IBM DS6000 Manual Do Utilizador

Página de 578
282
 
IBM System Storage DS6000 Series: Copy Services with IBM System z
Figure 23-8 shows more than a single FICON/FCP director at both sites, for availability 
reasons. Note that the Master and Subordinate paths are defined over a local SAN fabric, 
represented by the FICON/FCP directors. 
23.7  Recovery scenario after primary site failure
This section covers the steps that you need to follow in a Global Mirror environment, when a 
primary site failure requires you to recover at the secondary site. This discussion does not 
cover the specifics concerning whether it is a single site host connectivity scenario or 
multi-site host connectivity scenario, as discussed in 23.6, “Connectivity between local and 
remote site” on page 279.
23.7.1  Normal Global Mirror operation
Figure 23-9 shows a simple configuration with Global Mirror active and running.
Figure 23-9   Normal Global Mirror operation
Host writes are replicated through Global Copy. Consistency Groups are created as tertiary 
copies. The B volumes are Global Copy secondary volumes, and they are also FlashCopy 
source volumes. The C volumes are the FlashCopy target volumes. 
23.7.2  Primary site failure
A failure at the local site prevents all I/O to the local storage disk subsystems; see 
Figure 23-10. This may have some impact on the formation of Consistency Groups because 
the entire process is managed and controlled by the Master storage disk subsystem that is 
also the primary disk subsystem—and it just failed, and cannot communicate any longer with 
its partners at the remote site.
FlashCopy
Global Copy
Primary
Primary
 A 
Primary
Primary
 A 
Primary
 A 
Primary
Primary
 A 
Primary
Primary
 A 
Tertiary
 C 
Primary
Primary
 A 
Primary
Primary
 A 
Secondary
 B 
3E00
3F00
Host 
Remote site
Local site
PENDING
PENDING