If a planned outage occurs, use the following process for failover
and restore operations at the intermediate site:
- At the local site, ensure that data consistency is achieved between
the A and B volume pairs. This process will help coordinate the A
volumes and B volumes consistency and allow consistent data to be
copied to the remote site.
You can use either one of the following
methods to create data consistency:
- Quiesce I/O processing to the A volumes at the local site.
- Freeze write activity to the Metro Mirror primary volumes by performing
the following steps:
- Freeze updates to the A volumes in Metro Mirror relationships
across the affected LSSs. This ensures that the B volumes are consistent
at the time of the freeze. (One command per storage unit or LSS is
required.) As a result of the freeze action, the following actions
are taken:
- I/O processing to the Metro Mirror volume pairs is temporarily
queued during the time when updates are frozen.
- The volume pairs that are associated with the source and target
LSSs are suspended. During this time, the storage unit collects data
that is sent to the A Metro Mirror volumes.
- The established paths between the logical subsystem (LSS) pairs
are disabled.
- Resume operations following a freeze. This operation is also called
a thaw operation and it allows I/O processing to resume for the specified
volume pairs.
- Issue a PPRC establish pair command with the failover, force,
and cascade options. Specify the remote site secondary devices (C
volumes) as the primaries and the local site Metro Mirror primary
devices (A volumes) as the secondaries.
- Issue CRECOVER or PRECOVER commands to the intermediate devices
(B volumes), for the Metro Mirror relationships (A volumes to B volumes).
- Resume host I/O at the intermediate site.
- Issue PPRC establish pair commands with the failback and force
options from the remote site to the local site (site C to site A).
- Wait for the first pass copy to complete from the remote to the
local site (site C to site A), and then start the incremental re-synchronous
change recording mechanism on the site B to site C pairs, with NOCOPY
specified in the INCRESYNC parameter.
- Use the RESSION command to stop Global Mirror at the intermediate
site.
- Suspend the Global Copy pairs between the intermediate and remote
sites (site B to site C pairs).
- Wait for all data to drain from the remote site back to the local
site (site C to site A). An alternate from doing queries to wait
for all out of sync data to drain is to convert the site C to site
A pairs to synchronous. When they become full duplex, all data has
been copied.
- Suspend the site C to site A PPRC pairs.
- Issue CRECOVER or PRECOVER commands to the remote devices (C volumes)
to the B to C volumes, for the suspended site C to site B relationships.
- Issue PPRC establish pair commands with failover and cascading
to reverse the direction of the site C to site A relationships. The
command is issued to the A devices.
- Issue PPRC establish pair commands with failback to the site A
to site C relationships.
- Establish Metro Mirror by using the incremental resync change
recording that has been in process since the site has swapped to the
intermediate site, with the force option, from the intermediate to
the local site (issue PPRC establish pair commands with option SYNC,
mode INCRES, and force from site B to site A).
- Issue PPRC establish pair commands to the site B to site A relationships
to start the incremental resync change recording (issue PPRC establish
pair commands with mode NOCOPY and INCRESYNC(START))
- Wait for the site B to site A relationships to become full duplex
and the site A to site C relationships to complete their first pass
copy.
- Start the Global Mirror, running from the local to the remote
site in order to update the aged consistency group at the remote site.
The following steps show how to transit back to the original
configuration (from site A to B to C):
- Quiesce all host I/O activity and suspend the site B to site A
relationships.
- Issue a PPRC establish pair command with the failover, force,
and cascade options. Specify the remote site devices (C volumes) as
the primaries and the intermediate site devices (B volumes) as the
secondaries.
- Issue CRECOVER or PRECOVER commands to the local devices (A volumes),
for the site B to site A relationships.
- Resume host I/O at the local site.
- Issue PPRC establish pair commands with the failback and force
options from the remote site to the intermediate site (site C to site
B).
- Wait for the first pass copy to complete from the remote to the
intermediate site (site C to site B), then start the incremental resynchronous
change recording mechanism on the site A to site C pairs, with NOCOPY
specified in the INCRESYNC parameter.
- Use the RSESSION command to stop Global Mirror at the local site.
- Suspend the Global Copy pairs between the local and remote sites
(site A to site C pairs)
- Wait for all data to drain from the remote site back to the local
site (site C to site B). An alternate from doing queries to wait
for all out of sync data to drain is to convert the site C to site
B pairs to synchronous. When they become full duplex, all data has
been copied.
- Issue CRECOVER or PRECOVER commands to the remote devices (C volumes),
for the suspended site C to site A relationships.
- Suspend the site C to site B PPRC pairs.
- Issue PPRC establish pair commands with failover and cascading
to reverse the direction of the site C to site B relationships. The
command is issued to the B devices.
- Issue PPRC establish pair commands with failback to the site B
to site C relationships.
- Establish Metro Mirror by using the incremental resync change
recording that has been in process since the site swap to the local
site, with the force option, from the local site to the intermediate
site (issue PPRC establish pair commands with option SYNC, mode INCRES,
and force from A to B).
- Issue PPRC establish pair commands to the site A to site B relationships
to start the incremental resync change recording (issue PPRC establish
pair commands with mode NOCOPY and INCRESYNC(START)).
- Wait for the site A to site B relationships to become full duplex
and the site B to site C relationships to complete their first pass
copy.
- Start the Global Mirror, running from the intermediate to the
remote site in order to update the aged consistency group at the remote
site.
You have changed the Metro/Global Mirror operations running from
site A to B to C.