IBM Flex System V7000 Expansion Enclosure 4939H29 Manuel D’Utilisation

Codes de produits
4939H29
Page de 610
8068ch09-Copy Servies.fm
Draft Document for Review January 29, 2013 12:52 pm
396
 
IBM Flex System V7000 Storage Node Introduction and Implementation Guide
Table 9-5   FlashCopy and Remote Copy interaction
If you are not using Global Mirror with Change Volumes, you can, for disaster recovery 
purposes, use the FlashCopy feature to create a consistent copy of an image before you 
restart a Global Mirror relationship.
When a consistent relationship is stopped, the relationship enters the consistent_stopped 
state. While in this state, I/O operations at the primary site continue to run. However, updates 
are not copied to the secondary site. When the relationship is restarted, the synchronization 
process for new data is started. During this process, the relationship is in the 
inconsistent_copying state. The secondary volume for the relationship cannot be used until 
the copy process completes and the relationship returns to the consistent state. When this 
situation occurs, start a FlashCopy operation for the secondary volume before you restart the 
relationship. While the relationship is in the Copying state, the FlashCopy feature can provide 
a consistent copy of the data. If the relationship does not reach the synchronized state, you 
can use the FlashCopy target volume at the secondary site.
9.4  Troubleshooting Remote Copy
Remote Copy (Global Mirror and Metro Mirror) has two primary error codes: 1920 or 1720. A 
1920 error is a congestion error. This error means that either the source, the link between 
source and target, or the target were not able to keep up with the rate of demand. A 1720 
error is a heartbeat or cluster partnership communication error. This error tends to be more 
serious, as failing communication between your cluster partners involves some extended 
diagnostic time.
9.4.1  1920 error
Let us focus first on the 1920 error. A 1920 error (event ID 050010) can have several triggers. 
Official probable cause projections are:
��
Primary 2145 cluster or SAN fabric problem (10%)
Component
Remote Copy primary site
Remote Copy secondary site 
FlashCopy source
Supported
Supported.
When the FlashCopy 
relationship is in the Preparing 
and Prepared states, the cache 
at the Remote Copy secondary 
site operates in write-through 
mode. This process adds 
additional latency to the already 
latent Remote Copy 
relationship.
FlashCopy destination
This combination is supported. 
It has several restrictions:
��
Issuing 
stop -force 
might 
cause the Remote Copy 
relationship to fully 
resynchronize. 
��
The code level must be 
Version 6.2.x or higher. 
��
The I/O group must be 
the same.
This combination is supported 
with the restriction that the 
FlashCopy mapping cannot be 
copying, stopping, or 
suspended. Otherwise, the 
restrictions are the same as at 
the Remote Copy primary site.