A client only gets these ''repair'' /short deadlined tasks when:

1. Returning results on a continuous basis within 48 hours
2. Returns valid results on a continuous basis, like 99.8% of the time.

The deadlines of "repair" jobs are 40% of the original deadline, so a 10 day standard task that fails is resend with a 4 day deadline.

Given what I'm reading there is no bug v.v. this matter (ignore that guess). Let the client process them in FIFO and they'll be back in time... within 48 hours as your clients have been doing... there IS no problem. WCG identified your clients as highly reliable and thus picked these rush jobs up in the cycle so as to let them match up quicker with any waiting quorum wingman and batch completion.

cheers and thank you for crunching at WCG.

edit: How many days of work are you caching, for if you cache 3 days or more you should not be getting these short deadline tasks. If you don't want them, set the "connect every xxx days" to for instance 3 days. Work will still be fetched and reported if an internet connection is detected.

Just to clarify: another task with a deadline of October 30 just finished and a new task with a deadline of October 30 was started ignoring the ten tasks due by October 24 and 27. The question is when will the tasks due by October 24 and 27 be given a chance to run?

Yet again: the only task with a completion target of October 27 ended and a new task with a completion target of October 30 was started ignoring the 10 tasks due by October 24 and 27, five each.
Again I ask the question when will the tasks due by October 24 and 27 be given a chance to run ?
There may not be a problem but I sure have the feeling that there IS.

Unless tasks are run in high priority (seen in the Tasks view, status column of the client), the order of operation is First In First Out (FIFO), meaning the deadline is NOT considered a priority unless the client calculates that by the time it gets to these tasks they wont finish before their deadline.

IF you've sorted the "Report Deadline" column on deadline it's hard to see in what order tasks arrived so they seem to process out of order. Unless the receipt timestamp is absolute identical, you can sort the My Grid > Result Status page in the *send* order (link presorted for your convinience). Compare that to the order of processing in the client and it will likely get clearer.

Today is the 20th, so if you are receiving *repair* jobs today, they're due on the 24th (all but HCC) or 23th (HCC). If they're not repair jobs they have a due date of the 30th. The tasks with the 29th you must have received yesterday. The tasks due on the 27th I thus suspect them to be Help Conquer Cancer tasks. The exception is currently Help Conquer Cancer. They have a standard deadline of 7 days and if they are repair jobs they get 2.8 days to return. WCG sends these in random order, but when it has a repair job in the feeder, it will send them ahead of other tasks that have standard deadline. That's all it does.

Given the fact that your client is receiving repair jobs, as I outlined in my last post, your client must have been returning work on a frequent basis within 48 hours. If the client misses to do that frequently it wont get them.

If tasks are missing deadline than at any time you visit the Result Status page there would be tasks that are listed with No Reply or Too Late. Have any been missing their deadline? If not, let it crunch and zero time of your donated cycles are wasted. If your machine is an 8 threaded I7, it easily handles 30-50 tasks a day depending on what science it's processing.

So let us ascertain. If you go to the advanced menu of BOINC Manager, then Preferences, Then Network tab. What are the "Connect about every..." and "Additional work buffer ..." values? If the sum of these 2 is less than 2.00 days you're all fine and all your tasks will complete long ahead of deadline.

-- SekeRob

I know that Sekerob works very hard and I certainly appreciate the efforts put in by you guys. Nevertheless I think that every problem should be addressed whenever possible. Keep in mind that may not always be possible.
In any event I think I now understand what my problem is. That is to say my problem vis a vis what WCG took as the approach to expedite work. And again I understand that this is a valiant attempt at getting the most out of our resources.
So, normally my PCs and Macs don't have much to do but once in a while because of the nature of the work I do I need them and I need them on full blast. Also once in a blue moon I leave for 2 or 3 days and I shut my PCs off because I do not want to run them unattended while I am away.
This throws off the assumption that my systems are a very dependable resource.
In retrospect, and while I can not be sure right now, I think that each of the three cases I mentioned earlier was associated with me leaving and/or running a very large load.
When I know beforehand that this will happen I turn on the flag "NO NEW TASKS" and then the systems finish their work and that is it.
But when these events are unexpected I do not get the chance to do that.
I will keep an eye out and see if this is indeed the case. I can make adjustments to avoid the problem. I do not think that WCG can do anything about that.