This article summaries the Chinese ATE activity: EMC real-time replication solution – RecoverPoint. The original thread is https://community.emc.com/thread/187392
Question1: What is the characteristic of RecoverPoint compared with other EMC backup and recovery products?
Answer: RecoverPoint does real-time replication, every writing operation will be written to the production volume and replica volume in real-time. Similarly, when using the replica volume to restore the data, can be accurate to the second level RPO.
Question2: For the capacity planning of secondary RecoverPoint site, does RecoverPoint have a software or formula to do calculation? When the production volume fails, how does RecoverPoint achieve the protection? For MS SQL, Oracle applications, how does RecoverPoint ensure the data consistency? Is the RecoverPoint operation based on disk block or filesystem?
Answer: The capacity planning of RecoverPoint is the calculation of the journal volume. For the detailed calculation and planning, you need to consult a pre-sales engineer. In general, the recommended size of the journal volume is 20% of the replica volume. When a failure happens on the production volume, can be restored to a specified point time of mirror through replica volume and snapshot. RecoverPoint support data consistency for Oracle, SQL Server and Exchange applications. RecoverPoint only make replication for writing operation and it is based on the block.
Question3: If both the customer’s sites (Production and Secondary) are all not EMC storages, which component needs the RecoverPoint solution to be added to?
Answer: There are some differences depending on the RecoverPoint version.
3.5 and previous versions, RecoverPoint supports four kinds Splitter.
1. Host Splitter (No additional hardware, install the Splitter on the host)
2. Storage Based Splitter (EMC storage)
3. Switch Splitter (Needs a CISCO MD9000 or Brocade 7600, some other models switches.)
4. Vplex Splitter (Needs a Vplex device, supports heterogeneous storage)
From 4.0 versions, RecoverPoint supports two kinds Splitter.
1. Storage Based Splitter (EMC storage)
2. Vplex Splitter (Needs a Vplex device, support heterogeneous storage)
Question4: RecoverPoint is suitable for what kind of scenario? Is it also suitable for database scenarios, such as Oracle or DB2?
Answer: RecoverPoint is a block level replication and it is influenced by the upper application that is very small. So, RecoverPoint has a wide applicability. Whether it is EMC storage, or non EMC storage, all can be added into RecoverPoint. Just the condition is different. EMC storage only need to upgrade to a specific firmware version, make the appropriate Splitter configuration, it can support the RecoverPoint replication. Non EMC storage needs to add Vplex or SAN-TAP/Brocade AP7600 switch to support RecoverPoint replication. For mainstream of large database, such as Oracle, SQL Server and DB2, RecoverPoint also have good support.
Question5: Any advantages RecoverPoint and MirrorView have individually on the two same EMC storage environments? Which disaster recovery performance better? Which architecture more flexible? Which can achieve one to many disasters?
Answer: MirrorView is a mirror, but RecoverPoint is a replication based on Continuous Data Protection (CDP) technology. For example, if the production LUN deleted a file at 12:00:00, the copy of the MirrorView also was deleted. So, you cannot restore the file. But RecoverPoint can restore the file to the production LUN through the mirror of the11:59:59. RecoverPoint can have four replicas and two simultaneous backup.
Question6: When write a data to production LUN, RecoverPoint also will write a data to the replica LUN. So, is the process synchronous or asynchronous?
Answer: When create replica, you can select the replica policy is synchronous or asynchronous. Synchronous replica is writing synchronously, asynchronous replica is writing asynchronously.
Question7: What is the format of RecoverPoint written data? Can read and write directly? If the production storage failure, how to restore the data through RecoverPoint? And please tell us about the working mechanism of RecoverPoint.
Answer: RecoverPoint replicated data is block level data. When the production storage failure, you can choose that restore data from disaster recovery site to the production site, or failover to disaster recovery site.
RecoverPoint working mechanism:
1. I/O split: Split a writing operation, then replicated to storage in local site through continuous data protection (CDP), or replicated to storage in remote site through continuous remote replication (CRR).
2. Transmission phase: RecoverPoint will compare the consistency between local site and remote site and transmit changed data to the remote site. Data transmission has synchronous transmission and asynchronous transmission.
3. Transmitted replication data will be the first to enter the journal volume of RecoverPoint, and then distributed to the final replica volume.
4. Once a production storage failure occurs, you can restore data from journal volume of the remote site.
Question8: Could you tell us about the RecoverPoint authorization? For example, a RecoverPoint configuration will include RecoverPoint hardware. Besides, what other licensed software need to be considered?
Answer: RecoverPoint's authority needs to consider several factors:
· Supported storage types (EMC storage, non EMC storage)
· Binding storage model
· The capacity of cluster
· The replication capacity of local and remote
· Journal volume
According to the different needs of customers and the different features supported, RecoverPoint authorization can be divided to RPSE, RPEX, and RPCL. The number of features supported is RPSE<RPEX<RPCL. Detailed information can reference RecoverPoint Administrator's Guide.
Author: Leo Li