I took a look and wasn't able to see any issues in particular - but I might not be looking in the right place. Can you give me the workunitid or resultid of one of the tasks that you see as stuck in PV?

Think it was one of your clog events. There was a bunch in PVal ever building, without wingman, this occurred after an Invalid (it was really a stuck task abort) which may have ticked something off. Then that whole pile changed to PVer, i.e. a wingman got issued [most annoying]. With the rate of results being churned, am now about 600 along, all issued with wingman and hundreds validated since, but through 10:31 UTC was still getting new with wingman on FAHV. Go figure.

Learning point (actually a revival from the memory crofts): Disable workfetching the moment an invalid is recorded, but I only notice if my tool ran a new result fetch and wait until whatever number is needed 10-20-100-200? turn valid, before allowing new work.