Just processed the data from Jobs 1533, 1534 & 1903 (replacement O3 data). A two-pane mosaic of the Veil nebula complex.
I did not inspect the new O3 in too much detail, I just wanted to see whether we had the basis of a good image now. @NoMoreCloudyNights may have some views on the quality.
I did experience some difficulties which forced me to change my usual workflow.
Normally, I would combine each filter for the whole mosaic separately, and then use ChannelCombination or PixelMath to create the colour version.
When I did this, I got an image with a very clear red cast on one panel and a green cast on the other panel. Very obvious and I could not find a way to get rid of it. So instead, I combined the channels for each panel then merged the two panels into the mosaic - all while keeping the data linear.
This gave me a mosaic with an even background, which I could then process - denoise, MaskedStretch, Starnet2 to separate stars from nebula, GeneralizedHyperbolicStretch, a bit of CurvesTransformation, a touch of LocalHistogramEqualisation and then add the stars back. Then out to Affinity to warm the red hydrogen a touch.
Not perfect by any means, and if you pixel peek there are artefacts, but a satisfactory first process.