The other thread is becoming too long for comfort, but I've been getting the dreaded 'no work from project' message again today 5th Nov, all day long, on 4 different computers.

It eventually trickles through enough that none of my pcs actually 'run dry', but it's getting close (I got around only 2 WU spare on each pc). The number of times I'm getting the message has actually increased throughout the day.

---

Team CFVault.com
http://www.cfvault.com

I increased the size of the work units 2x to try to reduce the rate of results going out. Lets see if that helps and if not, our memory upgrade should in the near term. I may increase the size some more for the next batch.

Cheers! I'll post in this thread if I still see those messages in a few days.

---

Team CFVault.com
http://www.cfvault.com

I increased the size of the work units 2x to try to reduce the rate of results going out. Lets see if that helps and if not, our memory upgrade should in the near term. I may increase the size some more for the next batch.

I'm getting these no work messages on all of my computers.

---

I'm getting these no work messages on all of my computers.

Are you getting any work? enough to keep your cpu(s) busy?

I get the message also, but it is only for a couple of minutes/retries and then the WUs are downloaded ok, so really I do not have a problem with the message.

---

You can increase the size of the WU's all you want David & people wil still get the No Work Message from time to time for any number of reasons. I see it all the time myself but I never run out of work so I don't pay much attention to it.

Woops double posted

---

Join us in Chat (see the forum) Click the Sig


Join UBT

Best way is to sign up to 2/3 projects and if you want Rosetta as the main one just set the resource share higher so it downloads more Rosetta than any others that are on your projects tab

---

Join us in Chat (see the forum) Click the Sig


Join UBT

deleted...no problems seen here...

[Edit] I botched my previously posted screenshot. My BOINC manager shows success yesterday and today. Are you dropping your internet connection? My aDSL used to do that when I had an internal aDSL modem. No problems with the external modem.

Click on the thumbnail and click again on the enlarged screenshot to see where to check the messages.


[/Edit]

---

Are you getting any work? enough to keep your cpu(s) busy?

I usually get enough to keep my CPUs busy, but one of them was idle a while ago as it could not get new work. I could download work units from other projects, but prefer not to as Rosetta is my pet project.

I think part of the problem is that the settings for Rosetta let people download hundreds of work units they won't finish for weeks. I won't single anyone out, but I have seen quite a few relatively slow hosts with well over 100 work units "in progress".

They might struggle to finish them within a month while there's other CPUs sitting idle, that could return the work units within a few hours. Who benefits from that situation?

Perhaps there's some way you can limit this apparent hoarding.

---

*** Join BOINC@Australia today ***

Yup I've notice the WU-hogs as well... If they are responsible for my constant 'no work from project' messages - then I'd really like if the admins could reduce the allowed # downloaded.

---

Team CFVault.com
http://www.cfvault.com

before blaming them... might as well be the ghost wu problem someone spoke about some time ago. the wu's might never have reached the host(s).

---

--
Florian
www.domplatz1.de

Yup I've notice the WU-hogs as well...

No work unit hogs allowed! ;)

Seriously, though, a restriction to a "reasonable" amount of downloaded work units makes a heck of a lot of sense. :)

---

Regards,
Bob P.

before blaming them...

That's why I won't single anyone out and why I called it "apparent" hoarding, although I haven't seen many threads about ghost WUs on Rosetta.

My point however is that the following factors combined can result in someone downloading hundreds of work units that won't be finished for a long time. I'm sure that waiting weeks for a work unit to be returned does not help the project. Why wait when there are plenty of hosts waiting to turn the WU around in a matter of hours.

• BOINC allows one to request up to 10 days worth of work units.
  
• Rosetta has a quota of 1,000 work units per CPU per day
  
• Rosetta WUs have a 1 month deadline
  

Compare that with Predictor and SETI.

I would shorten the deadline and reduce the quota. 2 weeks should be ample, even on the recommended minimum system.

---

*** Join BOINC@Australia today ***

before blaming them...

That's why I won't single anyone out and why I called it "apparent" hoarding, although I haven't seen many threads about ghost WUs on Rosetta.

I seem to recall that the latest versions of BOINC have eliminated the ghost WU problem....if I recall correctly! :)

---

Regards,
Bob P.

...Why wait when there are plenty of hosts waiting to turn the WU around in a matter of hours...
I would shorten the deadline and reduce the quota. 2 weeks should be ample...

You have my vote; I concur at least in principle.

---

Larger WUs and reduced quotas are fine, but
Please don't shorten the deadline by much, unless you need to for science reasons. At least until the BOINC client stops asking for way too much work given resource share and work estimates and deadlines (on multi-project machines with anything more than a .1 day connection interval). With longer deadlines at least there's a chance to keep crunching on the project when the client would otherwise force itself into EDF for no good reason.

p.s. not sure how any of the above is relevant to getting no-work-from-project when there are 60K+ results waiting to be sent. This tends to clear itself up after retries, so is memory constraint the cause?

p.s. not sure how any of the above is relevant to getting no-work-from-project when there are 60K+ results waiting to be sent. This tends to clear itself up after retries, so is memory constraint the cause?

With the project undergoing maintenance shortly (see announcement on project home page), it appears they are implementing some solution, which is great news! :)

---

Regards,
Bob P.