IBM System storage DS6000 Series Redbook page 273
Copy services with ibm system z
Hide thumbs
Also See for System storage DS6000 Series:
- Redbooks (406 pages) ,
- Attachment manual (298 pages) ,
- Introduction and planning manual (119 pages)
Table of Contents
Advertisement
Database
Subsystem
1
Primary
Hold I/O
Storage Disk
Subsytem 1
2C00
A
2C00
A1
A
Primary
Primary
Primary
Primary
Log
Redrive
Storage Disk
Subsystem 3
2D00
A
2D00
A
A
Primary
A2
Primary
Primary
Primary
Primary
Primary
Primary
4
DB
Storage Disk
Subsystem 5
2D00
A
2E00
A
A
Primary
A3
Primary
Primary
Primary
Primary
Primary
Primary
RECON
Figure 22-3 Synchronous data replication, freeze, and restart without recovery required
The sequence of events in Figure 22-3 is the following:
1. Update intent to logging file.
2. Link between A2 and B2 becomes unavailable, or the storage disk subsystem with A2
fails, and a rolling disaster develops.
3. The first attempt to write to the database volume A2 will return a unit check. The host
re-drives the I/O. When only the links fail, the subsystem triggers an
(ELB). This is also called an
4. Within this automation window, or ELB window, a freeze operation removes all related
paths between both sites for the selected LSSs. From now on, no data is replicated any
longer between the selected LSSs.
5. After the ELB expires, or a RUN command is issued to end the freeze period, the I/O is
re-driven, and it may successfully complete when the pair is suspended, or may fail when
CRIT(HEAVY) is specified or the storage disk subsystem 3 is not available any longer.
6. When the I/O fails, then a
A3. Note volume A3 is no longer replicated due to the previous freeze. Note also that
volume A2 and B2 are still identical and at the very same level of data currency —that of
the moment previous to the error.
7. The primary site eventually becomes unavailable and a switch over to the secondary site
is carried out.
8. Database subsystem restart discovers
data is consistent —neither the
B3. Then, the database subsystems continues immediately after the restart phase is
finished.
Fail over to secondary site
6
Recover A2
3
4
Synchronous
Replicate
I/O
5
4
2
Synchronous
Replicate
4
Synchronous
Replicate
automation window
recovery required
is written to the database configuration file on
no recovery required
recovery required
Database
7
Subsystem
Secondary
Storage Disk
Subsystem 2
3C00
B1
1
Log'
Storage Disk
Subsystem 4
3D00
A
B2
Primary
Primary
Primary
DB'
Storage Disk
Subsystem 6
3E00
A
B3
Primary
Primary
Primary
RECON'
extended long busy
.
indication because all related
indication in step 6 was transmitted to
Chapter 22. Global Mirror overview
Restart
8
8
8
249
Advertisement
Table of Contents
