configuring a multi-node setup - cisco.com · •threeservicenodes •oneinventorydatabase...

Configuring a Multi-Node Setup

This chapter contains the following sections:

• About the Multi-Node Setup, page 1

• Minimum System Requirements for a Multi-Node Setup, page 3

• Guidelines and Limitations for a Multi-Node Setup, page 10

• Best Practices for a Multi-Node Setup, page 11

• Summary of Steps for Configuring a Multi-Node Setup, page 11

• Creating the Inventory Database, page 12

• Creating the Monitoring Database, page 13

• Creating the Primary Node, page 13

• Creating a Service Node, page 14

• System Tasks, page 15

• Backing Up with a Multi-Node Setup, page 19

• Restoring with a Multi-Node Setup, page 20

• Removing a Service Node, page 21

• Options for Migrating from a Standalone Appliance to a Multi-Node Setup, page 21

About the Multi-Node SetupThemulti-node setup is supported for Cisco UCSDirector on a 64-bit operating system only.With a multi-nodesetup, you can scale Cisco UCS Director to support a larger number of VMs than is supported by a singleinstallation of Cisco UCS Director. This setup includes the following nodes:

• One primary node

• One or more service nodes

• One monitoring database

Cisco UCS Director Shell Guide, Release 5.3 1

• One inventory database

Figure 1: Multi-Node Topology

For a multi-node setup, you have to install the license on the primary node only.Note

A multi-node setup improves scalability by offloading the processing of system tasks, such as inventory datacollection, from the primary node to one or more service nodes. You can assign certain system tasks to oneor more service nodes. The number of nodes determines how the processing of system tasks are scaled.

Node pools combine service nodes into groups and enable you to assign system tasks to more than one servicenode. This provides you with control over which system task is executed by which service node or group ofservice nodes. If you have multiple service nodes in a node pool and one of the service nodes is busy when asystem task needs to be run, Cisco UCS Director uses a round-robin assignment to determine which servicenode should process that system task. if all service nodes are busy, you can have the primary node run thesystem task.

However, if you do not need that level of control over the system tasks, you can use the default task policyand add all of the service nodes to the default node pool. All system tasks are already associated with thedefault task policy, and the round-robin assignment will be used to determine which service node shouldprocess a system task.

If you want to have some critical tasks processed only by the primary node, you can assign those tasks to alocal-run policy.

For more information about how to configure the primary node and service nodes, and how to assign systemtasks, see the Cisco UCS Director Administration Guide.

Cisco UCS Director Shell Guide, Release 5.32

Configuring a Multi-Node SetupAbout the Multi-Node Setup

http://www.cisco.com/en/US/products/ps13050/prod_maintenance_guides_list.html

Primary NodeA multi-node setup can have only one primary node. This primary node contains the license for Cisco UCSDirector.

The primary node also contains the configuration for the node pools and the service nodes, along with the listof which system tasks can be offloaded to the service nodes for processing.

Service NodesA multi-node setup can have one or more service nodes. The number of service nodes in a multi-node setupdepends upon the number of devices and VMs you plan to configure andmanage through Cisco UCSDirector.

Database NodesThe inventory and monitoring databases are created from the Cisco UCS Director MySQL database. The datathat Cisco UCS Director collects is divided between the two databases. The multi-node setup segregates thedata collection which is historically very heavy on the database into a separate database.

Inventory Database

A multi-node setup can have only one inventory database. This database contains the following:

• Physical and virtual accounts and their related inventory data

• Data used in the normal operation of Cisco UCS Director for all supported features

Monitoring Database

A multi-node setup can have only one monitoring database. This database contains the data that Cisco UCSDirector uses for historical computations, such as aggregations and trend reports.

The parameters of themonitoring database depend upon the number of devices andVMs you plan to configureand manage through Cisco UCS Director

Minimum System Requirements for a Multi-Node SetupThe minimum system requirements for a multi-node setup depends upon the number of VMs that need to besupported by Cisco UCS Director. The following table describes the number of VMs supported by eachdeployment size.

Number of VMs SupportedDeployment Size

5,000 to 10,000 VMsSmall

10,000 to 20,000 VMsMedium

20,000 to 50,000 VMsLarge


Configuring a Multi-Node SetupPrimary Node

Minimum System Requirements for a Small Multi-Node SetupThe small multi-node setup supports from 5,000 to 10,000 VMs.We recommend that this deployment includethe following nodes:


• Two service nodes



For optimal performance, reserve additional CPU and memory resources.Note

Minimum Requirements for each Primary Node and Service Node

Minimum Supported RequirementElement

4vCPU

16 GBMemory

100 GBHard disk

Minimum Requirements for the Inventory Database


4vCPU

30 GBMemory

100 GB (SSD Type Storage)Hard disk

Minimum Requirements for the Monitoring Database


4vCPU

30 GBMemory

100 GB (SSD Type Storage)Hard disk


Configuring a Multi-Node SetupMinimum System Requirements for a Small Multi-Node Setup

Minimum Memory Configuration for Cisco UCS Director Services on Primary and Service Nodes

ParameterFile LocationRecommendedConfiguration

Service

-Xms -Xmx/opt/infra/broker/run.sh512 MBbroker

-Xms -Xmx/opt/infra/client/run.sh512 MBclient

-Xms -Xmx/opt/infra/controller/run.sh512 MBcontroller

-Xms -Xmx/opt/infra/eventmgr/run.sh512 MBeventmgr

-Xms -Xmx/opt/infra/idaccessmgr/run.sh512 MBidaccessmgr

-Xms -Xmx/opt/infra/inframgr/run.sh8 GBinframgr

JAVA_OPTS="$JAVA_OPTS-Xmsm -Xmxm"

/opt/infra/web_cloudmgr/apache-tomcat/bin/catalina.sh

2 GBTomcat

Minimum Configuration for the Inventory and Monitoring Databases

Minimum Supported ConfigurationComponent

1000thread_cache_size

1000max_connections

100innodb_lock_wait_timeout

128 MBquery_cache_size

24576 MBinnodb_buffer_pool_size

10,000max_connect_errors

20connect_timeout

64innodb_read_io_threads

64innodb_write_io_threads

Minimum System Requirements for a Medium Multi-Node SetupThemediummulti-node setup supports between 10,000 and 20,000 VMs.We recommend that this deploymentinclude the following nodes:



Configuring a Multi-Node SetupMinimum System Requirements for a Medium Multi-Node Setup

• Three service nodes






8vCPU

30 GBMemory

100 GBHard disk



8vCPU

60 GBMemory

100 GB (SSD type storage)Hard disk



8vCPU

60 GBMemory




Service

-Xms -Xmx/opt/infra/broker/run.sh1 GBbroker




Service

-Xms -Xmx/opt/infra/client/run.sh1 GBclient

-Xms -Xmx/opt/infra/controller/run.sh1 GBcontroller


-Xms -Xmx/opt/infra/idaccessmgr/run.sh1 GBidaccessmgr




4 GBTomcat

Minimum Inventory Database Configuration



2000max_connections



43,008 MBinnodb_buffer_pool_size


20connect_timeout



Minimum Monitoring Database Configuration



2000max_connections








20connect_timeout



Minimum System Requirements for a Large Multi-Node SetupThe large multi-node setup supports between 20,000 and 50,000 VMs. We recommend that this deploymentinclude the following nodes:


• Six service nodes






8vCPU

60 GBMemory

100 GBHard disk



8vCPU

120 GBMemory


Configuring a Multi-Node SetupMinimum System Requirements for a Large Multi-Node Setup





8vCPU

120 GBMemory




Service

-Xms -Xmx/opt/infra/broker/run.sh2 GBbroker

-Xms -Xmx/opt/infra/client/run.sh2 GBclient

-Xms -Xmx/opt/infra/controller/run.sh2 GBcontroller


-Xms -Xmx/opt/infra/idaccessmgr/run.sh2 GBidaccessmgr




8 GBTomcat

Minimum Inventory Database Configuration



4000max_connections




Configuring a Multi-Node SetupMinimum System Requirements for a Large Multi-Node Setup




20connect_timeout



Minimum Monitoring Database Configuration



4000max_connections





20connect_timeout



Guidelines and Limitations for a Multi-Node SetupBefore you configure a multi-node setup for Cisco UCS Director, consider the following:

• The multi-node setup is supported for Cisco UCS Director on a 64-bit operating system only.

• Your multi-node setup can have only one primary node.

• You must plan the location and IP addresses of your nodes carefully as you cannot reconfigure the typesof most nodes later. You can only reconfigure a service node as a primary node. You cannot make anyother changes to the type of node. For example, you cannot reconfigure a primary node as a service nodeor an inventory database node as a monitoring database node.

• You have to install the license on the primary node only.


Configuring a Multi-Node SetupGuidelines and Limitations for a Multi-Node Setup

• After you configure the nodes, the list of operations available in the shelladmin changes for the servicenodes, the inventory database node, and the monitoring database node.

Best Practices for a Multi-Node SetupBefore you configure a multi-node setup for Cisco UCS Director, consider the following best practices:

• Tomaximize output andminimize network latency, we recommend that the primary node, service nodes,inventory database nodes, and monitoring database nodes reside on the same host.

• The primary node, inventory database nodes, and monitoring database nodes can reside on the samehost, network or physical location, and the service nodes can reside on another network or physicallocation, as long as the primary nodes and service nodes are accessible by the databases.

•We recommend that you assess the network latency as part of your multi-node deployment plan andplace a service node closer to the physical and virtual infrastructures, and associate any relevant systemtasks to run on the local service nodes. The primary node handles offloading the execution of systemtasks to the appropriate service node.

• You can offload system tasks to available service nodes by associating the service nodes with a servicenode pool.

Summary of Steps for Configuring a Multi-Node SetupThis procedure provides a high-level summary of the steps involved in configuring a multi-node setup.

Plan your multi-node setup carefully. You can only change a service node to a primary node. You cannotchange any other type of node after you configure it. For example, you cannot reconfigure a primary nodeas a service node or an inventory database node as a monitoring database node.

Note

Step 1 Deploy a Cisco UCS Director VM for each node in the multi-node setup.See Installing Cisco UCS Director on VMware vSphere.

For example, if your multi-node setup, includes a primary node, an inventory database, a monitoring database, and threeservice nodes, deploy six Cisco UCS Director VMs.

Step 2 In the Cisco UCS Director shelladmin, configure the nodes in the following order:

1 Creating the Inventory Database, on page 12

2 Creating the Monitoring Database, on page 13

3 Creating the Primary Node, on page 13

4 Creating a Service Node, on page 14

You must create and start the inventory database and the monitoring database nodes before you configure theprimary node and any service nodes.

Note


Configuring a Multi-Node SetupBest Practices for a Multi-Node Setup

Step 3 Update the license file in the Cisco UCS Director node that will be the primary node.You do not need to update the license file on any other node.

See Updating the License.

Step 4 In Cisco UCS Director on the primary node, configure the system tasks as follows:a) Create one or more node pools if you want to control the assignment of system tasks by service node, or accept the

default node pool.See Creating a Node Pool, on page 16.

b) Create one or more system task policies if you want to control the assignment of system tasks by service node, oraccept the default task policy.See Creating System Task Policy, on page 16.

c) Configure the service nodes.d) Configure the primary node.e) Assign the system tasks to system policies if you want to control the assignment of system tasks by service node.

See Assigning a System Policy to a System Task, on page 18.

For more information, see System Tasks, on page 15.

Creating the Inventory Database

Step 1 Log into the Cisco UCS Director shelladmin on the inventory database node.Step 2 From the Cisco UCS Director Shell menu, choose Configure Multi-Node Setup and press Enter.Step 3 From the menu, choose Configure Multi Node Setup (Advanced Deployment) and press Enter.Step 4 When prompted, press 1 to configure the current node.Step 5 When prompted, press y and then select the option to configure the node as the inventory database node.Step 6 From the menu, choose Configure Inventory Database and press Enter.Step 7 When prompted, press Enter to continue.Step 8 To verify that the services for the inventory database are up and running, choose Display Services Status and

press Enter.You should see the following lines:

2838 ? 00:00:00 mysqld_safe3172 ? 3-02:51:38 mysqld

After you return to the shelladmin, the menu options change to those available for an inventory database node.


Configuring a Multi-Node SetupCreating the Inventory Database

Creating the Monitoring Database

Step 1 Log into the Cisco UCS Director shelladmin on the monitoring database node.Step 2 From the Cisco UCS Director Shell menu, choose Configure Multi-Node Setup and press Enter.Step 3 From the menu, choose Configure Multi Node Setup (Advanced Deployment) and press Enter.Step 4 When prompted, press 1 to configure the current node.Step 5 When prompted, press y and then select the option to configure the node as the monitoring database node.Step 6 From the menu, choose Configure Monitoring Database and press Enter.Step 7 When prompted, press Enter to continue.Step 8 To verify that the services for the monitoring database are up and running, choose Display Services Status

and press Enter.You should see the following lines:

2838 ? 00:00:00 mysqld_safe3172 ? 3-02:51:38 mysqld

After you return to the shelladmin, the menu options change to those available for a monitoring database node.

Creating the Primary NodeBefore You Begin

The inventory and monitoring databases must be up and running before you create the primary node.

Do not run any daemons on the primary database.Note

Step 1 Log into the Cisco UCS Director shelladmin on the primary node.Step 2 From the Cisco UCS Director Shell menu, choose Configure Multi-Node Setup and press Enter.Step 3 From the menu, choose Configure Multi Node Setup (Advanced Deployment) and press Enter.Step 4 When prompted, press 1 to configure the current node.Step 5 When prompted, press y and then select the option to configure the node as the primary node.Step 6 From the menu, choose Configure Primary Node and press Enter.Step 7 At the Provide Inventory DB IP prompt, enter the IP address assigned to the Cisco UCS Director VM for the inventory

database.This step registers the VM as a primary node with the inventory database.


Configuring a Multi-Node SetupCreating the Monitoring Database

Step 8 At the ProvideMonitoring DB IP prompt, enter the IP address assigned to the Cisco UCSDirector VM for the monitoringdatabase.This step registers the VM as a primary node with the monitoring database.

Step 9 When prompted, press Enter to continue.Step 10 To verify that the services for the primary node are up and running, choose Display Services Status and press

Enter.You should see the following lines:

Service Status PID---------- ---------- -----broker RUNNING 25677controller RUNNING 25715eventmgr RUNNING 25749client RUNNING 25808idaccessmgr RUNNING 25854inframgr RUNNING 25911TOMCAT RUNNING 25967websock RUNNING 26009

Node Type : primaryInventory DB( 172.29.109.135:3306 ) status : UPMonitor DB( 172.29.109.144:3306 ) status : UPPress return to continue ...

After you return to the shelladmin, the menu options change to those available for a primary node.

Creating a Service NodeBefore You Begin

The inventory and monitoring databases must be up and running before you create the primary node.

Step 1 Log into the Cisco UCS Director shelladmin on the service node.Step 2 From the Cisco UCS Director Shell menu, choose Configure Multi-Node Setup and press Enter.Step 3 From the menu, choose Configure Multi Node Setup (Advanced Deployment) and press Enter.Step 4 When prompted, press 1 to configure the current node.Step 5 When prompted, press y and then select the option to configure the node as the secondary node.Step 6 From the menu, choose Configure Service Node and press Enter.Step 7 At the Provide Inventory DB IP prompt, enter the IP address assigned to the Cisco UCS Director VM for the inventory

database.This step registers the VM as a service node with the inventory database.


Configuring a Multi-Node SetupCreating a Service Node

Step 8 At the ProvideMonitoring DB IP prompt, enter the IP address assigned to the Cisco UCSDirector VM for the monitoringdatabase.This step registers the VM as a service node with the monitoring database.

Step 9 When prompted, press Enter to continue.Step 10 To verify that the services for the service node are up and running, choose Display Services Status and press

Enter.You should see the following lines:

Service Status PID---------- ---------- -----broker RUNNING 25677controller RUNNING 25715eventmgr RUNNING 25749client RUNNING 25808idaccessmgr RUNNING 25854inframgr RUNNING 25911TOMCAT RUNNING 25967websock RUNNING 26009

Node Type : serviceInventory DB( 172.29.109.135:3306 ) status : UPMonitor DB( 172.29.109.144:3306 ) status : UPPress return to continue ...

After you return to the shelladmin, the menu options change to those available for a service node.

Step 11 Repeat this procedure for every service node.

System TasksThe System Tasks tab displays all the system tasks that are currently available in Cisco UCS Director.However, this list of system tasks is linked to the type of accounts that you have created in Cisco UCSDirector.For example, if you have logged in for the first time, then only a set of general system-related tasks or VMwarerelated tasks are visible on this page. As and when you add accounts, such as rack accounts, or Cisco UCSManager accounts, system tasks related to these accounts are populated on this page.

Following are the tasks that you can complete from the System Tasks page:

• View the available systems tasks—You can use the Expand andCollapse options to view all the systemtasks that are available on this page. The tasks are categorized according to the accounts available inCisco UCS Director. For example: Cisco UCS Tasks, NetApp Tasks and so on.

• Disable and enable system tasks—In circumstances when there are multiple processes or tasks runningon the appliance, you can choose to disable a system task. If you do so, then until such time that youmanually enable it, the system task will not run. This will affect the data that is populated in other reports.For example, if you disable an inventory collection system task, then reports that require this data maynot display accurate data. In this case, you will have to manually run an inventory collection process,or enable the system task.


Configuring a Multi-Node SetupSystem Tasks

For more information, see Disabling or Enabling a System Task, on page 19.

In a single node setup, where there is only one server, all system tasks will run on this server. In a multi-nodesetup, where there are multiple servers configured, all system tasks, by default, run on the primary server.However, you can specify system tasks to run on the secondary servers. Following is the recommended stepsto perform this task:

1 Ensure that the secondary servers are available in Cisco UCS Director as nodes. If the servers are notavailable, then you must add the servers as nodes. For more information, see Creating a Service Node,on page 17.

2 Create a node pool from the available servers. For more information, see Creating a Node Pool, on page16.

3 Create a system task policy, and associate it with a node policy. For more information, see Creating SystemTask Policy, on page 16.

4 Associate a node pool with the system task policy. For more information, see Assigning a Node Pool toSystem Task Policy, on page 17.

5 Select a system task, and associate it with a system-task policy. For more information, see Assigning aSystem Policy to a System Task, on page 18.

Creating a Node Pool

Step 1 On the menu bar, choose Administration > System.Step 2 Choose the Service Nodes tab.Step 3 Click the Service Node Pools icon. The Service Node Pool dialog box displays.Step 4 Click the + (plus) icon. The Add Entry to Service Node Pools dialog box displays.Step 5 In the Name field, enter the node pool name.Step 6 (Optional) In the Description field, enter a description of the node pool name.Step 7 Click Submit. The node pool is created.

Creating System Task PolicyYou can group system tasks into a system task policy to later determine which system tasks are running onwhich node.


Configuring a Multi-Node SetupCreating a Node Pool

A new system task policy is created when an administrator decides when one or more policies are neededin addition to the default system task policy.

Note

Step 1 On the menu bar, choose Administration > System.Step 2 Choose the System Task Policy tab.Step 3 Click the Add icon. The Add dialog box displays.Step 4 In the Name field, enter the name that you gave the system task policy.Step 5 (Optional) In the Description field, enter a description of the system task policy.Step 6 From the Node Pool drop-down list, choose the node pool to which this system task policy belongs.Step 7 Click Submit. The selected node pool now belongs to the newly created system task policy.

Assigning a Node Pool to System Task Policy

Step 1 On the menu bar, choose Administration > System.Step 2 Choose the System Task Policy tab.Step 3 Select an existing system task policy from the Name column and click the Edit icon. The Edit dialog box displays.

If the default system task policy is used, you can assign service nodes to this policy. See the Creating SystemTask Policy, on page 16 section, if you want to configure a policy that is different from the default.

Note

Step 4 Select a node pool to which this System Task Policy belongs from the Node Pool drop-down list, choose the node poolto which this system task policy belongs.

Step 5 Click Submit. The selected node pool now belongs to the system task policy.

Creating a Service Node

Step 1 On the menu bar, choose Administration > System.Step 2 Choose the Service Nodes tab.Step 3 Click the Add (+)Step 4 In the Service Node dialog box, complete the following fields:

DescriptionName

Enter the name of the service node.Node Name field


Configuring a Multi-Node SetupAssigning a Node Pool to System Task Policy

DescriptionName

Service is described as the role of this node.Role field

The default-service-node-pool is chosen.Service Node Pool drop-down list

Enter either the DNS name or IP address of the servicenode.

This field cannot use the Primary Node’s IPaddress. Ensure that a valid Service Node DNSname or IP address is entered.

Note

DNS Name field

The description of the of the service node.Description field

Choose either http (default) or https.Protocol drop-down list

The default TCP port for the Hypertext Transfer Protocol(HTTP) 80 is entered by default. Enter a different TCP portif necessary.

Port field

The infraUser user name is entered by default. TheinfraUser is a user account created by default. To find thisuser account on the menu bar, choose Administration>Users and Groups. Choose the Login Users tab to findthe infraUser user account in the Login Name column.

The InfraUser user name is not the defaultadministrator user to login to the UCS Directorsystem itself.

Note

Another user name can be added to this field. This user’sAPI key is used to authenticate with the Service Node.

UserName field

Step 5 Click Submit.

Assigning a System Policy to a System Task

Step 1 On the menu bar, choose Administration > System.Step 2 Choose the System Task tab.Step 3 Choose a folder that contains system task(s). Click the folder arrow to expand its tasks.

128 system tasks areavailable.

Note


Configuring a Multi-Node SetupAssigning a System Policy to a System Task

Step 4 Choose the task and click theManage Task icon. TheManage Task dialog box displays.Step 5 From the Task Execution drop-down list, choose Enable.Step 6 From the System Task Policy drop-down list, choose a system policy.Step 7 Click Submit. The system task is assigned to the selected system policy.

Executing System Tasks

There are few system tasks that cannot be run remotely on the service node. You can also assign a systempolicy remotely from the local host or primary node.

Note

Step 1 On the menu bar, choose Administration > System.Step 2 Choose the System Tasks tabStep 3 Choose a task from the list.Step 4 Click Run Now. The selected task updates its information.

Disabling or Enabling a System Task

Step 1 On the menu bar, choose Administration > System.Step 2 Choose the System Task tab.Step 3 Choose a folder that contains system task(s). Click the folder arrow to expand its tasks.

128 system tasks areavailable.

Note

Step 4 Choose the task and click the Manage Task icon. TheManage Task dialog box displays.Step 5 To disable a system task, from the Task Execution drop-down list, choose Disable.Step 6 To enable a system task, from the Task Execution drop-down list, choose Enable.Step 7 Click Submit.

Backing Up with a Multi-Node SetupThis procedure provides a high-level summary of the steps required to back up a multi-node setup.


Configuring a Multi-Node SetupExecuting System Tasks

The backup procedure requires you to first stop the Cisco services. This must be done before you work withtheMonitoring and Inventory databases. Then you back up the databases, after which you re-start the services,starting with the Primary node.

Before You Begin

To back up the databases, you use an FTP server, for which you will need:

• An FTP server account and the IP address

• The FTP server's login credentials

• Backup file name(s)

SUMMARY STEPS

1. Shut down the services for both the Primary and Service nodes. Use the, Stop Services Shell potion.2. Back up the Monitoring database. Use the Backup Database Shell option.3. Backup the Inventory database. Use the Backup Database Shell option.4. Start the Services (Service nodes first, followed by Primary node). Use the Start Services Shell option.

DETAILED STEPS

Step 1 Shut down the services for both the Primary and Service nodes. Use the, Stop Services Shell potion.Step 2 Back up the Monitoring database. Use the Backup Database Shell option.Step 3 Backup the Inventory database. Use the Backup Database Shell option.Step 4 Start the Services (Service nodes first, followed by Primary node). Use the Start Services Shell option.

Restoring with a Multi-Node SetupThis procedure provides a high-level summary of the steps required to restore a multi-node setup.

The restore procedure requires you to first stop the Cisco services. This must be done before you work withthe Monitoring and Inventory databases. Then you restore the databases, after which you re-start the services,starting with the Primary node.

Before You Begin

To restore the databases, you use an FTP server, for which you will need:

• An FTP server account and the IP address

• The FTP server's login credentials

• Backup file name(s)


Configuring a Multi-Node SetupRestoring with a Multi-Node Setup

SUMMARY STEPS

1. Shut down the services for both the Primary and Service nodes. Use Shell option 3, Stop Services.2. Restore the Monitoring database. Use the Shell Restore Database option.3. Restore the Inventory database. Use the Shell Restore Database option.4. Start the Services (Service nodes first, followed by Primary node). Use Shell option 4, Start Services.

DETAILED STEPS

Step 1 Shut down the services for both the Primary and Service nodes. Use Shell option 3, Stop Services.Step 2 Restore the Monitoring database. Use the Shell Restore Database option.Step 3 Restore the Inventory database. Use the Shell Restore Database option.Step 4 Start the Services (Service nodes first, followed by Primary node). Use Shell option 4, Start Services.

Removing a Service NodeThis procedure provides a high-level summary of the steps required to remove a service node from amulti-nodesetup.

The removal procedure requires you to first stop the Cisco services. This must be done before you remove aservice node from Monitoring and Inventory databases.

Step 1 Shut down the services for both the Primary node and the Service node(s). Use Shell option 3, Stop Services.Step 2 Start the services on the Service node(s) that you want to be active and operational. Use Shell option 4, Start Services.Step 3 Start the services on the Primary node. Use Shell option 4, Start Services.Step 4 Using the Cisco UCS Director GUI, delete the Service node that you want to remove.

This removes the respective Service node from the database.

Options for Migrating from a Standalone Appliance to aMulti-Node Setup

All requirements and configuration for a multi-node setup must be completed before you attempt to migratefrom a standalone appliance to a multi-node setup. Please refer to the requirements and instructions pertainingto the multi-node setup you are planning to implement before proceeding.

Refer to the most recent Release Notes and ReadMe file for up-to-date information about support for migratingfrom a previous release of a Cisco UCS Director standalone appliance to a multi-node setup on the currentrelease. Single to multi-node migration from some earlier releases to the current release may not be supported.For more information on supported migration paths, see the Upgrading Cisco UCS Director chapter.


Configuring a Multi-Node SetupRemoving a Service Node

You must upgrade to Cisco UCS Director, Release 4.0 or later, before you can migrate to a multi-nodesetup and backup the database.

Note

You can choose between the following options for migrating a Cisco UCS Director standalone appliance toa multi-node setup:

• You restore a backup of the entire database from Cisco UCS Director, Release 4.0, to the Cisco UCSDirector, Release 4.1 monitoring database and inventory database nodes. This option is similar to thedatabase backup and restore you can perform through shelladmin.

• You back up the database from Cisco UCS Director, Release 4.0, separately for the inventory databaseand monitoring database nodes. If you choose this option, only selected database tables are backed upand restored to the monitoring and inventory databases. This option requires you to execute customscripts from /opt/infra.

• You perform a back up that involves using custom scripts twice, once for the inventory database andonce for the Monitor database. The backed up databases will have tables that are built selectively forthe node type. You then copy the backed up appliance file to the relevant node and execute the customscript again. But this time you execute the script in each Inventory and Monitor node.

Detailed instructions for these options follow in this section.

Migrating a Release 4.0 Standalone Appliance to a Multi-Node Setup with aFull Database Backup

This procedure restores all database tables to the inventory database and the monitoring database.

Before You Begin

You must have an FTP server with valid credentials to perform a backup and restore of the database.

Step 1 In the shelladmin, upgrade the Cisco UCS Director, Release 4.0, standalone appliance to Cisco UCS Director, Release4.1 with the Apply Patch option.For more information, see the following:

• Upgrading from Cisco Unified Infrastructure Controller 3.x 32-bit to Cisco UCS Director, Release 4.1 64-bit

• Upgrading from Cisco UCS Director, Release 4.0 to Release 4.1

Step 2 In the shelladmin, choose Backup Database to backup the full database of the upgraded Cisco UCS Director.Step 3 Deploy a Cisco UCS Director VM for each node in the multi-node setup.

For example, if your multi-node setup, includes a primary node, an inventory database, a monitoring database, and threeservice nodes, deploy six Cisco UCS Director VMs.

Step 4 In the shelladmin, configure the nodes in the following order:




Configuring a Multi-Node SetupMigrating a Release 4.0 Standalone Appliance to a Multi-Node Setup with a Full Database Backup




Note

Step 5 In the shelladmin, choose Stop services to stop the Cisco UCS Director services.Step 6 Choose Restore database to import and restore the database backup that you took in Step 2.Step 7 When prompted, choose Start services to start the Cisco UCS Director services.

After you log in to the primary node, all data from the database in the Cisco UCS Director, Release 4.0, standaloneappliance should be available.

Migrating a Release 4.0 Standalone Appliance to a Multi-Node Setup with aSelected Database Backup

This procedure backs up and restores only the selected database tables required for each of the inventorydatabase and the monitoring database.

Before You Begin

You must have an FTP server with valid credentials to perform a backup and restore of the database.

Step 1 In the shelladmin, upgrade the Cisco UCS Director, Release 4.0, standalone appliance to Cisco UCS Director, Release4.1 with the Apply Patch option.For more information, see the following:

• Upgrading from Cisco Unified Infrastructure Controller 3.x 32-bit to Cisco UCS Director, Release 4.1 64-bit

• Upgrading from Cisco UCS Director, Release 4.0 to Release 4.1

Step 2 In the shelladmin, choose Login as Root to log in to the upgraded Cisco UCS Director.Step 3 Back up the database tables required for the inventory database, as follows:

a) Navigate to the /opt/infra folder.

cd /opt/infra

b) From Infra, run the dbInfraBackupRestore.sh script.

# ./dbInfraBackupRestore.sh backup

This script takes a database backup with the tables required for the inventory database.

c) After the script has been executed, check for an output file in /tmp folder namedinfra_database_backup.tar.gz.

Step 4 Back up the database tables required for the monitoring database, as follows:


Configuring a Multi-Node SetupMigrating a Release 4.0 Standalone Appliance to a Multi-Node Setup with a Selected Database Backup

a) Navigate to the /opt/infra folder.

cd /opt/infra

b) From Infra, run the dbMonitoringBackupRestore.sh script.

# ./dbMonitoringBackupRestore.sh backup

This script takes a database backup with the tables required for the monitoring database.

c) After the script has been executed, check for an output file in /tmp folder namedmonitoring_database_backup.tar.gz.

Step 5 Deploy a Cisco UCS Director VM for each node in the multi-node setup.For example, if your multi-node setup, includes a primary node, an inventory database, a monitoring database, and threeservice nodes, deploy six Cisco UCS Director VMs.

Step 6 In the shelladmin, configure the nodes in the following order:






Note

Step 7 In the shelladmin, choose Stop services to stop the Cisco UCSDirector services on the primary node and all servicenodes.

Step 8 Restore the database to the inventory database, as follows:a) Copy the infra_database_backup.tar.gz to the /tmp folder on the node that hosts the inventory database.b) On the inventory database node, navigate to the /opt/infra folder.

cd /opt/infra

c) From Infra, run the dbInfraBackupRestore.sh script.

# ./dbInfraBackupRestore.sh restore

The script restores the database backup from the backup file in the /tmp folder .

Step 9 Restore the database to the monitoring database, as follows:a) Copy the monitoring_database_backup.tar.gz to the /tmp folder on the node that hosts the monitoring

database.b) Navigate to the /opt/infra folder.

cd /opt/infra


Configuring a Multi-Node SetupMigrating a Release 4.0 Standalone Appliance to a Multi-Node Setup with a Selected Database Backup

c) From Infra, run the dbMonitoringBackupRestore.sh script.

# ./dbMonitoringBackupRestore.sh restore

The script restores the database backup from the backup file in the /tmp folder .

Step 10 Choose Start services on the primary node and all service nodes to start the Cisco UCS Director services.After you log in to the primary node, all data from the database in the Cisco UCS Director, Release 4.0, standaloneappliance should be available.

Migrating a Release 4.1 Standalone Appliance to a Multi-Node Setup withSelective Database Table Backup

To perform this migration to a multiple node setup, your back up involves using custom scripts twice, oncefor the inventory database and once for the Monitor database. The backed up databases have tables that arebuilt selectively for the node type. To restore to the multi-node setup, you copy the backed up appliance fileto the relevant node and execute the custom script again in each Inventory and Monitor node.Steps for Backup and Restore for Standalone Multi node

The DB backup step involves executing custom script which needs to beexecuted twice, one for Inventory DB and other for Monitor DB.The backed up DB will have selective tables that are relevant to the node type.

DB Backup

1st DB Backup for Inventory Node- Login as root user

cd /opt/infraFrom Infra execute the dbInfraBackupRestore.sh as shown below#. /dbInfraBackupRestore.sh backupThis script will take a backup with necessary tablesfor inventory node.Once the script is executed, check for the output filein /tmp folder with name as infra_database_backup.tar.gz

- 2nd DB Backup for Monitor Node- Login as root usercd /opt/infraFrom Infra execute the dbMonitoringBackupRestore.sh as shown below#. / dbMonitoringBackupRestore.sh backupThis script will take a backup with necessary tables for monitor node.Once the script is executed, check for the output file in /tmp folderwith name as monitoring_database_backup.tar.gz

DB Restore

Stop services on Primary Node & Service Node. In case of more thanone Services nodes, Stop services on all the nodes.

To restore the DB on Inventory and Monitor node, user needs tocopy the backed up file from backup appliance to the relevant node.After the files are placed on the nodes, user needs to execute thecustom script again. But this time the script needs to beexecuted individually in each node (Inventory and Monitor nodes).


Configuring a Multi-Node SetupMigrating a Release 4.1 Standalone Appliance to a Multi-Node Setup with Selective Database Table Backup

- 1st DB Restore for Inventory Node- Login as root userCopy the backup file (infra_database_backup.tar.gz) from the backupappliance to /tmp folder in the Inventory node appliance.In the Inventory node appliancecd /opt/infraFrom Infra execute the dbInfraBackupRestore.sh as shown below#. /dbInfraBackupRestore.sh restoreThis script will restore the DB from backup file(infra_database_backup.tar.gz) placed in /tmp folder.

- 2nd DB Restore for Monitor Node- Login as root userCopy the backup file (monitoring_database_backup.tar.gz) from the backupappliance to /tmp folder in the Monitor node appliance.In the Monitor node appliancecd /opt/infraFrom Infra execute the dbMonitoringBackupRestore.sh as shown below#. / dbMonitoringBackupRestore.sh restoreThis script will restore the DB from backup file(monitoring_database_backup.tar.gz) placed in /tmp folder.

Before You Begin

Step 1 Start the database backup with the Inventory node.a) Login as root user.b) cd /opt/infrac) From infra, execute the dbInfraBackupRestore.sh script as shown below.

#. /dbInfraBackupRestore.sh backup

This script will make a backup with the necessary tables for the Inventory node.

d) After the script has executed, check in the /tmp folder for the output file, which will have a name something likeinfra_database_backup.tar.gz.Although actual file names may vary. This is the given name for the purposes of this explanation.

Step 2 Continue the database backup with the Monitor node.a) Login as root user.b) cd /opt/infrac) From infra, execute the dbMonitoringBackupRestore.sh as shown below.

#. / dbMonitoringBackupRestore.sh backup

This script will make a backup with the necessary tables for the Monitor node.

d) After the script has executed, check in the /tmp folder for the output file, which will have a name something likemonitoring_database_backup.tar.gz.

Troubleshooting TipsThis ends the database backup procedures. Next, you restore the database on each Inventory and Monitor node. Thisinvolves copying each backed up file from the backup appliance to the appropriate node. After the files are placed onthe nodes, execute the script again. This time, however, you execute the script individually in each node (that is, in eachInventory and Monitor node).

Step 3 Start the database restore process by stopping services.a) Stop services on the Primary node.b) Stop services on the Service node(s). If there are multiple Services nodes, stop services on all of them.

Step 4 Restore the database on each Inventory and Monitor node.



a) Copy each backed up file from the backup appliance to the appropriate node.b) After the files are placed on the nodes, execute the script again. But this time, execute the script individually in each

node (in each Inventory and Monitor node).

Step 5 Start the first database Restore process with the Inventory Node.a) Login as root user.b) Copy the backup file, infra_database_backup.tar.gz, from the backup appliance to the /tmp folder in the Inventory

node appliance.c) In the Inventory node appliance, use cd /opt/infrad) From infra, execute the dbInfraBackupRestore.sh script as shown below.#. /dbInfraBackupRestore.sh restore

This script will restore the database from the backup file, infra_database_backup.tar.gz, placed in /tmp folder.

Step 6 Continue the database Restore process with the Monitor Node.a) Login as root user.b) Copy the backup file,monitoring_database_backup.tar.gz, from the backup appliance to the /tmp folder in the

Monitor node appliance.c) In the Monitor node appliance, use cd /opt/infrad) From infra, execute the dbMonitoringBackupRestore.sh script as shown below.#. / dbMonitoringBackupRestore.sh restore

This script will restore the database from the backup file,monitoring_database_backup.tar.gz, placed in /tmp folder.

What to Do Next

Test thoroughly.



configuring a multi-node setup - cisco.com · •threeservicenodes •oneinventorydatabase...

Documents