Hi Lockleys

If you don't mind I'm going to start another thread for members to tell us about computers they've come across that crash lots of models. The new thread should probably really be started in Number Crunching but I'm going to create it in the Science section so it's near this Misconfiguration thread. There used to be a thread to report computers but let's start afresh.

It doesn't help members posting about model crash problems to have reports about other problem computers interspersed.

Here is the new thread.

---

Cpdn news

Hello. Received this message.

Your computer (host # 1120510) described below appears to have a misconfigured BOINC
installation and is crashing models. Would you please have a look at it?

If you need assistance, please post in this thread on our BOINC forums and we will suggest a way to fix the problem. You may post in any language:

Really not sure what to do. I did recently update my Boinc client.

http://climateapps2.oerc.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=1120510

Joe

If you look at the stderr errors on the task webpages for those tasks that are crashing, you'll find a "no space left on device" message within them. This suggests that either the disk is filling up, or more likely the boinc data directory size is exceeding that set in your preferences.

You may need to increase the size of disk space allowed for your boinc data in your boinc preferences, or you may need to clean out old directories for failed tasks (if they exist).

Joe, when you've followed Geophi's advice could you please let us know so we can ask the CPDN programmers to reenable your computer's model downloads.

---

Cpdn news

Joe

My guess is that you've got a chain of events:

You've crashed a huge number of models, each of which has left a small amount of debris. This would account for the "out of space" problem.

And a possible reason for all of the crashes, or at least the recent ones, is right there in your post: I did recently update my Boinc client.

What has this got to do with it? The answer is in a sticky post at the top of the Macintosh section of this board: Tasks crash after upgrading to BOINC 6.12.26

hello guys,
i got an email from your server, telling me that my computer is crashing some models and i might have misconfigured BOINC.

i'm asked to include this link: http://climateapps2.oerc.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=1286679

unfortunately, i don't know how to fix these issues. i'm thankful for any help!

please tell, if you need any further information.

thanks,
daniel

Daniel, that is for your Windows computer which doesn't seem to have contacted the project since July. If it is no longer crunching, just ignore the message. If it is still crunching and just not being sent work because it has been blacklisted for crashing every model it is sent I am not sure. Is there a windows equivalent of the 64 bit problem with linux? If so that is what I would suspect but having been windows free this century you will need someone else to help sort it out!

Good luck

Dave

Daniel

The error messages on the pages for the failed models says:

Could not launch model process. Last Error=216

It may be that your antivirus program is quarantining some of the many files associated with the models.
If so, then exclude both parts of BOINC from both manual and automatic scanning.

Daniel

Some questions:

Did you use Windows Compatibility Mode for your BOINC install?
And if so, which one?

Are you running the model on the same computer that you used to download it?

I'll try to add an exclusion to the antivirus program and hope that helps.

Did you use Windows Compatibility Mode for your BOINC install?
And if so, which one?

Not that i know. I'm participating in other projects as well.

Are you running the model on the same computer that you used to download it?

Yes

thanks for your help

Daniel

In that case, your computer has somehow downloaded the wrong file or files.

See this listing for the meaning of error 216: System Error Codes (0-499)

ERROR_EXE_MACHINE_TYPE_MISMATCH

216 (0xD8)

This version of %1 is not compatible with the version of Windows you're running. Check your computer's system information and then contact the software publisher.

The next thing to try is:

1) Set the project to No new tasks (in the Projects tab).
2) Reset the project (in the Projects tab)

This SHOULD delete ALL of the files under /projects/climateprediction.net, including the 2 models still listed as "running".

Once you've checked that the folder is empty, set it back to receiving work and let us know, so that we can have your computer unblocked.

Then we'll have to wait to see what happens.

Then we'll have to wait to see what happens.

And given how little work there is at the moment, it may be a long wait :(

les,
i reset the project and manually deleted all files in climateprediction's project directory that were left over.

i'm ready to give it another shot :)

hello,here am I :
http://climateapps2.oerc.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=1208986
I recently posted in "windows" forum:http://climateapps2.oerc.ox.ac.uk/cpdnboinc/forum_thread.php?id=7703&nowrap=true#47596
Then realised I had received a mail about lots of crashes,resulting in a blocked computer,i didn't see it at first
There is actually a problem with climateprediction tasks,with almost 100% errors.I participate also in Rosetta and SETI.Quite a lot of errors with Rosetta also,but I don't know if they are of the same nature
Seti seems to work fine,i see no errors,which I find strange.
Boinc and the three projects worked fine for some time .
I run Boinc 7.0.25,64 bits.This is not the latest version and there is a reason to this.When I tried to update a few months ago,there was a serious bug with Boinc crashing and not running at all. I had to come back to actual version.I didn't try further to update, can this be a lead for the crashes of models?.
I'm absolutely no expert in computers,neither in english (french is mother language)
Since I'm aware there is a problem I Tried to configure Boinc differently :no restrictions for cpu,restart of project,.
What can be done now to unban the computer,because I can't try if changes made are helping.
Before I restarted of the climateprediction project,there were two ongoing tasks remaining in Boinc,I to ran them and they both seemed to work properly,as far as I could tell.
Thanks for helping,

---

Hi Philippe,
One of the moderators will get back to you about the unblocking. It is worth making sure the BOINC directories are excluded from any virus scans as these can put a lock on a file when BOINC wants to write to it which will cause the model to crash. Also make sure that in Tools >computing preferences>Disk and Memory usage, "leave applications in memory while suspended" is ticked. Both of these can make a vast difference in the number of tasks that crash. Looking at the tasks that crash on this computer I suspect the latter will help most.

Good luck

Dave

hello Dave,
Well Boinc config was as you advised,I also checked the Boinc directories for viruses,no problem.
I keep spirit high,
Thanks

---

It is not virus's in the boinc directories that is the problem but the antivirus software checking them. The anti-virus software should give you the option to exclude the boinc directories from the check. The other thing that is worth doing is setting boinc so there is no restriction on boinc activity as suspending computation and then resuming it also makes task failures more likely.

okay ,thanks Dave,so,I think I can do that exclusion of Boinc directories from antivirus program checks.That is "C/Program files/Boinc" and also "C/Documents../All users/Boinc".Exclude both?
There is no restriction in Boinc activity,also GPU.
Actually,both SETI and Rosetta projects are working fine
Climateprediction still banned.
thank you,

---

Yes, exclude both folders.

All the models on host 1208986 that I looked at had large numbers of "Suspended CPDN Monitor - Suspend request from BOINC..." entries in the Stderr report. As Dave suggests, the BOINC settings can have a large effect, particularly on the larger, longer HADCM3N models. As well as the virus check exclusion, it is a good idea to change the BOINC preference "While processor usage is less than X percent (0 means no restriction)" so that X is zero. (The text will be different if your BOINC Manager runs in French.) That will eliminate the entries in Stderr. However, you should only do this if you are happy for BOINC to run most of the time.

If you post back here, then I'll ask for the machine restriction to be removed.

Thanks for trying to sort the problems out.

Hello Iain,
yes,all indications and I received on this forum have been applied to Boinc,it's been a week already
Rosetta and Seti projects are running fine.
Thanks,

---