beautypg.com

I/o throttling, Job command processing, Logs – HP P6000 Continuous Access Software User Manual

Page 85: Log overview and states, I/o throttling job command processing logs, Log state

background image

I/O throttling

I/O throttling refers to suspending some DR groups during a log merge or full copy event. By
suspending noncritical DR groups, the critical DR groups complete the merge or full copy sooner.

Throttling of I/O after logging

When logging, and not in the failsafe mode, remote replication automatically resumes when the
links are restored. If there are several DR groups with large logs, they can compete for remote
replication bandwidth, thus slowing the overall merge or full copy between the two sites.

By suspending the merge or full copy of the noncritical DR groups, the critical data is available
sooner at the destination. After the critical DR groups finish merging or copying, the remote
replication of the noncritical DR groups can be resumed.

Job command processing

When a job is run, some job commands may specify features that are not available on the target
array. See

Controller software versions - remote replication

. Rather than fail the job, the replication

manager ignores the unsupported feature. The following job commands are processed in this
manner.

Remarks

Job commands

These arguments are ignored if the feature is not available.

AutoSuspend On Link Down
Destination RAID level

CreateDrGroup
CreateDrGroupFromHostVolume

Source Log Disk Group
Destination Log Disk Group
Maximum Log Disk Size

This job command is ignored if the feature is not available.

ForceFullCopyDrGroup

This job command is ignored if the feature is not available.

SetDrGroupAutoSuspend

This job command is ignored if the feature is not available.

SetDrGroupFailsafeOnLinkdownPowerup

When a virtual disk is a remote copy, the Access Type = None (No read)
argument is ignored if the feature is not available.

PresentStorageVolume

Logs

Log overview and states

A DR group log is a designated virtual disk that stores the host data that is written to the virtual
disks in a source DR group. The contents of the log are replicated to the destination virtual disks
to synchronize them with their sources.

The process of storing host data, in the exact order received in the DR group, is called logging.
The process of replicating the logged data to the destination DR group, in the same order as
received, and synchronizing with the sources is called merging or DR group normalization. See

Logging

,

Merging

, and

DR group normalization

.

When a DR group log becomes full, a full copy is automatically initiated. See full copy mode and
log size.

DR group concepts

85