HP 3PAR Remote Copy Software - Manual (Page 53)

2.37 Error Handling
InForm OS Version 2.3.1 Remote Copy User’s Guide
In addition to taking snapshots, the offset and length of the I/O will fail to complete on the
secondary volumes because the link failures were recorded in the nonvolatile memory. The
failed I/O must be recorded separately because the snapshot is taken after that I/O has been
written for the primary volumes but before it was acknowledged and completed on the
secondary. When replication is restarted for the volume, the failed I/O is read from the
primary volume and sent to the system that holds the secondary volume. In addition, all
differences between the base volume and the snapshot taken when the replication was
stopped are also sent over in order to resynchronize the secondary volume with the primary
volume.
■ If the fail_wrt_on_err policy is enabled for a synchronous volume group, failing host
writes does not cause the mirror to become out of sync. A write error is returned to all host
writes as long as the links are down. When the links come back up, the writes are
completed normally. This policy can be enabled with the setrcopygroup command.
2.13.2.2 Asynchronous Periodic Groups
For asynchronous periodic groups, when a double-link failure occurs, replication is stopped on
both storage servers. No new snapshots are created. If a resynchronization was in progress
when the failure occurred, the snapshots on the secondary system are automatically promoted.
2.13.3 Remote Copy Failure Timeouts
There are two primary timeouts used by Remote Copy regarding link failures:
■ a link heartbeat timeout, which is used to declare an unresponsive link as down.
■ a target timeout, which is used to declare a target system down after all links have gone
down.
The value of these timeouts varies depending on the mode of the groups per target. If the
Remote Copy configuration contains asynchronous periodic mode volume groups, then the
link heartbeat timeout is 60 seconds and the target timeout is 200 seconds. If the Remote Copy
configuration contains synchronous mode volume groups, the target timeout is 15 seconds and
the heartbeat timeout is 10 seconds.
NOTE: For asynchronous periodic groups, when the links are restored, replication
can be restarted by issuing the startrcopygroup command. The volumes are
resynchronized at the next scheduled resynchronization time. To resynchronize
prior to the next scheduled time, use the syncrcopy command to manually
initiate immediate resynchronization.