Check that the cp-31 spray arms are not plugged is another issue I encounter with my machine. Likely not the root of your current problems.
Consider pulling the test paper from the middle of the box; paper on the outside sometimes performs differently.
Process an unexposed piece. It should come out white.
Process a flashed piece; it should come out black. These two tests are a worst case test of screwed chems.
I know this is wierd, but I run my tank sequence as water prewet, dev, blix, and then on to the wash/dry module. The digital thermostat died in my tank 1, and so I tried it this way. It seems to help the developer last longer, because the dye overcoat gets washed off before the paper hits the developer.
I also ran into red staining when I first made the switch to dev in tank 2 because of blix residue in the lines of the second tank pump circuit despite a good rinse. I moved the first rollers into tank 2, and scrubbed the former blix rollers and frame and transfer rack clean to use as the pre-wet tank set in tank 1.
I was able to solve the problem beciase I remebered the red staining I would get back in my tube processing days from blix residue getting left in the tube from a poor rinse and wipe out and then mingling with the developer on the next use of the tube.
I also took the CP31 machine apart and removed the replenisher lines, since at the time I was not auto replenishing. All sorts of mungy chem residue loved to hide in the replenisher lines when I first got my (very dirty when I aquired it) Fujimoto.
I have learned it takes very small amounts of blix in the developer to really screw it up. I take care to fill the blix tank first, wipe up any splashes, and then fill the developer tank next door. I use separate labelled funnels for each tank.
I hiope these ideas help you in your trouble shooting.