IBM DS6000 Manual Do Utilizador

Chapter 18. Global Copy options and configuration 

The steps in the procedure shown in Figure 18-3 are the following:

1. Quiesce the application updates.

2. Synchronize the volume pairs by one of these methods:

– Perform the catch-up by doing a 

go-to-sync operation

. This can be done with the 

Storage Manager as shown in 18.2.3, “Go-to-sync using the DS Storage Manager” on 
page 208. The volume pair leaves the copy pending state and reaches the full duplex 
state. At this time, if the pairs were not immediately suspended, primary write updates 
would be synchronously transmitted to the recovery site.

– Perform the catch-up by waiting until all application updates are transferred to the 

secondary site.

You can monitor the number of out-of-sync tracks with the DS Storage Manager or with the 
DS CLI as shown in 18.2.5, “Display out-of-sync tracks” on page 209.

3. Suspend the Global Copy pairs after they reach the full duplex state. If you use 

Consistency Groups, issue do a freeze operation.

At this point, we have a set of consistent secondary volumes.

4. You can resume the application. Updates are not copied to the secondary volumes 

because the pairs are suspended. The secondary volumes remain consistent, and the 
application does not experience any response time impact.

5. Perform a FlashCopy on the secondary volumes.

6. Resume Global Copy mode for the copy pair.

For applications recovery based on point-in-time copies, you must plan for appropriate 
checkpoints to briefly quiesce the application and synchronize the volumes pairs. When the 
recovery of the application is done, you must remember that, while in an active Global Copy 
relationship, the secondary volumes always have a current fuzzy copy of the primary 
volumes, so you have to keep the tertiary volumes where you did a FlashCopy of the last 
globally consistent catch-up. Therefore, this tertiary copy does not reflect the current update, 
but rather any updates up until the last global catch-up operation.

Asynchronous cascading PPRC is a function of the IBM TotalStorage Enterprise Storage 
Server (ESS), where the secondary volume of a Global Copy or Metro Mirror volume pair 
becomes the primary volume of another Global Copy or Metro Mirror relationship. Cascading 
allows you to set a remote copy relationship from a source volume through an intermediate 
volume to a target volume. 

The two volume pairs in a cascaded relationship can be either Global Copy or Metro Mirror 
pairs with one exception: the first pair (source to intermediate) cannot be Global Copy if the 
second pair (intermediate to target) is Metro Mirror. The two pairs can be established in any 
order; that is, first and second here do not refer to the order of establishment.

Cascading can be useful when you are migrating data from an ESS to the DS6000. If the 
volumes to be migrated are mirrored between two ESSs with Metro Mirror, you can set up a 
cascaded Global Copy from the secondary ESS to the DS6000. Then, you can migrate the 
volumes without having to delete the existing volume pairs.

Cascading can also be used if you need to migrate data from an ESS Model F20 to a 
DS6000. This cannot be done directly with Global Copy because the ESS F20 only supports 
ESCON remote copy links, while the DS6000 only supports FCP links. The migration can be