Zero downtime backup and HP StorageWorks Continuous Access EVA

Data Protector A.05.10 - ZDB for EVA in the CA Configuration
10
For definition of other technical terms related to EVA and CA EVA, refer to the CA EVA documentation, and to the Command View
EVA online help.
3.4 High Availability and Data Safety in CA EVA
3.4.1 Data Replication and Failover
On CA EVA, the DR groups are configured in such a way that your data located on the source virtual disks on the local disk array
(visible from the application system) is constantly replicated to the destination virtual disks on the remote disk array. If a disaster of the
source virtual disks (with original data) occurs, an unplanned failover is automatically triggered by the EVA controllers. During the
failover, the destination virtual disks of the "failed over" DR groups on the remote disk array take the role of the corresponding source
virtual disks of the same DR groups on the local disk array, and the direction of data replication reverses. After the failover, in order for
you resume with normal operation on the application system, you should establish a connection between this system and the remote
disk array, and restart the system.
After the original data is recovered, a planned failover (induced by the user) reverts the data replication direction to the initially
configured direction.
3.4.2 CA EVA Logs
The disk array containing the source virtual disks has some storage space dedicated for logging purposes. In circumstances where data
replication in CA EVA is not possible, due to broken connection between the source and destination disks, or some other reason, new
data and changes to the existing data on the application system are logged to the log space on the local disk array. Each DR group
configured on the disk array has its own log. During the logging process, the status of the DR groups with the source virtual disks being
replicated is set to "logging". After the connection between the disk arrays is reestablished, the contents of the log are merged with the
contents of the corresponding destination virtual disks on the remote disk array, so that the redundancy of data is restored. For the
duration of this activity, the status of the involved DR groups is set to "merging". After merging is complete, the status is set back to
"normal".
If the interruption of data replication in CA EVA is long-lasting, the free storage space reserved for the DR group logs on the local disk
array may run out. In this case the logs cannot hold all the changes to the original data. After the connection between the disk arrays is
reestablished, all original data in the involved DR groups has to be copied over. During copying, the status of the DR groups with the
source virtual disks being replicated is set to "copying", and reset to "normal" after the operation is complete.
data replication group
(DR group)
a grouping of up to eight copy sets that have
common characteristics and share a common CA
EVA log
source virtual disk a virtual disk that contains original data which is
replicated to a destination virtual disk
destination virtual disk a virtual disk that is the recipient of replicated
data from a source virtual disk.
failover In general, the meaning of this term is
context-specific. In this white paper, failover is
only used to denote DR group failover.
DR group failover an operation that reverses the direction of data
replication in a DR group
Ter m Definition