Last night my observations had stopped after the meridian flip (yes I know, I should run via scheduler...)
I had a look at the logs, to figure out what happened. I was running a sequence of 200 120s Hα exposures, directly from the Capture tab, with guiding via PHD2.
I used settings 'refocus every 64min' and 'refocus if ΔT>1°'.
At 02:12 a timer-based refocus occurred. Guiding is suspended during AF, as I use an OAG and sometimes the image moves a bit when changing focus direction.
All went fine, including the 20s settle after AF (as the guide star sometimes drifts away from focuser movement)
Right after that, at 02:21, it was time for a meridian flip that started one minute later.
After slewing to the other side, alignment was started. This unfortunately switched my filter from Hα to Red (*)
Alignment went fine without issues, autoguiding was restarted (02:23:57) and the next exposure initiated (02:24:10)
This switched back to Hα, which in turn again triggered an autofocus run
(I have set AF on filter change)
Also this AF run finished successfully. But when settling after re-enabling the guiding, already after 4 seconds it reported a failed settling, and aborted the sequence.
PHD2 itself needed some 20s to catch the star back and after that continued to guide for the rest of the night....
So I somehow suspect the problem was starting the guiding but immediately pausing it again, leaving it with a large error (some 7 pixels)
I had checked the PHD2 log, and indeed it had reported a settle failure right after the first exposure
Here's the part from the PHD2 log file:
INFO: SETTLING STATE CHANGE, Settling started
1,4.041,"Mount",-5.575,4.081,-5.095,5.074,-2.802,4.059,260,E,315,S,,,989036,489.46,0
2,8.046,"Mount",-3.526,1.777,-3.310,2.420,-1.820,1.936,169,E,150,S,,,1012387,457.94,0
3,12.113,"Mount",-2.078,0.182,-2.045,0.577,-1.125,0.462,105,E,36,S,,,1005718,446.58,0
INFO: Server received PAUSE
INFO: Server received RESUME
INFO: SETTLING STATE CHANGE, Settling failed
4,200.581,"Mount",-6.844,-1.689,-6.988,-0.346,-3.844,0.000,357,E,0,,,,1012080,406.98,0
So PHD2 was interrupted while still settling after the change of the guide star, and likely the settling period timed out during the AF operation.
Not sure who's to blame there - should PHD2 reset the settling state when it gets paused, or shouldn't EKOS send a pause while PHD2 is still settling?
In doubt I attach the logs (kstars and PHD2)
(*) I always disable the filter selector in Alignment and Focus tabs, as I want it to always use the current filter. Unfortunately it is not possible to save this setting, after each restart it is active again. And if I forget to disable that - see above