1. Introduction to Administering Oracle Solaris Cluster
2. Oracle Solaris Cluster and RBAC
3. Shutting Down and Booting a Cluster
4. Data Replication Approaches
5. Administering Global Devices, Disk-Path Monitoring, and Cluster File Systems
7. Administering Cluster Interconnects and Public Networks
Overview of Administering the Cluster
How to Change the Cluster Name
How to Map Node ID to Node Name
How to Work With New Cluster Node Authentication
How to Reset the Time of Day in a Cluster
SPARC: How to Display the OpenBoot PROM (OBP) on a Node
How to Change the Node Private Hostname
How to Add a Private Hostname for a Non-Voting Node on a Global Cluster
How to Change the Private Hostname on a Non-Voting Node on a Global Cluster
How to Delete the Private Hostname for a Non-Voting Node on a Global Cluster
How to Put a Node Into Maintenance State
How to Bring a Node Out of Maintenance State
How to Configure Load Limits on a Node
Performing Zone Cluster Administrative Tasks
How to Remove a File System From a Zone Cluster
How to Remove a Storage Device From a Zone Cluster
How to Uninstall Oracle Solaris Cluster Software From a Cluster Node
Troubleshooting a Node Uninstallation
Unremoved Cluster File System Entries
Unremoved Listing in Device Groups
Creating, Setting Up, and Managing the Oracle Solaris Cluster SNMP Event MIB
How to Enable an SNMP Event MIB
How to Disable an SNMP Event MIB
How to Change an SNMP Event MIB
How to Enable an SNMP Host to Receive SNMP Traps on a Node
How to Disable an SNMP Host From Receiving SNMP Traps on a Node
Running an Application Outside the Global Cluster
How to Take a Solaris Volume Manager Metaset From Nodes Booted in Noncluster Mode
How to Save the Solaris Volume Manager Software Configuration
How to Purge the Corrupted Diskset
How to Recreate the Solaris Volume Manager Software Configuration
10. Configuring Control of CPU Usage
11. Patching Oracle Solaris Cluster Software and Firmware
12. Backing Up and Restoring a Cluster
13. Administering Oracle Solaris Cluster With the Graphical User Interfaces
You can perform other administrative tasks on a zone cluster, such as moving the zone path, preparing a zone cluster to run applications, and cloning a zone cluster. All of these commands must be performed from the voting node of the global cluster.
Note - The Oracle Solaris Cluster commands that you run only from the voting node in the global cluster are not valid for use with zone clusters. See the appropriate Oracle Solaris Cluster man page for information about the valid use of a command in zones.
Table 9-2 Other Zone-Cluster Tasks
|
You can delete a specific zone cluster or use a wildcard to remove all zone clusters that are configured on the global cluster. The zone cluster must be configured before you remove it.
phys-schost# clresourcegroup delete -F -Z zoneclustername +
Note - This step is performed from a global-cluster node. To perform this step from a node of the zone cluster instead, log into the zone-cluster node and omit -Z zonecluster from the command.
phys-schost# clzonecluster halt zoneclustername
phys-schost# clzonecluster uninstall zoneclustername
phys-schost# clzonecluster delete zoneclustername
Example 9-11 Removing a Zone Cluster From a Global Cluster
phys-schost# clresourcegroup delete -F -Z sczone +
phys-schost# clzonecluster halt sczone
phys-schost# clzonecluster uninstall sczone
phys-schost# clzonecluster delete sczone
A file system can be exported to a zone cluster using either a direct mount or a loopback mount.
Zone clusters support direct mounts for the following:
UFS
Vxfs
QFS standalone shared file system
ZFS (exported as a data set)
Zone clusters can manage loopback mounts for the following:
UFS
Vxfs
QFS standalone shared file system
QFS shared file system
PxFS on UFS
PxFS on Vxfs
For instructions on adding a file system to a zone cluster, see Adding File Systems to a Zone Cluster in Oracle Solaris Cluster Software Installation Guide.
The phys-schost# prompt reflects a global-cluster prompt. This procedure provides the long forms of the Oracle Solaris Cluster commands. Most commands also have short forms. Except for the long and short forms of the command names, the commands are identical.
phys-schost# clresource delete -F -Z zoneclustername fs_zone_resources
phys-schost# clresource delete -F fs_global_resources
Use the -F option carefully because it forces the deletion of all the resources you specify, even if you did not disable them first. All the resources you specified are removed from the resource-dependency settings of other resources, which can cause a loss of service in the cluster. Dependent resources that are not deleted can be left in an invalid state or in an error state. For more information, see the clresource(1CL) man page.
Tip - If the resource group for the removed resource later becomes empty, you can safely delete the resource group.
phys-schost# clzonecluster configure zoneclustername
phys-schost# clzonecluster configure zoneclustername
clzc:zoneclustername> remove fs dir=filesystemdirectory
clzc:zoneclustername> commit
The file system mount point is specified by dir=.
phys-schost# clzonecluster show -v zoneclustername
Example 9-12 Removing a Highly Available File System in a Zone Cluster
This example shows how to remove a file system with a mount-point directory (/local/ufs-1) that is configured in a zone cluster called sczone. The resource is hasp-rs and is of the type HAStoragePlus.
phys-schost# clzonecluster show -v sczone ... Resource Name: fs dir: /local/ufs-1 special: /dev/md/ds1/dsk/d0 raw: /dev/md/ds1/rdsk/d0 type: ufs options: [logging] ... phys-schost# clresource delete -F -Z sczone hasp-rs phys-schost# clzonecluster configure sczone clzc:sczone> remove fs dir=/local/ufs-1 clzc:sczone> commit phys-schost# clzonecluster show -v sczone
Example 9-13 Removing a Highly Available ZFS File System in a Zone Cluster
This example shows to remove a ZFS file systems in a ZFS pool called HAzpool, which is configured in the sczone zone cluster in resource hasp-rs of type SUNW.HAStoragePlus.
phys-schost# clzonecluster show -v sczone ... Resource Name: dataset name: HAzpool ... phys-schost# clresource delete -F -Z sczone hasp-rs phys-schost# clzonecluster configure sczone clzc:sczone> remove dataset name=HAzpool clzc:sczone> commit phys-schost# clzonecluster show -v sczone
You can remove storage devices, such as SVM disksets and DID devices, from a zone cluster. Perform this procedure to remove a storage device from a zone cluster.
phys-schost# clresource delete -F -Z zoneclustername dev_zone_resources
phys-schost# clzonecluster show -v zoneclustername ... Resource Name: device match: <device_match> ...
phys-schost# clzonecluster configure zoneclustername clzc:zoneclustername> remove device match=<devices_match> clzc:zoneclustername> commit clzc:zoneclustername> end
phys-schost# clzonecluster reboot zoneclustername
phys-schost# clzonecluster show -v zoneclustername
Example 9-14 Removing an SVM Disk Set From a Zone Cluster
This example shows how to remove an SVM disk set called apachedg configured in a zone cluster called sczone. The set number of the apachedg disk set is 3. The devices are used by the zc_rs resource that is configured in the cluster.
phys-schost# clzonecluster show -v sczone ... Resource Name: device match: /dev/md/apachedg/*dsk/* Resource Name: device match: /dev/md/shared/3/*dsk/* ... phys-schost# clresource delete -F -Z sczone zc_rs phys-schost# ls -l /dev/md/apachedg lrwxrwxrwx 1 root root 8 Jul 22 23:11 /dev/md/apachedg -> shared/3 phys-schost# clzonecluster configure sczone clzc:sczone> remove device match=/dev/md/apachedg/*dsk/* clzc:sczone> remove device match=/dev/md/shared/3/*dsk/* clzc:sczone> commit clzc:sczone> end phys-schost# clzonecluster reboot sczone phys-schost# clzonecluster show -v sczone
Example 9-15 Removing a DID Device From a Zone Cluster
This example shows how to remove DID devices d10 and d11, which are configured in a zone cluster called sczone. The devices are used by the zc_rs resource that is configured in the cluster.
phys-schost# clzonecluster show -v sczone ... Resource Name: device match: /dev/did/*dsk/d10* Resource Name: device match: /dev/did/*dsk/d11* ... phys-schost# clresource delete -F -Z sczone zc_rs phys-schost# clzonecluster configure sczone clzc:sczone> remove device match=/dev/did/*dsk/d10* clzc:sczone> remove device match=/dev/did/*dsk/d11* clzc:sczone> commit clzc:sczone> end phys-schost# clzonecluster reboot sczone phys-schost# clzonecluster show -v sczone
Perform this procedure to uninstall Oracle Solaris Cluster software from a global-cluster node before you disconnect it from a fully established cluster configuration. You can use this procedure to uninstall software from the last remaining node of a cluster.
Note - To uninstall Oracle Solaris Cluster software from a node that has not yet joined the cluster or is still in installation mode, do not perform this procedure. Instead, go to “How to Uninstall Oracle Solaris Cluster Software to Correct Installation Problems” in the Oracle Solaris Cluster Software Installation Guide.
The phys-schost# prompt reflects a global-cluster prompt. Perform this procedure on a global cluster.
This procedure provides the long forms of the Oracle Solaris Cluster commands. Most commands also have short forms. Except for the long and short forms of the command names, the commands are identical.
See Table 8-2.
Note - Ensure that you have removed the node from the cluster configuration by using clnode remove before you continue with this procedure.
phys-schost# claccess allow -h hostname
Specifies the name of the node to be added to the node's authentication list.
Alternately, you can use the clsetup(1CL) utility. See How to Add a Node to the Authorized Node List for procedures.
phys-schost# clzonecluster uninstall -F zoneclustername
For specific steps, How to Remove a Zone Cluster.
On a SPARC based system, run the following command.
# shutdown -g0 -y -i0ok boot -x
On an x86 based system, run the following commands.
# shutdown -g0 -y -i0 ... <<< Current Boot Parameters >>> Boot path: /pci@0,0/pci8086,2545@3/pci8086,1460@1d/pci8086,341a@7,1/ sd@0,0:a Boot args: Type b [file-name] [boot-flags] <ENTER> to boot with options or i <ENTER> to enter boot interpreter or <ENTER> to boot with defaults <<< timeout in 5 seconds >>> Select (b)oot or (i)nterpreter: b -x
If the Java ES product registry contains a record that Oracle Solaris Cluster software was installed, the Java ES installer shows the Oracle Solaris Cluster component grayed out and does not permit reinstallation.
Run the following command, where ver is the version of the Java ES distribution from which you installed Oracle Solaris Cluster software.
# /var/sadm/prod/SUNWentsysver/uninstall
For more information about using the uninstall command, see Chapter 8, Uninstalling, in Sun Java Enterprise System 5 Update 1 Installation Guide for UNIX.
If the uninstalled node is connected to a storage device that uses Fibre Channel interfaces, no termination is necessary.
Tip - If you use a loopback file interface (lofi) device, the Java ES uninstaller automatically removes the lofi file, which is called /.globaldevices. For more information about migrating a global-devices namespace to a lofi, see Migrating the Global-Devices Namespace.
This section describes error messages that you might receive when you run the clnode remove command and the corrective actions to take.
The following error messages indicate that the global-cluster node you removed still has cluster file systems referenced in its vfstab file.
Verifying that no unexpected global mounts remain in /etc/vfstab ... failed clnode: global-mount1 is still configured as a global mount. clnode: global-mount1 is still configured as a global mount. clnode: /global/dg1 is still configured as a global mount. clnode: It is not safe to uninstall with these outstanding errors. clnode: Refer to the documentation for complete uninstall instructions. clnode: Uninstall failed.
To correct this error, return to How to Uninstall Oracle Solaris Cluster Software From a Cluster Node and repeat the procedure. Ensure that you successfully complete Step 7 in the procedure before you rerun the clnode remove command.
The following error messages indicate that the node you removed is still listed with a device group.
Verifying that no device services still reference this node ... failed clnode: This node is still configured to host device service " service". clnode: This node is still configured to host device service " service2". clnode: This node is still configured to host device service " service3". clnode: This node is still configured to host device service " dg1". clnode: It is not safe to uninstall with these outstanding errors. clnode: Refer to the documentation for complete uninstall instructions. clnode: Uninstall failed.
If you did not use the installer program to install or upgrade Sun Cluster 3.1 or 3.2 software that you now want to remove, there is no uninstall script to use for that software version. Instead, perform the following steps to uninstall the software.
Note - The Oracle Solaris Cluster 3.3 software is installed with the installer program, so these steps do not apply to that version of the software.
# cd /directory
# scinstall -r
# mv /var/sadm/install/productregistry /var/sadm/install/productregistry.sav
This section describes how to create, set up, and manage the Simple Network Management Protocol (SNMP) event Management Information Base (MIB). This section also describes how to enable, disable, and change the Oracle Solaris Cluster SNMP event MIB.
The Oracle Solaris Cluster software currently supports one MIB, the event MIB. The SNMP manager software traps cluster events in real time. When enabled, the SNMP manager automatically sends trap notifications to all hosts that are defined by the clsnmphost command. The MIB maintains a read-only table of the most current 50 events. Because clusters generate numerous notifications, only events with a severity of warning or greater are sent as trap notifications. This information does not persist across reboots.
The SNMP event MIB is defined in the sun-cluster-event-mib.mib file and is located in the /usr/cluster/lib/mib directory. You can use this definition to interpret the SNMP trap information.
The default port number for the event SNMP module is 11161, and the default port for the SNMP traps is 11162. These port numbers can be changed by modifying the Common Agent Container property file, which is /etc/cacao/instances/default/private/cacao.properties.
Creating, setting up, and managing an Oracle Solaris Cluster SNMP event MIB can involve the following tasks.
Table 9-3 Task Map: Creating, Setting Up, and Managing the Oracle Solaris Cluster SNMP Event MIB
|
This procedure shows how to enable an SNMP event MIB.
The phys-schost# prompt reflects a global-cluster prompt. Perform this procedure on a global cluster.
This procedure provides the long forms of the Oracle Solaris Cluster commands. Most commands also have short forms. Except for the long and short forms of the command names, the commands are identical.
phys-schost-1# clsnmpmib enable [-n node] MIB
Specifies the node on which the event MIB that you want to enable is located. You can specify a node ID or a node name. If you do not specify this option, the current node is used by default.
Specifies the name of the MIB that you want to enable. In this case, the MIB name must be event.
This procedure shows how to disable an SNMP event MIB.
The phys-schost# prompt reflects a global-cluster prompt. Perform this procedure on a global cluster.
This procedure provides the long forms of the Oracle Solaris Cluster commands. Most commands also have short forms. Except for the long and short forms of the command names, the commands are identical.
phys-schost-1# clsnmpmib disable -n node MIB
Specifies the node on which the event MIB that you want to disable is located. You can specify a node ID or a node name. If you do not specify this option, the current node is used by default.
Specifies the type of the MIB that you want to disable. In this case, you must specify event.
This procedure shows how to change the protocol for an SNMP event MIB.
The phys-schost# prompt reflects a global-cluster prompt. Perform this procedure on a global cluster.
This procedure provides the long forms of the Oracle Solaris Cluster commands. Most commands also have short forms. Except for the long and short forms of the command names, the commands are identical.
phys-schost-1# clsnmpmib set -n node -p version=value MIB
Specifies the node on which the event MIB that you want to change is located. You can specify a node ID or a node name. If you do not specify this option, the current node is used by default.
Specifies the version of SNMP protocol to use with the MIBs. You specify value as follows:
version=SNMPv2
version=snmpv2
version=2
version=SNMPv3
version=snmpv3
version=3
Specifies the name of the MIB or MIBs to which to apply the subcommand. In this case, you must specify event.
This procedure shows how to add an SNMP host on a node to the list of hosts that will receive trap notifications for the MIBs.
The phys-schost# prompt reflects a global-cluster prompt. Perform this procedure on a global cluster.
This procedure provides the long forms of the Oracle Solaris Cluster commands. Most commands also have short forms. Except for the long and short forms of the command names, the commands are identical.
phys-schost-1# clsnmphost add -c SNMPcommunity [-n node] host
Specifies the SNMP community name that is used in conjunction with the hostname.
You must specify the SNMP community name SNMPcommunity when you add a host to a community other than public. If you use the add subcommand without the -c option, the subcommand uses public as the default community name.
If the specified community name does not exist, this command creates the community.
Specifies the name of the node of the SNMP host that is provided access to the SNMP MIBs in the cluster. You can specify a node name or a node ID. If you do not specify this option, the current node is used by default.
Specifies the name, IP address, or IPv6 address of a host that is provided access to the SNMP MIBs in the cluster.
This procedure shows how to remove an SNMP host on a node from the list of hosts that will receive trap notifications for the MIBs.
The phys-schost# prompt reflects a global-cluster prompt. Perform this procedure on a global cluster.
This procedure provides the long forms of the Oracle Solaris Cluster commands. Most commands also have short forms. Except for the long and short forms of the command names, the commands are identical.
phys-schost-1# clsnmphost remove -c SNMPcommunity -n node host
Removes the specified SNMP host from the specified node.
Specifies the name of the SNMP community from which the SNMP host is removed.
Specifies the name of the node on which the SNMP host is removed from the configuration. You can specify a node name or a node ID. If you do not specify this option, the current node is used by default.
Specifies the name, IP address, or IPv6 address of the host that is removed from the configuration.
To remove all hosts in the specified SNMP community, use a plus sign (+) for host with the -c option. To remove all hosts, use the plus sign (+) for host.
This procedure shows how to add an SNMP user to the SNMP user configuration on a node.
The phys-schost# prompt reflects a global-cluster prompt. Perform this procedure on a global cluster.
This procedure provides the long forms of the Oracle Solaris Cluster commands. Most commands also have short forms. Except for the long and short forms of the command names, the commands are identical.
phys-schost-1# clsnmpuser create -n node -a authentication \ -f password user
Specifies the node on which the SNMP user is added. You can specify a node ID or a node name. If you do not specify this option, the current node is used by default.
Specifies the authentication protocol that is used to authorize the user. The value of the authentication protocol can be SHA or MD5.
Specifies a file that contains the SNMP user passwords. If you do not specify this option when you create a new user, the command prompts for a password. This option is valid only with the add subcommand.
You must specify user passwords on separate lines in the following format:
user:password
Passwords cannot contain the following characters or a space:
; (semicolon)
: (colon)
\ (backslash)
\n (newline)
Specifies the name of the SNMP user that you want to add.
This procedure shows how to remove an SNMP user from the SNMP user configuration on a node.
The phys-schost# prompt reflects a global-cluster prompt. Perform this procedure on a global cluster.
This procedure provides the long forms of the Oracle Solaris Cluster commands. Most commands also have short forms. Except for the long and short forms of the command names, the commands are identical.
phys-schost-1# clsnmpuser delete -n node user
Specifies the node from which the SNMP user is removed. You can specify a node ID or a node name. If you do not specify this option, the current node is used by default.
Specifies the name of the SNMP user that you want to remove.