I have been running AM3P models for a couple weeks as the server would not let me get a FAMOUS model. Finally at 10 Jul 2010 3:15:40 UTC I was able to download a FAMOUS model to run with a AM3P.

This morning the last AM3P was reported in and I tried getting another FAMOUS model but only received the following message.

7/11/2010 9:25:24 AM climateprediction.net Sending scheduler request: Requested by user.
7/11/2010 9:25:24 AM climateprediction.net Reporting 1 completed tasks, requesting new tasks
7/11/2010 9:25:27 AM climateprediction.net Scheduler request completed: got 0 new tasks
7/11/2010 9:25:27 AM climateprediction.net Message from server: No work sent
7/11/2010 9:25:27 AM climateprediction.net Message from server: (reached daily quota of 1 tasks)

I am in the UTC -4 time zone so it was just over 34 hours since I recieved the FAMOUS model. To me a day has always had 24 hours. Did this change for servers running on UTC time?

I had no problem getting FAMOUS models till the day that they had many end in downloading errors like THIS ONE.

If I am not able to get another how to I move one from my laptop so I have two runnning on this system? The laptop has no problems getting additinal work. I alreay have one running on this system so I not want to move the whole BOINC folder, I want to also keep the one here running.

All I really want to do is run the FAMOUS model that I have and another on the 2nd core of this computer.

VIP has no work for a few days so this would be a good time to let two CPDN models run rether than other programs.

---

Keep on crunching Pizza@Home

To move a model, you need to move the entire BOINC folder, along with all sub-folders.
It's the same deal as with making backups.

The daily quota problem is actually with the new version of BOINC, which has changed for a lot of things.
I guess that it's still being worked on, so for the moment, it's the luck of the draw.

---

Backups: Here

You can't move just one model from the laptop and add it to the FAMOUS you already have on your desktop. The entire contents of the Boinc Data folder have to be moved. So before doing that you'd have to delete the current contents of the desktop's Boinc Data file, or back them up to be crunched later before deleting them. The folder contents you transfer have to be put into an empty folder.

What do the climateprediction.net preferences in your account say for model selection? Which model types say Yes and which say No? Before any change of your selection please tell us what it says now so we know what isn't working for you. Well, we know that your current selection works for the laptop but not for the desktop.

In the desktop's Boinc manager, go to Advanced > Preferences > Network usage tab. Check that the additional work buffer is set to the maximum 10 days. If you have to increase the buffer click OK.

If the desktop has a lot of work for any other project sometimes you can get work by suspending the tasks from the other project(s).

(You've probably already tried these ideas.)

---

Cpdn news

I think that Les has it with the new version of BOINC as that will allow a different quota for each type of unit. When I try getting any other unit I get the following as they had none.

7/11/2010 1:11:27 PM climateprediction.net Requesting new tasks
7/11/2010 1:11:41 PM climateprediction.net Scheduler request completed: got 0 new tasks
7/11/2010 1:11:41 PM climateprediction.net Message from server: No work sent
7/11/2010 1:11:41 PM climateprediction.net Message from server: No work available for the applications you have selected. Please check your preferences on the web site.

When I set the preferences to FAMOUS I get:

7/11/2010 1:51:21 PM climateprediction.net Requesting new tasks
7/11/2010 1:51:22 PM climateprediction.net Scheduler request completed: got 0 new tasks
7/11/2010 1:51:22 PM climateprediction.net Message from server: No work sent
7/11/2010 1:51:22 PM climateprediction.net Message from server: (reached daily quota of 1 tasks)

Guess I have to wait till they get time to go through all the server code to see what is causing this.

Hope that when I complete the one that is running that I can get another at that time as I do not want to stop using this system for CPDN. But other projects would like the time.

---

Keep on crunching Pizza@Home

Ray,

I've had this problem too, since upgrading from 6.10.17 to 6.10.56.

My guess is that if you wait one full day after getting the "reached daily quota of 1 tasks' message, the next time we get to midnight, UTC, the client will download a task. This is based on a sample of 2, though, so just a guess. The first step (getting the 'reached daily quota' message) has just happened again, so I will see if the pattern is followed.

I had the additional work buffer set at 1 day, but I have set it back to 4 to ensure there will be work ready to go--provided the models don't crash. Maybe I will put it up to 10, as mo suggests, to allow for FAMOUS's crash-proneness.

Greg

The new day should begin at midnight your time, not 00:00:00 UTC for the whole world.

Ray, which of your three computers can't get new work?

I can understand why your quota was reduced to 1. There were downloading errors though of course they weren't your fault. The default quota is 4 per core per day. Each crashed model reduces the quota by 1. Each successful completion increases it by 1. So 3 errors reduced the quota to 1 per core per day.

You should still have been able to download 1 or 2 models: 1 per core. But I don't think you have been able to download any new models to the problem computer on Sunday.

The server's new Boinc version is not behaving properly with respect to the quota of some computers. Milo has wanted to manually change the quota of particular computers. The new Boinc cancelled Milo's action at least three times.

I can't understand why the new Boinc version behaves correctly for certain computers but incorrectly for others.

---

Cpdn news

It is the AMD Athlon II 235 that is having problems. All the Intel systems can get two when needed.

---

Keep on crunching Pizza@Home

Well, my guess was wrong. I didn't have to wait a day this time:
12/07/2010 11:14:49 a.m. climateprediction.net Sending scheduler request: To send trickle-up message.
12/07/2010 11:14:49 a.m. climateprediction.net Requesting new tasks
12/07/2010 11:14:57 a.m. climateprediction.net Scheduler request completed: got 2 new tasks
12/07/2010 11:14:59 a.m. climateprediction.net Started download of famous_ulaj_1199_200_006660438.zip
12/07/2010 11:14:59 a.m. climateprediction.net Started download of dump_s109_20a_1199.gz
12/07/2010 11:15:02 a.m. climateprediction.net Finished download of famous_ulaj_1199_200_006660438.zip
12/07/2010 11:15:02 a.m. climateprediction.net Started download of dump_s109_20o_1199.gz
12/07/2010 11:15:03 a.m. climateprediction.net Sending scheduler request: To fetch work.
12/07/2010 11:15:03 a.m. climateprediction.net Requesting new tasks
12/07/2010 11:15:05 a.m. climateprediction.net Scheduler request completed: got 0 new tasks
12/07/2010 11:15:05 a.m. climateprediction.net Message from server: No work sent
12/07/2010 11:15:05 a.m. climateprediction.net Message from server: (reached daily quota of 1 tasks)
12/07/2010 11:15:25 a.m. climateprediction.net Finished download of dump_s109_20a_1199.gz
12/07/2010 11:15:25 a.m. climateprediction.net Started download of famous_un8c_1599_200_006662951.zip
12/07/2010 11:15:26 a.m. climateprediction.net Finished download of famous_un8c_1599_200_006662951.zip
12/07/2010 11:15:26 a.m. climateprediction.net Started download of dump_s13t_20a_1599.gz
12/07/2010 11:15:54 a.m. climateprediction.net Finished download of dump_s13t_20a_1599.gz
12/07/2010 11:15:54 a.m. climateprediction.net Started download of dump_s13t_20o_1599.gz
12/07/2010 11:16:29 a.m. climateprediction.net Finished download of dump_s109_20o_1199.gz
12/07/2010 11:17:16 a.m. climateprediction.net Finished download of dump_s13t_20o_1599.gz

So a request to get new work embedded in a trickle-up is fine, but a separately issued request for work gets an error. I had 3 cores eligible for new work and got work for 2. I don't know why the second request for work was issued - maybe for the third core? That a 'bare' request for work fails could be a problem if you have no trickles going up. Maybe this is Ray's problem?

Then Boinc decided to defer communications for 24 hours, so no trickles are going up -- 'Trickle up message pending, Communication deferred 23:50:52' on the Project tab. As I write, there are 25 trickle_up files waiting to go, the first one created at 11:23. But the 5% milestone zip files are being uploaded:

12/07/2010 5:10:05 p.m. climateprediction.net Started upload of famous_ulps_1799_200_006660987_1_4.zip
12/07/2010 5:16:06 p.m. climateprediction.net Finished upload of famous_ulps_1799_200_006660987_1_4.zip

The scheduler was probably overloaded. Boinc will have tried a few times to upload the trickles and deferred further attempts for increasing periods. But a CPDN model of mine sent a trickle up an hour ago at its first attempt.

The files upload to upload servers none of which seem to have had recent problems.

---

Cpdn news

I've been getting these messages for about a week now on one machine, and now on this one too.


7/11/2010 11:42:16 PM climateprediction.net Sending scheduler request: To fetch work.
7/11/2010 11:42:16 PM climateprediction.net Requesting new tasks
7/11/2010 11:42:17 PM climateprediction.net Scheduler request completed: got 0 new tasks
7/11/2010 11:42:17 PM climateprediction.net Message from server: No work sent
7/11/2010 11:42:17 PM climateprediction.net Message from server: (reached daily quota of 3 tasks)

Both machines (1036631 and 1070830) are shared 50-50 with CPDN and CPDN Beta, so they need only work for two and four cores, respectively, and at the moment they still have it. In a few days, the current work will start to finish, so we'll see if any new work is forthcoming. In the past (more than a week ago), they would fill up as if all cores were working 100% for each project, something that still happens with Beta. I have BOINC set to "connect every 0.1 days" and "Additional work buffer" of 0 days, so I would expect them to ask for new work only as they are nearing completion of the current work.

Greg
When that happens,set CPDN to NNW and update it. After that the trickles will go in normally. I have had that several times when I can not get a new model.

Looks like I will only be able to get a new one when I return the one that I have running. Was looking forward to running two of these. Here are my last two request for work, a little over a day apart.

7/12/2010 1:20:53 AM climateprediction.net Requesting new tasks
7/12/2010 1:20:54 AM climateprediction.net Scheduler request completed: got 0 new tasks
7/12/2010 1:20:54 AM climateprediction.net Message from server: No work sent
7/12/2010 1:20:54 AM climateprediction.net Message from server: (reached daily quota of 1 tasks)
7/13/2010 2:22:55 AM climateprediction.net Requesting new tasks
7/13/2010 2:22:57 AM climateprediction.net Scheduler request completed: got 0 new tasks
7/13/2010 2:22:57 AM climateprediction.net Message from server: No work sent
7/13/2010 2:22:57 AM climateprediction.net Message from server: (reached daily quota of 1 tasks)

I think the word daily has to be taken out, just have quota.

I can't understand why the new Boinc version behaves correctly for certain computers but incorrectly for others.

No.v
I think that the download failures set that off, boinc seems to think that it can not send them to us after that.

I am getting what has become the usual message:

climateprediction.net Message from server: No work sent
climateprediction.net Message from server: (reached daily quota of 1 tasks)

Isn’t there some way to disable this quota software in CPDN. It is obvious that this quota system DOES NOT WORK! I realize that it was probably created to deal with the mis-configured computers that download and immediately crash WU’s, but, it is annoying the heck (not my first choice of words, but, it is child friendly) out of everyone else.

I also wonder way my quota is still at 1 per day? I thought that daily quotas were supposed to increase by 1 with each successful completions. I have not crashed a WU in about 2 weeks now (knock wood) and have successfully competed 2 Famous WU’s this week, so why hasn’t my quota gone up?

---

What is being reported on computer pages, and what is being used by BOINC, are 2 different things with the new server version.
These problems are known about, and fixes are being tried.

Anyone who wants to work/comment on the BOINC server code is welcome to do so in the appropriate place on the BOINC site. The code is available for all to see. Just sign up to the developers area.

PS
The answer to the title of this thread is: Not enough.

---

Backups: Here

The current problem with work fetch on certain computers is unrelated to the problem of the misconfigured computers.

Milo wrote a script for misconfigured computers. His script a) sends an email to the owner telling them there's a problem with all the crashed models and inviting them to post on the forum to fix the problem b) sets the misconfigured computer's daily quota to -1. Boinc can't recover from this minussing. Only if the person posts and says they've fixed things does Milo then manually reset the quota to 1 and the person can try crunching again.

Well, that's what used to happen before the Boinc upgrade.

After Milo upgraded the server's Boinc version the script still sent out the email but refused to set the misconfigured computers' quota to -1.

The script for misconfigured computers was only ever applied to the particular computers that we reported for crashing multiple models. I think Milo had to type in the computer IDs manually. The script was never applied to all computers.

There's something wrong with the Boinc quota system in this new Boinc version that only seems to affect work fetch for certain (properly configured) computers. Whether the real cause is only the quota or something else as well I don't know.

---

Cpdn news

Jim, could you please give us a link to the computer that can't get new work?

---

Cpdn news

I'm not Jim, but I have a similar problem, and perhaps a clue to the answer. Take a look at this computer's tasks. Seems like it has a lot to work on, no? But the only valid tasks are the four issued on June 27-30, one of which, 11045389, just finished. The other three 11041787, 11040358, and 11040062 are still running, due to finish in the next few days. These are all like the last slab models issued before they were discontinued. There are 20 other tasks listed as in progress which never made it to my machine. Evidently they figure in the reluctance of the server to issue it more work. Thank goodness the Beta site doesn't have these issues, looks like it will take over as these slabs finish and I get no more work from the main site.

-Gene

Lots of people are getting that quota message. Usually if you wait a while (several hours) it will give you a replacement WU.

---

Thanks for the concern, but, the problem has solved itself, at least for the moment. The sever finally relented and granted me a new WU several hours later. I’ve noticed that it usually does this after several hours of waiting. It was however the same calendar day, at least here in the Eastern Time Zone of the U.S. It might have had something to do with it passing midnight GMT.

---

Genes, are you saying that if you suspend all Beta models on computer 100830 which has 8 cores you then have 3 CPDN models running, 5 idle cores and the server won't send new work? I'm assuming that those 5 cores aren't busy with tasks from other projects.

---

Cpdn news