
13
Write mode option
The DR Groups write mode is specified during the creation of a DR group. The options are
synchronous and asynchronous. Testing is performed by using synchronous replication write mode,
which means that the array acknowledges I/O completion after the data is cached on both the local
and destination arrays.
Failsafe mode option
Failsafe mode is only available when a DR group is being replicated in synchronous mode. It
specifies how host I/Os are handled if data cannot be replicated between the source and destination
array. Testing is performed with Failsafe mode disabled.
The failsafe options are as follows:
Failsafe enabled: All host I/O to the DR group is stopped if data cannot be replicated between the
source array and destination array. This ensures that both arrays always contain the same data (RPO
of zero). A failsafe-enabled DR group can be in one of two states.
o Locked (failsafe-locked): Host I/O and remote replication stops because data cannot be
replicated between the source and destination array.
o Unlocked (failsafe-unlocked): Host I/O and remote replication resumes once replication
between the arrays is re-established.
Failsafe disabled: If replication of data between the source and destination array is interrupted, the
host continues writes to the source array, but all remote replication to the destination array stops and
I/Os are put into the DR group write history log until remote replication is re-established.
Replication log options
The size of the DR group write history log and the disk group locations on source and destination
arrays are specified during DR group creation. By default, the write history log is 100 GB for each DR
group. There is a write history log location for both source and destination arrays, because the source
and destination roles are reversed when a failover occurs.
The DR group-write history log stores data when replication to the destination DR group is stopped-
This occurs because the destination DR group is unavailable or suspended. The process is called
logging. When replication resumes, the contents of the log are sent to the destination virtual disks in
the DR group. This process of sending I/Os contained in the write history log to the destination array
is called merging. Because the data is written to the destination in the order that it is written to the log,
merging maintains an I/O-consistent copy of the DR group data at the destination.
Normalization
The method of synchronizing source and destination virtual disks is called normalization. When a DR
group is first created, or when a new member is added to the DR group, a full-copy normalization
occurs. All data on each Vdisk member of the DR group is then copied from the source array to the
destination array, bringing the two arrays into synchronization. Normalizations copy data from the
source array to the destination array in 128 KB blocks. Even though the array prioritizes front end
I/O requests over normalization activity, a performance impact can be incurred when adding multiple
Vdisks to a DR group; therefore it is best to create the DR groups during off-peak hours of operation.
Normalization also occurs when a write history log overflows. This may happen if the replication link
is down for a longer time than the log size can accommodate. When a write history log overflows,
the controller invalidates the log contents and marks the DR group for normalization. In some cases,
normalization will be optimized to copy only blocks that are written before the write history log
overflows, not a full-copy normalization. The array tries to optimize the full copy whenever possible.
Once the DR group is created, additional virtual disks can be added to the DR group by selecting the
virtual disk and by using the Data Replication tab on the Vdisk properties, as shown in Figure 5.
Komentarze do niniejszej Instrukcji