The Replication feature lets you create a redundant failover device for one CFA (primary CFA) using another CFA (secondary CFA). There are two primary modes of replication, Standard replication, and Cloud replication.

When you enable replication, you dedicate the entire RAID on the secondary CFA to a copy of the data for the clients you have set to replicate.

In addition, CFA byte-level replication allows for only the changes to a file to be transferred. Byte-level replication will significantly reduce the time and bandwidth required for replication.

info A third subtab (Pending Jobs on the primary CFA, Replicated Jobs on the secondary CFA) will appear after replication has been configured on the CFA.

In order to use replication, you need:

• Two CFAs to configure as the primary and secondary CFAs. Both CFAs must be running the exact same version of the firmware that supports the Enhanced Replication or an account on the Infrascale Cloud Portal.

info It is recommended that with physical CFAs they both have the same size raids since if the RAID fills up on one CFA it will disable the replication service, and could cause the secondary to fall behind on replication.
• (Optional) A license obtained from Infrascale to enable the byte-level replication feature. Byte-Level Replication will further optimize the transfer of backup file data by computing deltas to send to the secondary CFA. Without Byte-Level Replication enabled, the data is simply copied in its entirety from one CFA to the other. This feature is enabled by default if you are using cloud replication, and does not need an additional license.

## Configuring replication

When initially configuring CFAs for replication, all of the jobs for clients set to replicate on the primary CFA will automatically be identified as replicable by default and will be copied to the Secondary. Depending on the number of jobs and the amount of data they contain, this could be a lengthy process. Therefore, it is highly recommended that both the Primary and secondary CFAs first be connected on a LAN or using a high-bandwidth connection.

After this initial job copy (seeding) is complete, the secondary CFA may be moved to its desired location. If a lower bandwidth connection is expected between the CFAs, Byte-level replication is recommended.

## Standard replication setup

This first section will detail setting up standard replication between two CFAs.

### Configuring the secondary CFA

Though it may seem out of order, the configuration process is easier when the secondary CFA is configured first, then the primary CFA. However, once you become accustomed to configuring CFAs for replication, you can configure them in either order successfully.

Log in to the secondary CFA Management Console, go to the Replication tab, and then click the Mode/Options subtab.

Select Secondary to switch modes, the following confirmation dialog will be displayed:

Click Yes to confirm the switch to the secondary replication, the following options will be displayed:

Both the primary and secondary CFA need to share a security key. You can generate a new key on the secondary CFA which will then be used when setting up the primary CFA. Select Generate new key to indicate that you would like to generate a new key.

info The new key will have to be manually transferred to the replication partner whenever it is changed.

Next, enter the network address of the CFA you will be using as the primary CFA into the Address field. You can use either a system name (which must be resolvable via DNS) or an IP address as shown in the example:

info You can also enter an optional description of the remote CFA you will be using as the primary CFA.

The value in the CFA Port field should be left at its default value (2300) unless you have a specific reason (firewalls, security, etc.) to change it. If you need to change it, click the down-arrow to expand the box labeled Advanced and access the port setting.

info If the value in the CFA Port field is changed, it will be necessary to change the primary’s port to the same value.

At this point you have entered enough information to make a connection to the primary CFA. Click Apply at the bottom of the panel.

While the primary CFA is being contacted, the UI panel will temporarily become disabled. This is normal and it may remain this way for up to a minute or more. Once the primary CFA has replied, the UI panel will again become active and the authentication key and remote (primary) CFA GUID fields will be populated as shown below.

info The actual text shown in the authentication key, address, and remote GUID fields will be unique to your CFA and will not match the example shown above.

The Configuration Status section at the bottom of the panel will keep you informed of any errors that may occur. During the initial setup process, it is normal for the first CFA being set up to receive message saying The remote CFA is not connected and will show an error symbol. This condition will be resolved once the primary CFA has been configured.

If you encounter any errors other than the Remote CFA is not connected error shown above (see Errors During Setup and Configuration for more information).

Now that the secondary CFA has the GUID of the primary CFA the replication service enabled.

Click Start Replication Service to enable replication.

Click Apply again.

This completes the basic configuration of the secondary CFA.

info Any error message generated when the first connection attempt was made may come back at this point.

### Configuring the primary CFA

Log into the primary CFA and go to Replication > Mode/Options, and then select Primary. After clicking Yes on the subsequent confirmation dialog, the options for the primary CFA will appear.

Next, make sure both Replication service buttons are greyed out while setting up the connection to the secondary CFA (you will come back here later to turn on the replication service once the connection to the secondary CFA has been confirmed).

Using the exact same authentication key that was generated when configuring the secondary CFA enter the key text into the Key field.

info The most accurate method to transfer the key is to copy/paste the text from the secondary CFA’s Key field directly into the primary CFA’s Key field. This makes it easier to ensure that the text is identical on both CFAs.

Enter the network address of the secondary CFA (the one you already set up) into the Address field. As with the secondary CFA configuration, you can also specify an optional description of the CFA you have designated as the secondary CFA.

You will also have to select the connection type under the advanced drop down options. The options are LAN mode and WAN mode. Set these according to the type of connection you will have between the CFAs.

LAN mode is designed to be used on local networks, it transfers larger blocks of data at one time, so it can achieve faster speeds, but has more to retransfer if something does not transfer correctly.

WAN mode is designed for slower connections, it transfers smaller blocks of data at one time, so it cannot achieve as much speed as LAN mode, but is more stable, and has less to retransmit if something does not transfer correctly.

As indicated at the beginning of this chapter, there is an option which will speed up the initial seeding process considerably. Expand the Compression box to display the list of compression options. Next, ensure that the neither option is enabled.

Although this form of compression is one of the key aspects to the enhanced performance of byte level replication over low bandwidth connections, it is unnecessary during the initial seeding process. In fact, it will slow down the seeding process since the overhead to calculate delta data is far greater than the benefit it could provide at this first stage of replication.

Once initial seeding is complete, you will need to return to the primary CFA’s bandwidth management options and turn the delta compression on by selecting Enable Byte Level Delta Compression and clicking Apply at the bottom of the panel. This is essential to unlock the full potential of the byte-level replication. See Starting and Monitoring Replication for more information.

You are now ready to complete the Primary’s connection configuration. Click Apply at the bottom of the panel. After a short delay, the UI panels should become re-enabled and the GUID of the remote (secondary) CFA should appear in the GUID field.

If an error appears in a pop-up or in the Configuration Status area at the bottom of the panel, and/or you do not see the GUID field populated as shown above, please refer to Errors during setup and configuration for more information.

Finally you will need to start the replication service on the primary CFA to complete setup.

Once you have started the service the connection status should show a green check mark, and display the message Remote CFA connected as of ....

There are some additional, optional controls that you have over the replication process and they are described below. You can adjust these settings at any time and do not have to deal with them at this point in the configuration process. However, it would be a good idea to read through and become familiar with what is available. You can then decide whether to set them up now or continue with starting the replication process.

info Once replication has been enabled, you will need to click Stop Replication Service in order to be able to modify any of these settings. Then once the settings have been changed and applied, click Start Replication Service again to resume replication with the new settings.

These settings can generally be left unchanged. However, if you are replicating over the internet by using port forwarding and NAT on the local and remote firewall this section of settings gives you the flexibility to change the ports as required by your firewall setup.

Web Services SSL – Check to enable the use of the SSL protocol for CFA recognition, defaults to on.

Web Services Port – TCP port used to contact remote CFA via web services for CFA recognition.

If port forwarding is not enabled on a firewall in between the two CFAs this value must be 443 if SSL is enabled, or 80 if SSL is disabled. If your firewall will translate some other port to the remote CFA’s port 80 or 443 this value can be changed to said port.

CFA Port – TCP port used by the replication service, this value and the corresponding value on the remote CFA need to match. The default value is 2300.

Propagate Manual Delete – Enabled by default. This option will remove jobs from the secondary that have been manually deleted from the primary. Disabling this option will allow the job to remain on the secondary until it is removed by its retention policy.

Bandwidth Management (optional) – The settings in this section allow you to tailor the effect of replication on your network.

Enable Byte Level Delta Compression – Toggle on/off the process of calculating and using only the changes in job data when replicating jobs to the Secondary CFA. This feature may require a Feature license key. Contact your Sales Representative if you require this functionality for replication over limited bandwidth connections.

Enable DDFS Assisted – Toggle on/off the process of replicating only the block files needed to recreate a backup job on the secondary.

Enable Bandwidth Limiting – Toggle on/off the bandwidth limiting feature. By default, the replication process will utilize as much network bandwidth between the Primary and Secondary CFAs as it can. You can limit the bandwidth available to replication by using the Bandwidth Management controls. The controls and their function are described below.

Limit (KB/s) – The maximum number of kilobytes per second that replication will use when transferring data from the primary CFA to the secondary CFA, when bandwidth limiting is enabled. For example, to use only 256 KBps, enter “256”. These values may be adjusted and depends on your internet bandwidth speeds.

info This is measured in kilobytes per second, while networks are typically measured in kilobits per second. To calculate what value to set this setting at to match a specific kilobit per second speed divide the kilobit per second size by 8. For example, if you wanted to limit the system to 800 kilobit per second you would set this setting to 100.

Specify Window for Limiting – Toggle on/off the ability to specify a start/stop time within which to limit replication bandwidth, this way you can limit bandwidth usage during business hours, but utilize the full connection when it will not impact production.

info The following options are only available on the primary CFA, though they control all communication between the primary and secondary CFAs.

Start Time – Hour and minute of the day to start limiting replication bandwidth

End Time – Hour and minute of the day to stop limiting replication bandwidth

Limit Bandwidth Over Weekend Days – Toggle on/off whether to use the specified limiting on Saturday and Sunday (“off” is the default meaning all available bandwidth is used over the weekend)

### Connection options

Some additional options are available to control the replication process.

Encryption – specify whether or not to encrypt data being sent between the Primary and Secondary CFAs, possible values are:

• None – data is not encrypted, this is default and should not be necessary if you are replicating over a secure LAN or VPN

• AES-256 – data is encrypted using the Advanced Encryption Standard 256-bit block cipher

After changing any of the options described above you must click Apply in order for the options to take effect.

## Cloud replication setup

This section will cover configuring your CFA to replicate to Infrascale cloud.

### Infrascale cloud

After registering the CFA, you can select Cloud under Mode/Options.

Select Cloud and you will be prompted to walk through the Cloud Replication Setup Wizard. You will have the option to create a new cloud setup, or an existing cloud setup in the event you need to replace the CFA that is no longer replicating.

After making your selection, the Cloud Replication Setup Wizard will complete and replication will be started automatically. It may take several minutes for a new cloud setup to come online and communicate with the primary CFA.

## Start or stop replication services

To start or stop the replication services on the CFA, click the appropriate button.

Stopping the replication service is needed to make any changes to any settings once replication has been set up. After making changes in any settings, click Apply at the bottom, then click Start Replication Service to resume replication with the new settings.

### Seeding progress

Once replication starts, the initial seeding process will begin immediately. Generally, the number of completed, error-free jobs on both the primary and secondary CFA should be the same when the primary CFA has been replicated successfully to the secondary CFA. It is possible that the counts may be off due to the erred jobs, or when new jobs have been created on the primary CFA and are actively being replicated to the secondary CFA.

With that in mind, use the Jobs group of the replication status panel to judge the overall state of replication between two CFAs.

The job count on the primary CFA indicates how many jobs can be replicated.

The job count on the secondary CFA indicates how many jobs have been replicated.

Therefore, there are two main conditions to look for:

• Job counts are equal (or nearly equal) – indicates that the primary CFA has been fully (or nearly so if counts are close) replicated to the secondary CFA

• Job count is greater on the primary CFA – indicates that there are jobs that have not yet been replicated to the secondary CFA

To judge when seeding is complete, monitor the job counts on both CFAs. When they are (nearly) equal and the connection statuses on both CFAs are idle, click Actions > Reconcile Secondary on the primary CFA. If no additional replicable jobs result from reconciling the secondary, seeding can be considered complete.

At this point, if the byte-level replication has been licensed, enable the byte-level delta compression as described in the previous section during the primary CFA configuration, and relocate the secondary CFA to its final destination.

### Replication status

There are some additional panels that provide more detailed information on the replication process. On both the primary and the secondary CFA, there is a general status panel available in Replication > Status.

This panel provides an overview of the connection status between two CFAs, and information on the data being processed and transferred as a result of the replication procedure.

On the primary CFA, the Replicated Jobs panel shows information relating to the jobs that have been replicated to the secondary CFA:

This page has the Effort column that will show one of three priorities for each job. The three priorities are Priority, Normal, and Best Effort.

Priority – When first starting replication the current recovery set for each job (latest full, latest differential, and subsequent incremental backups) will be set to Priority. Jobs in this status will be replicated first.

Normal – Any backup taken after the recovery set that is set to Priority is set to Normal. These jobs will replicate in the normal mode, if there are no higher priority jobs pending replication.

Best Effort – Any jobs older than the recovery set that is set to Priority, will be set to Best Effort. Best Effort jobs will be queued one at a time, and will only replicate if there are no higher priority jobs pending replication.

On the secondary CFA, the Pending Jobs panel provides information on the jobs that have been identified as needing to be transferred to the secondary CFA, and the jobs on which they are dependent.

### Errors during setup and configuration

If an error occurs during setup and configuration of the primary or the secondary CFA, it will be displayed in either a pop-up message or within the Configuration Status area at the bottom of the options panel.

Although unlikely, some errors that you might encounter are as follows:

Error message Description
Could not contact CFA Check that the address and port are correct and that network connectivity is available. Click Test Connection to test any changes and click Apply once corrections have been made
Connection refused Check that the CFA you are connecting to has been set up correctly and the replication service has been started on the CFA
Password mismatch Check that the authentication key specified in the Key field matches exactly between two CFAs, correct the key and click Apply once the keys have been adjusted to match
Version mismatch Check that both CFAs are running the same software version, the first two values must be the same. Minor versions will be able to replicate if they are not matched, e.g. 6.7.4.10 will replicate with 6.7.5.15

If you need to restart setup and configuration of the primary and/or secondary CFA, click Clear at the bottom of the options panel. Once completed, return to the beginning of the primary/secondary configuration section and repeat the steps for configuration. If you continue to experience problems, please contact the Infrascale Support.

## Selective replication

Sometimes you do not want to replicate everything from one CFA to another. This might be because the secondary CFA has a smaller RAID, or there is not enough bandwidth between the CFAs to keep up with all clients.

### Turning the replication on and off for a client

To turn replication on or off for a client:

1. Go to Clients > Edit, and select the client.

2. In the Offsite group, select or clear the Replicate checkbox.

If you select Purge Remote Jobs, the jobs from this client, which have already been replicated to the secondary CFA, will be deleted. If cleared, the already replicated jobs will remain on the secondary CFA, but more jobs from this client will not be added to the replication queue.

info If you select to purge all jobs from the remote CFA, and later decide to enable replication for a client again, it will replicate all existing jobs for that client. Transferring the first job or jobs for a client can take significantly larger amounts of bandwidth than it takes to transfer subsequent jobs because there is nothing for the byte delta replication to compare to.