Hi,

in the last days I got no work from SETI@home. Strange, because I got no errormessage. The log says:

Do 1 Nov 11:08:15 2007|SETI@home|Sending scheduler request: Requested by user
Do 1 Nov 11:08:15 2007|SETI@home|(not requesting new work or reporting completed tasks)
Do 1 Nov 11:08:20 2007|SETI@home|Scheduler RPC succeeded [server version 511]
Do 1 Nov 11:08:20 2007|SETI@home|Deferring communication for 11 sec
Do 1 Nov 11:08:20 2007|SETI@home|Reason: requested by project

Why do I got no new work? Does anybody of you have an Idea?

Greetz

Frank

---

in the last days I got no work from SETI@home. Strange, because I got no errormessage. The log says:

Do 1 Nov 11:08:15 2007|SETI@home|Sending scheduler request: Requested by user
Do 1 Nov 11:08:15 2007|SETI@home|(not requesting new work or reporting completed tasks)
Do 1 Nov 11:08:20 2007|SETI@home|Scheduler RPC succeeded [server version 511]
Do 1 Nov 11:08:20 2007|SETI@home|Deferring communication for 11 sec
Do 1 Nov 11:08:20 2007|SETI@home|Reason: requested by project

Are you running any other projects? Does the log say nothing about available memory or disk space?

---

Hi.

Yes, I ran other Projects. Rosetta and Einstein. No probs there. I share the computing time equally, so every project gets 33,3 %.

Unfortunately I just got this I have copied out of the BOINC-clients message board.

However, the SETI-PArt of the HArddisk use ist the smallest. Can I raise this and will this hel§ How do I do this if possible/necessary?

Greetz

Brig

---

Yes, I ran other Projects. Rosetta and Einstein. No probs there. I share the computing time equally, so every project gets 33,3 %.

In that case itÕs most likely a [[Long_Term_Debt]] issue: BOINC thinks itÕs given S@h more than its share of time recently, so wonÕt download more work until the other projects have caught up. You can tell whether or not this is the case by looking in the client_state.xml file.

In that case itÍs most likely a [[Long_Term_Debt]] issue: BOINC thinks itÍs given S@h more than its share of time recently, so wonÍt download more work until the other projects have caught up. You can tell whether or not this is the case by looking in the client_state.xml file.

Thanks for the quick answer. As a noob in this I really appreciate the support.

I have looked into the xml-file but didn't got the clou. At least I didn't know what to look for. ;-)

The long term debt could be the solution. I started the work for Rosetta and Einstein recently, so the workunits done for s@h outnumber the units for the other two projects 6 to 1. Will that mean, that s@h units will be downloaded when the other two projects catched up?

Greetz

Frank

---

The long term debt could be the solution. I started the work for Rosetta and Einstein recently, so the workunits done for s@h outnumber the units for the other two projects 6 to 1. Will that mean, that s@h units will be downloaded when the other two projects catched up?

Yes, but BOINC doesnÕt look at the total credit or anything like that, just the time spent on active projects. The time you spent on S@h before attaching to the other projects doesnÕt count against it.

I should have provided a little more detail: the client_state.xml file can be a little intimidating. After an introductory part that contains data on the host and the time it spends working on BOINC, there will be a section for each project. The parts youÕre interested in are within the <project> tags at the beginning of each section; about twenty lines down in each one you should see a number within a <long_term_debt> tag. This is the time in seconds that the project is owed by the resource pool, so to speak: the project with the highest number here will be the next one to be asked for work (unless it has requested a communication deferral). The total of all long_term_debt values should be zero. (Likewise for short_term_debt, which BOINC uses to decide which projectÕs task gets crunched next, of those from which it has already downloaded tasks.)

It may not be necessary, but for safetyÕs sake when poking around in BOINCÕs files I like to open a copy IÕve made in another location, in case the text editor protects the file from being written while itÕs openÑclient_state.xml gets updated (and backed up as client_state_prev.xml) every few minutes. The quickest way to find the values weÕre after is to do a search for the long_term_debt tags, skipping over all the details of the science apps, work units, &c.

"I've also got this question", and the answer for me was also to look at my long term debt. One of my computers had an absurd -6e6 magnitude number as a long term debt, with Einstein I presume. The result was weeks without Seti downloads. So I zero'd the parameter and immediately got downloads. So we'll see what happens next.

This debt thing is a neat feature. If I were a project owner and very ambitious I would include in my science client a little process call which edits the client state file and makes everyone else's project own me a bundle. There doesn't seem to be any protection for this sort of hack. I assume my long term debt was corrupted during one of these upgrades/optimized client installs, but you never know!?