I'm runninng BOINC Manager 5.8.11 on a PowerBook G4, Mac OS 10.4.8. My network connection is via EVDO wireless and is not on all the time.

When I finish a task and connect to the Internet, the result is uploaded, and when I press Update, the result is reported, but no new work is downloaded. Here is the sequence of messages:

Tue Feb 20 21:42:19 2007|SETI@home|[file_xfer] Started upload of file 09ap00aa.21686.29762.111092.3.138_3_0
Tue Feb 20 21:42:24 2007|SETI@home|[file_xfer] Finished upload of file 09ap00aa.21686.29762.111092.3.138_3_0
Tue Feb 20 21:42:24 2007|SETI@home|[file_xfer] Throughput 4691 bytes/sec
Tue Feb 20 21:44:33 2007|SETI@home|Sending scheduler request: Requested by user
Tue Feb 20 21:44:33 2007|SETI@home|Reporting 1 tasks
Tue Feb 20 21:44:43 2007|SETI@home|Scheduler RPC succeeded [server version 507]
Tue Feb 20 21:44:43 2007|SETI@home|Deferring communication for 11 sec
Tue Feb 20 21:44:43 2007|SETI@home|Reason: requested by project
Tue Feb 20 22:18:16 2007|SETI@home|Sending scheduler request: Requested by user
Tue Feb 20 22:18:16 2007|SETI@home|(not requesting new work or reporting completed tasks)
Tue Feb 20 22:18:22 2007|SETI@home|Scheduler RPC succeeded [server version 507]
Tue Feb 20 22:18:22 2007|SETI@home|Deferring communication for 11 sec
Tue Feb 20 22:18:22 2007|SETI@home|Reason: requested by project

Why does it say, "not requesting new work" in response to my scheduler request? I haven't clicked the "No new tasks" button, the percentages listed under my computer are high (60%-ish).

No problems with my other computer, which is connected via cable modem and is also running 5.8.11. I think this problem began after I installed 5.8.11. No problems with this wireless setup before that. But now, it seems the only way I can get new work is to detach from the project and then attach again - which seems drastic. This is a problem with both SETI and Einstein.

Thanks.

---

Your Computer has enought work to compute for some time indipendendly and did not need to download any new work.

---

Your Computer has enought work to compute for some time indipendendly and did not need to download any new work.

No work was shown under Tasks when the message "not requesting new work" was sent. Therefore, I thought the machine had no work. Can tasks be present but not listed?

Shortly after I posted my question, the machine did download some work.

---

No work was shown under Tasks when the message "not requesting new work" was sent. Therefore, I thought the machine had no work. Can tasks be present but not listed?

“Never say never,” but no, that shouldn’t happen. Are you attached to any other projects?

Perhaps you tried to update the project while BOINC was in some kind of deferral state—it will normally request more work (and report any completed results) when a current task is nearly done. I haven’t tried v5.8.x yet, but I gather the work scheduler was given an overhaul. Keep an eye out for further occurrences of this kind, and post back if you see any.

---

I'm getting the same "(not requesting new work or reporting completed tasks)" message now and I'd like to know if there's a way to clear what the SETI server thinks my computer is working on from its queues.

In the middle of working on some units, my iBook regurgitated on a bunch of lousy software (Microsoft Offic x. V for Mac), so I ended up wiping the drive and reinstalling the OS. So, I know there are no SETI units on this machine. How do I get the server to zero out what it thinks I have on it?

Thank you.

---

I'm getting the same "(not requesting new work or reporting completed tasks)" message now and I'd like to know if there's a way to clear what the SETI server thinks my computer is working on from its queues.

Well, nothing’s going to get sorted out properly for a few days because of thumper’s demise. See the front-page news; it’s probably best to suspend the project in BOINC Manager until Monday or so, because you’re unlikely to get any work anyway. Check the home page and the Technical News forum for progress on the replacement server.

Meantime, I’ll ask the obligatory first questions the above message prompts: Are you running any other projects? What does the Projects tab show for S@h’s status? Are there any WUs in the Tasks tab?

In the middle of working on some units, my iBook regurgitated on a bunch of lousy software (Microsoft Offic x. V for Mac), so I ended up wiping the drive and reinstalling the OS. So, I know there are no SETI units on this machine. How do I get the server to zero out what it thinks I have on it?

If your results are truly gone, there’s no way to clear them from your end. When their deadlines expire they’ll be labelled “No reply”, and sent to someone else if required to make a quorum. Nothing to worry about. Regardless, not requesting work is a decision of your BOINC client, and AFAICT the status of WUs on the server is irrelevant to it.

I'm getting the same "(not requesting new work or reporting completed tasks)" message now and I'd like to know if there's a way to clear what the SETI server thinks my computer is working on from its queues.

Well, nothing’s going to get sorted out properly for a few days because of thumper’s demise. See the front-page news; it’s probably best to suspend the project in BOINC Manager until Monday or so, because you’re unlikely to get any work anyway. Check the home page and the Technical News forum for progress on the replacement server.

I was aware of the server downtime, so I wasn't concerned at first. But when my older computer picked up some new WUs today, and my iBook didn't, that's when I decided to take a look as to why. I can understand the continuing situation, so I'll suspend my older computer's work until the server situation is improved.

Meantime, I’ll ask the obligatory first questions the above message prompts: Are you running any other projects? What does the Projects tab show for S@h’s status? Are there any WUs in the Tasks tab?

Well, right now my iBook's BOINC manager has no projects, tasks, or WU's status at all. It's a blank slate. I was running only the SETI project with it and in an attempt to try to clear the information the server contained for this particular computer, I detached from the project. Upon looking at my statistics after the detachment, I see the server still thinks I have this iBook attached to the SETI project, which in reality, I don't.

But prior to my detachment, the manager displayed only information that I was attached to the SETI project. Otherwise there was nothing else; no tasks, no transfers, but a bunch of messages saying that my client wasn't asking for any more work (as stated in my original post).

In the middle of working on some units, my iBook regurgitated on a bunch of lousy software (Microsoft Offic x. V for Mac), so I ended up wiping the drive and reinstalling the OS. So, I know there are no SETI units on this machine. How do I get the server to zero out what it thinks I have on it?

If your results are truly gone, there’s no way to clear them from your end. When their deadlines expire they’ll be labelled “No reply”, and sent to someone else if required to make a quorum. Nothing to worry about. Regardless, not requesting work is a decision of your BOINC client, and AFAICT the status of WUs on the server is irrelevant to it.

OK. I'll wait for two things to happen before I try to get SETI running on this iBook again. I'll wait for the server replacement work to be completed, and I'll wait until the server's stats for this particular machine shows that the deadlines for the old (but no longer existant) WUs have expired.

Thanks for your responses. Please have a pleasant evening.

---

Why does it say, "not requesting new work" in response to my scheduler request? I haven't clicked the "No new tasks" button, the percentages listed under my computer are high (60%-ish).

No problems with my other computer, which is connected via cable modem and is also running 5.8.11. I think this problem began after I installed 5.8.11. No problems with this wireless setup before that. But now, it seems the only way I can get new work is to detach from the project and then attach again - which seems drastic.

I had (and am still having) the same problem (see "Even Before Thumper Died" from 07 May). Ever since I installed 5.8.15 the number of work units my fastest computer keeps in queue dropped. (Consequently, when Thumper died, my 400MHz laptop had several days of work left but this one ran out almost right away.)

I've tried setting the # of days to something much smaller than the '10' I was using and nothing changed. It still only downloaded one or two at a time. Hardly more than twenty hours of work was cached at any one time while both my laptops keep several days worth (with deadlines weeks in the future).

Right now I have one work unit that this machine is working on with a report deadline of 11 June. It has less than ten hours left until completion but BOINC is still telling me "not requesting new work".

Can't say as I 'get it'...

---

IBA# 11465
http://imagesdesavions.com

I just had this problem with World Community Grid but SETI continued to process data. I tried every thing I could think of other than reloading the software or detaching from the project. I then detached and reatached the project and World Community Grid started functioning again. I suspect something got messed up in the information that tracks the project but I have no idea what. Someone who works with the software may have a better idea what is going on but also note that a project reset did not clear the problem.

---

This behavior is OK. The BOINC client downloads work and computes them, depending on the resource shares you have set up. If a project has done enought work, the BOINC client computes the other project.

---

I am into each project for 100% of the work and the only time I didn't have work for both projects, SETI would not give me work. SETI had been up and working will for almost two weeks before this problem happened. I can think of no good reason for the work load to be shifted under the above conditions. I suspected it might be priority processing but after 3 days of unbalanced processing, something is not right. I have run two projects for aroumd 9 months and have not seen this unexplained work hold before. Also not noted before, I am running 5.8.17 software.

---

Sorry, I did not excatly understand what you mean.

BOINC balances via a short and a long term debt between the projects. So, it could be if a project (for example project B) gets more work, because one project (example project A) is offline or has no work, the BOINC client only downloads work from a project (B) wich has work. Later, if both projects has work, the BOINC client gives more attention to the project (A) which has had no work before.
The goal is to keep your client bussy, and later to fit your resource shares.

There are some articels in the BOINC Wiki which describes this in detail :
http://boinc-wiki.ath.cx/Work_Scheduler
http://boinc-wiki.ath.cx/Resource_Share
http://boinc-wiki.ath.cx/Work_Buffer

---

In my case, WCG just stopped requesting work but both projects had been running fine for the last two weeks. If there was an imbalance, it should have cleared in a few hours. I also understand how the work is balanced and I even have had a lockout because of a software bug that caused a work unit to take over 30 hours of processing time and was about to hit a deadline. Most of my work units run between 1 and 8 hours with SETI taking about 4 hours or less. With a dual core processor and the current level of Boinc running two projects, normal is one project on each core. If I don't see that configuration, I know something is wrong and I need to go looking for the problem. Besides that, if a project lacks work, they shouldn't get more processing power than normal when they do get work. That type of thinking could put many times the load on the server than normal and cause problems like the ones SETI has been seeing. I need to write a profile because I have had 35 years of working with software and have become very good at spotting software flaws. I also have also learned when to keep my mouth shut because I didn't know what I was talking about. In this case, something went strange in the software. I suspect the problem is in Boinc but it is possible some other program stepped on a file used by Boinc to keep track of work balances. I have not looked at the Boinc software because it would take me a while to come up to speed. Most of my work is with Assembler and sometime I get a little Fortran (not for a while) for fun. The biggest problem with the internet is that you have to judge a person and their abilities by a few lines that they have written. People are far more complex and can take a life time to really get to know them.

---

The balancing is done with a short term debt and a long term debt...
I did not see any problems here, only the normal LTD behavior of the BOINC client. The BOINC client is designed to configure and let it running. The BOINC client did everything (STD/LTD scheduling) for you (meat the deadlines, fit the resource shares, debts,....).

---

I reviewed the links you provided and I see nothing that indicates the catch up is delayed. From what I read, as soon as work units are available, Boinc starts to even thing out. In my case, it had two weeks to get thing cleaned up and then it locked WCG out of processing. Also, with my settings, I have not seen work delayed even when a project didn't provide work for a while. What I have seen is when work becomes available, both projects run at the same time. Also, my system is turned off for the night and balancing may not be written to disk. The other question is how close do the link match reality. I have learned that documentation is always out of date and how you would like it to work is not always how it ends up. The only true way to understand what code is up to is to look at it and see what it is really doing. I admit I have not looked at the code, but a feature that acts this strange should be documented better than it seems to be. Note: Old programmer saying, it's not a bug, it's a design feature.

---

Both, the download and processing of the WUs are delayed and handeled by the BOINC client as the BOINC client mean to do it right. The things you mean affects the STD. But you forget the LTD, which handle the schedules and downloads on a very long time base. - So there is NO problem, as you think, the BOINC client is excatly doing what it was desinged for.
If you turn of your system this did not effect any status writing or the state of the WU. The state would be written periodicly to the disk and readen, at the start of the BOINC client.

---

I hope this is not what your problem is, but I just learned this recently. If S@H is set to run when computer is inactive, and you only monitor it when it is active, and try to download new WUs when the computer is active, BOINC Manager 5.8.17 won't download new WUs until the computer becomes inactive, and then it does download new WUs.
Hope that helps, or someone else figures out an answer for you. I wish you well.

Todd