Several technical details of 3par disaster recovery

1. During synchronous remote replication, if there is a problem with the replication link, the synchronization is interrupted. The 3PAR can be configured to the following two coping modes:

• Failsafe mode: Two array hangs, does not provide service to host

• default mode: The primary array switches to class Async mode, which stores the changed data in the CPG 's remaining storage space where data VV resides, waits for the replication link to resume, and then resumes synchronizing the data. If the link outage time is too long, causing the primary array CPG remaining storage space to be filled with the change data, the data synchronization will be completely interrupted. After the link is restored, the remote array needs to resynchronize all the primary array data.

2. Data consistency group issues:

since upper-level applications often require more than 3par of VV storage data, in the remote replication process, it is important to ensure that the image data write order of these VV on the remote array is consistent with the primary array. So 3par introduces the concept of data consistency groups by combining these vv into a data consistency group to achieve: not only is the target-side write order of each VV in the group consistent with the source, the group VV The write order between the sources is exactly the same as the source end.

Similarly, for this reason, if you configure a double-live, an application that uses all VV should be within a data consistency group. This allows all VV master paths for an application to not be assigned to two 3par . Similarly, to provide application reliability, all VV for an application is best allocated to one store.

Several technical details of 3par disaster recovery