Realtime Content and Database Replication for Alfresco

Question

I'm currently researching on how to replicate physical files (store contents) and the meta data (database) of Alfresco. This is of course a safety measure in case of server failure or whatsoever.

Currently i am running Alfresco's Database on PostgreSQL Engine, and by far, have learned PostgreSQL's WAL and Stream replication. Of which i believe, i can use in terms on replicating Alfresco's meta data (database) real time.

The next problem i face now, is as to how i can replicate alfresco's repository/physical files (store contents) in real time ?

i am currently looking at Alfresco's Built-in Replication Job. But as far as i have read, it is "scheduled" and not in real time. And, it needs another instance of Alfresco running on the "SLAVE" Server.

So my question is:

Does Alfresco's Built-in Replication Job cover both the Physical/Repository Files (store contents) and meta data (database) contents ?

or

what is/are the viable ways of replication Alfresco's Physical/Repository Files (store contents) and meta data (database) contents in real time ?

Answer 1

The replication service can be used to replicate objects from one Alfresco server to another at the object level, not the file system and database level. So, of course there are files and database records that are created when an object is replicated, but the those are by-products of the object being created in the replication target.

The replication service is really used to make it easier for objects in a particular path to be read by people in another office. When they read the object they get it locally. When they click "Edit" in Share they will be redirected back to the source Alfresco server.

Long story short, the replication service is in no way, shape, or form, to be used to replicate data for backup or disaster recovery.

If you are running on EC2 or a local filer that supports it, it should be enough to take volume snapshots.

Otherwise, you could use something like rsync scheduled with cron.

But this approach sounds risky. I'm not sure how you will ensure that your database is kept in sync with your file system, which is a requirement for your Alfresco repo to remain consistent.