HP StorageWorks Storage Mirroring user's guide (T2558-96073, February 2008)

13 - 1
13 X Managing Replication Sets
Managing Replication Sets
Depending on how you created your connection, you may have created your replication set by browsing your source and
selecting which items to include or Storage Mirroring may have created it for you by automatically selecting specific items
from the source. In either case, behind the scenes, when a replication set is created, a series of rules are defined that identify
the volumes, directories, files, and/or wild card combinations that will be replicated to the target. Each rule includes:
Path—The path including volume, drive, directory, file, and/or wild card
Include—If the specified path is to be included in the files sent to the target
Exclude—If the specified path is not to be included in the files sent to the target
Recursive—If the rule should automatically be applied to the subdirectories of the specified path
For example, a replication set rule might be
volume\directory\* inc, rec
This specifies that all files contained in the volume\directory path are included in the replication set. Because recursion is set,
all files and subdirectories under volume\directory are also included. A complete replication set becomes a list of replication
set rules.
Replication sets offer flexibility tailoring Storage Mirroring to your environment. For example, multiple replication sets can be
created and saved for a source to define a unique network configuration. There may be three replication sets - Critical Data,
User Data, and Offsite Data. Critical Data could be configured to replicate, in real-time, to an onsite high-availability server.
Offsite Data is replicated across a WAN and, therefore, is configured to queue changes until a sufficient amount of data is
changed to justify transmission. At that point, the connection is made and stays active until all the data is transmitted. User
Data is not replicated throughout the day, but a nightly changed file mirror copies only blocks of data that are different
between the source and target server prior to a nightly tape backup operation being run on the target server. Each of these
replication sets can be automated to transmit as needed, thus protecting your entire environment.
Keep in mind the following notes when creating and working with replication sets and connections.
Limitations
Replication set rules are limited in length meaning that the entire volume\directory\filename including slashes, spaces,
periods, extensions, cannot exceed 259 characters.
Storage Mirroring can mirror, replicate, verify, and restore paths up to 32,760 characters, although each individual
component (file or directory name) is limited to 259 characters. Paths longer than 32,760 characters will be skipped
and logged to the Storage Mirroring log file and the Windows Event Viewer.
Do not name replication sets or select a target location using illegal characters. Illegal characters include the following.
If your source contains both a long file name and a separate file with the corresponding 8.3 file name, the files must
be kept in two different locations on the source. This is because the operating system, not an application, controls
the conversion from long file names to 8.3 file names. If the files are stored together in the same location on the
source, the potential exists for the wrong file to get updated on the target.
Error checking and avoidance
Replication sets contain error checking to avoid inadvertent overwrites of the replication set rules. When replication
sets are modified, a generation number is associated with the modifications. The generation number is incremented
anytime the modifications are saved, but the save is not allowed if there is a mismatch between the generation
number on the source and the Management Console. You will be notified that the replication set could not be saved.
This error checking safeguards the replication set data in the event that more than one client machine is accessing
the source’s replication sets.
Storage Mirroring will not replicate the same data from two different replication sets on your source. The data will
only be replicated from one of the replication sets. If you need to replicate the same data more than once, connect
the same replication set to multiple targets.
Do not connect more than one replication set to the same location on a target. You could overwrite or corrupt your
data.
quotation mark " forward or backward slash \ / forward or backward angle bracket < >
colon : pipe or vertical bar | asterisk *
question mark ? period .