The journal sequence number indicates the primary data volume write sequence that the primary storage system has created for each journal group. The journal data is transferred to the secondary storage system asynchronously with the host I/O. The secondary storage system updates the secondary data volume in the same order as the primary data volume according to the sequence number information in the journal.
Note: URz processing continues uninterrupted if the SVP reboots or even if the SVP fails.
Types of Journal
In addition to the journal data for updating, the primary storage system sends control information to the secondary storage system. This control information indicates when volume pair status changes and when a primary storage system
Journal Processing at the Secondary Storage System
When a primary storage system receives a read journal command from a secondary storage system, the primary storage system sends the journal data to the secondary storage system. The secondary storage system’s initiator ports act as host processor channels and issue special I/O operations, called remote I/Os (RIOs), to the primary storage system. The RIO transfers the journal data in FBA format using a single channel command. The primary storage system can send several journal data using a single RIO, even if their sequence numbers are not contiguous. Therefore, the journal data are usually sent to the secondary storage system in a different order than the journal data were created at the primary storage system. The secondary storage system ensures that the journal data are applied to the secondary data volume in the correct sequence. This method of remote I/O provides the most efficient use of primary storage
Note: You must make sure that your channel extenders are capable of supporting remote I/O. For further details, please contact your Hitachi account team.
Storing Journal at the Secondary Storage System
A secondary storage system receives the journal data that is transferred from a primary storage system according to the read journal command. The journal data will be stored into the cache at first, and then into the journal volume.
Chapter 2 About Universal Replicator Operations |
