I have the Resource Share percentage set to 10% in the individual project preferences, for all 12 projects I'm running.

For some reason MilkyWay stands out with the RS stuck at 100%. I have rechecked this a hundred times, reset the project, detatched-reattached, even loaded Boinc on an entirely new computer, and MilkyWay remains steadfast in its stated desire to use 100% of my "resources".

I have to wonder if this is just a minor bug, or is this project just being "piggish". I realize they have a whole lot of data to be processed. Suffice to say it is as "infinite" as the universe they are trying to map. But THIS "resource" really belongs to me and it is not theirs for the taking.

So I will assume this is a bug, and I sincerely hope they will fix it soon. Otherwise, I am going to have to stop running MilkyWay. Don't want to but that's the way I roll.

n2n

Maybe your other projects don't have any WU available and that's why MW is using 100%. I've noticed the projects I'm a part of have been low on WUs lately.

Hi,

First thing to remember is, setting all of your projects to 100% is the same thing as setting them all to 10%. These are relative settings.

Do you use an account manager such as BAM (BoincStats)? If so, check the resources settings there.

Do you have multiple venues set at Milkyway such as Work, Home, School? If so, make sure the correct one is set the correct way.

I have Milkyway set to 10% and it works correctly.

-Mike

Hi,

I've found that that latest batch of MW WU's seem to be targeting "High Priority" in that they estimate a long 'to completion time' but a very short "required by" date (kind of like 273 estimated hours to completion by the morning after next). That has a habit of putting the WU into a "High Priority" Status in BOINC and nothing else will run until the MW WU completes (I generally abort when I see those types running). How do I prevent MW from hogging the CPU runtime (by forcing BOINC to set a "High Priority" status) ?

I'll say this gently......
IT AIN'T GONNA HAPPEN
They refuse to fix the problems.
They refuse to admit the problems.
I truly wish they would just take care of these issues but whoever is in "charge" says "NO WAY" for some unknown reason. I just don't get it.

Here's an over 300 day old post I started that relates to this very same thing.

HERE

I came back to see if anything has changed since then and I can see from just the first few posts on page one of the Number Crunching forum, nothing has.
And....looks like it's not going to anytime soon.

Yo-

---

A lot of the perceived problems that are heaped on MWs doorstep, are in fact aspects they can do little about. Many assume a lot of aspects are "Project" issues, they are not, its a BOINC infrastructure issue. Its when the BOINC layout and split of responsibility is fully understood, that it comes clear where the problems originate, and its rarely Project Staff.

BOINC Infrastructure
Some 80%+ of the whole BOINC infrastructure is provided by BOINC - not Project teams. BOINC hands out the various types of Server Software to the Projects to run on their hardware at Project locations. The updating and maintenance of that software is a BOINC responsibility, Project Staffs cannot amend Core Code. They can and do use the multitude of configuration options, but amending Core - despite being against the licence - is also fraught with issues as the next BOINC Server update may well overwrite any unofficial code changes.

So in terms of the overall layout, all the BOINC Servers, Schedualars, BOINC Wrapper to run Project Wus inside etc are all provided to Projects, they do not maintain them. In operations Projects are responsible for generating work units and presenting them to the work schedular. At that point, by necessity, all Project responsibility ends. BOINC is then responsible for picking up the WUs, transporting them to the Crunchers BOINC Manager, and firing up the WU exe file.

Project Responsibility
On completion of Crunching, BOINC picks up the result in BOINC Manager and sends it back to the Projects. At that stage the projects re-enter the process and do whatever they need to do with the WU results.

So if a WU keeps falling over/corrupted, yup its Project staffs to sort out where its happening, if necessary letting BOINC know of any clash with infrastructure software - but thats it, the rest is out of their hands apart from using server configuration options to help. But if those options do not resolve issues, its down to BOINC to develop and send out solutions.

Every issue on this thread is a BOINC issue, not a Project one. Of course Project Staffs try to do what they can, its in their interests to do so, but there are limits, they dont own or maintain the BOINC Infrastruture/Server code. In any case they have their hands full doing Project work, BOINC is contracted to provide everything else.

That lot's probably going to cause an avalanche, but whether it does or not, reality will not change. At the end situation will still be the same, these are BOINC issues - not Project ones.

So everything visible to us is all BOINC, the BOINC Manager, the facilities inside the BOINC Manager, tracking WUs, timing tracking, yaddie yadda, not Project ....

Regards
Zy