Search <book_title>...

Storage Foundation and High Availability Solutions 7.3.1 HA and DR Solutions Guide for Microsoft SharePoint 2010 - Windows

Last Published: 2017-11-05

Product(s): InfoScale & Storage Foundation (7.3.1)

Notes and recommendations for cluster and application configuration

Review the Hardware compatibility list (HCL) and Software Compatibility List (SCL) at:
https://sort.veritas.com/documents
Note:
Solutions wizards cannot be used to perform Disaster Recovery, Fire Drill, or Quick Recovery remotely on Windows Server Core systems.
The DR, FD, and QR wizards require that the .NET Framework is present on the system where these operations are to be performed. As the .NET Framework is not supported on the Windows Server Core systems, the wizards cannot be used to perform DR, FD, or QR on these systems.
Refer to the following Microsoft knowledge database article for more details:
http://technet.microsoft.com/en-us/library/dd184075.aspx
Shared disks to support applications that migrate between nodes in the cluster. Campus clusters require more than one array for mirroring. Disaster recovery configurations require one array for each site. Replicated data clusters with no shared storage are also supported.
If your storage devices are SCSI-3 compliant, and you wish to use SCSI-3 Persistent Group Reservations (PGR), you must enable SCSI-3 support using the Veritas Enterprise Administrator (VEA).
See the Storage Foundation Administrator's Guide for more information.
SCSI, Fibre Channel, iSCSI host bus adapters (HBAs), or iSCSI Initiator supported NICs to access shared storage.
A minimum of two NICs is required. One NIC will be used exclusively for private network communication between the nodes of the cluster. The second NIC will be used for both private cluster communications and for public access to the cluster. Veritas recommends three NICs.
NIC teaming is not supported for the VCS private network.
Static IP addresses are required for certain purposes when configuring high availability or disaster recovery solutions. For IPv4 networks, ensure that you have the addresses available to enter. For IPv6 networks, ensure that the network advertises the prefix so that addresses are autogenerated.
Static IP addresses are required for the following purposes:
- A minimum of one static IP address for each physical node in the cluster.
- One static IP address per cluster used when configuring Notification or the Global Cluster Option. The same IP address may be used for all options.
- For Volume Replicator replication in a disaster recovery configuration, a minimum of one static IP address per site for each application instance running in the cluster.
- For Volume Replicator replication in a Replicated Data Cluster configuration, a minimum of one static IP address per zone for each application instance running in the cluster.
Configure name resolution for each node.
Verify the availability of DNS Services. AD-integrated DNS or BIND 8.2 or higher are supported.
Make sure a reverse lookup zone exists in the DNS. Refer to the application documentation for instructions on creating a reverse lookup zone.
DNS scavenging affects virtual servers configured in SFW HA because the Lanman agent uses Dynamic DNS (DDNS) to map virtual names with IP addresses. If you use scavenging, then you must set the DNSRefreshInterval attribute for the Lanman agent. This enables the Lanman agent to refresh the resource records on the DNS servers.
See the Cluster Server Bundled Agents Reference Guide.
In an IPv6 environment, the Lanman agent relies on the DNS records to validate the virtual server name on the network. If the virtual servers configured in the cluster use IPv6 addresses, you must specify the DNS server IP, either in the network adapter settings or in the Lanman agent's AdditionalDNSServers attribute.
If Network Basic Input/Output System (NetBIOS) is disabled over the TCP/IP, then you must set the Lanman agent's DNSUpdateRequired attribute to 1 (True).
You must have write permissions for the Active Directory objects corresponding to all the nodes.
If you plan to create a new user account for the VCS Helper service, you must have Domain Administrator privileges or belong to the Account Operators group. If you plan to use an existing user account context for the VCS Helper service, you must know the password for the user account.
If User Access Control (UAC) is enabled on Windows systems, then you cannot log on to VEA GUI with an account that is not a member of the Administrators group, such as a guest user. This happens because such user does not have the "Write" permission for the "Veritas" folder in the installation directory (typically, C:\Program Files\Veritas). As a workaround, an OS administrator user can set "Write" permission for the guest user using the Security tab of the "Veritas" folder's properties.
For a Replicated Data Cluster, install only in a single domain.
Route each private NIC through a separate hub or switch to avoid single points of failure.
NIC teaming is not supported for the VCS private network.
Verify that your DNS server is configured for secure dynamic updates. For the Forward and Reverse Lookup Zones, set the Dynamic updates option to "Secure only". (DNS > Zone Properties > General tab)
This is applicable for a Replicated Data Cluster configuration.
This is applicable for a Replicated Data Cluster configuration. You can configure single node clusters as the primary and secondary zones. However, if using a shared storage configuration, you must create the disk groups as clustered disk groups. If you cannot create a clustered disk group due to the unavailability of disks on a shared bus, use the vxclus UseSystemBus ON command.
To configure a RDC cluster, you need to create virtual IP addresses for the following:
- Application virtual server; this IP address should be the same on all nodes at the primary and secondary zones
- Replication IP address for the primary zone
- Replication IP address for the secondary zone
Before you start deploying your environment, you should have these IP addresses available.