I just happened to notice this and thought some people might find it interesting. I have work units but nothing is running, and this is NOT the Boinc.exe bug where Boinc.exe takes up all processor power. Instead, task manager says 100% of the CPU is being used by system idle processes.

http://img121.echo.cx/img121/2479/seti9qp.jpg

---

-----

Are you using a laptop on batteries?

---

maybe you got a "Female" version of the client and she just wanted some time for herself? otherwise it might be a windows/system glitch

I just happened to notice this and thought some people might find it interesting. I have work units but nothing is running, and this is NOT the Boinc.exe bug where Boinc.exe takes up all processor power. Instead, task manager says 100% of the CPU is being used by system idle processes.

http://img121.echo.cx/img121/2479/seti9qp.jpg

I've heard reports of this before, and I cannot figure out why this is occuring. It does seem to fix itself after an hour.

---

BOINC WIKI

Maybe related:

On my secondary cruncher I noticed something new, today. I had two wu's from climate one of which was at 7% and a predictor wu which had a computational error. I wanted to report the predictor wu and hit Update, but nothing happened. Then I noticed lot's of strange errors in the messages tab as it wanted to connect with pirates which is down at the moment. Naturally it can't connect to pirates but it didn't just say "connect failed" as usual - but before I could get a closer look at the error messages, boinc manager lost connection to boinc.

At the time the host was crunching away at the climate wu - but now cpu usage wnt down to zero. I tried to re-connect with boinc, but no luck. Then I exited boinc and re-started it. Pretty soon the pirates error messages started again, boinc manager lost connection and cpu usage went down to zero again. After a couple of retries at this more - I thought it might be the pirates thing that did this. Btw no attempts to connect to other attached projects were made by boinc all this time - only pirates. So I suspended pirates. Well, it stopped the error messages from pirates, but it continued loosing connection and stop crunching. And still no attempts to connect to other projects, not even deference messages.

Although the host had been crunching away at that climate wu for 4 cpu days I then decided to abort it and see what would happen - I aborted the climate wu it hadn't been crunching on as well. And THAT seems to have done the trick - at least connection hasn't been lost since, connection to other projects have been attempted (and wu's downloaded) and it's crunching away.

BOINC cc 4.43 - host 995439 - climate wu in question

---

Maybe related:

On my secondary cruncher I noticed something new, today. I had two wu's from climate one of which was at 7% and a predictor wu which had a computational error. I wanted to report the predictor wu and hit Update, but nothing happened. Then I noticed lot's of strange errors in the messages tab as it wanted to connect with pirates which is down at the moment. Naturally it can't connect to pirates but it didn't just say "connect failed" as usual - but before I could get a closer look at the error messages, boinc manager lost connection to boinc.

At the time the host was crunching away at the climate wu - but now cpu usage wnt down to zero. I tried to re-connect with boinc, but no luck. Then I exited boinc and re-started it. Pretty soon the pirates error messages started again, boinc manager lost connection and cpu usage went down to zero again. After a couple of retries at this more - I thought it might be the pirates thing that did this. Btw no attempts to connect to other attached projects were made by boinc all this time - only pirates. So I suspended pirates. Well, it stopped the error messages from pirates, but it continued loosing connection and stop crunching. And still no attempts to connect to other projects, not even deference messages.

Although the host had been crunching away at that climate wu for 4 cpu days I then decided to abort it and see what would happen - I aborted the climate wu it hadn't been crunching on as well. And THAT seems to have done the trick - at least connection hasn't been lost since, connection to other projects have been attempted (and wu's downloaded) and it's crunching away.

BOINC cc 4.43 - host 995439 - climate wu in question

Are you certain that the second CPDN WU did not put you into a no more work situation?

---

BOINC WIKI

Are you certain that the second CPDN WU did not put you into a no more work situation?

No, but the first one had been crunching away for 4 cpu days while the second one was there during which time it had finished both predictor and seti wu's. The messages tab only displayed the "overcommited" and "nearly overcommited" messages

---

Are you certain that the second CPDN WU did not put you into a no more work situation?

No, but the first one had been crunching away for 4 cpu days while the second one was there during which time it had finished both predictor and seti wu's. The messages tab only displayed the "overcommited" and "nearly overcommited" messages

These messages mean no more work will be downloaded until the problem is cleared.

---

BOINC WIKI

These messages mean no more work will be downloaded until the problem is cleared.

And it was cleared soon enough to try connecting to pirates ;) But still, it has run with no problems since I aborted the climate wu's. I don't believe that the scheduler or overcommitment has anything to do with boinc suddenly going "numb" - but I'm no expert ;) I do believe, however, that whatever caused boinc stopping to work, caused the wierd messages I got, and caused the scheduler not to try other projects when it discovered that no work could be fetched from pirates AND the update attempts I made on predictor not to be processed

---

I just happened to notice this and thought some people might find it interesting. I have work units but nothing is running, and this is NOT the Boinc.exe bug where Boinc.exe takes up all processor power. Instead, task manager says 100% of the CPU is being used by system idle processes.

http://img121.echo.cx/img121/2479/seti9qp.jpg

Matt if you are running Predictor Clients versions above 4.28 abort them. They are hanging and nothing runs..

---

I'd rather speak my mind because it hurts too much to bite my tongue.

American Spirit BBQ Proudly Serving those that courageously defend freedom.