Hi,
this morning the client finished it's first 10k Timestep, but no 'trickle' was reported. A file 'trickle.out' existed in boinc/slots/2 for a while. I issued an 'update' manually, while another project (einstein) was active. I realised then/thereafter, that both projects were active - i could start either einstein or cpdn graphic! Both projects cpu-time was increasing! I exited BOINC and restarted. Short while after restart, a 'visual fortran runtime error' was reported with cpdn-client (I made a screenshot). The file 'trickle.out' is gone since, progress now is reset to zero (while total cpu is not), time remaining is at 600 hrs, what is less than before (maybe that item is meant to be total boinc uptime until model finishes - cpdn equally shares with two other projects, and 1800 hrs was, what was shown before)

Again exited and restarted BOINC - immediately got the same runtime error again. BOINC then restarts the client, so it is up and running - but i'm not optimistic with this situation.

stdout/stderr are empty, no messages with BOINC concerning this.

What am i to do?

Thx!

Frank

---

Welcome to the project, Frank.

Trickles are at 10,802 timesteps. See bottom of the FAQ.

I can't help with the rest. I only run CPDN.
1st because it's what I consider important, but also because it keeps the processor running full time as it is.
From reading other peoples posts over the months, I've concluded that time switching several dc projects
causes too many problems. BOINC, in any version, is too buggy to be reliable for this.

Les

The trickle.out file gets moved to the projects/climateprediction.net directory (as trickle_up*.xml), and will be sent up to the server in the next scheduler request. Unless you've got BOINC networking disabled that should be straight away.

As with Les, I'm only running CPDN, so don't have any personal experience of BOINC switching projects. I recall some people did have problems with projects not shutting down under Windows 98 (which I notice you're running) when BOINC switches project, but I'm not sure if that's still the case.

---

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

> Welcome to the project, Frank.

Hi, Les!

I'm with CPDN since Sept. 2003 ;-)

But in fact i retired from grid computing a while ago and start over with BOINC now.
>
> causes too many problems. BOINC, in any version, is too buggy to be reliable
> for this.

I'll give it a try. I was in BOINC Beta almost from the beginning, when with v2.25 BOINC wouldn't start anymore on my main machine. Nearly hopping mad, i took a time out :-)

What I see now isn't that ugly. I'm curious at the other projects (Einstein & Predictor are already running here w/o any problems). I hope to see how things are improving. Process swapping is quite a simple task - they'll manage it one day or another ;-)

Thx for caring
Frank

---

> The trickle.out file gets moved to the projects/climateprediction.net
> directory (as trickle_up*.xml), and will be sent up to the server in the next
> scheduler request.

Sorry for being not accurate:

A file 'trickle_whatsoever.something' was at the machine this morning.

After the crash, it was gone.

I 'thought' it was in slots/2 and i 'thought' it's name was trickle.out. But now i remember at least, it had a xml-extension!

It's definitely gone now, it didn't show up in the 'Transfer' section of BOINC, it never reached the server (otherwise, I should be able to find it, shouldn't I?)

Something went wrong with it ...

> Unless you've got BOINC networking disabled

No, NC was enabled - but not available when that TS (10.8xx) was done. The machine came back online at TS 13.000 - is that crucial?

Thx for caring,
Frank

edit:
> Windows 98 (which I notice you're running)

pleading guilty; really should throw it away ;-)

---

> I'm with CPDN since Sept. 2003 ;-)

> But in fact i retired from grid computing a while ago and start over with BOINC now.

In that case, Welcome back Frank.

Trickle_up...... files accumulate in the project directory. It doesn't matter how long they stay there.
When you are next on line they ALL get uploaded.

HOWEVER! Thyme Lawn has recently descovered a problem with uploading a large number of trickles.

See <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=1887"> this</a> thread.
There is a link near the end of the thread coming back here to the Windows Q &amp; A to the orignal discovery message.

Les