SRDF adaptive copy Mode

Have you had a look at the "modes of operation" section of the "EMC Symmetrix Remote Data Facility (SRDF) for VMAX... Product Guide"?

Page 19/20 explains the difference between Adaptive Copy and Asynchronous. As it states in this overview, Adaptive Copy can be more than one IO out of Synchronous, therefore data consistency at the R2 site is not guaranteed.

Adaptive copy is a good 'catch up' mode, to perform bulk copy between R1 and R2 when there are many tracks to synchronise.

Only Synchronous mode affects write performance of the host as the host does not receive a write acknowledgement until the data has been sent to the R2.

As Jasonc mentioned above. There are much details description about SRDF Adaptive Copy mode in SRDF product Guide.

Generally, adaptive copy mode is used for sync large amounts of data with minimal host impact as SRDF adaptive copy modes have a fast application response time because they return a positive acknowledgment to the host as soon as data is written to cache in the primary Symmetrix system during adaptive copy process. it is recommended that you use this mode temporarily to transfer the bulk of your data to target (R2) devices, and then switch to a full SRDF mode (synchronous or asynchronous) to ensure full data protection.

Thanks a lot for your answer.

i believe its not 100% guarantee that all the data will send to remote server. because we are not getting confirmation acknowledge ment as it does in Synchronous

VM , Data will be sent over but adaptive mode does not give you guarantee for the consistency of the data , means data may not be sent over in the write dependent order.

Hope this helps  clearing your doubts.

Thanks,

Shubh

Which you should choose really depends on your requirements for RPO/RTO and Link speed. 

If you are running SRDF/Sync,  all writes from the host to the source storage array are sent to the target storage array before being acknowledged to the host - this ensures you have every piece of data the host knows about on the target replica array. The downside of this is your host has some delay while it waits for the storage array to copy the data from source to target... If you have a large, low latency pipe, and a requirement for near-zero RPO this is a great solution.

SRDF/A and SRDF/S are both consistent,  where as ACP is generally used for your initial bulk copy for a new replication session.   With ACP each and every write made from host to array is copied to the target array - so if the host writes data to the same track (i.e. overwrites data) it will copy both writes across the link.

SRDF/A copies on a 30 second cycle.  That means if two writes happen to the same location within the 30 seconds window, it will only be copied once.

Which mode is best for host connected to the Source with best performance?

It depends on your requirement , whether you need  Synchronous mode of replication or Async mode of replication. You always get better performance in Async mode .

I understand Asynchronous will not have Gold Copy

1. SRDF/S provides near zero RPO and will keep all DR data consistent with Production. The applications will wait for Ack of the writes and this will cause a increased response time and performance Impact.
2. SRDF/AC doesn’t require Ack of writes, so there wont be any wait on writes. The performance is at its best for SRDF write traffic. The disadvantage is, there wont be consistency of data for DR. This mode is usually used to catch up track changes quickly if the replication was suspended or stopped of long duration.
3. SRDF/A is the best of both. There wont be response time delays as there is no IO write ACK. The Data will be controlled as delta sets to keep consistency across RDF groups.