I\'ve been running the model for a while now and no credits show up on the project web site. I\'ve looked in all the usual places and can\'t see any errors. Stats show model to be 1.38% complete.

Machine is running Fedora Core 5.

Credits are granted for trickles received by the servers. No trickles, no credit.
And the server says that you haven\'t sent any trickles.

Perhaps you have a comms problem, and all of the trickle_up_ files are accumulating in the folders of the two models that you have attempted.

Credits are granted for trickles received by the servers. No trickles, no credit.
And the server says that you haven\'t sent any trickles.

Perhaps you have a comms problem, and all of the trickle_up_ files are accumulating in the folders of the two models that you have attempted.

Always a possibility! Are there any logs I can check. I\'ve looked in most of the likely places but can\'t see anything but as I\'m new to CPDN I\'m not sure what to look for.

There are firewalls in between but can\'t see any problems in the logs there. What is required? Server, UDP/TCP, port number? HTTP to anywhere should be OK.

Is there a script or moddule that does the trickle?

Thanks.

---

Seems your post dropped by the boards, so to speak, Nigel.

Are you running from Command Line or the GUI? If the GUI, there should indications under the Messages tab that connections were attempted but failed. In either case, unsent Trickles accumulate as boinc/cpdn/Projects/climateapps.../trickle_up.... (I\'m no longer running anything in Linux so am not sure of file names. [TCM runs faster in Windoze, unlike earlier Slab & Sulphur Cycle compiles.]) Unsent Trickles will go up \'automagically\' as soon as connection is made to the Server; same is true of the larger accumulated 10-year uploads.

---

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

I have this problem too.
Climateprediction has been working now for 100 hours, completing around 2.5% - but no trickles have been sent, and clearly I did not receive any credit.
I have also Seti@Home working on that machine, and it\'s correctly communicating with the servers to report completed work.
There is no trickle_up... message inside the boinc folders, neither any warning/error message in the messages tab of the graphical interface.
Any ideas?

---

Welcome to the Boards, ClassicaL. You\'ve been with CPDN for a long time.

Is there a difference in firewall or other settings that could interfere with Internet connection between your two machines? One has a good record, the other no contact, as you say.

At 2.5%, Trickles should have been uploaded; that they are not in CPDN Folders suggests that they were uploaded. There\'s been some server grief in recent days (I\'ve experienced some contact problems, even when the status shows all servers \"up\"); perhaps Oxford has your reports but hasn\'t been able to post credit for them yet.

Please give it a day or two. If your credits don\'t appear, please post again.

---

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

I\'ve just looked at one of the stats sites, and it looks as though the credits program may not have restarted after the problem on the 27th.

I\'ve posted a note about it.

However, you DID say that no trickles had been sent. If this really is the case, then you can\'t expect to get credits.
No trickles, no credits.

Have a look in the Messages tab, (or stdoutdae.txt in the BOINC folder), and see if it says Reason: To send trickle-up message.
If not, then maybe your program has already gotten into a loop, and is repeating the same year over and over.

I confirm that no trickles have been sent. The model is not looping, years go on (I\'m on 1926 now) and the percentage completed moves up.

Skimming through stdoutdae.txt file, I found the following lines connected with trickles:

Trickling yearly means for 1925
Building file list for 1925 in dataout -- 7 entries
adding: hadcm3lbm_al91_25263368_0_meana_1925_pd.nc (deflated 83%)
adding: hadcm3lbm_al91_25263368_0_meana_1925_pg.nc (deflated 83%)
adding: hadcm3lbm_al91_25263368_0_meano_1925_opt.nc (deflated 80%)
adding: hadcm3lbm_al91_25263368_0_meano_1925_pi.nc (deflated 85%)
adding: hadcm3lbm_al91_25263368_0_meano_1925_pj.nc (deflated 80%)
Creating trickle file trickle_hadcm3lbm_al91_25263368_0_1925.zip for upload...

A few lines later there is:

Trickling yearly means for 1925
Building file list for 1925 in dataout -- 0 entries
Creating trickle file trickle_hadcm3lbm_al91_25263368_0_1925.zip for upload...

As regards suspicious entries, I found this:

Error file dataout/al91fo.pjc5c10 has unknown file type
Error in getting file type
Error file dataout/al91fo.pic5c10 has unknown file type
Error in getting file type
Error file dataout/al91fo.pfc5c10 has unknown file type
Error in getting file type
Error file dataout/al91fa.phc5c10 has unknown file type
Error in getting file type

Maybe I should just cancel the execution and start another model from scratch?

---

I\'ve run SETI too with no problem. Of course SETI does not use the trickle mechanism. I\'ve re-installed BOINC 5.4.9 and started a new model. Now at 1.16%, 07/10/1922 and no sign of trickles. No comminication messages, good or bad, in boinc manager messages tab. I haven\'t found any messages indicating a communication problem. Model files download ok (using port 80?).

---

ClassicaL and Nigel

These are the lines that you should see (and look for), in the messages tab (and in stdoutdae.txt):

5/10/2006 10:53:02 AM|CPDN Seasonal Attribution Project|Sending scheduler request to http://climateapps3.atm.ox.ac.uk/hadam3_cgi/cgi
5/10/2006 10:53:02 AM|CPDN Seasonal Attribution Project|Reason: To send trickle-up message
5/10/2006 10:53:02 AM|CPDN Seasonal Attribution Project|Note: not requesting new work or reporting results
5/10/2006 10:53:07 AM|CPDN Seasonal Attribution Project|Scheduler request to http://climateapps3.atm.ox.ac.uk/hadam3_cgi/cgi succeeded

Note: You\'ll have a different date/time/project_name, and probably a different upload server name in the last line.

If you DON\'T see these, then you HAVE A PROBLEM.

Usually it means that BOINC isn\'t getting through to the server, in which case the trickle_up files accumulate in the folder with the model\'s name under, (for the default location): C:\\Program Files\\BOINC\\projects\\climateprediction.net\\model_name
These XML files will have the name: trickle_up_hadcm3 .........

boinc.exe needs access to the web on ports 80 and 443, local access on ports 1043 and 31416.
(I think that some are now obsolete for BOINC 5.* versions)

The next thing to talk about is the Network activity option in BOINC.
In 5.2.8, which is what I\'m still using, there are 3:
always available
based on preferences
suspended

If you don\'t have something special setup in your Account preferences, the first one should be selected.

**************


I\'ll leave it there, and wait for feedback from each of you.

I\'ve finally had a successful trickle. Because of space issues I had moved the projects directory to another file system and linked to it with a symbolic link. After freeing up some space and moving the projects directory back it works. It seems something does not like symbolic links.

I ran a packet capture while it was doing the and the only traffic was between port 80 on climateapps2.oucs.ox.ac.uk. and my local port.

Thanks for your help.

---

Because of space issues I had moved the projects directory to another file system and linked to it with a symbolic link. After freeing up some space and moving the projects directory back it works. It seems something does not like symbolic links.

Intersting enough, that\'s exactly what I did too. I guess you found the problem... I\'ll try to remove symbolic links, and let you know what happens.

---

I confirm that removing the symlink solved the problem, trickles now work properly.

---

I\'m happy to hear that you\'re both back in business.