HP StorageWorks XP Remote Web Console Software - Manual (Page 24)

– If a pair in journal group 3 is suspended, the consistency time will be 14:03.
– If a pair in journal group 4 is suspended, the consistency time will be 14:04
If a failure occurs in a primary storage system and then you want to recover from the failure,
restore journal data with time stamps later than the consistency time of the extended consistency
group to the secondary data volumes. For example, in the case described in
Figure 4 (page
23) , the consistency time of the extended consistency group is 15:00, and therefore you must
restore the following data to secondary data volumes:
◦ Data with the time stamp 15:02 in journal group 2
◦ Data with the time stamp 15:03 in journal group 3
◦ Data with the time stamp 15:04 in journal group 4
If a failure occurs in a secondary storage system and then you want to recover from the failure,
compare the consistency times of all journal groups in the extended consistency group, and
then identify the oldest consistency time. Next, restore all data with time stamps later than the
oldest consistency time to the secondary data volume. For example, in the case described in
“Time Stamps of Data that Have Not Been Restored to Secondary Data Volumes” (page 23) ,
the consistency time of journal group 2 is the oldest among journal groups 1 to 4. Since the
consistency time of journal group 2 is 14:02, you must restore the following data to secondary
data volumes:
◦ Data with the time stamp 15:00 in journal group 1
◦ Data with the time stamp 15:02 in journal group 2
◦ Data with the time stamp 14:03 and data with the time stamp 15:03 in journal group 3
◦ Data with the time stamp 14:04 and data with the time stamp 15:04 in journal group 4
Functionality of Host
Host I/O Time Stamp
If you plan to establish Universal Replicator for Mainframe journal groups, the I/O time-stamping
function must be installed on the host processor at the primary site. The I/O time-stamp, which is
provided by MVS DFSMSdfp, is the same time-stamp that is used by Compatible XRC pairs. The
I/O time-stamping function should also be installed on the host processor at the secondary site, so
that time stamps can be used when copying data in the reverse direction.
NOTE: If the system at the primary and/or secondary site consists of several CPU complexes, a
SYSPLEX timer is required to provide a common time reference for the I/O time-stamping function.
ERC
ERC, which transfers information between host processors at the primary and secondary sites, is
a critical component of any disaster recovery effort. You can configure ERC using channel-to-channel
communications, NetView technology, or other interconnect technologies, depending on your
installation requirements and standards. Neither Universal Replicator for Mainframe nor the Universal
Replicator for Mainframe remote console software provides ERC between the primary and secondary
sites.
When Universal Replicator for Mainframe is used as a data migration tool, ERC is recommended
but is not required. When Universal Replicator for Mainframe is used as a disaster recovery tool,
ERC is required to ensure effective disaster recovery operations. When a Universal Replicator for
Mainframe pair is suspended due to an error condition, the primary storage system generates
sense information which results in an IEA491E system console message. This information should
be transferred to the primary site via the ERC for effective disaster detection and recovery.
24 About Universal Replicator for Mainframe Operations