HP MPX200 Multifunction Router - Manual (Page 20)

indicates congestion at the array controller. Thus, the MPX200 may require automated throttling
while trying to maximize migration performance by increasing concurrent I/Os. To control
automatic throttling and pacing of migration I/O, use the Enable I/O Pacing option.
•
Enable I/O Pacing: This feature is applied only to a source array. The MPX200 intelligently
manages concurrent migration I/Os to maximize overall migration throughput. If a Queue
Full or Busy condition is detected, the router throttles the migration I/O until it detects either
that the array queue is full or a busy condition. After the condition is cleared, it starts issuing
additional migration I/Os. This method maximizes host and migration I/O performance.
To achieve pacing, the router uses a configured, concurrent I/O limit and an internal counter
(current concurrent I/O limit, which is less than or equal to the configured limit) and a set of
steps for automatic throttling and pacing of migration I/O. The user sets the configured limit.
Data migration job groups
The MPX200 uses the concept of job groups to associate data migration jobs with user-defined
groups. A job group allows better management of data migration jobs. You can create a maximum
of 32 job groups that are shared between the two blades on a chassis. Both the HP mpx Manager
and the CLI provide options for removing and editing job groups.
Creating job groups is an opportunity to organize your data migration jobs. One typical
organizational model is creating groups that relate to application type or server class. For example,
you could classify a data migration job related to the Microsoft Exchange application as part of
the “Exchange” group and a data migration job related to a payroll application as part of the
“Payroll” group. Data migration jobs are then tracked separately within their respective groups.
Group information for each data migration job is recorded in the data migration log, see
“Data
migration job log” (page 64) .
If you do not define a group, all jobs are assigned to the default group, Group 0. You cannot
delete Group 0.
Data migration jobs
DMS processes data migration jobs according to a schedule. You can schedule a maximum of
512 jobs (256 jobs per blade) to run at any time. This section covers job attributes, migration
types, job scheduling, job states and job failover and failback.
Job attributes
Data migration jobs include the following attributes:
•
Migration type: Data migration jobs can be either online (local or remote) or offline. For more
information, see
“Migration types” (page 21) .
•
Source and destination LUN: For an offline migration job, you can configure a single source
LUN to migrate to one or multiple destination LUNs. For an online migration job, you can
configure a single source LUN to migrate to only one destination LUN. Any specified destination
LUN can be part of a single data migration job.
•
Job groups: For better manageability, you can configure data migration jobs to belong to a
specific, user-defined job group. By default, a job is assigned to a default group, Group 0.
For more information, see
“Data migration job groups” (page 20) .
•
Scheduling: You can configure data migration jobs to start immediately, start at a specified
time, or to use a priority-based serial scheduling. For more information, see
“Job scheduling”
(page 21) .
20 Data migration objects