Setup / Deployment Configuration

[Harry Catharell]

There seem to be a number of topics on the forum which loosely refer to this subject line but none that specifically follow our production topology so I thought it easier to try to be specific and describe what we do in order to establish whether MirrorSync is for us :-)

For one of our clients, we currently have FMS13 running their internal system on a server inside their network

There is a second FMS13 server sitting in a DMZ outside their network but accessible from the internal server

Data is moved from the solution on the internal server to the server in the DMZ

This data comes from only 10-12 tables in the internal system and we have written business logic into the internal solution that identifies which data needs to moved and when it needs to be moved [this part is all fine and working without issue]

Remote users who are running FM Go on iPads or FM Pro on desktops will run synchronisations with the data hosted on the server in the DMZ

At this point in time the data moves one way - from the server in the DMZ to the remote users

To all intents and purpose, the datafile on the DMZ server is 'headless' - i.e. It has no user interface and is merely a repository for the data which needs to be synchronised with remote users

The remote users have a file which has a full user interface and logic to manage and work with the data that is brought down from the server in the DMZ

The schema of the datafile in the DMZ and of the remote user file is almost exactly the same - there are minor differences which allows the remote file to be able to cope with some of the local business logic required but the core tables are exactly the same

Can we setup MirrorSync to map the syncing of data from the file 'A' hosted on the DMZ to the remote file 'B' on various user 'devices' ?

A 'nice to have' would be: Having two distinct files between server and remote users, are there still options to allow downloads of new versions of the user file

Hope that I have been clear enough but please ask questions where I may not have been

Cheers

Harry

Edited April 28, 2015 by Harry Catharell

[evanseeds]

Harry, from what I'm gathering from your post, MirrorSync should support everything you're asking about. Specifically, for syncing from the server file to a slightly modified offline file, we offer the ability to sync with a 'separate mobile file', just as you describe. You can choose this option in the spoke portion of the configuration, which will then lead you to a dialog where you map the fields in one to fields in the other. 

Additionally, this does support downloading new versions of the mobile file. MirrorSync can generate download links for databases hosted on FileMaker Server, and when navigated to, those links generate a up to date download of the database file.

You didn't mention how you're accomplishing the server-to-server data movement, but if you're currently using MirrorSync to accomplish that, you will need to perform a bit of a workaround to sync the same file server-to-server and also server-to-client. You can read about this procedure at this documentation page.

Please let me know if you have any more questions!

[Harry Catharell]

Hi Evan (apologies if that is not your name :-)

Thanks for the feedback

We pretty much came to that conclusion and managed to get a test bed setup in house yesterday which worked - albeit with some errors and tweaks required and we will be finessing this today

In the meantime - is there a help area which describes versioning ?

How do we manage the setup and distribution of a new 'spoke'

Is this just a matter of replacing the spoke file on the FM Server and then modifying the configuration ?

Also as an FYI - we manage the server to server data movement through FileMaker itself

All of our scripted processes manage the recognition of data which requires being 'moved' and then we have a FileMaker drone that handles the necessary add, update, delete actions

Thanks

Harry

[evanseeds]

In the meantime - is there a help area which describes versioning ?

How do we manage the setup and distribution of a new 'spoke'

Is this just a matter of replacing the spoke file on the FM Server and then modifying the configuration ?

That's the general idea, yes. For the most part, making changes to the hosted spoke database is as simple as making those changes in FileMaker, making any necessary changes in MirrorSync, then repasting the script steps at the end. As long as you aren't removing fields from the sync, your existing offline copies should maintain the ability to sync, though they won't receive data from any added fields or tables. 

[evanseeds]

I realized I was a little unclear about the following point: MirrorSync does not have any feature that automatically downloads or notifies users regarding new versions, so it does not perform any versioning itself. However, the download links we generate will always download the current version from FileMaker Server, so if your users have that link, you can have them download a new copy whenever there have been changes to the hosted file. 

[Harry Catharell]

Thanks for the feedback, Evan

Understood and much appreciated

Cheers

Harry

[Harry Catharell]

One add-on question:

If we had an issue with a new version and wanted to revert to a previous one, any thoughts or a process to follow in how would we need to handle this with different configurations in the mix as well ?

Cheers

Harry

[Jesse Barnum]

The download link always gets the most recent version on the server, so if you wanted to roll back to an older version, you could either 1) roll back the version on the server and have users trigger the download link, or 2) send them the older version without using the download link.