American Megatrends ManageTrends 2.7, Руководство пользователя

Страница: 264 / 367

StorTrends® 1300 User’s Guide (StorTrends iTX version 2.7) 246
Once JAR is enabled and is in action, it faithfully replicates every incoming I/O to the
recovery site allowing a continuous spectrum of recovery points. However, if there is a
big disparity between the incoming and outgoing rates, then the Journal size increases
and eventually a situation may arrive when the journal will overflow and the replication
cannot proceed. As seen from the diagram, simple flow control strategies are
implemented by the stack to throttle the incoming pipe to mitigate this effect. If this first
level of defense fails due to, say, extended periods of link failure then JAR agent
seamlessly migrates to Snap Assisted Replication. There are other instances like Journal
file failure, Volume Rollback etc. which also cause this transition. Eventually, over time,
when the Journal catches up again, a reverse transition brings it back to JAR. Elaborate
instrumentations are provided for monitoring and controlling the JAR operations. For
example, a Journal can be manually stopped to push out the replication to SAR.
Journal and Journal Files
Journal is a storage pool created using logical drives (LD). One or more Journal Files are
created in a Journal. Each Journal File represents a Consistency Group (CG). Consistency
Group is nothing but a group of related volumes which work together to carter storage for
database or similar application.
Journaling provides a mechanism to log main volumes’ (part of a CG) IO sequentially
into Journal File. Along with volumes’ IO, it also journals Snapshot events, Windows
events and Custom events in chronological order. StorTrends iTX supports upto a
maximum of 8 Journals and 64 Journal Files.