Vault Replication

[Mr.Caro]

Hello everyone,

I am new in the replication world, and I have a couple of questions. I have built a test enviroment with connected workgroups, as it follows:

- Workgroup 1

    - Publisher: SQL 2008 and Windows 2008

    - Subscriber: Remote SQL Database; Windows 2008

- Workgroup 2

     - Subscriber: Local SQL Database (SQL 2008), replicated from Publisher in Workgroup 1; Windows 2008

All the Servers are running Vault Professional 2014. At the beginning, I thought that the information was not instantly replicated, but now I see that is that way. So, my questions are:

- Is there a way to stop this instantaneous replication?

- If the information is instantly replicated, what can I use for Scheduled Replication?

Thanks for your help.

CARLOS

[paul.gunn]

Hi Carlos,

For database replication (connected workgroups), information is replicated continuously. This is necessary to facilitate ownership changes, etc. Scheduled replication applies only to replication of the physical files, which are stored separately from SQL and replicated via HTTP calls between servers.

Hope this helps,

Paul

[Mr.Caro]

Hi Paul,

thanks a lot for your fast answer. I think I'm starting to get it. So, databases are replicated continuously, and filestores are only replicated with Scheduled Replication right? Let's put an example:

A subscriber makes a new version of a file. The Publisher knows that there is a new version of the file, and in the File Status it shows "it needs to be updated", but it is not automatically done. When the scheduled replication starts to work, will the file in the Publisher be updated, or will the file in the Subscriber go to the previous version? If I stablish the scheduled replication from the publisher to the subscriber, the last version in the subscriber will be erased right?

And now, i have another question: is it possible to stablish replication for some folders from the subscriber to the publisher?

Kind regards,

CARLOS

[paul.gunn]

Hi,

Scheduled replication will replicate each version of the file separately - both the previous and latest if not already replicated.The files are never erased automatically - as it is sometimes necessary to access previous versions of a file. If you were to purge the older versions of the file then the physical files would be cleaned up after that point as a background task.

It is possible to select what folders you want replicated (this affects the file replication only). That would be the 'replicated folders' command in adms console. Each site can have a different set of folders selected for replication.

Paul

[Mr.Caro]

Hi Paul,

that was really helpful, thanks a lot! When you say that each site can have a different set of folder to be replicated, you mean each subscriber, even if it is not in the same workgroup right? And, to fnish, one more question.

The Replication must be scheduled from the Publisher. But, In the ADMS Console, under "File Stores", I can choose whether I configure it clicking in the Vault I want to replicate under the Publisher, or under the subscriber. Which one is correct? Or both of the are OK? When I change it under the Publisher, for example, under the Subscriber remains the same, and the other way.

With this I am done. Thanks a lot!

Kind regards,

CARLOS

[paul.gunn]

Hi Carlos,

Each adms server (listed under the Filestores node in adms console) has its own replication schedule and folder selection. So one site could replication all files more or less continuously - whereas other sites might not replicate at all. From a file replication point of view, it doesn't matter if a site is on the publisher or subscriber - file replication is bi-directional and all sites are treated equally.

So the behavior you are seeing is right - changing schedule of a site on the pubisher will not affect the schedule for other sites. And vice versa.

Paul

[ryan · Accepted solution]

In addition to what Paul said, the reason you are seeing both site vaults listed under filestores is because you can have a seperate replication schedule for each.  What you have to remember about file replication is that it works under a "pull" model - in other words, each server has a schedule to "pull" missing files from the other server.  So, when you see Unreplicated Files/Size in ADMS Console you are seeing the data that exists on the other server but is not currently on the one you selected.  Therefore, for the Publisher Filestore Vault you may have a schedule for it to "pull" data from the Subscriber and similarly, on the Subscriber Filestore Vault you may have a different schedule for it to "pull" data from the Publisher.  Although, you may not need a schedule on the publisher to pull the data from the subscriber since you should have the backup configured to do this automatically.

P.S.  The only reason you have to set the schedule in the Publisher ADMS is because by default it "owns" the administrative settings.   If you were to transfer the admin ownership from Publisher to Subscriber then you would need to set schedule on Subscriber ADMS.

Ryan