My total credits for one host flatlined since Sept 15. The Stats graph in BOINC manager shows 9 data points (little circles) at the same value.

I see no error messages in the log. I see these in the text file (stderrdae.txt):

2006-09-07 09:08:24 [climateprediction.net] Scheduler request failed: HTTP internal server error
2006-09-10 04:26:59 [climateprediction.net] Scheduler request failed: HTTP internal server error
2006-09-10 04:28:55 [climateprediction.net] Scheduler request failed: HTTP internal server error
2006-09-10 04:30:56 [climateprediction.net] Scheduler request failed: HTTP internal server error
2006-09-10 04:32:01 [climateprediction.net] Project is down
2006-09-17 16:33:07 [climateprediction.net] Scheduler request failed: HTTP internal server error

But note that I kept getting credit up to the 15th, and almost one data point per day since then but with no credit.

The model date and times are incrementing normally in the display, and CPU% complete is advancing in BOINC manager GUI. Other projects are running fine, and this project is running fine on another computer.

The computer ID is 462145.
I believe the corresponding results ID is \"http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=462145\". As you can see, the last result is 9/15. Yet I see the following in stdoutdae.txt:

2006-09-25 09:27:21 [climateprediction.net] Sending scheduler request to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi
2006-09-25 09:27:21 [climateprediction.net] Reason: To send trickle-up message
2006-09-25 09:27:21 [climateprediction.net] (not requesting new work or reporting completed tasks)
2006-09-25 09:27:26 [climateprediction.net] Scheduler request succeeded
2006-09-25 09:34:13 [climateprediction.net] Sending scheduler request to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi
2006-09-25 09:34:13 [climateprediction.net] Reason: To send trickle-up message
2006-09-25 09:34:13 [climateprediction.net] (not requesting new work or reporting completed tasks)
2006-09-25 09:34:18 [climateprediction.net] Scheduler request succeeded

It would be nice to see the WUID, Result ID, computer ID, and etc. in the logs, in case that\'s what\'s messed up somehow.

Thanks,
Jeff

---

The date of the 15th is suspicious. It\'s about when the latest Windows update occurred.
This may have messed up your firewall settings, stopping BOINC from getting through.
Have a look in the folder for the model in question, and see if there are any files whose names start with: trickle_up_

If not, it\'s possible that your computer finally uploaded them on the 25th at 9.34am, and the data just hasn\'t shown up yet on the server.

Thanks for the reply.

Other projects are getting through fine, so I doubt it\'s a firewall issue. There are no trickle_up files in my \"Program Files\\BOINC\\projects\\climateprediction.net\" folder. Furthermore, there are no errors in the log indicating communications problems.

Furthermore, when results are NOT reported (which often happens for CPDN, simply because the model hasn\'t reached a suitable point), the stats graph shows no little circle representing a data point. In my case, I\'m getting little circles almost daily (same frequency as before, roughly) but no advance in credits. This could be a red herring; perhaps the circles don\'t represent reports but rather runs of some CPDN process of gathering stats. I saw 9 such data points when I posted two days ago; now I see 11.

On my model, progress has increased by over 1%, and the date & time keep marching along -- I\'m up to 01/06/1972.

Right now my best hypothesis is that somehow the computer ID is goofed up so the credit is simply going into a wrong bin. If that\'s the case, it really doesn\'t matter and I\'ll let it run to completion rather than discarding months of CPU time and starting over. But it might indicate that something\'s not working correctly. Also, I use the stats to make sure that things are working, and I\'d lose that.

Thanks again,
Jeff

---

Hmm: I see these messages in the log:

2006-09-26 06:07:55 [climateprediction.net] Task hadcm3lbm_3uo2_05135855_1 exited with zero status but no \'finished\' file
2006-09-26 06:07:55 [climateprediction.net] If this happens repeatedly you may need to reset the project.
2006-09-26 06:07:55 [---] Rescheduling CPU: application exited
2006-09-26 06:07:55 [climateprediction.net] Restarting task hadcm3lbm_3uo2_05135855_1 using hadcm3lb version 508


It does indeed happen repeatedly, about twice per day. (It should be an error message so it would stand out better: it\'s difficult to find meaningful messages amidst thousands of log lines!) That\'s a shame: several months of CPU time wasted.

---

Les posted:
The messages: exited with zero status but no \'finished\' file , are nothing to worry about. You can just ignore them.
They seem to be triggered by Windows updating it\'s internal clock from an internet time clock. This happens about every 4 hours.
http://www.climateprediction.net/board/viewtopic.php?p=50303#50303


There seem to be other causes, unually, all that\'s done is trash-up the Messages log. Usually.

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Les posted:
The messages: exited with zero status but no \'finished\' file , are nothing to worry about. You can just ignore them.
They seem to be triggered by Windows updating it\'s internal clock from an internet time clock. This happens about every 4 hours.
http://www.climateprediction.net/board/viewtopic.php?p=50303#50303


There seem to be other causes, unually, all that\'s done is trash-up the Messages log. Usually.

OK, thanks for pointing out that these messages are a red herring. Thus not being error messages.

I\'m wasting a lot of CPU time running a model that seems to be operating properly but might not be reporting results. Should I cancel it and start over? If I don\'t hear in a couple days I\'ll just go ahead and reset (wasting a 33% completed work unit). Meanwhile I\'m going to suspend the CDPN project.

---

Hi,

Could you paste in more from the messages tab? (preferably a couple of days worth of messages).

If the model date is still going up it should be able to fix it.

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

Could you paste in more from the messages tab? (preferably a couple of days worth of messages).

OK, I\'ll resume it and post in a couple days.

Thanks,
Jeff

---