Managing HP X9000 Network Storage System Remote Replication Application Note (TA768-96049, November 2011)

Replication and X9000 file system snapshots
File system snapshots are presented in a .snapshot directory under the root of the snapped
directory. Continuous replication replicates only the current version of a file and skips snapshots
completely. Run-once replication allows individual snapshots to be replicated. Snapshots on the
source are sparse, but are fully populated when they are replicated. Be sure to plan file system
sizes on the target cluster accordingly. There are no special restrictions on snapping source and
target file systems and directories, allowing you to have multiple recovery points using replicas
and snapshots.
Planning considerations
When planning for continuous replication, consider the following:
• X9000 replication keeps source and target file systems synchronized. All changes on the
source are replicated to the target, including creation or deletion of files and directories,
whether planned or accidental. File system snapshots can be used on the source and target
clusters to protect against accidental file deletion.
• If you change the attributes of a file or directory, only the attribute changes are replicated.
• When you change part of a previously replicated file on the source, the entire file is replicated
again, not just the changed part of the file.
It is important to understand how applications modify the files stored on the cluster. Applications
typically behave in one of the following ways:
• The application rarely changes files and most files are replicated only once. Examples are
backup and archiving applications.
• The application completely replaces the old file when saving changes. Some applications
create a local temporary copy of a file in memory or on disk while you are working on it. The
application then overwrites the old version with the new version when saving changes.
Applications such as Microsoft Word and PowerPoint do this.
Because the whole file is new, it is a candidate for replication after updates, regardless of the
replication technology used.
• The application updates ranges of blocks in the file or appends data. On X9000 systems, this
can cause a large file to be replication after only small changes have been made. Some
research and scientific applications, technical applications, and video editing applications
do this.
This application behavior requires further research into how the application is used. When
using applications that work with large files, users often check a file out to local storage (that
is, they make a local copy). The user works on the copy and checks the file back in (saves it
back to the X9000) occasionally through the day or when finished. In this case, the file is
updated on the X9000 only when the user checks the file back in and only then will the file
need to be replicated. At most, the file will be replicated a few times a day. If this situation
does not apply and you are concerned about large amounts of data being replicated, it may
be appropriate to consider periodic run-once replication, possibly combined with snapshots,
instead of continuous replication.
The HP X9000 File Serving Software File System User Guide provides further details on the X9000
remote replication software, including advice on system and network sizing and how to perform
site failover and failback.
Running replication tasks
Each file serving node runs its own replication and synchronization processes, independent of and
in parallel with other nodes in the cluster. The source-side management console monitors the
Overview 5