HP StorageWorks Business Copy EVA V3.x Administrator Guide (T3680-96001, April 2005)

14 Planning and best practices
Managing replication activities
Develop operational guidelines and best practices to manage the following.
Simultaneous replication operations
Minimize the number of replication requests to the same array at the same time. Too many simultaneous
local replication events can reduce array performance.
Avoid making simultaneous multiple replication requests to the same virtual disk. Multiple replication
operations on the same virtual disk not only slow performance, but in the case of automated jobs, can lead
to job failures. For example, if the maximum number of snapshots per virtual disk is exceeded when the job
is running, the job will fail.
Do not run two or more simultaneous jobs that interact with the same host. If possible, design a single job
that includes the combined tasks of the separate jobs. When scheduling replication jobs, consider the
timing relative to other jobs on the same array, host, or management server. Tune and balance demands to
maximize performance.
Snapshot and snapclone rules
When a local replication interface indicates that a storage volume (virtual disk) does not support
replication, or if a script fails in a replication job, one of the following rules has probably been violated.
Snapshot rules
• All snapshots of a given virtual disk must have the same allocation policy
• No more than seven snapshots of a given virtual disk can exist at one time
• When managing array resources, snapshots are counted as a virtual disks.
• Snapshots must be created in the same disk group and disk family as the source disk.
Snapclone rules
• Snapclones are created in the same disk group as the source by default but can be created in
another disk group.
• Snapclone cannot be created if the source has a snapshot.
Common snapshot and snapclone rules
Neither snapshots nor snapclones can be created when the source:
• Is itself a snapshot
• Is unsharing or being deleted
Job dependencies
Avoid changing storage and host configurations while replication jobs are running. For example, do not
change an array configuration using HP Command View EVA while the replication manager is running
jobs. Changing resources can lead to job failures and require manual intervention to restore resources to
operational readiness.
Ensure that planned downtime for hosts is coordinated with replication jobs. A job will fail if a referenced
host is unavailable when the job is run.
Ensure that planned management server shutdowns are coordinated. Stopping a management server:
• Stops any local replication applications on the server
• Causes running jobs to fail
• Prevents scheduled jobs from starting
Ensure that network connections between a management server and host agents are maintained,
especially while jobs are running. Jobs that interact with hosts fail if the network connection is not
maintained throughout the job.