. . Here it was up at 9:05pm AEDT (10:05am UTC) which makes it 2:05am Berkeley time, but down at 9:10 pm or 2:10am Berkeley. So while very repetitive this phenomenon is not punctual ...

. . As someone also remarked, I sure hope this time is a lot shorter outage than last week.

. . This is really making me want to spoof the BOINC client too ... I grow weary of constantly running out of work ...

Stephen

:(

I have one computer only, I do this for science not for EGO..

I have only one too. I do for the science and the fun.
Bunkering, spoofing or whatever are just different ways to have fun.

WoW is an event, a celebration, not a competition, that is what some still not understand.

No matter how much you could crunch, the important is participate in the party.

BTW this is exactly the same philosophy of SETI@Home anyone could participate, no matter the size of his cruncher or farm, from a single Pi, to a large minning cruncher, all are welcomed to the party.

What can cause the huge spikes in Munin graphics for Average turnaround? Average usually is around 30-40 hours but the spikes are around 200-350 hours. Someone is returning a lot of long overdue tasks in batches? It must be huge batches for the average to jump so high.

WOW event, bunkering.

Keeping ahead of the competition by having one machine in reserve, use GPU spoofing, download several thousand tasks before the event starts, set machine to not report tasks, release a few thousand at the start, download replacement tasks, set the machine to not report. Crunch these tasks, when done, release these in bulk to "jump ahead" of the competition. Rinse and repeat till the end of event, when all tasks, probably 4-6 thousand are released.

To do this successfully hide machines while event is on. Keeps the competition guessing.

PS I am not doing this it is what I think happens.