So DSLR and mount are running nicely on INDI on an RPi2, last thing to get to work is guiding.
I use a webcam (PCVC840 acting as SPC900) and the indi_v4l2_ccd driver. PHD2 runs locally on the RPi2, using INDI to access the webcam.

All seems fine, driver connects, I can change settings etc and I can start the image capture loop, but at some point the chain stops. The final message from the indi server is:

At the same time some system messages are logged, dmesg gives:

Does anybody have a clue on what's happening? It seems that the webcam device driver gets stuck, and the interface to the indi_v4l2_ccd service is shut down...

Quite strange error. As the driver can't communicate to the device anymore. Only idea from my side is to ask you if you're connect via USB directly to the RPi ? Usually better to use powered USB hub to avoid a lot of strange error.
Philippe

Apparently it has to do with a USB issue, the function usb_control_msge() in the PWC driver gets an error and as a result the connection to the webcam is lost.
I am not sure whether it is actually a HW problem, since I am using a Pi2, which is supposed to have a better USB implementation. Other USB devices seem to work fine.

Yesterday I had it running for 7 hours before it got stuck. I have changed the settings so that it only gets a snapshot every few seconds.

do you use any usb extension cables? Or maybe an USB device on your rp2 sucks to much power from the rp2 (then try an seperate powered usb hub)*. I had a lot of connection issues with my DSLR-USB cable. You could try to replace cables, change usb port, or just test the camera alone without any other usb-devices to trace down the problem. Sometimes its just the wrong usb port. That sounds stupid but for my setup I am shure one specific usb port doesnt like my wlan stick.

*If I remember correctly proper grounding can also be a matter (potential equalization between mount and rp2)

I am pretty sure now that it is the USB interface. It may be the cable, but I would be very disappointed about USB when this is the case, camera runs fine on a laptop. I doubled the USB power to 1.2W, which didn't help either.
What I do suspect is a non-resilient or plain shaky driver implementation, either the PWC (webcam) driver or the USB layer or the bad chemistry between the two. I think you don't notice as long as things are quiet, but when you start loading the HW with other processes glitches start to happen.

The 4 USB ports are equipped with keyboard, mouse, WLAN and the webcam. The RPi2 is running Ubuntu Mate 15.10, with full GUI.
PHD2 or lin_guider is running on one display, with 320x240 image size, which can be OK for hours. Lin_guider seems smoother and able to capture higher image rates probably because there is no INDI in between.
Things tend to go wrong however when I either start using e.g. Firefox on the other display or when I increase the image size in the guider.

I would be very interested to know whether somebody has got a webcam (esp. Philips) running on a Pi, and even more in how this was managed.

Edit: there is no mount connected yet, nor the DSLR, just the webcam and the guider SW....

I just stumbled upon some posts mentioning similar error message:
www.forum-raspberrypi.de/Thread-raspbian-syslog-laeuft-voll
github.com/raspberrypi/linux/issues/624

an suggestion was to setin /boot/cmdline.txt but dont ask me what it actually does... I have no clue
could be some kind of kernel problem

Thanks for the suggestions.
The German one has to do with ever growing syslog,
The github one has to do with masking off certain USB interrupts. I could give it a try, but it boils down to ignoring something is wrong...

I am running another test with lin_guider in 320x240 mode, doing nothing else. Still runs after a couple of hours. I would rather use PHD2 though, ARM / INDI version by Patrick Chevalley, which can do the guiding directly through the INDI/EQMod device. Maybe I need to dedicate my second Pi to guiding and use the other one for the rest and the INDI server ...