Please post logs if you have, that's the only way to tell.

Just ran a short sequence with logging to file enabled.
Basically, got the same behavior except the guider calibration failed in the very beginning too.
Attaching the log.

Timeline;
20:45:41 Guide calibration starts but never finishes - RA forward goes wrong (opposite) way, RA reverse continues in the same direction. I aborted.
20:47:14 Restarted calibration. Lost guiding star. Failed. My fault. Increased exposure time.
20:48:15 Restarted calibration. Now OK.
21:15:51 Meridian Flip.
21:18:00 (auto) Align. result: 16 arcsec.
21:18:26 Guide calibration starts but never finishes - RA forward goes wrong (opposite) way, RA reverse continues in the same direction. I aborted.
21:20:06 Restarted calibration. Now OK.
21:22:59 Stopped guiding to check Align.
21:23:06 First Capture & Solve gives 4' error -> can this be the result of RA going in the the same direction for too long?

Thanks!

I finally was able to test the meridian flip behavior with my HEQ5-Pro mount when not guiding. I set it to 10 degrees (40 minutes) after culmination (i.e. Flip if HA > 10 degrees) and it started immediately and finished successfully. I was imaging in Ha and Ekos changed to L for the post-flip alignment. Then it switched back to Ha and Ekos then continued imaging where it left off right before the meridian flip. This was all with the latest stable release of KStars on Ubuntu 20,04 on an RPi 4 with 4 Gb of memory. Thanks for the fix @Jasem!

Just had another flip (first one since my last post).
Same thing - the flip goes well but the subsequent guider calibration fails - RA forward goes wrong (opposite) way.
Calibration needs to be stopped and restarted - then it works.

That was with SM 1.5.7/KS3.5.2 Stable.
INDI server on RPI4, KS/Ekos on Windows client.

Right now I am running everything on RPI4 and going to have the flip in 20 min. Will report back.

Usually these are not issue with the meridian flip itself but post actions which could fail for any reason. I will test tonight myself, but I've tested this so many times and it depends on how good the image is and the number of good stars detected.

Jasem - thanks!
I just had a flip and yes the flip itself went well.
Align went well.
But the Guide calibration again went in the opposite direction.
I let it finish and RA guiding immediately went out of control.
I stopped guiding. Restarted/calibrated and now it's guiding just fine.

Weather is good. The sky is clear and crisp. Enough good guiding stars.
The target is well above the horizon.
Guiding is better that my average.

This has been consistent for me for the past two-three months.
Sorry but I don't remember exactly when the change happened. It used to work just fine.

Jasem,
I am not sure what happened but the meridian flip used to work for me a few months ago.
Now, after the flip, RA  ALWAYS goes in the wrong direction during calibration.
When stopped and restarted - all is well.