Should have a separate database for feature services for editing

Question

We currently have our enterprise gdb divided into three. GISdata is our editing database and we have 6 editors that could be editing at one time, GISPUB is our read-only database for desktop viewers, crystal web, and web services that are consumed by our non-GIS enterprise applications, and GISRaster is for all of our imagery. We use versioning and the GDB's reconciled and compressed nightly.

How do you handle data that is published as feature services that will be used for editing? Do you keep it in your editing database or do you have a separate database for feature service editing?

Thanks in advance for any input.

Asrujit_SenGupta · Answer

natgeo1984‌ Along with looking for Best Practices, it will be a good idea if you create a new Post detailing your requirements, the available applications to you, current configuration(if any). That way you will get more meaningful suggestions for your scenario.

NatalieBuchwald · Answer

This isn't really a true "reply" but I couldn't figure out how to message you directly (maybe that's not possible, maybe it is - sorry if it is, I am new to this and in a time crunch). When you said:

"We currently have our enterprise gdb divided into three."

What did you mean? Do you mean that you have three separate databases running? One for editing, one for read only/publishing web services, and one to store imagery? If you have three separate databases running, is the one for read only/publishing a replica of the editing one? Is this also how you handle permissions?

My apologies if these are trivial questions, I recently started a new position and have been tasked with creating a new enterprise database - I have been an editor, built applications, etc but have never managed a database - I'm just trying to do my homework before I screw anything up too badly.

JuneAcosta · Answer

We are currently building ArcGIS Enterprise 10.5.1 in our test environment, so we have a list of collector apps to push out once we go-live. We use replication and versioning so the enterprise gdb gets synchronized, compress, re-index, and analyzed nightly with a python script. The script will disconnect all connections before running so we can get as close as possible to state '0'. I try to do as much as I can to make sure our users have the best performance from the GDB.

I posted this question to find out how other admins are setting up the gdb and if someone has some suggestions because of their "lessons learned". My biggest concern is maintenance and performance. If I publish these feature services from our editing gdb then I will need to modify the nightly job to disconnect/re-publish the connection for these AGS services. Maybe I am too concerned with compressing down to state '0' and it's really not a big deal. Now my next question is, " do I publish the feature services from a version?". I'll have to do more research and most likely will need to post the question on a difference forum.

Thanks for responding to my question!