•
To optimize performance, HP recommends that LUNs under migration are balanced across two
controllers of a storage array.
Choosing the right DMS options
Follow these guidelines when choosing DMS options:
•
To get optimum MPX200 performance, schedule a maximum of eight jobs to run simultaneously.
•
To sequentially run migration jobs, use the Serial Scheduling feature, which allows the migration
jobs to start automatically after the jobs with higher priority values are completed; no user inter-
vention is required. Serial Scheduling helps ensure optimum bandwidth usage without filling up
array I/O queues, which may occur if jobs run simultaneously.
•
Use the array bandwidth feature when applications are still using different LUNs of the array while
DMS is migrating the data. This ensures that the MPX200 uses limited array bandwidth for the
data migration and does not impact application performance during data migration.
General precautions
Note the following:
•
Make sure that hosts connected to source and destination LUNs are either zoned out completely,
or that LUN masking is changed appropriately, such that the LUNs being migrated are not touched
by any host other than the MPX200 until the migration is complete.
•
In high availability (HA) configurations where the LUNs are visible from both MPX200 ports, make
sure that both ports from each MPX200 blade are configured under a single host or host group
entity in the array management software. This configuration ensures that all MPX200 ports from
the same virtual port group (VP group) see the same set of LUNs as having the same LUN ID.
CAUTION:
Failure to follow this configuration can lead to unpredictable or erroneous behavior.
For a dual blade configuration for the MPX200, add the same VP group WWPNs from both
blades as one host entry. If you need to migrate more than 255 LUNs, you may create additional
host entries in the array using WWPNs from additional VP groups in the MPX200.
•
Migration logs require the following handling:
• Always clear migration logs at the start of the project.
• Export migration logs onto your system after the project completes.
• Migration logs wrap after 1,024 migration log entries.
•
If the source array controllers are configured in redundant fabrics, configure one MPX200 port
into Fabric A and the second port into Fabric B.
•
When using the serial scheduling feature, configure similar size jobs with the same priority.
•
Array reconfiguration precautions include the following:
• If the LUN presentation from the array to the MPX200 is changed, click the
Refresh
button two
or three times, waiting a few seconds between each click, to see the changes.
•
Remove unused arrays for the following reasons:
• The maximum number of arrays configured at any time is seven.
• Arrays stored in persistence consume resources even if the array is offline and no longer
needed.
• After the migration is complete, HP recommends that you remove the arrays.
MPX200 Multifunction Router
147
Summary of Contents for Storageworks 8100 - enterprise virtual array
Page 20: ...20 ...
Page 30: ...MPX200 Multifunction Router overview 30 ...
Page 47: ...Figure 27 iSCSI controller options Remove controller MPX200 Multifunction Router 47 ...
Page 48: ...Managing the MPX200 using HP Command View EVA 48 ...
Page 59: ...Figure 32 Virtual disk properties Figure 33 Host details MPX200 Multifunction Router 59 ...
Page 112: ...MPX200 iSCSI configuration rules and guidelines 112 ...
Page 219: ...Load Balancing Enabled Array License Not Applied MPX200 Multifunction Router 219 ...
Page 248: ...Offline Data Migration 248 ...
Page 258: ...Diagnostics and troubleshooting 258 ...
Page 306: ...Command referencecommand reference 306 ...
Page 330: ... Blade number 1 or 2 Simple Network Management Protocol setup 330 ...
Page 368: ...Saving and restoring the MPX200 configuration 368 ...
Page 402: ...Data migration best practices 402 ...
Page 408: ...408 ...