I had not noticed, but I have the same problem: no work allowed when non-BOINC use above 75%. I changed it to 0, but nothing happened. However, I updated and it shows a delay of an hour; perhaps I have to wait until the project gets updated? I will check again in an hour.Under activity, I have of course use GPU always.

---

Yes, the problem lies in whether the preferences are default, or school, or home etc. Once the default has been changed, everything is OK.

---

Thanks, getting rid of separate preferences for work seem to have solved the problem.

Hello to all,
my problem :
computer 1276888 has last contact 21. Aug.2013
computer 1177973 had last contact 26 Jul 2013

what can I do ?

---

Both computers have a full set of models, so they don't need more.
The 8 core machine is returning trickles, so it doesn't seem to have a problem.

Which leaves the 4 core computer.
In the messages window, now called the Event Log, accessible from the menu (Advanced->Event Log, or CTRL + SHIFT + E), there will be one or more lines about what is wrong.

Please copy these lines and post them here.

This is the information after a new restart:

21/08/2013 22:49:30 | climateprediction.net | Restarting task hadcm3n_u0ss_2020_40_008334159_3 using hadcm3n version 607 in slot 4
21/08/2013 22:49:30 | climateprediction.net | Restarting task hadcm3n_o0fp_1980_40_008398168_0 using hadcm3n version 607 in slot 1
21/08/2013 22:49:30 | climateprediction.net | Restarting task hadcm3n_o188_1980_40_008402717_0 using hadcm3n version 607 in slot 2
21/08/2013 22:49:30 | climateprediction.net | Restarting task hadcm3n_000n_1980_40_008402221_0 using hadcm3n version 607 in slot 0
21/08/2013 22:49:37 | climateprediction.net | update requested by user
21/08/2013 22:49:41 | climateprediction.net | Sending scheduler request: Requested by user.
21/08/2013 22:49:41 | climateprediction.net | Sending trickle-up message
21/08/2013 22:49:41 | climateprediction.net | Requesting new tasks for CPU and NVIDIA
21/08/2013 22:50:41 | climateprediction.net | Scheduler request failed: Failure when receiving data from the peer
21/08/2013 22:50:55 | | Project communication failed: attempting access to reference site
21/08/2013 22:50:57 | | Internet access OK - project servers may be temporarily down.

---

Oh dear.
I think this is a good news/bad news thing.

The bad news is that the problem may be an obsolete url for all of those models.
The good news is that, if so, it's just a matter of a lot of edits in client_state.xml

(The background is:
Back in 2011 those of our servers that were in oucs (computing services), were moved to oerc (e-research), which is where 'our' 2 project people work. (Other side of the main road, and a bit south.)
Redirects were then put in place so that either address worked. It looks like these are now missing, (possibly not before time), so all references will need to be edited to change them from oucs to oerc.)

1) Suspend BOINC.
2) Exit from BOINC.
3) Make a copy of client_state.xml, and put it in a safe place.
4) Open the original with Notebook, and start searching for oucs.
5) If found, change it to oerc.
6) When finished, save the file.
7) Restart BOINC.
8) Unspend BOINC.
9) Watch what happens.

This has already been tried by at least one person, and it works.

If you've made contact with the server in the past hour, then you'll have to wait until the end of the back off period.

For us rank amateurs, this is a bad thing. To start, where do I find client_state.xml? However, I did not get a "scheduler request failed" message, so am I OK to just sit it out? I show a miriad of trickles received, although no change in points yet.
Thanks

---

To start, where do I find client_state.xml?

This file is in the folder where BOINC is installed. Because I have been defenestrated for the past thirteen years, I don't know where that is in windows but should be fairly easy to find.

I will check the files on my own two computers shortly though I don't remember any strange messages recently.

BOINC is a complex beast, and this thread is touching on many different parts of it.

Forget credits, that's a different problem which is still being worked on.
There's a new thread on that for people who want to worry.

The problem with the wrong url shouldn't have affected ANYONE. It's been 2 years now, long enough for people to have changed their bookmarks, and for running models to have been completed or abandoned.
So I don't know why some ARE being affected.

Those who don't want to dig into the files MAY be able to just do a Project reset. This will kill running models, and delete the associated program and data files. I don't know about this; it may work, it may not.
And it should NOT be necessary for the majority of people.

client_state.xml is a very important file. It's BOINC's "To Do list".
And it lives in the BOINC data folder, one level up from the projects folder.

The BOINC FAQ Service has a section that describes where the various BOINC folders are: BOINC 6.2 - 6.12.

The Data directory (where client_state.xml lives), is listed in one of the first 10 messages at BOINC startup, in the Event Log viewer.

Advanced View -> Advanced -> Event Log... -> Scroll to the top

Is it possible that the error is just a connectivity error? BOINC on my machine has had trouble connecting to the climateprediction project the last couple days, and I couldn't even connect to climateprediction forums or project status pages.

The Data directory (where client_state.xml lives), is listed in one of the first 10 messages at BOINC startup, in the Event Log viewer.

Shows I don't read the events log very carfuly! :)

client_state.xml is a very important file. It's BOINC's "To Do list".
And it lives in the BOINC data folder, one level up from the projects folder.

A little more on Les� advice to check the ProgramData folder. If you have never looked for ProgramData before then you will have to make it visible. Windows hides this folder by default.

To make it visible type �folder options� into the search box. Click on it in search and click �view� tab. Then click on �show hidden files, folders and drives.� this will make the ProgramData folder visible.

Hi Les,
I did as You explained. All the oucs have been changed to oerc.
her are the results : Malaria is going well (as before the changes) and climate has these :

23/08/2013 19:22:07 | malariacontrol.net | Starting task wu_3301_176_1222834_0_1377107601_1 using openMalariaA version 658 in slot 4
23/08/2013 19:22:09 | malariacontrol.net | Started upload of wu_2899_327_1222832_0_1377107472_1_0
23/08/2013 19:22:09 | malariacontrol.net | Started upload of wu_2899_327_1222832_0_1377107472_1_1
23/08/2013 19:22:10 | malariacontrol.net | Finished upload of wu_2899_327_1222832_0_1377107472_1_0
23/08/2013 19:22:10 | malariacontrol.net | Finished upload of wu_2899_327_1222832_0_1377107472_1_1
23/08/2013 19:28:14 | climateprediction.net | project resumed by user
23/08/2013 19:28:17 | climateprediction.net | Sending scheduler request: To send trickle-up message.
23/08/2013 19:28:17 | climateprediction.net | Reporting 1 completed tasks
23/08/2013 19:28:17 | climateprediction.net | Requesting new tasks for NVIDIA
23/08/2013 19:29:11 | climateprediction.net | Scheduler request failed: Failure when receiving data from the peer
23/08/2013 19:29:26 | | Project communication failed: attempting access to reference site
23/08/2013 19:29:29 | | Internet access OK - project servers may be temporarily down.
23/08/2013 19:31:32 | malariacontrol.net | project suspended by user
23/08/2013 19:31:33 | climateprediction.net | Restarting task hadcm3n_o0fp_1980_40_008398168_0 using hadcm3n version 607 in slot 1
23/08/2013 19:31:33 | climateprediction.net | Restarting task hadcm3n_o188_1980_40_008402717_0 using hadcm3n version 607 in slot 2
23/08/2013 19:31:33 | climateprediction.net | Restarting task hadcm3n_000n_1980_40_008402221_0 using hadcm3n version 607 in slot 0

---

Have you rebooted your PC recently? Sometimes the IP addresses are cached locally & so won't be updated until a reboot.

---

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

Yes I have rebooted several times,but nothing changed, going to reset the project now.

---

I see that credit totals are updating at least on the �statistics� page. One more step back to full normal. Good work guys.

Nah, looks like the old ones

I installed Boinc on a new computer in June this year and added CPDN using the command line method, as the graphic method didn't work. Yet client_state.xml contains links to oucs, though only for the executable files. For the data the links refer to oerc.