Dave\'s computer is here. Its Boinc upgrade was clearly a success and the new model has trickled three times.

Don\'t worry about that model download failure; everyone has one of these from time to time. I think you could allow new tasks for CPDN now and let both cores work on models if you want to. Keep an eye on what\'s happening though.

It would probably be a good idea to update your AVG.

---

Cpdn news

It appears I have a configuration problem as well.
I have updated the client to 6.10.43 and checked my preferences.

Any help on what I need to do will be appreciated.

The link:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=896382

Barry M

Hi Barry

Normally, the stderr out for each of the models shows what has gone wrong, but not in your case. So I did a Google using: \"Exit status -226\".
This produced a few threads, mostly from about a year ago, and on several projects.
This post from Einstein may explain it:

The task failures seem to be indeed a problem with the less than 100% setting of CPU time used, since Exit status -226 (from your STDERR log) translates to

ERR_TOO_MANY_EXITS -226

An application has exited prematurely (unexpectedly) more than 99 times without generating a checkpoint, so giving up on that task.


(quoted from BOINC FAQ Service)

You don\'t say which version of BOINC you had before, or if you were using less than 100% of processors, but that may have been the cause of the crashes.

---

Backups: Here

Hi Barry

Here\'s the computer.

Clicking on the tasks shows that this computer has done sterling work for CPDN in the past including successful 160-year HadCM models. But you can see from the tasks pages that since about October 2009 all the models have been failing with either -226 errors as Les says (too many exits) or -185 errors (create process failed & access is denied). I suspect that the cause of both may be the same. The models can\'t get started or can\'t reach the first checkpoint. For each model you need to click on stderr out to see its messages.

As Geophi said above to another member with the -185 error, the Boinc FAQ says for this error:

ERR_RESULT_START -185

This is an error that will occur:
- if BOINC couldn\'t start the application
- if files are missing
- upon catch of other error returns
- on nonzero exit or signal
- if exceeded resource limit
- as catch-all for resume/start errors

It suggests a computer reboot. I don\'t believe your computer hasn\'t been rebooted since last October so that doen\'t help.

But the -226 \'too many exits\' message is like the older one saying \'exited with zero status & no finished file\'. Here\'s the Boinc FAQ about that with Jorden\'s advice.

* He mentioned AVs blocking the model. You need to get both Boinc folders into your AV\'s trusted zone/excluded from scanning.

* As has been mentioned above, the AVG antivirus has caused problems for many members as Les said here. The problem may not only have been with uploads. If you have AVG please update it.

* At least one member with constant model starting problems has found that a complete reinstall of Boinc fixed the problem. As this computer has no useful CPDN work on it I\'d uninstall its current Boinc before reinstalling (Start > Settings > Control panel > Add and remove programs). There are two Boinc folders.

When you\'ve checked the AV settings and reinstalled Boinc please post back and let us know because Milo will have to allow your computer to download models again ie reset its daily quota. We\'ll then see whether the suggestions work.

Barry, until this problem is fixed it would be useful if you could unhide your computers by editing your CPDN account. Then if by chance the Boinc reinstall gives your computer a new ID number as happened to Dave512 who posted above we\'ll still be able to find its web page easily.

---

Cpdn news

I have completely uninstalled BOINC and the install folders. I have also changed the hide computer setting to show them. I will download the newest BOINC client again and install it. Having done all that I now see this system is getting a spoolsv.exe error. I don\'t know if this could be related but I am going to do a repair of Win XP-SP3 - to see if that will fix it since a system restore didn\'t help. If anyone has an idea of how to fix a spollsv.exe error - I am all ears.



[quote]Hi Barry

Clicking on the tasks shows that this computer has done sterling work for CPDN in the past including successful 160-year HadCM models. But you can see from the tasks pages that since about October 2009 all the models have been failing with either -226 errors as Les says (too many exits) or -185 errors (create process failed & access is denied). I suspect that the cause of both may be the same. The models can\'t get started or can\'t reach the first checkpoint. For each model you need to click on stderr out to see its messages.

Google produces quite a few results, including this: What is spoolsv.exe?

---

Backups: Here

Having done all that I now see this system is getting a spoolsv.exe error. I don\'t know if this could be related but I am going to do a repair of Win XP-SP3 - to see if that will fix it since a system restore didn\'t help. If anyone has an idea of how to fix a spollsv.exe error - I am all ears.

It depends on the exact error, but the most likely cause is a corrupt or incompatible printer driver.

Try http://support.microsoft.com/kb/324757/en-us.

---

"The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer

Did a complete system restore. Newest BOINC client installed. Old files that started running with a April 15 completion date have been aborted. BOINC client and Manager added to Kaspersky. No errors now on boot up. All printers working fine. I think I am ready to try this again. What has to happen now.

Having done all that I now see this system is getting a spoolsv.exe error. I don\'t know if this could be related but I am going to do a repair of Win XP-SP3 - to see if that will fix it since a system restore didn\'t help. If anyone has an idea of how to fix a spollsv.exe error - I am all ears.

It depends on the exact error, but the most likely cause is a corrupt or incompatible printer driver.

Try http://support.microsoft.com/kb/324757/en-us.

What has to happen now.

I\'ve re-enabled your host - please try to connect again and we\'ll see if it\'s working.

What has to happen now.

I\'ve re-enabled your host - please try to connect again and we\'ll see if it\'s working.

It reconnected and downloaded 2 WU. I suspended all other tasks temporarily to see if any of the new tasks would run. They do. However The other tasks have a short deadline and are now running as a high priority. I\'ll keep an eye on the computer for awhile to see it restarts again. I also made sure Kaspersky was fully updated.

That\'s looking good, Barry. I hope you\'ll soon get the printer problem fixed too.

It seems as if your models are now getting past the first checkpoint which they couldn\'t previously. If I were you I\'d keep an eye on those models when possible. We should all keep an eye on HadSM graphics a couple of times a week anyway just in case they develop the iceworld problem with graphics of one colour only. If this happens it will be an error in the model, not the fault of the computer.

I see that the reinstall also seems to have created a new computer record for you. Here\'s your list. #3 looks like the old ID with the errors and #2 (1070796) is its new ID.

You may wish to merge the two computer records but I\'d wait at least a couple of weeks before doing this to check that everything is working well and trickles are uploading.

Thanks for taking the time to fix the problem. It would be helpful to us if you could leave your computers unhidden for as long as possible as we need to see whether the advice we give does work.

---

Cpdn news

je ne sais pas quel est le problème ?

Bonsoir Michel, bienvenu au forum.

L\'administrateur de ClimatePrediction (CPDN) qui s\'appelle Milo vous a envoyé un courrier parcequ\'on s\'est rendu compte que tous les modèles de climat faillissent sur votre ordinateur depuis quelque temps. C\'était pour vous donner l\'occasion de faire quelques pas pour éviter ce problème. Votre ordinateur n\'est pas le seul qui ait eu ce problème et je crois qu\'il doit fonctionner parfaitement sauf en ce qui concerne l\'exécution des modèles.

La page web CPDN de votre ordinateur se voit ici:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=1026829

Vous remarquerez que l\'ordinateur n\'a pas reçu de crédit (qui est virtuel d\'ailleurs); ça veut dire qu\'aucun modèle n\'a pu démarrer comme il faut bien que 127 modèles aient essayé.

Vous verrez aussi que Milo a fixé le quota cotidien de l\'ordinateur à -1, c\'est-à-dire qu\'aucun nouveau modèle peut être téléchargé jusqu\'à ce que le problème soit résolu. Je ne crois pas que la résolution soit difficile.

Si vous faites clic sur le lien 127 vous verrez la liste de vos modèles. Les voici:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=1026829

Si vous faites clic sur le numéro des tâches dans la colonne à gauche et puis sur la page web des modèles vous faites clic sur stderr out +, vous verrez que les faillites ont plusieurs codes et messages d\'erreur. Mais leur origine est toujours le même.

Jusqu\'ici les noms des pistes des fichiers des modèles CPDN ont toujours contenu des espaces blancs. Il y a quelques ordinateurs Windows qui ne reconnaissent pas ces espaces. Les tâches ne trouvent donc pas les fichiers qu\'il leur faut pour s\'exécuter. Elles ne peuvent jamais démarrer.

Votre ordinateur a ce problème. Il serait possible d\'éditer certains fichiers de votre ordinateur pour éliminer le problème mais si vous n\'êtes pas programmeur professionnel je vous conseille de ne pas l\'essayer. Moi je ne le ferais pas.

Que faire? CPDN lancera bientôt des modèles qui s\'appellent FAMOUS où le problème des espaces blancs sera résolu. Votre ordinateur pourra donc les exécuter.

Dans l\'attente de ces nouveaux modèles qui sont encore en état béta (je pense qu\'on les lançera dans un mois au plus tard) je vous conseille d\'attacher l\'ordinateur à un ou deux autres projets où en principe vous n\'auriez pas les mêmes faillites. On le fait à travers le gestionnaire Boinc > Outils > Attacher à un projet. Il y en a beaucoup d\'intéressants. Voici la liste des projets (seulement en anglais autant que je sache mais ça vous donnera une idée):

http://boinc.ssl.berkeley.edu/projects.php

Chez CPDN vous devriez éditer des détails de votre compte personnel pour recevoir seulement des modèles FAMOUS à l\'avenir. Ce message est déjà très long; je vous expliquerai demain comment éditer votre compte. Ce n\'est pas difficile.

---

Cpdn news

Comment éditer le choix de modèles qu\'on veut télécharger:

Dans le menu bleu sur cette page à gauche cliquez sur Taking part in CPDN. Un sous-menu s\'ouvre.

Cliquez sur Your account.

Cliquez sur ClimatePrediction.net preferences. Vous verrez la liste des tâches disponibles.

En bas cliquez sur le bouton Edit Preferences.

Dans les cases à côté de chaque modèle il faut choisir FAMOUS et laisser les autres cases en blanc.

En bas de la page cliquez sur Update Preferences. Vous pouvez fermer votre compte si vous voulez.

Dans le gestionnaire Boinc choisissez l\'onglet Projets. Sélectionnez ClimatePrediction puis cliquez sur le bouton Sélectionner. Le serveur CPDN saura maintenant que vous voulez seulement exécuter les FAMOUS.

Quand les FAMOUS se lanceront au projet CPDN vous les verrez sur la page du serveur ici:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/server_status.html


Quand vous aurez édité vos préférences dites-le-nous svp. A ce moment-là Milo pourra faciliter à nouveau votre quota quotidien.

---

Cpdn news

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=1068093

Hi,
I\'ve recieved an email with regards to a misconfigured BOINC installation.
Without any further information I\'ve uninstalled and reinstalled the BOINC application.

Can you tell me if A) my problem has been fixed now or B) if you know what the problem is and I\'ll resolve it by following your instructions.

Thanks

Hi Dregnet, welcome to the forum.

Here\'s an active link to the computer\'s CPDN web page:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=1068093

You can see that Milo, the project admin, has set its daily quota to -1 which means the computer can\'t download models at the moment. At soon as you say you\'ve fixed the problem Milo will reallow the quota.

If you click on the link to Tasks at the bottom of the computer\'s web page, then click on any of the tasks and on a task\'s own web page click on + beside stderr out you\'ll see the error messages your models have crashed with. Eg this task:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=11406272

The significant phrase is CreateProcess() failed - Access is denied. (0x5). It means something is preventing access to the model\'s files. Probably your firewall or your AV.

There\'s an explanation in these FAQs:

http://boincfaq.mundayweb.com/index.php?language=1&view=448&sessionID=81c20a76df6aa5846bc136c23eb79620

If you can check that please let us know. Thanks for posting.

Edit: Don\'t disable your firewall or AV while connected to the internet. It\'s usually possible to change their settings while they\'re running.

---

Cpdn news

Comment éditer le choix de modèles qu\'on veut télécharger:

Dans le menu bleu sur cette page à gauche cliquez sur Taking part in CPDN. Un sous-menu s\'ouvre.

Cliquez sur Your account.

Cliquez sur ClimatePrediction.net preferences. Vous verrez la liste des tâches disponibles.

En bas cliquez sur le bouton Edit Preferences.

Dans les cases à côté de chaque modèle il faut choisir FAMOUS et laisser les autres cases en blanc.

En bas de la page cliquez sur Update Preferences. Vous pouvez fermer votre compte si vous voulez.

Dans le gestionnaire Boinc choisissez l\'onglet Projets. Sélectionnez ClimatePrediction puis cliquez sur le bouton Sélectionner. Le serveur CPDN saura maintenant que vous voulez seulement exécuter les FAMOUS.

Quand les FAMOUS se lanceront au projet CPDN vous les verrez sur la page du serveur ici:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/server_status.html


Quand vous aurez édité vos préférences dites-le-nous svp. A ce moment-là Milo pourra faciliter à nouveau votre quota quotidien.

préférences modifiées

préférences modifiées

Merci - quand plus de FAMOUS seront disponibles, votre ordinateur pourra les télécharger.

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=981194

It seems I have a problem. Please advise

---

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=981194

It seems I have a problem. Please advise

It seems that your computer has the problem described in this thread: Exit status -2. Last error = 193.

Before approximately 20 February 2010, your machine was running models normally: since that time there has been a problem in which BOINC cannot start properly because of Windows path handling. This problem may be affecting other applications on your computer and not just BOINC (e.g. some services), so it is well worth fixing.

Have a look at that thread and if you need any further advice then post back here. When you post back to say that the problem is fixed then Milo will re-enable the downloading of models to that computer.