedure.
If yes, relocate the quorum device function to this storage array.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
You need to replace a host adapter if your host adapter fails, if it becomes unstable, or if you want to upgrade to a newer version. These procedures define Node A as the node with the host adapter that you plan to replace.
Choose the procedure that corresponds to your cluster configuration.
If your cluster is x86 based, Oracle RAC services are not supported. Follow the instructions in How to Replace a Host Adapter When Using Failover and Scalable Data Services Only.
Cluster Configuration |
Instructions |
---|---|
Sun Cluster failover and scalable data services only, using the recommended HBA configuration |
How to Replace a Host Adapter When Using Failover and Scalable Data Services Only |
Oracle Parallel Server/Real Application Clusters (OPS/RAC) only, using the recommended HBA configuration |
SPARC: How to Replace a Host Adapter When Using Oracle Real Application Clusters Only |
Both failover and scalable data services and OPS/RAC, using the recommended HBA configuration | |
All clusters using a single, dual-port HBA to provide both paths to shared data |
The first three procedures in this section assume that you are using the recommended HBA configuration: two redundant hardware paths to shared data. If you choose to use a single HBA configuration, see Configuring Cluster Nodes With a Single, Dual-Port HBA in Sun Cluster 3.0-3.1 Hardware Administration Manual for Solaris OS for the risks and restrictions of that configuration and use How to Replace a Host Adapter When Using a Single, Dual-Port HBA to Provide Both Paths to Shared Data.
This procedure relies on the following prerequisites and assumptions.
You are not using Oracle Real Application Clusters. If you are using Oracle Real Application Clusters, follow the instructions in SPARC: How to Replace a Host Adapter When Using Oracle Real Application Clusters Only or SPARC: How to Replace a Host Adapter When Using Both Failover and Scalable Data Services and Oracle Real Application Clusters.
Except for the failed host adapter, your cluster is fully operational and all nodes are powered on.
If your nodes are configured for dynamic reconfiguration, see the Sun Cluster system administration documentation, and skip steps that instruct you to shut down the node.
Does your cluster use Oracle Real Application Clusters?
If yes, follow the instructions in SPARC: How to Replace a Host Adapter When Using Oracle Real Application Clusters Only or SPARC: How to Replace a Host Adapter When Using Both Failover and Scalable Data Services and Oracle Real Application Clusters.
If no, proceed to Step 2.
Determine the resource groups and device groups that are running on Node A.
Record this information because you use this information in Step 17 of this procedure to return resource groups and device groups to Node A.
# scstat |
(Optional) If necessary, move all resource groups and device groups off Node A.
# scswitch -S -h nodename |
Determine if the storage device that is attached to the failed host adapter contains a quorum device.
# scstat -q |
If no, skip this step.
If yes, add a new quorum device on a different storage device and remove the old quorum device.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
Detach the Solstice DiskSuite/Solaris Volume Manager submirrors or VERITAS Volume Manager plexes on the storage array.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Record the details of disk groups and volumes affected by the failed host adapter.
Record this information because you use it in Step 16 of this procedure to reattach submirrors on the storage array. For more information about how to determine which submirrors or plexes are affected, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Is Node A enabled with the Solaris DR feature?
For more information, see your Sun Cluster system administration documentation.
Shut down Node A.
For the full procedure about how to shut down and power off a node, see your Sun Cluster system administration documentation.
Power off Node A.
Replace the failed host adapter.
For the procedure about how to remove and add host adapters, see the documentation that shipped with your nodes.
Is Node A enabled with the Solaris DR feature?
For more information, see your Sun Cluster system administration documentation.
Power on Node A.
x86: Set the HBA ports to ensure that each array has a unique SCSI address.
For instructions on setting SCSI initiator IDs in x86 based systems, see x86: How to Install a Storage Array in a New X86 Based Cluster.
Boot Node A into cluster mode.
For more information about how to boot nodes, see your Sun Cluster system administration documentation.
If necessary, upgrade the host adapter firmware on Node A.
PatchPro is a patch-management tool that eases the selection and download of patches required for installation or maintenance of Sun Cluster software. PatchPro provides an Interactive Mode tool especially for Sun Cluster. The Interactive Tool makes the installation of patches easier. PatchPro's Expert Mode tool helps you to maintain your configuration with the latest set of patches. Expert Mode is especially useful for obtaining all of the latest patches, not just the high availability and security patches.
To access the PatchPro tool for Sun Cluster software, go to http://www.sun.com/PatchPro/, click Sun Cluster, then choose either Interactive Mode or Expert Mode. Follow the instructions in the PatchPro tool to describe your cluster configuration and download the patches.
For third-party firmware patches, see the SunSolveSM Online site at http://sunsolve.ebay.sun.com.
Reattach the Solstice DiskSuite/Solaris Volume Manager submirrors or VERITAS Volume Manager plexes on the storage array.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
(Optional) If necessary, return the resource groups and device groups you identified in Step 2 to Node A.
# scswitch -z -g resource-group -h nodename # scswitch -z -D device-group-name -h nodename |
For more information, see your Sun Cluster system administration documentation.
Did you relocate the quorum device in Step 4?
If no, this step completes the procedure.
If yes, and you want the cluster configured as it was before replacing the host adapter, relocate the quorum device function back to this storage array.
For procedures about how to add and remove quorum devices, see your Sun Cluster system administration documentation.
This procedure relies on the following prerequisites and assumptions.
You are using Oracle Real Application Clusters only. If you are using other failover or scalable data services, follow the instructions in How to Replace a Host Adapter When Using Failover and Scalable Data Services Only or SPARC: How to Replace a Host Adapter When Using Both Failover and Scalable Data Services and Oracle Real Application Clusters.
Except for the failed host adapter, your cluster is fully operational and all nodes are powered on.
If your nodes are configured for dynamic reconfiguration, see the Sun Cluster system administration documentation, and skip steps that instruct you to shut down the node.
Does your cluster use failover or scalable data services (other than Oracle Real Application Clusters)?
If yes, follow the instructions in How to Replace a Host Adapter When Using Failover and Scalable Data Services Only or SPARC: How to Replace a Host Adapter When Using Both Failover and Scalable Data Services and Oracle Real Application Clusters.
If no, proceed to Step 2.
Determine the Oracle instance that is running on Node A.
# ps -ef | grep oracle |
Shut down the Oracle Real Application Clusters instance and any other process on Node A that should be stopped before shutting down the node.
For information about how to shut down and restart an Oracle instance in the RAC environment, refer to your Oracle documentation.
Determine if the storage device that are attached to the failed host adapter contains a quorum device.
# scstat -q |
If no, skip to Step 5.
If yes, add a new quorum device on a different storage device and remove the old quorum device.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
Detach the VERITAS Volume Manager plexes on the storage array attached to the failed host adapter.
For more information, see your VERITAS Volume Manager documentation.
Record the details of disk groups and volumes affected by the failed host adapter
Record this information because you use it in Step 15 of this procedure to reattach plexes on the storage array. For more information about how to determine which plexes are affected, see your VERITAS Volume Manager documentation.
Is Node A enabled with the Solaris dynamic reconfiguration (DR) feature?
For more information, see your Sun Cluster system administration documentation.
Shut down Node A.
For the full procedure about how to shut down and power off a node, see your Sun Cluster system administration documentation.
Power off Node A.
Replace the failed host adapter.
For the procedure about how to remove and add host adapters, see the documentation that shipped with your nodes.
Is Node A enabled with the Solaris dynamic reconfiguration feature?
For more information, see your Sun Cluster system administration documentation.
Power on Node A.
Boot Node A into cluster mode.
For more information about how to boot nodes, see your Sun Cluster system administration documentation.
If necessary, upgrade the host adapter firmware on Node A.
PatchPro is a patch-management tool that eases the selection and download of patches required for installation or maintenance of Sun Cluster software. PatchPro provides an Interactive Mode tool especially for Sun Cluster. The Interactive Tool makes the installation of patches easier. PatchPro's Expert Mode tool helps you to maintain your configuration with the latest set of patches. Expert Mode is especially useful for obtaining all of the latest patches, not just the high availability and security patches.
To access the PatchPro tool for Sun Cluster software, go to http://www.sun.com/PatchPro/, click Sun Cluster, then choose either Interactive Mode or Expert Mode. Follow the instructions in the PatchPro tool to describe your cluster configuration and download the patches.
For third-party firmware patches, see the SunSolveSM Online site at http://sunsolve.ebay.sun.com.
Reattach the VERITAS Volume Manager plexes on the storage array to their respective volumes.
For more information, see your VERITAS Volume Manager documentation.
(Optional) Bring your Oracle Real Application Clusters instance online. This is the instance you identified in Step 2.
For information about how to shut down and restart an Oracle instance in the RAC environment, refer to the Oracle 9iRAC Administration Guide.
(Optional) Did you relocate a quorum device in Step 4?
If no, this step completes the procedure.
If yes, and you want your configuration to use the same quorum structure after the host adapter replacement, relocate the quorum device function back to this storage array.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
This procedure relies on the following prerequisites and assumptions.
You are using both Oracle Real Application Clusters and failover or scalable data services.
If you are using only Oracle Real Application Clusters, follow the instructions in SPARC: How to Replace a Host Adapter When Using Oracle Real Application Clusters Only. If you are using only failover and scalable data services, follow the instructions in How to Replace a Host Adapter When Using Failover and Scalable Data Services Only.
Except for the failed host adapter, your cluster is operational and all nodes are powered on.
If your nodes are configured for dynamic reconfiguration, see the Sun Cluster system administration documentation, and skip steps that instruct you to shut down the node.
Does your cluster use both failover or scalable data services and Oracle Real Application Clusters?
If no, follow the instructions in SPARC: How to Replace a Host Adapter When Using Oracle Real Application Clusters Only or How to Replace a Host Adapter When Using Failover and Scalable Data Services Only.
If yes, proceed to Step 2.
Determine the Oracle instance that is running on Node A.
# ps -ef | grep oracle |
Shut down the Oracle Real Application Clusters instance identified in Step 2.
For information about how to shut down and restart an Oracle instance in the RAC environment, refer to the Oracle 9iRAC Administration Guide.
Determine the resource groups and device groups that are running on Node A.
Record this information because you use it in Step 18 of this procedure to return resource groups and device groups to Node A.
# scstat |
Move all resource groups and device groups off Node A.
# scswitch -S -h nodename |
Determine if the storage device that is connected to the failed host adapter contains a quorum device.
# scstat -q |
If Node A contains a quorum device, add a new quorum device on a different storage device and remove the old quorum device.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
Detach the VERITAS Volume Manager plexes on the storage array that is attached to the failed host adapter.
For more information, see your VERITAS Volume Manager documentation.
Record the details of plexes that are affected by the failed host adapter.
Record this information because you use it in Step 18 of this procedure to reattach plexes on the storage array. For more information about how to determine which plexes are affected, see your VERITAS Volume Manager documentation.
Is Node A enabled with the Solaris dynamic reconfiguration (DR) feature?
For more information, see your Sun Cluster system administration documentation.
Shut down Node A.
For the full procedure about how to shut down and power off a node, see your Sun Cluster system administration documentation.
Power off Node A.
Replace the failed host adapter.
For the procedure about how to remove and add host adapters, see the documentation that shipped with your nodes.
Is Node A enabled with the Solaris DR feature?
For more information, see your Sun Cluster system administration documentation.
Power on Node A.
Boot Node A into cluster mode.
For more information about how to boot nodes, see your Sun Cluster system administration documentation.
If necessary, upgrade the host adapter firmware on Node A.
PatchPro is a patch-management tool that eases the selection and download of patches required for installation or maintenance of Sun Cluster software. PatchPro provides an Interactive Mode tool especially for Sun Cluster. The Interactive Tool makes the installation of patches easier. PatchPro's Expert Mode tool helps you to maintain your configuration with the latest set of patches. Expert Mode is especially useful for obtaining all of the latest patches, not just the high availability and security patches.
To access the PatchPro tool for Sun Cluster software, go to http://www.sun.com/PatchPro/, click Sun Cluster, then choose either Interactive Mode or Expert Mode. Follow the instructions in the PatchPro tool to describe your cluster configuration and download the patches.
For third-party firmware patches, see the SunSolveSM Online site at http://sunsolve.ebay.sun.com.
Reattach the VERITAS Volume Manager plexes on the storage array to their respective volumes.
For more information, see your VERITAS Volume Manager documentation.
(Optional) If necessary, return the resource groups and device groups you identified in Step 5 to Node A.
# scswitch -z -g resource-group -h nodename # scswitch -z -D device-group-name -h nodename |
For more information, see your Sun Cluster system administration documentation.
(Optional) Bring your Oracle Real Application Clusters instance online. This is the instance you identified in Step 2.
For information about how to shut down and restart an Oracle instance in the RAC environment, refer to the Oracle 9iRAC Administration Guide.
(Optional) If you relocated a quorum device in Step 7 and you want your configuration to use the same quorum structure after the host adapter replacement, relocate the quorum device function back to this storage array.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
In the following example, a two-node cluster is running Oracle Real Application Clusters and VERITAS Volume Manager. In this situation, you begin the host adapter replacement by determining the Oracle instance name.
# ps -ef | grep oracle oracle 14716 14414 0 14:05:47 console 0:00 grep oracle oracle 14438 1 0 13:05:44 ? 0:02 ora_lmon_tpcc1 . . . oracle 14434 1 0 13:05:43 ? 0:00 ora_pmon_tpcc1 oracle 14458 1 0 13:05:50 ? 0:00 ora_d000_tpcc1 |
This output identifies the Oracle Real Application Clusters instance as tpcc1.
Shutting down the Oracle Real Application Clusters instance on Node A involves several steps, as shown in the following example.
# su - oracle Sun Microsystems Inc. SunOS 5.9 Generic May 2002 $ ksh $ ORACLE_SID=tpcc1 $ ORACLE_HOME=/export/home/oracle/OraHome1 $ export ORACLE_SID ORACLE_HOME $ sqlplus " /as sysdba " SQL*Plus: Release 9.2.0.4.0 - Production on Mon Jan 5 14:12:28 2004 Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved. Connected to: Oracle9i Enterprise Edition Release 9.2.0.4.0 - 64bit Production With the Partitioning, Real Application Clusters, OLAP and Oracle Data Mining options JServer Release 9.2.0.4.0 - Production SQL> shutdown immediate ; Database closed. Database dismounted. ORACLE instance shut down. SQL> exit Disconnected from Oracle9i Enterprise Edition Release 9.2.0.4.0 - 64bit Production With the Partitioning, Real Application Clusters, OLAP and Oracle Data Mining options JServer Release 9.2.0.4.0 - Production $ lsnrctl LSNRCTL for Solaris: Version 9.2.0.4.0 - Production on 05-JAN-2004 14:15:09 Copyright (c) 1991, 2002, Oracle Corporation. All rights reserved. Welcome to LSNRCTL, type "help" for information. LSNRCTL> stop Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=IPC)(KEY=EXTPROC))) The command completed successfully LSNRCTL> |
After you have stopped the Oracle Real Application Clusters instance, check for a quorum device and, if necessary, reconfigure the quorum device.
When you are certain that the node with the failed adapter does not contain a quorum device, proceed to determine the affected plexes.
Record this information for use in reestablishing the original storage configuration. In the following example, c2 is the controller with the failed host adapter.
# vxprint -ht -g tpcc | grep c2 dm tpcc01 c2t0d0s2 sliced 4063 8374320 - dm tpcc02 c2t1d0s2 sliced 4063 8374320 - dm tpcc03 c2t2d0s2 sliced 4063 8374320 - dm tpcc04 c2t3d0s2 sliced 4063 8374320 - dm tpcc09 c2t8d0s2 sliced 4063 8374320 - dm tpcc10 c2t9d0s2 sliced 4063 8374320 - sd tpcc02-01 control_001-01 tpcc02 0 41040 0 c2t1d0 ENA . . . sd tpcc03-06 temp_0_0-02 tpcc03 2967840 276480 0 c2t2d0 ENA sd tpcc03-04 ware_0_0-02 tpcc03 2751840 95040 0 c2t2d0 ENA |
From this output, you can easily determine which plexes and subdisks are affected by the failed adapter. These are the plexes you detach from the storage array.
/usr/sbin/vxplex -g tpcc det control_001-02 /usr/sbin/vxplex -g tpcc det temp_0_0-02 |
After the plexes are detached, you can safely shut down the node, if necessary.
Proceed with replacing the failed host adapter, following instructions that accompanied that device.
After you replace the failed host adapter, and Node A is in cluster mode, reattach the plexes and replace any quorum device to reestablish your original cluster configuration.
This procedure relies on the following prerequisites and assumptions.
You are using a single, dual-port HBA to provide the redundant paths to your shared data.
For an explanation of the limitations and risks of using this configuration, see Configuring Cluster Nodes With a Single, Dual-Port HBA in Sun Cluster 3.0-3.1 Hardware Administration Manual for Solaris OS.
Except for the failed host adapter, your cluster is operational and all nodes are powered on.
Your nodes are not configured for dynamic reconfiguration.
You cannot use dynamic reconfiguration for this procedure when using a single HBA.
Node A is the node with the failed host adapter.
If you are using scalable and failover services, determine the resource groups and device groups that are running on Node A.
Record this information because you use it in Step 18 of this procedure to return resource groups and device groups to Node A.
# scstat |
Record the details of metadevices that are affected by the failed host adapter.
SPARC: If you are using Oracle Real Application Clusters, shut down all RAC instances running in your cluster.
For information about how to shut down and restart an Oracle instance in the RAC environment, refer to the Oracle 9iRAC Administration Guide.
Shut down the cluster.
For the full procedure about how to shut down a cluster, see your Sun Cluster system administration documentation.
Power off Node A.
Replace the failed host adapter.
For the procedure about how to remove and add host adapters, see the documentation that shipped with your nodes.
Power on Node A.
Boot all nodes into cluster mode.
For more information about how to boot nodes, see your Sun Cluster system administration documentation.
If necessary, upgrade the host adapter firmware on Node A.
PatchPro is a patch-management tool that eases the selection and download of patches required for installation or maintenance of Sun Cluster software. PatchPro provides an Interactive Mode tool especially for Sun Cluster. The Interactive Tool makes the installation of patches easier. PatchPro's Expert Mode tool helps you to maintain your configuration with the latest set of patches. Expert Mode is especially useful for obtaining all of the latest patches, not just the high availability and security patches.
To access the PatchPro tool for Sun Cluster software, go to http://www.sun.com/PatchPro/, click Sun Cluster, then choose either Interactive Mode or Expert Mode. Follow the instructions in the PatchPro tool to describe your cluster configuration and download the patches.
For third-party firmware patches, see the SunSolveSM Online site at http://sunsolve.ebay.sun.com.
Perform any volume management maintenance procedures that are necessary to fix any metadevices affected by this procedure.
For more information, see your volume manager software documentation.
(Optional) If necessary, return the resource groups and device groups you identified in Step 5 to Node A.
# scswitch -z -g resource-group -h nodename # scswitch -z -D device-group-name -h nodename |
For more information, see your Sun Cluster system administration documentation.
(Optional) Bring all Oracle Real Application Clusters instance online.
For information about how to shut down and restart an Oracle instance in the RAC environment, refer to the Oracle 9iRAC Administration Guide.
Adding a disk drive enables you to increase your storage space after a storage array has been added to your cluster.
(Sun StorEdge Multipack Enclosures Only) SCSI-reservations failures have been observed when clustering storage arrays that contain a particular model of Quantum disk drive: SUN4.2G VK4550J. Do not use this particular model of Quantum disk drive for clustering with storage arrays.
If you must use this model of disk drive, you must set the scsi-initiator-id of Node A to 6. If you use a six-slot storage array, you must set the storage array for the 9-through-14 SCSI target address range. For more information, see the Sun StorEdge MultiPack Storage Guide.
This procedure relies on the following prerequisites and assumptions.
Your storage array contains an empty disk slot.
Your nodes are not configured with dynamic reconfiguration functionality.
If your nodes are configured for dynamic reconfiguration, see the Sun Cluster system administration documentation, and skip steps that instruct you to shut down the node.
Locate an empty disk slot in the storage array for the disk drive you are adding.
Identify the empty disk slot and note the target number. Refer to your storage array documentation.
Install the disk drive.
For the procedure about how to install a disk drive, see your storage array documentation.
On all nodes that are attached to the storage array, configure the disk drive.
# cfgadm -c configure cN |
On all nodes that are attached to the storage device, probe all devices and write the new disk drive to the /dev/rdsk directory.
# devfsadm |
Depending on the number of devices that are connected to the node, the devfsadm command can require at least five minutes to complete.
On all nodes, verify that the entries for the disk drive have been added to the /dev/rdsk directory.
# ls -l /dev/rdsk |
If necessary, use the format(1M) command or the fmthard(1M) command to partition and label the disk drive.
From any node, update the global device namespace.
If a volume management daemon such as vold is running on your node, and you have a CD-ROM drive that is connected to the node, a device busy error might be returned even if no disk is in the drive. This error is an expected behavior.
# scgdevs |
On one node, verify that a device ID has been assigned to the disk drive.
The new device ID that is assigned to the new disk drive might not be in sequential order in the storage array.
# scdidadm -L |
Perform volume management administration to add the new disk drive to the configuration.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
If you want this new disk drive to be a quorum device, add the quorum device.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
Removing a disk drive can allow you to downsize or reallocate your existing storage pool. You might want to perform this procedure in the following scenarios.
You no longer need to make data accessible to a particular node.
You want to migrate a portion of your storage to another storage array.
For conceptual information about quorum, quorum devices, global devices, and device IDs, see the Sun Cluster concepts documentation.
This procedure relies on the following prerequisites and assumptions.
You do not need to remove the entire storage array.
If you need to remove the storage array, see How to Remove a Storage Array.
You do not need to replace the storage array's chassis.
If you need to replace your storage array's chassis, see How to Replace the Chassis.
Your nodes are not configured with dynamic reconfiguration functionality.
If your nodes are configured for dynamic reconfiguration, see the Sun Cluster system administration documentation, and skip steps that instruct you to shut down the node.
Identify the disk drive that you are removing. Also identify the slot that the disk drive needs to be removed from.
If the disk error message reports the drive problem by device ID, use the scdidadm -l command to determine the Solaris device name. To list all configurable hardware information, use the cfgadm -al command.
# scdidadm -l deviceID # cfgadm -al |
Determine if the disk you want to remove is a quorum device.
# scstat -q |
If possible, back up the metadevice or volume.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Perform volume management administration to remove the disk drive from the configuration.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
On all nodes, unconfigure the disk drive. Remove references to the disk drive from the operating system and clustering environment.
# cfgadm -c unconfigure cN::dsk/cNtXdY |
Remove the disk drive.
For the procedure about how to remove a disk drive, see your storage documentation. For a list of storage documentation, see Related Documentation.
On both nodes, delete the paths to the disk drive that you plan to remove.
# devfsadm -C |
On both nodes, remove all obsolete device IDs.
# scdidadm -C |
You need to replace a disk drive if the disk drive fails or when you want to upgrade to a higher-quality or to a larger disk.
For conceptual information about quorum, quorum devices, global devices, and device IDs, see the Sun Cluster concepts documentation.
This procedure relies on the following prerequisites and assumptions.
Your cluster is operational.
Your system is not running Oracle Real Application Clusters.
If your system is running Oracle Real Application Clusters, see How to Replace a Disk Drive With Oracle Real Application Clusters.
(Solstice DiskSuite/Solaris Volume Manager Only) If the disk drive failure prevents Solstice DiskSuite/Solaris Volume Manager from reading the disk label, you have a backup of the disk-partitioning information.
Your nodes are not configured with dynamic reconfiguration functionality.
If your nodes are configured for dynamic reconfiguration, see the Sun Cluster system administration documentation, and skip steps that instruct you to shut down the node.
Identify the failed disk drive.
# scdidadm -o diskid -l cNtXdY |
Is the disk drive that you want to remove a quorum device?
# scstat -q |
If no, proceed to the next step.
If yes, add a new quorum device on a different storage device. Remove the old quorum device.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
If possible, back up the metadevice or volume.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
(Solstice DiskSuite/Solaris Volume Manager Only) Does your disk drive failure prevent Solstice DiskSuite/Solaris Volume Manager from reading the disk label?
If yes, use the disk-partitioning information that you saved. You saved this disk-partitioning information when you performed one of the following tasks.
Installed your storage array in an initial cluster as outlined in Step 10 of SPARC: How to Install a Storage Array in a New SPARC Based Cluster
Added the storage array to an operational cluster.
If no, save the disk-partitioning information now if you have not already done so. Use this information when you partition the new disk drive.
# prtvtoc /dev/rdsk/cNtXdYsZ > filename |
Do not save disk-partitioning information under /tmp because you will lose this file when you reboot. Instead, save this file under /usr/tmp.
Replace the failed disk.
Determine which node owns the device group.
# scstat -D |
If you are using VERITAS Volume Manager, remove the disk drives from the VERITAS Volume Manager control on a node that does not have ownership of the device group.
# vxdisk offline cNtXdY # vxdisk rm cNtXdY |
On a node that does not have ownership of the device group, suspend activity on the SCSI bus.
# cfgadm -x replace_device cX::disk/cXtYdZ |
When prompted, type y to suspend activity on the SCSI bus.
After SCSI bus activity stops, replace the disk and type y at the prompt.
After replacing the disk, warning messages might be displayed. Ignore these messages.
Switch the primary for the device group.
# scswitch -z -D disk-device-group -h node |
If you are using VERITAS Volume Manager, remove the disk drives from the VERITAS Volume Manager control on a node that does not have ownership of the device group.
# vxdisk offline cNtXdY # vxdisk rm cNtXdY |
Suspend activity on the device group's original primary.
# cfgadm -x replace_device cX::disk/cXtYdZ |
When prompted, type y to suspend activity on the SCSI bus.
When prompted to proceed with the disk replacement, type y to proceed.
Warning messages might be displayed. Ignore these messages.
On all nodes that are attached to the storage device, run the devfsadm(1M) command to probe all devices and to write the new disk drive to the /dev/rdsk directory.
# devfsadm |
Depending on the number of devices that are connected to the node, the devfsadm command can require at least five minutes to complete.
Label the new disk drive by using the format command.
(Solstice DiskSuite/Solaris Volume Manager Only) From any node that is connected to the storage device, partition the new disk drive by using the partitioning you saved when you installed or added the storage array.
# fmthard -s filename /dev/rdsk/cNtXdYsZ |
From all nodes that are connected to the disk drive, update the DID database and driver.
# scdidadm -R deviceID |
After running scdidadm —R on the first node, each subsequent node that you run the command on might display a warning. Ignore this warning.
From any node, confirm that the failed disk drive has been replaced by comparing the new physical DID to the physical DID that was identified at the beginning of this procedure.
If the new physical DID is different, you successfully replaced the failed disk drive with a new disk drive.
# scdidadm -o diskid -l cNtXdY |
Perform volume management administration to add the disk drive back to its diskset or disk group.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
If you want this new disk drive to be a quorum device, add the quorum device.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
You need to replace a disk drive if the disk drive fails or when you want to upgrade to a higher-quality or to a larger disk.
For conceptual information about quorum, quorum devices, global devices, and device IDs, see the Sun Cluster concepts documentation.
While performing this procedure, ensure that you use the correct controller number. Controller numbers can be different on each node.
If the following warning message is displayed, ignore the message. Continue with the next step.
vxvm:vxconfigd: WARNING: no transactions on slave vxvm:vxassist: ERROR: Operation must be executed on master |
This procedure relies on the following prerequisites and assumptions.
Your cluster is operational.
Your system is running Oracle Real Application Clusters.
If your system is not running Oracle Real Application Clusters, see How to Replace a Disk Drive Without Oracle Real Application Clusters.
Your nodes are not configured with dynamic reconfiguration functionality.
If your nodes are configured for dynamic reconfiguration, see the Sun Cluster system administration documentation, and skip steps that instruct you to shut down the node.
On each node, check the status of the failed disk drive.
# vxdisk list |
On each node, identify the failed disk's instance number.
You use this instance number in Step 13.
# ls -l /dev/dsk/cWtXdYsZ # cat /etc/path_to_inst | grep "device_path" |
Ensure that you do not use the ls command output as the device path. The following example demonstrates how to find the device path and how to identify the sd instance number by using the device path.
# ls -l /dev/dsk/c4t0d0s2 lrwxrwxrwx 1 root root 40 Jul 31 12:02 /dev/dsk/c4t0d0s2 -> ../../devices/pci@4,2000/scsi@1/sd@0,0:cb41d # cat /etc/path_to_inst | grep "/pci@4,2000/scsi@1/sd@0,0" "/node@2/pci@4,2000/scsi@1/sd@0,0" 60 "sd" |
If you are using Solaris 10, the node@2 portion of this output is not present. Solaris 10 does not add this prefix for cluster nodes.
On one node, identify the failed disk's device ID number.
You will use the device ID number in Step 15 and Step 16.
# scdidadm -L | grep cXtYdZ |
On one node, identify the failed disk's disk ID number.
You will use the disk ID number in Step 17.
# scdidadm -o diskid -l cXtYdZ |
Determine if the disk drive that you want to replace is a quorum device.
# scstat -q |
If the disk drive is not a quorum device, skip this step.
If the disk drive that you want to replace is a quorum device, add a new quorum device on a different storage device and remove the old quorum device.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
On each node, remove the failed disk from VERITAS Volume Manager control.
# vxdisk offline cXtYdZ # vxdisk rm cXtYdZ |
On each node, verify that you removed the disk entry.
# vxdisk list |
Remove the failed disk from the storage array.
For the procedure about how to remove a disk drive, see your storage documentation. For a list of storage documentation, see Related Documentation.
On each node, unconfigure the failed disk.
# cfgadm -c unconfigure cX::dsk/cXtYdZ |
On each node, remove the paths to the disk drive that you are removing.
# devfsadm -C |
On each node, verify that you removed the disk.
# cfgadm -al | grep cXtYdZ # ls /dev/dsk/cXtYdZ |
Add the new disk to the storage array.
For the procedure about how to add a disk drive, see your storage documentation. For a list of storage documentation, see Related Documentation.
On each node, configure the new disk.
Use the instance number that you identified in Step 2.
# cfgadm -c configure cX::sd_instance_Number # devfsadm |
Verify that you added the disk.
# ls /dev/dsk/cXtYdZ |
On one node, update the device ID numbers.
Use the device ID number that you identified in Step 3.
# scdidadm -R DID_number |
Verify that you updated the disk's device ID number.
Use the device ID number that you identified in Step 3.
# scdidadm -L | grep DID_number |
Verify that the disk ID number is different than the disk ID number that you identified in Step 4.
On each node, add the new disk to the VERITAS Volume Manager database.
# vxdctl enable |
Verify that you added the new disk.
# vxdisk list |grep cXtYdZ |
Determine the master node.
# vxdctl -c mode |
Perform disk recovery tasks on the master node.
Depending on your configuration and volume layout, select the appropriate VERITAS Volume Manager menu item to recover the failed disk.
# vxdiskadm # vxtask list |
Upgrade your disk drive firmware if you want to apply bug fixes or enable new functionality.
This procedure relies on the following prerequisites and assumptions.
Your cluster is operational.
Your nodes are not configured with dynamic reconfiguration functionality.
If your nodes are configured for dynamic reconfiguration, see the Sun Cluster system administration documentation, and skip steps that instruct you to shut down the node.
Is the disk drive on which you want to upgrade the firmware a quorum device?
# scstat -q |
If no, proceed to Step 2.
If yes and you need a quorum device to maintain quorum, choose another disk drive to configure as the quorum device. Then remove the old quorum device.
For procedures about how to add and remove quorum devices, see the Sun Cluster system administration documentation.
If possible, back up the metadevices or volumes that reside in the disk drive.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Perform volume management administration to remove the disk drive from the configuration.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Install the firmware.
For more information about how to install firmware, see the patch installation instructions.
Perform volume management administration to add the disk drive back into the configuration.
For more information, see your Solstice DiskSuite/Solaris Volume Manager or VERITAS Volume Manager documentation.
Upgrade your host adapter firmware if you want to apply bug fixes or enable new functionality.
This procedure relies on the following prerequisites and assumptions.
Your cluster is operational.
This procedure defines Node A as the node with which you begin working. Node B is the remaining node.
Determine the resource groups and device groups that are running on the node on which you plan to upgrade the host adapter firmware.
# scstat |
Move all resource groups and device groups off Node A.
# scswitch -S -h from-node |
Perform the firmware upgrade.
For the procedure about how to upgrade your host adapter firmware, see the patch documentation.
(Optional) Return the resource groups and device groups to Node B.
# scswitch -z -g resource-group -h nodename # scswitch -z -D device-group -h nodename |