Cluster Server 7.4.2 Configuration and Upgrade Guide - Linux

Last Published:
Product(s): InfoScale & Storage Foundation (7.4.2)
Platform: Linux
  1. Section I. Configuring Cluster Server using the script-based installer
    1. I/O fencing requirements
      1. I/O fencing requirements
        1.  
          Coordinator disk requirements for I/O fencing
        2.  
          CP server requirements
        3.  
          Non-SCSI-3 I/O fencing requirements
    2. Preparing to configure VCS clusters for data integrity
      1. About planning to configure I/O fencing
        1.  
          Typical VCS cluster configuration with disk-based I/O fencing
        2.  
        3.  
          Recommended CP server configurations
      2. Setting up the CP server
        1.  
          Planning your CP server setup
        2.  
          Installing the CP server using the installer
        3.  
          Setting up shared storage for the CP server database
        4.  
          Configuring the CP server using the installer program
        5. Configuring the CP server manually
          1.  
            Configuring the CP server manually for HTTPS-based communication
          2.  
            Generating the key and certificates manually for the CP server
          3.  
            Completing the CP server configuration
        6. Configuring CP server using response files
          1.  
            Response file variables to configure CP server
          2.  
            Sample response file for configuring the CP server on single node VCS cluster
          3.  
            Sample response file for configuring the CP server on SFHA cluster
        7.  
          Verifying the CP server configuration
    3. Configuring VCS
      1.  
        Overview of tasks to configure VCS using the product installer
      2.  
        Starting the software configuration
      3.  
        Specifying systems for configuration
      4.  
        Configuring the cluster name
      5.  
        Configuring private heartbeat links
      6.  
        Configuring the virtual IP of the cluster
      7.  
        Configuring VCS in secure mode
      8.  
        Setting up trust relationships for your VCS cluster
      9. Configuring a secure cluster node by node
        1.  
          Configuring the first node
        2.  
          Configuring the remaining nodes
        3.  
          Completing the secure cluster configuration
      10.  
        Adding VCS users
      11.  
        Configuring SMTP email notification
      12.  
        Configuring SNMP trap notification
      13.  
        Configuring global clusters
      14. Completing the VCS configuration
        1.  
          Verifying the NIC configuration
      15.  
        About Veritas License Audit Tool
      16. Verifying and updating licenses on the system
        1.  
          Checking licensing information on the system
        2.  
          Replacing a VCS keyless license with another keyless license
        3.  
          Replacing a VCS keyless license with a permanent license
    4. Configuring VCS clusters for data integrity
      1. Setting up disk-based I/O fencing using installer
        1.  
          Initializing disks as VxVM disks
        2.  
          Configuring disk-based I/O fencing using installer
        3.  
          Refreshing keys or registrations on the existing coordination points for disk-based fencing using the installer
        4. Checking shared disks for I/O fencing
          1.  
            Verifying Array Support Library (ASL)
          2.  
            Verifying that the nodes have access to the same disk
          3.  
            Testing the disks using vxfentsthdw utility
      2. Setting up server-based I/O fencing using installer
        1.  
          Refreshing keys or registrations on the existing coordination points for server-based fencing using the installer
        2. Setting the order of existing coordination points for server-based fencing using the installer
          1.  
            About deciding the order of existing coordination points
          2.  
            Setting the order of existing coordination points using the installer
      3.  
        Setting up non-SCSI-3 I/O fencing in virtual environments using installer
      4.  
        Setting up majority-based I/O fencing using installer
      5.  
        Enabling or disabling the preferred fencing policy
  2. Section II. Automated configuration using response files
    1. Performing an automated VCS configuration
      1.  
        Configuring VCS using response files
      2.  
        Response file variables to configure VCS
      3.  
        Sample response file for configuring Cluster Server
    2. Performing an automated I/O fencing configuration using response files
      1.  
        Configuring I/O fencing using response files
      2.  
        Response file variables to configure disk-based I/O fencing
      3.  
        Sample response file for configuring disk-based I/O fencing
      4.  
        Response file variables to configure server-based I/O fencing
      5.  
        Sample response file for configuring server-based I/O fencing
      6.  
        Response file variables to configure non-SCSI-3 I/O fencing
      7.  
        Sample response file for configuring non-SCSI-3 I/O fencing
      8.  
        Response file variables to configure majority-based I/O fencing
      9.  
        Sample response file for configuring majority-based I/O fencing
  3. Section III. Manual configuration
    1. Manually configuring VCS
      1.  
        About configuring VCS manually
      2. Configuring LLT manually
        1.  
          Setting up /etc/llthosts for a manual installation
        2.  
          Setting up /etc/llttab for a manual installation
        3.  
          About LLT directives in /etc/llttab file
        4.  
          Additional considerations for LLT for a manual installation
      3.  
        Configuring GAB manually
      4. Configuring VCS manually
        1.  
          Configuring the cluster UUID when creating a cluster manually
      5. Configuring VCS in single node mode
        1. Disabling LLT, GAB, and I/O fencing on a single node cluster
          1.  
            Disabling LLT, GAB, and I/O fencing
        2.  
          Enabling LLT, GAB, and I/O fencing
      6.  
        Starting LLT, GAB, and VCS after manual configuration
      7.  
        About configuring cluster using VCS Cluster Configuration wizard
      8.  
        Before configuring a VCS cluster using the VCS Cluster Configuration wizard
      9.  
        Launching the VCS Cluster Configuration wizard
      10.  
        Configuring a cluster by using the VCS cluster configuration wizard
      11.  
        Adding a system to a VCS cluster
      12. Modifying the VCS configuration
        1.  
          Configuring the ClusterService group
    2. Manually configuring the clusters for data integrity
      1. Setting up disk-based I/O fencing manually
        1.  
          Identifying disks to use as coordinator disks
        2.  
          Setting up coordinator disk groups
        3.  
          Creating I/O fencing configuration files
        4.  
          Modifying VCS configuration to use I/O fencing
        5.  
          Verifying I/O fencing configuration
      2. Setting up server-based I/O fencing manually
        1.  
          Preparing the CP servers manually for use by the VCS cluster
        2.  
          Generating the client key and certificates manually on the client nodes
        3. Configuring server-based fencing on the VCS cluster manually
          1.  
            Sample vxfenmode file output for server-based fencing
        4.  
          Configuring CoordPoint agent to monitor coordination points
        5.  
          Verifying server-based I/O fencing configuration
      3. Setting up non-SCSI-3 fencing in virtual environments manually
        1.  
          Sample /etc/vxfenmode file for non-SCSI-3 fencing
      4. Setting up majority-based I/O fencing manually
        1.  
          Creating I/O fencing configuration files
        2.  
          Modifying VCS configuration to use I/O fencing
        3.  
          Verifying I/O fencing configuration
        4.  
          Sample /etc/vxfenmode file for majority-based fencing
  4. Section IV. Upgrading VCS
    1. Planning to upgrade VCS
      1.  
        About upgrading to VCS 7.4.2
      2.  
        Upgrading VCS in secure enterprise environments
      3.  
        Supported upgrade paths
      4.  
        Considerations for upgrading secure VCS 6.x clusters to VCS 7.4.2
      5.  
        Considerations for upgrading VCS to 7.4.2 on systems configured with an Oracle resource
      6.  
        Considerations for upgrading CP servers
      7.  
        Considerations for upgrading CP clients
      8.  
        Using Install Bundles to simultaneously install or upgrade full releases (base, maintenance, rolling patch), and individual patches
    2. Performing a VCS upgrade using the installer
      1.  
        Before upgrading VCS using the script-based installer
      2.  
        Upgrading VCS using the product installer
      3.  
        Upgrading to 2048 bit key and SHA256 signature certificates
      4. Tasks to perform after upgrading to 2048 bit key and SHA256 signature certificates
        1.  
          Deleting certificates of non-root users after upgrading to 2048 bit key and SHA256 signature certificates
        2.  
          Re-establishing WAC communication in global clusters after upgrading to 2048 bit key and SHA256 signature certificates
        3.  
          Re-establishing CP server and CP client communication after upgrading to 2048 bit key and SHA256 signature certificates
        4.  
          Re-establishing trust with Steward after upgrading to 2048 bit key and SHA256 signature certificates
      5.  
        Upgrading Steward to 2048 bit key and SHA256 signature certificates
    3. Performing an online upgrade
      1.  
        Limitations of online upgrade
      2.  
        Upgrading VCS online using the installer
    4. Performing a phased upgrade of VCS
      1. About phased upgrade
        1.  
          Prerequisites for a phased upgrade
        2.  
          Planning for a phased upgrade
        3.  
          Phased upgrade limitations
        4.  
          Phased upgrade example
        5.  
          Phased upgrade example overview
      2. Performing a phased upgrade using the product installer
        1.  
          Moving the service groups to the second subcluster
        2.  
          Upgrading the operating system on the first subcluster
        3.  
          Upgrading the first subcluster
        4.  
          Preparing the second subcluster
        5.  
          Activating the first subcluster
        6.  
          Upgrading the operating system on the second subcluster
        7.  
          Upgrading the second subcluster
        8.  
          Finishing the phased upgrade
    5. Performing an automated VCS upgrade using response files
      1.  
        Upgrading VCS using response files
      2.  
        Response file variables to upgrade VCS
      3.  
        Sample response file for full upgrade of VCS
  5. Section V. Adding and removing cluster nodes
    1. Adding a node to a single-node cluster
      1. Adding a node to a single-node cluster
        1. Setting up a node to join the single-node cluster
          1.  
            Installing VxVM or VxFS if necessary
        2.  
          Installing and configuring Ethernet cards for private network
        3.  
          Configuring the shared storage
        4.  
          Bringing up the existing node
        5.  
          Creating configuration files
        6.  
          Starting LLT and GAB
        7.  
          Reconfiguring VCS on the existing node
        8.  
          Verifying configuration on both nodes
    2. Adding a node to a multi-node VCS cluster
      1.  
        Adding nodes using the VCS installer
      2. Manually adding a node to a cluster
        1.  
          Setting up the hardware
        2.  
          Installing the VCS software manually when adding a node
        3. Setting up the node to run in secure mode
          1.  
            Configuring the authentication broker on node sys5
        4.  
          Configuring LLT and GAB when adding a node to the cluster
        5. Configuring I/O fencing on the new node
          1.  
            Preparing to configure I/O fencing on the new node
          2. Configuring server-based fencing on the new node
            1.  
              Adding the new node to the vxfen service group
          3.  
            Starting I/O fencing on the new node
        6.  
          Adding the node to the existing cluster
        7.  
          Starting VCS and verifying the cluster
        8. Adding a node using response files
          1.  
            Response file variables to add a node to a VCS cluster
          2.  
            Sample response file for adding a node to a VCS cluster
    3. Removing a node from a VCS cluster
      1. Removing a node from a VCS cluster
        1.  
          Verifying the status of nodes and service groups
        2.  
          Deleting the departing node from VCS configuration
        3.  
          Modifying configuration files on each remaining node
        4.  
          Removing the node configuration from the CP server
        5.  
          Removing security credentials from the leaving node
        6.  
          Unloading LLT and GAB and removing Veritas InfoScale Availability or Enterprise on the departing node
  6. Section VI. Installation reference
    1. Appendix A. Services and ports
      1.  
        About InfoScale Enterprise services and ports
    2. Appendix B. Configuration files
      1.  
        About the LLT and GAB configuration files
      2.  
        About the AMF configuration files
      3. About the VCS configuration files
        1.  
          Sample main.cf file for VCS clusters
        2.  
          Sample main.cf file for global clusters
      4.  
        About I/O fencing configuration files
      5. Sample configuration files for CP server
        1.  
          Sample main.cf file for CP server hosted on a single node that runs VCS
        2.  
          Sample main.cf file for CP server hosted on a two-node SFHA cluster
        3.  
          Sample CP server configuration (/etc/vxcps.conf) file output
      6. Tuning LLT variables for FSS environments
        1.  
          Tuning LLT variables for RDMA links
        2.  
          Tuning LLT variables for Ethernet links
        3. Configuring parallel ports for I/O shipping
          1.  
            To tune the LLT variables for I/O shipping
    3. Appendix C. Configuring LLT over UDP
      1. Using the UDP layer for LLT
        1.  
          When to use LLT over UDP
      2. Manually configuring LLT over UDP using IPv4
        1.  
          Broadcast address in the /etc/llttab file
        2.  
          The link command in the /etc/llttab file
        3.  
          The set-addr command in the /etc/llttab file
        4.  
          Selecting UDP ports
        5.  
          Configuring the netmask for LLT
        6.  
          Configuring the broadcast address for LLT
        7.  
          Sample configuration: direct-attached links
        8.  
          Sample configuration: links crossing IP routers
      3. Manually configuring LLT over UDP using IPv6
        1.  
          The link command in the /etc/llttab file
        2.  
          The set-addr command in the /etc/llttab file
        3.  
          Selecting UDP ports
        4.  
          Sample configuration: direct-attached links
        5.  
          Sample configuration: links crossing IP routers
      4.  
        LLT over UDP sample /etc/llttab
      5. About configuring LLT over UDP multiport
        1.  
          Manually configuring LLT over UDP multiport
        2.  
          Enabling LLT ports in firewall
        3.  
          Disabling the UDP multiport feature
    4. Appendix D. Configuring LLT over TCP
      1.  
        Using the TCP layer for LLT
      2. Manually configuring LLT over TCP using IPv4
        1.  
          Broadcast address in the /etc/llttab file
        2.  
          The link command in the /etc/llttab file
        3.  
          The set-addr command in the /etc/llttab file
        4.  
          Selecting TCP ports
        5.  
          Configuring the netmask for LLT
        6.  
          Configuring the broadcast address for LLT
        7.  
          Sample configuration: direct-attached link
        8.  
          Sample configuration: link crossing IP routers
      3. Manually configuring LLT over TCP using IPv6
        1.  
          The link command in the /etc/llttab file
        2.  
          The set-addr command in the /etc/llttab file
        3.  
          Selecting TCP ports
        4.  
          Sample configuration: direct-attached links
        5.  
          Sample configuration: links crossing IP routers
      4.  
        LLT over TCP sample /etc/llttab
    5. Appendix E. Migrating LLT links from IPv4 to IPv6 or dual-stack
      1.  
        About migrating the LLT links from IPv4 to IPv6 or to a dual-stack network
      2.  
        Review the current configuration
      3.  
        Meet the prerequisites for migration
      4.  
        Adding a new node to an existing cluster
      5.  
        Migrating LLT links to IPv6 or dual-stack when LLT is configured over UDP using IPv4
      6.  
        Migrating LLT links to IPv6 or dual-stack when LLT is configured over TCP using IPv4
    6. Appendix F. Using LLT over RDMA
      1.  
        Using LLT over RDMA
      2.  
        About RDMA over RoCE or InfiniBand networks in a clustering environment
      3.  
        How LLT supports RDMA capability for faster interconnects between applications
      4.  
        Using LLT over RDMA: supported use cases
      5. Configuring LLT over RDMA
        1.  
          Choosing supported hardware for LLT over RDMA
        2.  
          Installing RDMA, InfiniBand or Ethernet drivers and utilities
        3. Configuring RDMA over an Ethernet network
          1.  
            Enable RDMA over Converged Ethernet (RoCE)
          2.  
            Configuring RDMA and Ethernet drivers
          3.  
            Configuring IP addresses over Ethernet Interfaces
        4. Configuring RDMA over an InfiniBand network
          1.  
            Configuring RDMA and InfiniBand drivers
          2.  
            Configuring the OpenSM service
          3.  
            Configuring IP addresses over InfiniBand Interfaces
        5. Tuning system performance
          1.  
            Tuning the CPU frequency
          2.  
            Tuning the boot parameter settings
        6. Manually configuring LLT over RDMA
          1.  
            Broadcast address in the /etc/llttab file
          2.  
            The link command in the /etc/llttab file
          3.  
            Selecting UDP ports
          4.  
            Configuring the netmask for LLT
          5.  
            Sample configuration: direct-attached links
        7.  
          LLT over RDMA sample /etc/llttab
        8.  
          Verifying LLT configuration
      6. Troubleshooting LLT over RDMA
        1.  
          IP addresses associated to the RDMA NICs do not automatically plumb on node restart
        2.  
          Ping test fails for the IP addresses configured over InfiniBand interfaces
        3.  
          After a node restart, by default the Mellanox card with Virtual Protocol Interconnect (VPI) gets configured in InfiniBand mode
        4.  
          The LLT module fails to start
    7. Appendix G. Configuring the secure shell or the remote shell for communications
      1.  
        About configuring secure shell or remote shell communication modes before installing products
      2.  
        Manually configuring passwordless ssh
      3.  
        Setting up ssh and rsh connection using the installer -comsetup command
      4.  
        Setting up ssh and rsh connection using the pwdutil.pl utility
      5.  
        Restarting the ssh session
      6.  
        Enabling rsh for Linux
    8. Appendix H. Installation script options
      1.  
        Installation script options
    9. Appendix I. Troubleshooting VCS configuration
      1.  
        Restarting the installer after a failed network connection
      2.  
        Cannot launch the cluster view link
      3.  
        Starting and stopping processes for the Veritas InfoScale products
      4.  
        Installer cannot create UUID for the cluster
      5.  
        LLT startup script displays errors
      6.  
        The vxfentsthdw utility fails when SCSI TEST UNIT READY command fails
      7.  
        The GAB program reports incorrect membership results with existing iptable rules
      8.  
        Issues during fencing startup on VCS cluster nodes set up for server-based fencing
    10. Appendix J. Sample VCS cluster setup diagrams for CP server-based I/O fencing
      1. Configuration diagrams for setting up server-based I/O fencing
        1.  
          Two unique client clusters served by 3 CP servers
        2.  
          Client cluster served by highly available CPS and 2 SCSI-3 disks
        3.  
          Two node campus cluster served by remote CP server and 2 SCSI-3 disks
        4.  
          Multiple client clusters served by highly available CP server and 2 SCSI-3 disks
    11. Appendix K. Upgrading the Steward process
      1.  
        Upgrading the Steward process

How LLT supports RDMA capability for faster interconnects between applications

LLT and GAB support fast interconnect between applications using RDMA technology over InfiniBand and Ethernet media (RoCE). To leverage the RDMA capabilities of the hardware and also support the existing LLT functionalities, LLT maintains two channels (RDMA and non-RDMA) for each of the configured RDMA links. Both RDMA and non-RDMA channels are capable of transferring data between the nodes and LLT provides separate APIs to their clients, such as, CFS, CVM, to use these channels. The RDMA channel provides faster data transfer by leveraging the RDMA capabilities of the hardware. The RDMA channel is mainly used for data-transfer when the client is capable to use this channel. The non-RDMA channel is created over the UDP layer and LLT uses this channel mainly for sending and receiving heartbeats. Based on the health of the non-RDMA channel, GAB decides cluster membership for the cluster. The connection management of the RDMA channel is separate from the non-RDMA channel, but the connect and disconnect operations for the RDMA channel are triggered based on the status of the non-RDMA channel

If the non-RDMA channel is up but due to some issues in RDMA layer the RDMA channel is down, in such cases the data-transfer happens over the non-RDMA channel with a lesser performance until the RDMA channel is fixed. The system logs displays the message when the RDMA channel is up or down.

LLT uses the Open Fabrics Enterprise Distribution (OFED) layer and the drivers installed by the operating system to communicate with the hardware. LLT over RDMA allows applications running on one node to directly access the memory of an application running on another node that are connected over an RDMA-enabled network. In contrast, on nodes connected over a non-RDMA network, applications cannot directly read or write to an application running on another node. LLT clients such as, CFS and CVM, have to create intermediate copies of data before completing the read or write operation on the application, which increases the latency period and affects performance in some cases.

LLT over an RDMA network enables applications to read or write to applications on another node over the network without the need to create intermediate copies. This leads to low latency, higher throughput, and minimized CPU host usage thus improving application performance. Cluster volume manager and Cluster File Systems, which are clients of LLT and GAB, can use LLT over RDMA capability for specific use cases.

More Information

Using LLT over RDMA: supported use cases