To deploy the COS 3.5.2 software, perform the following tasks in order:
2. Setting Up the External DNS Server
4. COS Installation on the C3x60
5. COS Installation on the CDE465
6. Viewing NTP Server Information
7. Viewing DNS Server Information
8. Configuring COS from the GUI
9. Automated COS Node Configuration (Optional)
10. Configuring Cassandra Database Maintenance
Note COS Release 3.5.2 and later support remote network installation of the COS client using the Intel Preboot Execution Environment (PXE) in combination with the Red Hat Enterprises network installation feature using NFS, FTP, or HTTP, PXE. For details, see PXE Network Installation.
Before deploying the COS, complete the following preliminary tasks:
Step 1 Prepare your COS VMware datacenter topologies and networks. COS requires the following networks:
Step 2 Download the COS Service Manager Open Virtual Appliance (OVA) file and load the image into a repository (HTTP server) that is accessible by vCenter.
Step 3 Determine the blade and VM layout that you will use, and your management IP address allocation scheme for the network interfaces.
Step 4 Install the external DNS and NTP servers.
COS 3.5.2 software currently can be deployed on the following hardware:
Note You can convert a C3160 to a C3260 in the field. For details, see Migrating a Cisco UCS C3160 Server to a Cisco UCS C3260 Server in the Cisco UCS C3260 Rack Server Installation and Service Guide.
For information about installing the hardware, see the following:
Figure 2-1 provides a view of the network architecture of a COS cluster based on the UCS C3x60.
Figure 2-1 C3x60 COS Cluster Architecture
The administrator is responsible for creating the transaction signature (TSIG) key for the TSIG algorithm on the external DNS server. Valid TSIG algorithms are:
The following procedure shows how to create the TSIG key for the hmac-md5 algorithm.
Step 1 Generate the TSIG key by entering the following command:
dnssec-keygen -a HMAC-MD5 -b 128 -n HOST testdns.com.
This command creates a key file and a private key file.
Step 2 Create the key file (in this example, testdns.com.key) in the /etc/ directory.
Step 3 Add the key file path to the /etc/named.conf file by inserting the following line:
Step 4 Configure the DNS zones (used when deploying the PAM) on the external DNS server as shown in the following examples.
a. Update the /etc/named.conf file with the details of the DNS zone (in this example, testdns.com) and reverse zone, for all interfaces.
In the following sample information:
– The DNS zone is related to the 172.20.216.xx subnet, which can be the Management interface.
– For the Data In and Data Out interfaces, similar information related to the reverse zone must be added.
– Data In is related to the 15.1.1.x subnet.
– Data Out is related to the 25.1.1.x subnet.
– testdns.com. is the key name.
Reverse Zone Details (Management Interface)
Reverse Zone Details (Data In Interface)
Reverse Zone Details (Data Out Interface)
b. Create db.* files in the /var/named/slaves directory in the external DNS server for the DNS zone and reverse zone, for all interfaces.
Sample db File for Reverse Zone of Management Interface (db.216.20.172)
Sample db File for Reverse Zone for Data In Interface (db.1.1.15)
Sample db File for Reverse Zone for Data Out Interface (db.1.1.15)
Step 5 Make sure that the PAM and the external DNS server are in sync for the time and date.
To deploy the PAM from the vCenter:
Step 2 Click File > Deploy OVF Template. The Source dialog appears.
Step 3 Select the location of the OVF template (for example, mendocino-shenandoah-301.16144) and click Next. The OVA Template Details dialog appears.
Step 4 Verify the OVF template details and click Next. The End User License Agreement dialog appears.
Step 5 Accept the End User License Agreement and click Next. The Name and Location dialog appears.
Step 6 Specify a name and location for the deployed template and click Next. The Deployment Configuration dialog appears.
Step 7 Select the MOS PAM deployment configuration and click Next. The Storage dialog appears.
Step 8 Select a destination storage file for the VM files and click Next. The Disk Format dialog appears.
Step 9 Confirm that Thin Provision disk format for virtual disk storage is selected and click Next. The Network Mapping dialog appears.
Step 10 For each Source Network listed, choose the appropriate Destination Network and click Next. The Properties dialog appears.
Step 11 Enter or choose the following information to customize the software solution for the deployment:
Step 12 Click Next, and then click Finish to deploy the OVA and create the PAM VM. The progress bar shows the status of the deployment.
Step 13 Verify connectivity and open an SSH session into the PAM VM.
When installed on a C3160, the previous (3.5.1) COS release created a crash partition on one of the SSDs at the rear of the chassis. With COS Release 3.5.2, the location of the crash partition depends on the node hardware, as follows:
These locations assume a fresh installation and not an upgrade (not supported in COS 3.5.2 in any case).
Note COS 3.5.2 is fully qualified for the UCS C3260 dual-node configuration. Qualification and production support for C3260 single-node 28- and 56-disk configurations is planned for a future COS release. The single-node options described below, while logically valid, are not currently supported.
The C3260 provides four 40 Gbps ports for video and management traffic. To separate the management traffic from the content data, each port on the system I/O controller (SIOC) is divided into two four virtual NICs (vNICs) and assigned to one of two user-defined virtual LANs (vLANs). The first vNIC on each port is limited to 1 Gbps and is assigned to the management vLAN. The remaining vNICs are assigned to the data vLAN.
Each vNIC appears in Linux as a physical network interface. A pre-installation script, described below, is provided to establish this vNIC and vLAN configuration.
Note You must configure the first downstream switch(es) to receive the traffic from the two defined vLAN IDs, and route the traffic through the switch appropriately.
Before installing COS on a Cisco UCS C3260 Rack Server, you must configure the C3260 for either single-node service (not fully supported in COS 3.5.2) or dual-node service, and configure the data and management vLAN IDs for one or both nodes, as needed. For a C3260 in dual-node configuration, you will then install COS on each of the two server nodes separately.
A Bash script named preinst_setup_UCSC-C3260.sh is provided to help configure the C3260. This script is located at the root of the full COS product ISO image, which you can download from the COS product support page.
Note You must execute the script from a remote Linux node with network access to the target C3260 CIMC. Execution occurs after the C3260 CIMC IP address is defined, but before COS is installed.
The help for this script appears as follows:
Note ● Option -c is the IP address defined for the CIMC.
Example - Dual-Node Configuration
This script removes all currently defined virtual RAID devices.
For C3260s in single-node configuration, all drives are assigned to the one node. For dual-node systems, the storage devices are divided equally between nodes. The drives in slots 1-28 are assigned to compute node 1, and the drives in slots 29-56 are assigned to compute node 2.
As described above, the vNICs are configured so that each 40 Gbps port on the SIOC has one 1 Gbps virtual management interface and three virtual data interfaces assigned to their provided vLAN.
Note In the CIMC, the vNICs are listed in their order of appearance in Linux. Thus, eth0 and eth1 are the management interfaces, and all other eth ports are the data interfaces.
Note You can convert a C3160 to a C3260 in the field. For details, see Migrating a Cisco UCS C3160 Server to a Cisco UCS C3260 Server in the Cisco UCS C3260 Rack Server Installation and Service Guide.
The C3160 provides a total of four 10 GbE ports for video data and management traffic. To separate the management traffic from content data, port 1 on each SIOC is divided into two vNICs, with the management vNIC limited at 1 Gbps and the data vNIC limited at 9 Gbps. The second port on each SIOC is configured as a 10 Gbps data vNIC. The management vNICs are assigned to a user-defined management vLAN ID. The remaining data vNICs are assigned to a user-defined vLAN ID.
Each vNIC appears in Linux as a physical network interface. A pre-installation script, described below, is provided to establish this vNIC and vLAN configuration.
Note You must configure the first downstream switch to receive the traffic from the two vLAN IDs and route the traffic through the switch appropriately.
A Bash script named preinst_setup_UCSC-C3160.sh is provided to help configure the C3160. This script is located at the root of the full COS product ISO image, which you can download from the COS product support page.
Note You must execute the script from a remote Linux node with network access to the target C3160 CIMC. Execution occurs after the C3160 CIMC IP address is defined, but before COS is installed.
The help for this script appears as follows:
When it runs, the script prompts for any parameters missing from the command line. Password values entered in response to such a prompt are not echoed to the screen.
Installing and configuring COS on the C3x60 involves the following procedures:
1. Configure the PAM using the COS Configuration Wizard
2. Prepare the C3x60 for COS installation
3. Install the COS ISO image on the C3x60
4. Configure the COS node network on the C3x60
5. Register COS nodes to the PAM
Step 1 Log into the COS Service Manager GUI as described in Using the COS Service Manager GUI.
Step 2 Follow the steps described in Using the COS Configuration Wizard to configure the PAM.
Note For this procedure, you must include the final step, . This step creates URLs that you use later to add COS nodes to the PAM.
Step 1 For the C3160 only, remove any hard drives from the rear horizontal tray (up to four possible).
Step 2 Connect the CIMC Ethernet port on SIOC1 (the rightmost RJ45 port on each SIOC) to your management network.
Step 3 Connect the 40 Gbps (for C3260) or 10 Gbps (for C3160) cables to your downstream switch.
Note Your video network switch should already be configured to support the data and management vLAN IDs.
Step 4 Connect the power cords to the server.
Step 5 Temporarily attach a monitor and keyboard to the server.
Step 6 Apply power to the server, and immediately after the function key banner appears (early in the power-up sequence), press F8 to enter Cisco CIMC Configuration.
Step 7 Under NIC Properties, set the CIMC IP, mask, and gateway as appropriate for your installation.
Step 8 Press F10 to save your changes, then press F5 to refresh the screen and confirm that the IP address is set properly.
Step 9 Press ESC to exit and reboot (the system will not yet fully reboot).
Step 10 Configure the server as follows:
Step 1 Attach the COS installation ISO file using one of the following methods:
Note When using an external DVD drive, installation may fail with missing files reported. If this occurs, the likely cause is that the DVD drive is drawing excessive current from the USB port of the C3x60. To work around this issue, use a DVD drive that has its own AC power adapter.
– For C3260: Navigation Button (page upper left) > Compute: Server X > Remote Management > Virtual Media > Add New Mapping.
– For C3160: Server > Remote Presence > Virtual Media > Add New Mapping.
Note Experience has shown that using the HTTP protocol produces the best results.
After the virtual media mapping is defined, select the CIMC BIOS > Config Boot Order option, define a CIMC Mapped DVD subtype mapping, and place it in boot position 1.
Note Optionally, you can set the boot device by pressing F6 early in the C3x60 power-up sequence and selecting the desired boot device.
Step 2 Reboot the C3x60 and confirm that it boots from the COS ISO image.
Installation is automatic, and typically requires 10-15 minutes to complete from DVD (installation from virtual media is typically somewhat slower).
Note The default system baud rate is 9600. To define a different baud rate, enter auto ks_baud_rate=<rate> at the installation boot: prompt before the 10 second auto-installation timeout occurs.
Step 3 When prompted to reboot, disconnect the installation DVD drive or disable the virtual installation media, and then press Enter to reboot the C3x60.
To configure the COS network on the C3x60:
Step 1 When the C3x60 finishes rebooting, log in to the console using the following credentials:
The COS initialization (cosinit) script launches automatically to set up the management network that will be located on bond0.
Note If you exit the cosinit script now, you can relaunch it manually from the /opt/cisco/cos/config folder.
Step 2 When asked if you want to continue, enter y and configure the management interface as follows (IP addresses shown below are examples only):
Note If you exit the cosinit script now, you can relaunch it manually from the /opt/cisco/cos/config folder.
Step 3 Continue the script by entering the hostname of the node and the details of the default gateway as follows (IP addresses shown below are examples only):
Step 4 Ping the gateway to validate the configuration:
After installing a COS image and rebooting the node, you are prompted to start configuration by running the cosinit script. This script performs initial setup and adds the node to the COS management system.
The cosinit script can be run with the following options:
When the cosinit script executes, the COS node is registered with the COS management system. You can configure the node using the COS Service Manager web GUI, add it to a COS cluster, and make it operational. For more information, see COS Node Initial Configuration.
Step 1 Use the URLs generated in Configure the PAM Using the COS Configuration Wizard to run cosinit on each COS node, thus adding each node to the PAM.
/opt/cisco/cos/config/cosinit -nameserver 97.98.99.1 -input https://service-mgr.cisco.com:8043/v2/cosnodeinstances/c3260
Step 2 In the COS Service Manager GUI, navigate to PAM > Services > Cloud Object Store.
Step 3 Click An unused COS Service, and then click the Overview tab to view the service details.
Step 4 Verify that the COS node services are running.
To create users and accounts for a COS node, install the Swauth package on any Linux server and use the commands described below.
Note For additional details on the CLIs, see COS Command Line Utilities and the Cisco Cloud Object Storage Release 3.5.1 API Guide.
The cos-swift command line utility is based on the Swift API, and can be used to manage storage accounts, storage containers, and storage objects in a COS cluster.
cos-swift [–t <auth-token>] [–a <storage-url>] [–v/– –verbose][–h/– –help <subcommand>] [subcommand <options>]
The cos-swauth command line utility is based on the Swauth API, and can be used to manage authentication accounts and users in a COS cluster.
Note For information about installing the CDE465 platform, see the Cisco Content Delivery Engine 465 Hardware Installation Guide.
When installing the COS software on the CDE645 platform for the first time, recommended practice is to use the Supermicro Intelligent Platform Management Interface (IPMI) as described below.
Note This procedure requires that IPMI first be enabled on the CDE465. If necessary, enable IPMI on the CDE465 by enter the BIOS, selecting IPMI management, and then rebooting the CDE465.
Step 1 Log in to the CDE IPMI as follows:
a. Enter the IPMI management IP address in the browser.
b. Enter the default Username admin and Password admin.
Step 2 Launch the KVM console as follows:
a. Choose Remote Control > Console Redirection.
a. Select Virtual Media > Virtual Storage.
b. Insert the COS ISO file into Device1.
Step 4 Cycle power to the CDE465, enter the BIOS, set the machine to boot from Device1: USB Drive, and then reboot the CDE465.
After rebooting, the CDE465 starts COS installation automatically without manual intervention.
Step 5 When prompted at the end of installation, reboot the CDE465 and remove the ISO file from Device1.
After the CDE465 has rebooted, configure the COS network on the CDE465 as follows:
Step 1 Log in to the console using the following credentials:
The COS initialization script (cosinit) launches automatically to set up the management network that will be located on bond0.
Note If you exit the cosinit script now, you can relaunch it manually from the /opt/cisco/cos/config folder.
Step 2 When asked if you want to continue, enter y and configure the management interface as follows (IP addresses shown below are examples only):
Note If you exit the cosinit script now, you can relaunch it manually from the /opt/cisco/cos/config folder.
Step 3 Continue the script by entering the hostname of the node and the details of the default gateway as follows (IP addresses shown below are examples only):
Step 4 Ping the gateway to validate the configuration:
Step 5 Verify that the file /etc/cosd.conf, has been updated by the cos-aic-client to uncomment the db host field (about two pages down), and to specify the IP address of the local management interface as the DB Host.
Step 7 Configure IP Pools as described in Configuring IP Pools.
Step 1 Provided that cosinit has run in an earlier procedure, you can now execute cosinit to skip the network configuration and begin PAM configuration, as shown in the following example:
Step 2 Specify the configuration of your hardware platform as follows:
Note The choices cosinit offers here are only those that COS supports for the hardware platform cosinit has detected. In this example, cosinit detects CDE465 hardware and so lists the sole CDE465 configuration option. If cosinit detects C3x60 hardware, it lists different configuration options accordingly.
Note The IP pool name should be the same as the one created in the COS Service Manager GUI.
Step 3 Enter the appropriate IP address of the COS Service Manager PAM, as follows:
Note The docserver port is 5087.
Note Enter a cluster name at this point only if you plan to use automated node configuration.
Note For Enter expected Cluster Size above, enter the expected number of nodes in the cluster that this node will join. This value is used later to calculate resiliency values for distributed erasure coding (DEC).
Step 4 Reboot the node if prompted.
Note If you did not enter a cluster name, cosinit does not prompt you to reboot the CDE465 as shown in this example, as reboot is not needed.
When cosinit executes, the following services are started automatically:
If the cluster name is provided, the command service cserver start -C is executed automatically when cserver starts. In addition, execution of cosinit results in the following:
Step 5 Confirm that the COS AIC client service is started as follows:
The COS node is now registered with COS Service Manager.
Note NTP servers are configured when the PAM is deployed. You cannot add or delete NTP servers using this procedure.
To view NTP server information using the COS Service Manager GUI:
Step 1 Log in to the GUI as described in Using the COS Service Manager GUI.
Step 2 Go to Infrastructure > Platform Services.
In the NTP Servers table, the following information is displayed for each NTP server:
– The name of the NTP server is a string of up to 30 characters. Acceptable characters include uppercase and lowercase letters, numbers, periods (.), dashes (-), and underscores (_).
– The name must not begin with a period (.), and it is not case-sensitive.
Note DNS servers are configured when the PAM is deployed.
To view DNS server information using the COS Service Manager GUI:
Step 1 Log in to the GUI as described in Using the COS Service Manager GUI.
Step 2 Go to Infrastructure > Platform Services.
In the DNS Servers table, the following information is displayed for each DNS server:
– The name of the DNS server is a string of up to 35 characters. Acceptable characters include uppercase and lowercase letters, numbers, periods (.), dashes (-), and underscores (_).
– The name cannot begin with a period (.), and it is not case-sensitive.
To configure COS using the COS Service Manager GUI:
Step 1 Log in to the GUI as described in Using the COS Service Manager GUI.
Step 2 Go to Infrastructure > Regions & Zones.
A zone is a set of Cloud platform components (compute, network, storage, and security) that are fate-shared. The zone can be mapped to the underlying Cloud platform provider, such as a datacenter in vCenter, an Availability Zone, or any other combination of fate-shared Cloud resource topologies. Each zone is associated with one Cloud Controller.
A region is made up of one or more zones. It is an abstract representation of the underlying Cloud platform. A region can be associated with a geographical region, one or more datacenters, or a service area. Release 3.5.2 of the COS software supports only one region.
Step 3 Expand a region to display the zones associated with that region.
Step 4 Click Add Row to add zones to the region, as needed.
Step 5 Go to Infrastructure > IP Address Pools and Networks to define IP pools for the COS node.
Step 6 Go to Infrastructure > COS Cluster and create a COS cluster.
Step 7 Go to Services > Cloud Object Stores, click the COS service instance, and create one COS endpoint.
Step 8 On the COS node, execute /opt/cisco/cos/config/cosinit and configure the IP pool name, DocServer IP or domain name (depending on PAM configuration; see note below) and Port (which is the COS Service Manager management IP). This will register the COS node with the COS Service Manager.
Note For a single PAM, the DocServer IP address is the same as the COS Service Manager IP address, so you can enter it here directly. For HA PAMs, however, you must enter the DocServer domain name here (for example, mgmt-docserver.cos.cisco.com) and then enter the NameServer IP address, to help locate the mgmt-docserver leader.
Step 9 Go to Infrastructure > COS Nodes and configure Cluster Name by choosing the created cluster name from the drop-down list.
Step 10 Ensure that the Admin State for the node is set to Inservice. Click Save.
Step 11 Go to Services > Cloud Object Stores, select the COS service instance, and then click the Overview tab to confirm that you can see the registered COS Node.
To configure the COS service instance template using the COS Service Manager GUI:
Step 1 Log in to the GUI as described in Using the COS Service Manager GUI.
Step 2 Go to Services > Cloud Object Storage. The Service Summary page opens, displaying a list of COS service instances.
Step 3 Click the COS service instance to be configured. The Service Definition tab for that instance opens, containing the General and Service Endpoints areas.
Step 4 The fields of the General area are listed below. Edit the field values as appropriate.
Step 5 There is no default COS endpoint when the PAM is brought up, so you must create one manually.
The following parameters are displayed for the service endpoint. Edit the fields as appropriate:
Step 6 Click Save to save your changes, or click Cancel to discard them and start over.
Before installing COS nodes, you must create a network and configure IP pools using the COS Service Manager GUI.
Step 1 Log in to the GUI as described in Using the COS Service Manager GUI.
Step 2 Go to Infrastructure > IP Address Pools and Networks and create a suitable network for the IP pools.
Step 3 To add a new IP pool, click IP Address Pools > Add Row.
Step 4 In the Name field, enter a unique name for the IP pool. In the Description field, enter a brief description for the IP pool.
Step 5 To associate an IP address range with the IP pool, click IP Address Ranges > Add Row.
Step 6 Enter appropriate values in the fields described below:
|
|
---|---|
Note If you try to initialize a COS node with an IP pool having fewer IP addresses available than are needed, the initialization fails and an Event is generated. For details, see COS AIC Server Events.
Step 7 Click Save to save your changes, or click Cancel to discard them and start over.
Note After you associate an IP pool with a COS node, do not edit or delete the pool before first dissociating the pool from the node.
To edit or delete configured IP pools using the COS Service Manager GUI:
Step 1 Log in to the GUI as described in Using the COS Service Manager GUI.
Step 2 Go to Infrastructure > IP Pools and Networks.
Step 3 Before editing or deleting an IP pool, ensure that it is not serving any COS nodes as follows:
a. Dissociate any COS node(s) associated with the IP pool to be edited or deleted.
b. Link the dissociated node(s) to another IP pool.
Step 4 Check the box against the name of the IP pool to be edited or deleted.
Step 5 Click Edit or Delete as appropriate.
When COS is installed, it creates a single default node cluster. You can configure the cluster using the COS Service Manager GUI.
To view or edit the settings of a COS node cluster from the COS Service Manager GUI:
Step 1 Log in to the GUI as described in Using the COS Service Manager GUI.
Step 2 Go to Infrastructure > COS Clusters.
The default COS node cluster is displayed.
Step 3 Edit the fields described below based on your deployment.
|
|
---|---|
Step 4 Click Save to save your changes, or click Cancel to discard them and start over.
A COS node may ship with a preinstalled image that can be upgraded when provisioning the node on a network. After installing the image and rebooting the node, you are prompted to start configuration by running the cosinit script.
run /opt/cisco/cos/config/cosinit
When it runs, the cosinit script performs the initial setup and adds the node to the COS management system.
The cosinit script can be run with the options shown below:
Running cosinit with the -skipnw option skips the management network configuration and registers the COS node to the COS Service Manager. You must manually input the IP pool name and select the C/F interfaces that use this IP pool.
When cosinit prompts for a cluster name, you can either skip this step or enter a cluster name that has been created on the PAM. If you enter a cluster name, cosinit helps you to enable all the C/F interfaces that are added to the IP pool, and to update COS node to cluster and inservice.
Running cosinit with the -input option enables cosinit to run with a config file. All parameters are read from the config file.
Executing the cosinit script completes the following tasks:
If the install_baud_rate file is not found under the /root directory, the operator is prompted to enter a baud rate during cosinit execution or specify the value in the configFile.
When the cosinit script is successfully executed:
The COS node is now registered with the COS management system. You can configure the node using the COS Service Manager GUI, add it to a COS cluster, and make it operational.
After installing and initializing a COS node, you must add the node to a COS cluster to begin servicing COS requests.
To add a node to the COS cluster using the COS Service Manager web GUI:
Step 1 Log in to the GUI as described in Using the COS Service Manager GUI.
Step 2 Go to Infrastructure > COS Nodes.
Step 3 Check the box beside the name of the node you want to edit and click Edit.
Step 4 In the Service Interfaces list, check the box beside the interfaces to add to the IP pool, and then enable them.
Step 5 From the Cluster Name list, choose the created COS cluster.
Step 6 Click Save to save your changes, or click Cancel to discard them and start over.
When a COS node is added to a COS cluster, the following files are written on to the node:
When the node Admin State is set to Inservice, all enabled service interfaces of that node are written to the DNS (internal or external).
Note To avoid possible bootstrapping issues with the Cassandra service, be sure to provide (or use a script to provide) a delay of at least two minutes between adding two nodes in sequence to the cluster.
If the replication factor is too high, use the following script on any one node in the cluster to adjust it:
sh /opt/cisco/cos-aic-client/cassandra/cassandra-adjust-replication.sh { 1 | 2 | 3 }
For the final argument, use 1 for one node, 2 for two nodes, and 3 for three or more nodes.
After installing and initializing a COS node, you can modify its configuration parameters using the COS Service Manager GUI as follows:
Step 1 Log in to the GUI as described in Using the COS Service Manager GUI.
Step 2 Go to Infrastructure > COS Nodes.
Step 3 Check the box against the name of the node you want to edit and click Edit.
Step 4 Edit the fields described below based on your requirements.
|
|
---|---|
Indicates whether the node is Inservice or under Maintenance |
Note The Admin State of a COS node must be set to Maintenance before removing the node from a COS cluster.
Step 5 To modify the service interfaces of the node, follow the procedure Configuring the Service Interface of a COS Node.
Step 6 Click Save to save your changes, or click Cancel to discard them and start over.
To modify the settings of a service interface of a functioning COS node using the COS Service Manager GUI:
Step 1 Log in to the GUI as described in Using the COS Service Manager GUI.
Step 2 Go to Infrastructure > COS Nodes.
Step 3 Check the box against the name of the node you want to edit and click Edit.
Step 4 Check the box against the name of service interface you wish to edit and click Edit.
Step 5 Edit the fields described below based on your requirements.
Step 6 Click Save to save your changes, or click Cancel to discard them and start over.
To use automated configuration, you provide a file to cosinit that includes a ClusterName and IP Pool reference for at least one service interface. This enables the system to configure the COS node without manual intervention through either the COS Service Manager GUI or the API.
Run cosinit with the input file:
/opt/cisco/cos/config/cosinit -input cosinit_auto
This results in one of the following:
The AIC Server handles the rest of the configuration automatically by assigning an IP address from the specified IP Pool and adding the node to the cluster. The AIC Client then writes the necessary configuration files to the COS node as usual. When configuration is complete, the AIC Client automatically starts the cassandra, cosd, and cserver cos node services.
You can verify the input by viewing the /tmp/.cosnodeinit file:
Note When using automated configuration of multiple COS nodes, configure the first node, and then wait before configuring the second node until the Cassandra database service to appear as Running in the GUI of the first node configured. Otherwise, there may be unexpected behavior in the seed list configuration for the Cassandra database of the nodes added after the first node.
Step 1 Verify that cassandra is running:
Step 2 Verify that cosd is functional:
The Cassandra database requires periodic maintenance with an anti-entropy repair that must be manually configured to execute every two days on each COS node. For more information on the Cassandra repair process, see the Cassandra 2.1 documentation available at:
www.datastax.com/documentation/cassandra/2.1/
To execute periodic repair, we recommend configuring a CRON job on each node, and then scheduling the CRON jobs on the nodes to begin at different times so as to avoid the repair running concurrently on more than one node at a time. As a reference, we recommend providing 15 minutes between scheduling of the repair operation on each node.