Справочник Пользователя для IBM Flex System V7000 Expansion Enclosure 4939H29

Модели
4939H29
Скачать
Страница из 610
Chapter 2. Introduction to IBM Flex System V7000 Storage Node 
75
Draft Document for Review January 29, 2013 12:52 pm
8068ch02-HW Intro.fm
Synchronous Remote Copy ensures that updates are committed at both the primary and the 
secondary before the application considers the updates complete; therefore, the secondary is 
fully up to date if it is needed in a failover. However, the application is fully exposed to the 
latency and bandwidth limitations of the communication link to the secondary. In a truly 
remote situation, this extra latency can have a significant adverse effect on application 
performance.
Special configuration guidelines exist for SAN fabrics that are used for data replication. It is 
necessary to consider the distance and available bandwidth total network links to determine 
the appropriate method to use. See 9.3, “Remote Copy” on page 380 for details on planning, 
configuring and using Remote Copy for replication functions.
With the Global Mirror method there is a design option that assists with low bandwidth for 
theIBM Flex System V7000 Storage Node and the other IBM virtual storage systems that 
aremirroring to it running 6.4.1 code level. This option uses change volumes associated with 
the primary and secondary volumes. These volumes are used to record changes to the 
remote copy volume, the FlashCopy relationship that exists between the secondary volume 
and the change volume, and between the primary volume and the change volume. This 
function is called Global Mirror cycling mode. Figure 2-20 shows an example of this function 
where you can see the relationship between volumes and change volumes.
Figure 2-20   Global Mirror with change volumes
In asynchronous Remote Copy, the application acknowledges that the write is complete 
before the write is committed at the secondary. Hence, on a failover, certain updates (data) 
might be missing at the secondary. The application must have an external mechanism for 
recovering the missing updates, if possible. This mechanism can involve user intervention. 
Recovery on the secondary site involves starting the application on this recent “backup” and 
then rolling forward or backward to the most recent commit point.
2.6.5  Copy Services configuration limits
In Table 2-8 on page 76, we describe the Copy Services configuration limits. For the most 
up-to-date list of these limits, see the following website:
Note: Though this feature was added in the 6.3 code level only 6.4.1code level is 
supported to mirror to the IBM Flex System V7000 Storage Node.