Configuring policy-based file replication
Before you begin
- Managed Server: For a server to be considered as managed server, it needs to be entered in the replication process configuration.
- Not a Managed Server: A non-managed server is one where the IP Address and user name/password of the server is not known by the replication process. Administrators can still select a non-managed server as the target by specifying the IP address along with the user name and password
Procedure
Navigate to , and click add to display the Policy Destination Type page.
Select the policy destination type:
- Select Managed Server to create a policy to replicate to a server that is managed by the replication process.
- Select Not a Managed Server to create a policy to replicate to a server that is not managed by the replication process.
Click next to display a destination type-specific Add Policy page.
The Add Policy page for a managed server replication destination displays as:
The Add Policy page for an unmanaged server replication destination is similar, with only the fields in the Destination section being different:
NoteAdministrators should be authorized to use the external server to access and store replication data.Enter the requested information:
Field/Item Description Identification Name: Allows you to specify the name of the replication policy. The name may not contain spaces or any of the following characters: \/<>"'!@#$%^%&*(){}[] +=?:;,~`|.' Source Source of the replication. Set this field only if you want to make a simple copy of a specific snapshot. Do not set this field if you are intending to run incremental replications. The source is identified using the following fields: - Server: Name of the server/cluster that has the source file system for this replication policy.
- EVS/file system: Name of the EVS and file system to which the replication source is mapped. Click change to change the EVS or file system.
- Path: Select a virtual volume from the drop-down list. Or select Directory and enter the path.
- Snapshot: Select a snapshot to migrate a file system from a snapshot taken at a specific point in time. Using a snapshot as a source allows you to replicate the snapshot rather than the live file system, eliminating the possibility of file changes during the replication.
Destination (for managed replication destinations) Destination of the replication (managed server): - Server: Name of the server/cluster that hosts the destination file system for this replication policy.
- EVS/file system: Name of the virtual server and file system to which the replication is mapped. Click change to change the EVS/file system.
- Path: Specify the directory path. Note that you may not specify a virtual volume as a path.
- Current Syslock status: Indicates if the file system is in Syslocked mode. When System Lock is enabled for the destination file system, a warning icon is displayed. NDMP has full access to the file system and can write to the syslocked file system during a backup or replication, but the file system remains in read-only mode to clients using the file service protocols (NFS, SMB, FTP, and iSCSI).
If the destination file system is not in syslock mode during a replication operation, clients may write to the file system, creating inconsistencies between the source and target of the replication.
During transfer of primary access operations, both the source file system and the destination file system are put into System Lock mode.
To manually enable or disable the Syslock mode for a file system, you must navigate to the File System Details page for the file system. For more information on Syslocked mode, see the File Services Administration Guide.
Destination (for unmanaged replication destinations) Destination of the replication (non-managed server): - File Serving IP Address / Host Name: Name of the server containing the target EVS/ file system. Click change to change the destination to a different server.
- File System: Name of the file system to which the replication is mapped. Click change to change the file system.
- Path: Specify the directory path. Note that you may not specify a virtual volume as a path.
- NDMP User Name: Name of the NDMP user for which the replication target was created.
- NDMP User Password: Password for the selected NDMP user.
Processing Options - Source Snapshot Rule Name: The snapshot rule for replication of the source file system.
- Destination Snapshot Rule Name: The snapshot rule to use for the snapshot of the destination file system following a successful replication.
- Pre-/Post-Replication Script: A user-defined script to run before or after each replication. Scripts must be located in /opt/smu/adc_replic/final_scripts. The permissions of the scripts must be set to "executable".
File Replication Rule Optional configuration parameters that allow tuning of replications to enable and disable specific functions or to optimize performance. Verify your settings, click OK to save, or cancel to decline.
Specifying IP addresses for an NDMP replication policy manually
You can manually specify the IP addresses to be used by the source and destination servers for the control and data connections of a replication policy.
A file replication policy requires the Service EVS IPs to be able to route data to each other. To specify exactly which EVS IP addresses to use, you can create a per-policy address configuration file.
The configuration file must be placed on the SMU in the following location:
/opt/smu/adc_replic/conf/replication_policies
Each of the policies has its own sub-directory. The configuration text file must be located in the parent directory above and called <policy name>_addresses. The policy names must match exactly.
The file contents consist of IP addresses and port numbers as follows:
SRC_CTRL_ADDRESS=IP-Address DEST_CTRL_ADDRESS=IP-Address SRC_DATA_ADDRESS=IP-Address DEST_DATA_ADDRESS=IP-Address DEST_DATA_PORT=Port-Number
IPv4 addresses should be specified in conventional dot notation: X.X.X.X
IPv6 address should be specified in conventional notation, for example: ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff
You do not need to specify all values. However, if you specify one DATA address, you must also specify the other DATA address and they must have a common IP family type (both IPv4 or both IPv6).
Connection errors
When attempting to add a new replication policy, a connection error may be indicated by “Unable to connect to <IP address>“ or “Error accessing <source/destination> server".
The “Unable to connect to“ message means one of the following:
- The server is not currently powered up or is temporarily disconnected from the network. The server must be available and properly connected when creating a replication policy.
- The NDMP service may be disabled. The replication uses the NDMP service which must be enabled when adding or running replications. Please use the NDMP configuration page (or the ndmp-status command) to enable and start the NDMP service.
- The gigabit Ethernet port providing access to the EVS which hosts the file system is not accessible from the NAS Manager. This may be the case if the network is set up with private subnetworks as commonly used with VLANs. In this case, the server may have been configured so that NAS Manager access is through the management ports instead of the ports set using the ndmp-management-ports-set command.
The "Error accessing server" message may occur as a result of restricting NDMP access using the ndmp-option command. The allowip and blockip options can be set such that the NAS Manager is not allowed to access the NDMP services using the standard routes. If the NDMP connection restrictions are definitely required, change the configuration of the server to allow NAS Manager access by way of the management ports using the ndmp-management-ports-set command. The NAS Manager connections then bypass the allowip/blockip checks.
The NAS Manager replication and data migration features use the NDMP service on the NAS server. The NDMP service is usually accessed by way of the IP address of the EVS which hosts the file system, this access usually happens through a gigabit Ethernet port. In some cases, the IP address is within a private subnetwork and is not accessible from the NAS Manager. When this is the case, the ndmp-management-ports-set command can be used to request that the NAS Manager access goes through the management ports and is then relayed to the NDMP service.
The ndmp-management-ports-set command takes two parameters which are the TCP ports. One is used to accept the incoming connection on the management port and one to pass the requests to the NDMP code. These must be ports that are not in use by any other service. In particular, these ports must not be the standard NDMP service port. The port numbers 10001 and 10002 usually work and, being next to the standard NDMP port 10000, can be useful in identifying the port usage.
Having set up the NDMP management ports this way, all NAS Manager replication and data migration NDMP accesses will be routed by way of the management port. Note that the actual data transfer connections involved are between the NAS server EVSs and do not run over the management connections. In particular, a replication between two NAS servers passes the data over a TCP connection between EVS IP addresses through the gigabit Ethernet ports. Therefore, the two EVSs must have IP addresses that can communicate with each other.