connection is a reference to the database connection in the
<connectionsString> section of the configuration file
under the <conection> section in the repository, a number of routes for replication can be configured e.g.
<routes>
<route to="repo" />
</routes>
section in the
config file for jobs
Jobs are scheduled activities that supplements the
real-time event-driven replication with regular batch synchronization.
The timestamp of the last load is maintained in both source
and destination databases to ensure that even database-restore does not result
in missing updates. Whilst the initial
load jobs are scheduled on regular basis, the last-load timestamp ensures that
the job id not rerun:
·
Load Job /Fastload Job
These job provide for the initial load of repositories
·
Change Job
This regular job performs the same function as realtime replications, but
ensures that updates are replicated regardless of whether the notification
addin is enabled
·
Delete Job
Best practice with a distributed environment is not to delete content, but mark
it with the status ‘Archived’ so that knowledge is never lost – this job
ensures that content that has been deleted at source does not become orphaned
in lifecycle/stakeholder views, by deleting diagrams/packages/elements from
dependant repositories. The connections
property of job can be used to ensure that t is only run for non-archive
repositories
·
Reload Job /
FastReload Job
The Enterprise Hub is designed to support real-time replication of connect, and
bulk initial load - this job exists to
enable changes to the status or element-type rules result in catch-up
replications.
Triggers are the actions to perform in response to changes (within the <connection .. /> section or for a whole repository within a <job><trigger ... /></job> section
The Change Job is designed to run intraday to supplement
the real-time event driven replication of for situations where the desktop
addin is not enabled. The “only
modified” parameter ensures that the recursive replication does to synchronise
elements that have not been changed through the desktop client (typically this
only applies when data is loaded through a tool that does not update the
modified date of content)
The fast-load threshold parameter of the Hub is used to switch from recursive replication when a very large number of objects have been updated, and fast-load would be quicker
Best practice with
a distributed environment is not to delete content, but mark it with the status
‘Archived’ so that knowledge is never lost – this job ensures that content that
has been deleted at source does not become orphaned in lifecycle/stakeholder views,
by deleting diagrams/packages/elements from dependant repositories. The
connections property of job can be used to ensure that t is only run for
non-archive repositories
The Fast-Load Job can be significantly faster than the recursive load of element at a time, but at the cost of holding most of a source model in memory during replication, provided the server has at least 16gb of memory
We recommend that it is only run at weekends, when other modelling is unlikely to be running
The job only runs if there is not a change replication date, which it creates after initial run
We recommend that it is only run at weekends, when other modeling is unlikely to be running, and only if there is not enough memory to run fast reload