I know there was some downtime on 19th March (12 days ago) but is this still happening now?

I usually have a few days worth of WU in hand for downtime periods but currently at the moment I have nothing on my main machines just SETI.

When pressing Update in BOINC I get these messages.

31/03/2010 15:52:10 Milkyway@home update requested by user
31/03/2010 15:52:11 Milkyway@home Sending scheduler request: Requested by user.
31/03/2010 15:52:11 Milkyway@home Not reporting or requesting tasks
31/03/2010 15:52:16 Milkyway@home Scheduler request completed

Looks like you are CPU crunching. BOINC will "add up" the WUs from all CPU projects when you request an update to check to see if there is time to crunch everything compared to deadlines. The process is fine in theory (preventing mega downloads that will never meet deadlines) but has its limitations when running multiple projects.

I suspect the total WUs on the machines is such that BOINC adds it all up and says "no more you will not finish them in time".

Two main ways to get round this that usually resolves multi project anomalies:

1. Look at preferences in BAM and increase the cache level
and/or
2. Suspend a CPU project whilst you update Milkyway

Looking at your BAM message, I suspect in your case increasing the cache will "force" a download, difficult to say from here. Be a little careful when you do this, as any other running project is liable to spot the increased cache and say "ooooh lovely, more room" and try to download some of its WUs and may beat you to it (!). Looks like you only have SETI lurking in queues so this should be ok if you suspend SETI whilst you update Milkyway.

BOINC is still steeped in the past of single projects, and whilst they are changing things, its not a quick process - regretably .... meanwhile mend and make do manually.

If its just SETI you run in addition to Milky Way, I suggest for the future, set a cachelevel that allows a good SETI buffer of X days, whatever size suits you, then set SETI for no more work. You then work through the buffer, refilling as you feel is needed, keeping it as no more work meantime. Whilst this goes on, keep milkyway running and available for downloads as it will only give a restricted ammount. This way your SETI cache is kept up, and there will be room to accept milkyway WUs as and when needed.

Regards
Zy

Thankyou, suspending SETI and doing an update did bring new WU down (6).

Suspending a project to force a download from another is never a good idea. It just makes the long term (LTD) debt worst which mean you are just delaying the problem. If you feel you are not getting enough work from a project the best way is to increase its resource share (RS). Boinc is design to handle multiple projects and it give each its fair share share of time based on the RS you set. This does not mean that you will always have work from each project on hand all the time, but it will always balance out in the long run. When crunching with a short deadline project in the mix you should set your cache to no more then that half the time of that deadline. Given reasonable settings, Boinc is capable of handling thing on its own.

---

Questions? Answers are in the BOINC Wiki.

Boinc V7.0.27
Win7 i5 3.33G 4GB, GTX470

The RS is 80% Milkyway, 20%SETI

I have just checked my other few clients and they are in the same position of just having SETI work and no Milkyway WU.

One of the clients had run out 4 days ago of Milkyway WU but was still just collecting SETI WU as early as 1 hour ago.

Suspending SETI and updating Milkyway got me 6 new WU and then resumed SETI.

Is your work buffer less than 2 days?

Running dry again...

Of course
Server Status/results ready to send = 0
may be relevant to this discussion.

---

Cheers,

PeterV

.

Oh, didn't know that was discussable. Thought more of it as a force of nature...

Like the fact that roofworkers need to cause a circuit break at least once a day.

Sigh...

Oh, didn't know that was discussable. Thought more of it as a force of nature...

Like the fact that roofworkers need to cause a circuit break at least once a day.

Sigh...

Sorry about all the crashes. Working as fast as we can to get everything moved over to the new application, which should fix the issues we're having with the server's file system not being able to keep up with you guys.

---