This is the computer, an XP 2500, A7N8X Motherboard, 512MB memory, 32MB shared with on board geforce4. Ran Climate, Predictor, Seti, LHC, Einstein for as long as each project has been up trouble free. Now since about March 3rd, every morning I find my computer in the process of crashing, have to reboot it, all my WUs for all projects have been cleared out.

I then would uninstall the application completly, manually delete the folder, reboot, then reinstall.

Originally happened with BOINC 4.19, I'm now using 4.25 and same errors.
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=128195

It all seems to go wrong when this happens
2005-03-10 19:57:19 [climateprediction.net] Unrecoverable error for result 0sck_000057131_0 ( - exit code -2 (0xfffffffe))

I have the st files and everything else all saved from the BOINC directory, so if anything is needed from them I can email them out.

Thank you for your assistance,
MIKE

---

Updated info,

After detaching the computer from Climate, it has run all other projects without any incidents.

---

I notice that your team is Overclockers.com
It is possible to get away with overclocking on other projects, but it's risky here without a VERY stable computer.

For some ideas, see my first post in <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=2026"> this</a> thread.

And as for comparing CP with other projects, see my first post <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=2210"> here.</a>

Les

The machines I run CP on are at work, and none of them are overclocked.

My 2 home machines that are overclocked have no errors whatsoever in any of the projects, Prime95 passed minimum of 24hrs.

I'll check out those threads, thansk for the help.

---

&gt; The machines I run CP on are at work, and none of them are overclocked.
&gt;
&gt; My 2 home machines that are overclocked have no errors whatsoever in any of
&gt; the projects, Prime95 passed minimum of 24hrs.

Another thing is to check the harddrive. The usual suspects: defrag, surface check, etc. Myself and others have also found that when a HD in a system starts to go it causes the model to crash.

---

Another machine, same problem occured.
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=137353

So far hardware is the exact same on both. Neither are overclocked. CPUs never break 55C in either machine, hardrives are defragged bi-weekly. Both also had new models on them when it started.

Both are on 24hrs a day, and spend about 14 hours at nite not being used. Screensaver is set to password protect. Both only got screwed up by Climate, and after sitting overnite, sitting during 45min lunches did not cause any issues. Climate then caused the other projects to crash as well.

Does anyone know what the "exit code -2 (0xfffffffe)" refers to?

I'm thinking something with disk access that its having issues with when the machines been sitting a while, but would like a more definite answer.

Thank you.

---

Add another NON-Overclocked machine to the list of machines doing this. I now have confirmed one of my suspicions, it happens when it downloads the latest WU with the latest application.

Notice how it happens right after succesfully completing a unit.

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=24526

So far errors only on machines that are on 24hrs a day and spend a lot of time idle, which WAS the perfect machine to run Climate on. My 2 machines at home should be finishing some Climate units in the next month or so, I'll try and watch for when that happens and see if the same disastorous results happen.

I'm also noticing more and more people with similar issues appearing on the boards here.

---

&gt; Add another NON-Overclocked machine to the list of machines doing this. I now
&gt; have confirmed one of my suspicions, it happens when it downloads the latest
&gt; WU with the latest application.
&gt;
&gt; Notice how it happens right after succesfully completing a unit.
&gt;
&gt; http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=24526
&gt;

Wow, what a timely post. It just happened to me. Stable PC that had never had a bad WU (8 Classic, 4 BOINC successes), uploaded results from a competed run. The first download of the next model resulted in an "upgrade" from 4.04 hadsm to 4.10. Then it had two WUs fail immediately. Not good. It's crunching on a new model now, but it just started so I hope it's stable.

&gt; Wow, what a timely post. It just happened to me. Stable PC that had never
&gt; had a bad WU (8 Classic, 4 BOINC successes), uploaded results from a competed
&gt; run. The first download of the next model resulted in an "upgrade" from 4.04
&gt; hadsm to 4.10. Then it had two WUs fail immediately. Not good. It's
&gt; crunching on a new model now, but it just started so I hope it's stable.
&gt;

So far I've only seen it happen on machines that are genearlly on 24hrs a day, but then again these machines are also finishing WUs quicker so I'll assume that they are getting the updated 4.10 hadsm first.

Everyone thats still running is a 4.03/4.04 client.

I don't think it has anything to do with machine instability as I personally put these 2 of the 3 machines that "Client Errored" together and being an overclocker, I tend to go way overboard when it comes to case cooling even if its not going to be overclocked. Stock heatsinks, but lotta extra fans in these machines.

---

And another one bites the dust
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=675810

After succesfully finishing its 4.03/4.04 WU, downloads a new 4.10, 75 minutes later crashes BOINC. At least I was there this time to make sure it didn't go and corrupt all my other projects.

Only thing all the machines have had in common is that they are on 24hrs a day. My brothers computer thats slightly overclocked is the only one still going but he shuts his off to sleep at nite.

---

I have the same problem
See: http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=2387
for details. same error doe. my hw is fine.

---

I did have a problem with 4.12 crashing within a couple of seconds on one of my systems last week (six models went that way in very quick succession). The system had been rock solid on 4.04.

I temporarily detached it from the project and ran the sulphur cycle alpha project instead. I reattached to climateprediction.net a couple of days ago and the system is now running 4.12 without any problems. It was suggested on the alpha boards that there might have been problems with my 4.12 download (have my doubts because of the MD5 checksums), but it might be worth trying a project reset to force the programs to be downloaded again.

---

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