The name of the developer is Dr. David Anderson. He can be found at the developer alpha mail list ;-)

The critical tab missing is network. Are your tasks still running High Priority?

Oops - sorry.

http://i702.photobucket.com/albums/ww27/Gandalf_PCHF/WCGNetpref.jpg

Are your tasks still running High Priority? Yes, Sir.

Welcome aboard, Sgt Joe.

SekeRob - done
"Connect about every xx days" temporarily to 0.00 days

If you run BOINCTasks [special multi-client BOINC Manager], you can easily see how much total computing time work is presently in your client buffer [project tab]. It's very well possible that work was called when the projected times were short, and then suddenly one ran 8 hours instead of the projected 4. The client applies this "ran 2x longer" to all work in the buffer, so when there was 1.8 days worth, the client thought there was suddenly was 3.6 days. With a steady cache of 1.8 days, it means that any task started is already ~1.8 days old with normal FIFO. With that DCF [Duration correction factor] having shot from 1.0 to 2.0, there's 3.6 days work. Try cram that in 1.8 days. Anyway, with BOINCTasks it is easy to trim the work queue by deleting the not yet started, soonest due tasks to the point that none will be overdue by the time the cache has emptied... i.e. if it's seen that you have too much work to complete before deadline, abort the soonest due [which buys some time for those due later].

If your client runs GPU work also, then your client may have fallen victim to the excessive over-caching bug. That too is fixed in the client 7 release [which WCG is working on to test and formally endorse]. All in summary, if the panic state persists, don't hesitate to abort any excess cached work. Don't let it sit there till it's overdue.