G'day - time for a mid-level n00b question regarding distributed INDI.

Same use-case as most everyone else - I want to get the OTA related services running on a device as close to the OTA as possible with as few cords as possible (in this case a Raspberry Pi running the cameras and an Ultimate Power Box for focuser control, dew management and device power along with 4x CMOS cameras and EFW.

The rest of the setup that's related to the management of the observatory itself I want "off mount" including the mount itself (Paramount), the dome (Nexdome), flat fielder (Flat Man clone) and observatory power, lighting and cooling control (eventually de-humidifier) - in this case Arduino Simpler Switcher but likely to move to something more robust over time. Currently using a second UPB instead.

I've read the great article about setting up via the command line in Jassem's document and gotten it all working but didn't like the idea of scripting and config management - it just seemed too fragile.

I then started down the path of using INDI Web Manager to do the same thing - it works great btw - FORGET having to hack startup files etc.

It was then I noticed that there's a "Remote:" field in the Profile Manager on EKOS (OSX) and after slapping my head for not realising a simple solution had been staring me in the face for month, I tried entering the details for my remote INDI server into that.

The good news, it works! The bad news, KINDA....

I found that provided I only have ONE remote device server in the field, it works fine. But the minute I put two or more in, it somehow parses and sends only the first device in the field to the remote INDI Web Manager. I have tried every combination of punctuation I can think of but with no success.

I am assuming I am doing something wrong, but what?

Here's a screen shot of the field in question. If I attempt to start the services then only the first in the list makes it into Web Manager carina.local.

If I copy and paste this into the web manager on "carina.local" and manually start (or use Autostart) then they're all up and running and thankfully EKOS sees them all and I have my observatory running just as I'd like it.

Bug or Layer-8 protocol problem? (I am guessing the latter).

In other news - this is obviously a great next step towards a simple way to run distributed services and "splitting" the job of data acquisition from the physical processes required to start, operate, check and close the observatory which is already somewhat de-coupled from the process of planning, prioritising and scheduling data capture.

In an ideal world I would be able to select a different server host for each device as I am setting up within EKOS profile manager - in fact I can see a day where each device itself natively has an INDI server embedded into the firmware and they can all be discovered on my network and I then compose them into "meta-services" that solve a particular problem.

Thinking out loud here, not sure where it's going other than it would sure be nice to be able to specify a host field for each device in the Profile Manager (and while we're at it have at least 4 more slots available for devices - mine has gone from being empty when I started in this hobby to being overflowing!)