We have just released the 5th version of RecoverPoint which offers an even deeper integration with XtremIO, if you are not familiar with RecoverPoint, it’s the replication solution for XtremIO and basically offers a scale out approach to replication. RecoverPoint can be used in it’s physical form (which is the scope of this blog post) or as a software only solution (RecoverPoint For Virtual Machines, RP4VMs)
Expanding an XtremIO volume is very simple from the CLI or UI. To expand a volume in XtremIO simply right-click the volume, modify it and set the new size.
Before RecoverPoint 5.0, increasing the size of a volume was a manual process. While XtremIO made it very easy to expand volumes, RecoverPoint was unable to perform the change in size. To increase the size of a volume, you would have to remove the volume
from RecoverPoint, log in to the XtremIO and resize the volume. Once the volume was resized, add the volume to RecoverPoint again. A volume sweep would be triggered by the new volume in RecoverPoint.
RecoverPoint 5.0 and above allows online expansion to Replication Set volumes without causing a Full Sweep resulting in journal loss. When both the production and copy volumes
are from an XtremIO array the Replication set can be expanded. Best practice is to perform the re-sizing on the copy first, then change production to match.
This example has a consistency group containing two replication sets. The selected replication set has a production volume and a single remote copy. Both are on XtremIO arrays and in different clusters.
Here is an example of expanding the size of a replication set. The first step is to expand the (remote) copy on the XtremIO array. The volume can be identified by the UID, which is
common to both RecoverPoint and XtremIO. Next we increase the size to 75 GB in this example.
Notice that now the Copy and Production volumes of the Replication Set are different sizes. Since we expanded the Copy volume first, the snapshots created during the time the
volumes are mismatched will still be available for Failover and Recover Production. Upon a Rescan of the SAN volumes the system will issue a warning and log an event. A rescan can be initiated by the user or it will happen during Snapshot creation.
Next we will expand the Production volume of the Replication Set. Once this is accomplished the user can initiate a rescan or wait until the next snapshot cycle.
After the rescan is complete, the Replication Set now contains the Production and the copy of the same size. Displayed is the Journal history, notice the snapshots and bookmarks are intact. Also there is a system generated bookmark after the resizing is complete.
RecoverPoint protection is policy-driven. A protection policy, based on the particular business needs of your company, is uniquely specified for each consistency group, each copy (and copy journal) and each link. Each policy comprises settings that collectively
govern the way in which replication is carried out. Replication behavior changes dynamically during system operation in light of the policy, the level of system activity, and the availability of network resources.
Some advanced protection policies can only be configured through the CLI.
Beginning with RecoverPoint 5.0, there is a new Snapshot Consolidation Policy for copies on the XtremIO array. The Goal of this new policy is to make the consolidation of
snapshots more user configurable.
The XtremIO array currently has a limitation of 8k volumes, snapshots and snapshot mount points. RecoverPoint 4.4 and below enforces the number of maximum number of snapshot, but in a non-changeable manner. For example, the user cannot change the time RecoverPoint will consolidate the snapshots on the XtremIO. This new policy will allow the user more control over how long the snapshots will be retained.
One to three consolidation policies can now be specified for each copy of a consistency
group that resides on an XtremIO array. By default and for simplicity reasons, consolidation policy is selected automatically based on number of snaps and required Protection Window.
The new CLI command config_detailed_snapshot_consolidation_policy will allow a much more detailed and concise consolidation policy for copies on an XtremIO array.
The config_copy_policy command will allow for the setting of the maximum number of snapshots.
Data Domain is an inline deduplication storage system, which has revolutionized disk-based backup, archiving, and disaster recovery that utilizes high-speed processing. Data Domain easily integrates with existing infrastructure and third-party backup solutions.
ProtectPoint is a data protection offering which brings the benefits of snapshots together with the benefits of backups. By integrating the primary storage with the protection storage, ProtectPoint reduces cost and complexity, increases speed, and maintains recoverability of backups.
RecoverPoint 4.4 introduced support for ProtectPoint 3.0 by enabling the use of Data Domain as a local copy for XtremIO protected volumes. The Data Domain system has to be registered with RecoverPoint using IP. Data transfer can be configured to use IP or FC. The Data Domain based copy, is only local and the link policy supports two modes of replication:
• Manual – bookmarks and incremental replication are initiated from the File System or
• Periodic – RecoverPoint creates regular snapshots and stores them on Data Domain.
There is no journal for the Data Domain copy during the Protect Volumes or Add Copy wizard. If the add Data Domain copy checkbox is selected, users will only select the Data Domain registered resource pool.
When using RecoverPoint 5.0 with ProtectPoint 3.5 and later, specific volumes of a consistency group can be selected for production recovery, while others volumes in the group are not recovered.
Displayed is an example of a use case for the new Partial Restore feature of RecoverPoint 5.0. In this example, this new feature allows for the recovery of only part of a database system
Only selected volumes are blocked for writing. Transfer is paused from XIO source > DD replica during recovery. During transfer, marking data is collected for the non-selected volumes, in case writes are being performed to them in production.
During partial restore:
• Transfer is paused to all copies, until the action completes
• At production – selected volumes are blocked, non-selected volumes remain accessible
• Only selected volumes are restored
After production resumes all volumes undergo short initialization (in case of periodic snapbased replication policy is configured for XIO->DD link).
Things to keep in mind about the partial restore option for ProtectPoint:
• Only selected volumes are blocked for writing
• Transfer is paused from XIO source -> DD replica during recovery
• During transfer, marking data is collected for the non-selected volumes, in case writes are being performed to them in production