Hi,

My 2xQuad Xeon (http://setiathome.berkeley.edu/show_host_detail.php?hostid=3816900)is almost out of work (should have more than 1200 units in cache), but is still not requesting new work!

12.05.2008 10:28:55 Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 0 completed tasks
Scheduler request succeeded: got 0 new tasks

I'm baffled by this - any idea anyone?

Morten

---

Morten Ross

Hi,

My 2xQuad Xeon (http://setiathome.berkeley.edu/show_host_detail.php?hostid=3816900)is almost out of work (should have more than 1200 units in cache), but is still not requesting new work!

12.05.2008 10:28:55 Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 0 completed tasks
Scheduler request succeeded: got 0 new tasks

I'm baffled by this - any idea anyone?

Morten

Is it set to No New Tasks?

---

Daniel

Is it set to No New Tasks?

No, it's not.

---

Morten Ross

Hi,

My 2xQuad Xeon (http://setiathome.berkeley.edu/show_host_detail.php?hostid=3816900)is almost out of work (should have more than 1200 units in cache), but is still not requesting new work!

12.05.2008 10:28:55 Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 0 completed tasks
Scheduler request succeeded: got 0 new tasks

I'm baffled by this - any idea anyone?

Morten

Is it set to No New Tasks?

Yesterday, 11 may, had quite the opposite, all 5 hosts stopped 'crunching' about 11:45 (UTC+1).
Enough WU's in cache, 3rd time this happens, for NO apparent reason!
Use BOINC 6.1.0. and the latest SSSE3/SSE4.1 optimized app.'s.
No machine problems, would be odd, 5 hosts stopping about the same time.
Network is OK.
Out off options @ the moment?

---

Hi,

My 2xQuad Xeon (http://setiathome.berkeley.edu/show_host_detail.php?hostid=3816900)is almost out of work (should have more than 1200 units in cache), but is still not requesting new work!

12.05.2008 10:28:55 Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 0 completed tasks
Scheduler request succeeded: got 0 new tasks

I'm baffled by this - any idea anyone?

Morten

Do you have any "suspended" tasks... I don't know about BOINC 5.10.45, but the version I'm running, 5.10.30 will not request any work if any tasks are suspended.

I've also heard that re-installing BOINC can "fix" this kind of problem.

HTH,
JDWhale

---

Hi,

My 2xQuad Xeon (http://setiathome.berkeley.edu/show_host_detail.php?hostid=3816900)is almost out of work (should have more than 1200 units in cache), but is still not requesting new work!

12.05.2008 10:28:55 Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 0 completed tasks
Scheduler request succeeded: got 0 new tasks

I'm baffled by this - any idea anyone?

Morten

This is most likely a case of "too large cache-size".

Taking a look on hostid-3816900 shows work downloaded 23.04 is returned today, this is 18-19 days turnaround-time, and this indicates you're likely running with a 20-day cache-setting. If isn't doing anything special, this likely means "connect every 10 days" + "additional 10 days cache".

Now, having a large "additional N days" isn't normally a problem, but the "connect every 10 days" basically means "work with less than 20 days till deadline is in deadline-trouble..."

If atleast one task in a project is in deadline-trouble, this shows-up in client as "Running, high priority", work-requests to project is blocked until not in deadline-trouble any longer, or idle cpu...

Your task-list shows atleast one task has deadline 17.05, so if "connect every N days" is set to larger than 2.5 days, you're in deadline-trouble, and work-request is currently being blocked.


To fix, if you have a permanent internet-connection, set "Connect every N days" to 0.1 days or something. You can still keep a fairly large cache by keeping "additional N days" at 10.

It's still possible work-requests gets blocked with these settings, but normally it's for a much shorter time than if you're running with a high "connect every N days"-setting.

---

"I make so many mistakes. But then just think of all the mistakes I don't make, although I might."

Hi,

My 2xQuad Xeon (http://setiathome.berkeley.edu/show_host_detail.php?hostid=3816900)is almost out of work (should have more than 1200 units in cache), but is still not requesting new work!

12.05.2008 10:28:55 Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 0 completed tasks
Scheduler request succeeded: got 0 new tasks

I'm baffled by this - any idea anyone?

Morten

Do you have any "suspended" tasks... I don't know about BOINC 5.10.45, but the version I'm running, 5.10.30 will not request any work if any tasks are suspended.

It also won't request any new work if you have a stalled download on the 'Transfers' tab.

I've also heard that re-installing BOINC can "fix" this kind of problem.

That's a bit drastic! Probably best to work through the other possibilities first, and keep this in reserve just in case....

I've heard of people who re-format their hard disks and re-install Windows for less than this: I call it the "sledgehammer and two short planks" school of computer maintenance.

Hi,

Down to only 2 WUs in the queue and the magic happened - now it's downloadning like never before, and will propably not stop before 800 WUs (which is the daily limit).

I have not changed my settings since 7 November 2007, so nothing on my part should have caused this. Furthermore, I have 9 other hosts where this has not happened.

Problem solved for now...

Morten

---

Morten Ross

Re-installing BOINC is not at all drastic. I have "upgraded" to 6.1.x and back to 5.x.y several times with no loss of work and only the slightest cost in time.

With the new "sandbox" technology you will have to reinstall to set permissions quite often if you make changes within the BOINC directories.

---

With the new "sandbox" technology you will have to reinstall to set permissions quite often if you make changes within the BOINC directories.

On a Mac I just correct the owner and group settings in the Finder’s Get Info window; I imagine Windows & Linux have similar functions—available from a command line at least, if not from a GUI.

---

I'd think a 20 day cache is a bit excessive O_O. Also from personal observation, the Boinc client will not request additional work without also reporting as completed a portion of your cache.

I'd think a 20 day cache is a bit excessive O_O. Also from personal observation, the Boinc client will not request additional work without also reporting as completed a portion of your cache.

You're correct on all counts.

The reason is that BOINC is looking at the due dates, and is "worried" about reporting on time if it isn't supposed to connect more than three times per month.

Setting "connect every 0.5 days" and then setting the additional cache will allow BOINC to "top up" every day.

---

It really depends on the deadlines. If you have "connect every X" set to 10 days, the CPU scheduler knows that it has to finish all work at least 10 days before the report deadline so that it can be reported before you disconnect your computer from the internet for 10 days. If you also have additional work set to 10 days, you have 20 days of work to get finished 10 days before the deadline.

If the deadlines are significantly longer than 30 days, there is unlikely to be any problem. However, if deadlines are less than 20 days, you are going to return work late. With mixed deadlines, it is possible to get the CPU scheduler in a bind with a queue that is too long.

If you connect to the internet more frequently than once every 10 days, please drop your connect every X setting to something more reflective of reality.

---

BOINC WIKI