Motorola 6806800C47B Manuel D’Utilisation
1
Checkpoint Service Programmer’s Reference (6806800C47B)
13
Introduction
1.1
Overview
The Checkpoint Service provides a facility for processes to record checkpoint data
incrementally, which can be used to protect an application against failures. When recovering
from fail-over or switch-over situations, or restart situations, the checkpoint data can be
retrieved, and execution can be resumed from the state recorded before the failure.
incrementally, which can be used to protect an application against failures. When recovering
from fail-over or switch-over situations, or restart situations, the checkpoint data can be
retrieved, and execution can be resumed from the state recorded before the failure.
Checkpoints are cluster-wide entities that are designated by unique names. A copy of the data
stored in a checkpoint is called a checkpoint replica, which is stored in the main memory rather
than on disk for performance reasons. A given checkpoint may have several checkpoint replicas
stored on different nodes in the cluster to protect it against node failures.To avoid accumulation
of unused checkpoints in the system, checkpoints have a retention time. When a checkpoint has
not been opened by any process for the duration of the retention time, the Checkpoint Service
automatically deletes the checkpoint.
stored in a checkpoint is called a checkpoint replica, which is stored in the main memory rather
than on disk for performance reasons. A given checkpoint may have several checkpoint replicas
stored on different nodes in the cluster to protect it against node failures.To avoid accumulation
of unused checkpoints in the system, checkpoints have a retention time. When a checkpoint has
not been opened by any process for the duration of the retention time, the Checkpoint Service
automatically deletes the checkpoint.
The CPSv service supports the following two types of update options:
z
Asynchronous update option
z
Synchronous update option
In the case of asynchronous update option, one of the replicas is designated as the active
replica. Data is always read from the active replica and there is no guarantee that all the other
replicas contain identical data. A write call returns after updating the active replica.
replica. Data is always read from the active replica and there is no guarantee that all the other
replicas contain identical data. A write call returns after updating the active replica.
In the case of synchronous update options the call invoked to write to the replicas returns only
when all replicas have been updated, i.e. either all replicas are updated or the call fails and no
changes are made to the replicas.
when all replicas have been updated, i.e. either all replicas are updated or the call fails and no
changes are made to the replicas.
The CPSv supports both collocated and non-collocated checkpoints. In case of checkpoints
opened with collocated and asynchronous update option, it is up to the application to set a
checkpoint to the active state. In all other cases the CPSv itself handles which checkpoint is
currently active.
opened with collocated and asynchronous update option, it is up to the application to set a
checkpoint to the active state. In all other cases the CPSv itself handles which checkpoint is
currently active.
The CPSv defined by SAF does not support hot-standby. This means that the currently stand-
by component is not notified of any changes made to the checkpoint. When the stand-by
component gets active, it has to iterate through the respective checkpoint sections to get up-to-
date. To overcome this drawback, the CPSv provides additional, non-SAF APIs which help to
notify the stand-by component of changes and thus facilitate the implementation of a hot-stand-
by.
by component is not notified of any changes made to the checkpoint. When the stand-by
component gets active, it has to iterate through the respective checkpoint sections to get up-to-
date. To overcome this drawback, the CPSv provides additional, non-SAF APIs which help to
notify the stand-by component of changes and thus facilitate the implementation of a hot-stand-
by.