I tried it quite some time ago with two sessions and simulators. But, it wouldn't keep the sessions in sync. Might very much be my fault, though.
I think it would be a great feature to support multiple long exposure cameras. But I guess it could be tricky to fiddle with the exposure pattern, like 1800 s for Ha in Cam 1, but 6x300 s L on Cam 2 and 3x600s R,G,B on Cam 3. The master-sessions would have to wait for respective slave session(s) to complete. Certainly manageable, but not a quick/easy implementation I assume. If one messes up the exposure pattern it might also cause a horrible duty-cycle.
Anyway, INDI/Ekos/KStars has received so many great features over the last few years, I keep my hopes up that the multi-cam feature will find it's way into the code eventually.

It's on my TODO list for quite a while, but as indicated by Niko it can get quite complicated. It's also a LOT of work to make the architecture of Ekos ready for N cameras (not just 2, but any arbitrary number). This also requires changes to how the telescope information is shared as well, so we're talking major architectural changes.

+1 for this. I'm help to a friend with an INDI based setup and he have in mind use two DSLRs with two telescopes in parallel (same mount).

Just a thought I had, but the support is already there and working for 2 cameras. The first camera is the main imaging camera, and the second is the guide camera. The guide camera still has all the basic functions such as repeated capture with settings such as exposure time and gain, and it always returns the image and goes to grab another one. It already does this in conjunction with the main imaging camera doing it's work. Would it be easier, as an intermediate step, to configure a third camera more like a second guide camera than a full fledged imaging system? It would not have all the features, but it seems like it could work. I am not certain if the architecture allows that any easier than what @knro was talking about though.

How would it handle dithering with different exposure lengths? For example, camara 1 takes 300 second exposures on the main scope, and 120 seconds with a piggybacked camera 2. Having the shorter exposure camera waiting for the longer shot to finish is inefficient. So ideally dithering will take place with the least amount of idle time according to a ratio of images between camera 1 and 2 to finish.
With my example Camera 1 fanishes 1 image in 300 seconds and camera 2 can take 2 in 240 seconds but would then idle 60 seconds a ratio of 1:2. Instead camera 1 can take two images in 600 seconds, and camera 2 can finish 5 with zero idle time, a ratio of 2:5. While that works in theory; the time to download each frame must also be considered. Each frame could take 10-20 seconds to download. That would have to be added to the time per frame and could change the most efficient ratio between camera 1, 2, or more cameras.

Certainly not very straightforward.

Or just don't start the exposure if there is not enough time left in the dithering timer/counter. There is no download contention either, this way. Yes, you would get fewer than the maximum frames on the waiting camera, but perfect is the enemy of something cool we don't have today.

There are two major challenges here, first is the synchronising of all the activities - alignment - focusers - filter wheels - rotators - weather - guiding - dithering - image acquisitions - meridian flip - limits - etc. Then there is the management of the schedules so as to minimise the waiting time, especally for the image devices.

The second is really up to the user but some tools to help would be useful. I can imagine running a rehearsal of the sequence using estimates for the various activities to get an idea of how much wasted time there is and when. A graphical display showing this would be good.

I'm not saying all this must be in place at the start but having a strategy for fitting everything into the system will I think be essential. Without that there's a risk that development will get bogged down in an endless series of refacatoring everything to handle an annotional scheduler option.

First, really loving kstars/ekos/indi - I'm a paid-up StellarMate RPi4 user and have been learning a LOT lately about moving into automated astronomy. My goal is mainly to do EAA but given the limitations of the platform for EAA I've been spending more time on astrophotography lately.

Apologies for re-opening the re-opened topic, but I have been thinking through the same question but for two different reasons.

1. simultaneously viewing the finder(s) and main scope
- mainly for initial alignment using a Meade LX-90. Using the finder scope for wide field "finding" and then the main scope (f/10 so REALLY tight FOV) I want to be able to see both simultaneously especially to be able to get the two scopes/cameras aligned on the same star. I have no interest in keeping the sequence, it's simple for alignment (I can't use the alignment tool, it causes my scope to do strange things with tracking).

2. I'd love to have a window open that's got a fish-eye view of the sky - of course I could do this with another window talking directly to the camera, but i like the idea of being able to see what's happening in teh sky above me especially when I am down in my office and can't see the sky only to find I am taking images of clouds! I saw two astronomers in Finland with a remote observatory who are doing something similar with a ZWO 120MM lying flat with the bundled fish-eye lens looking up to help them see what's happening in teh sky above - seems like a cool idea.

A long way of saying that I can see use cases for where N cameras (webcam of the observatory, 2-3 ultra-wide, wide and narrow) FOV cameras might be in operation simultaneously either for telemetry (looking out for cable snags, etc) or dithering etc could be super helpful.

I also say this knowing I don't have the skills to offer to help with the refactoring/architecting so it's easy to be the back seat driver!

Thanks again for everything you guys are doing to make this incredible platform!

So yes this is in principle possible. You can enable video streaming for multiple cameras simultaneously. If there is no active capture sequence going on, you can simply click the live streaming button and switch between cameras. If an active session in progress, you can go to INDI Control Panel --> Camera Driver --> Streaming and turn on the streaming from there.

@xthestreams, those are great use-cases for more cameras. The full sky camera tied into Ekos would be nice, but I wonder if it would actually make more sense to be part of the dome module vs the capture module. Perhaps even a new AUX cameras module?

With two ZWO cameras, I can currently set one as the main capture and one as the guider. The guider module will allow a continuous loop of images in its setup, as will the focus module. You might be able to use it, or the focus module to align your camera with the main.

Agreed, adding the camera view would be part of the "security" feature of the observatory (dome) tab.

-Eric

Great suggestion - perhaps it's my settings or how kstars plays with OS X (will never get used to calling in macOS, blech), the multi-window thing tends to be clunky so I usually avoid it (eg: the Indi control panel getting obscured by the main Ekos window and you can't bring one to the front without essentially moving the other out of the way) - so my muscle memory has become too close things down to stop other things from popping up where they're not wanted.

I'm hoping (praying, starting a large dry fire, sacrificing a goat and/or carrot for the vegans reading this) that tonight is clear skies here in Melbourne and I'll be able to do a test of the ideas shared here.