Just had a boinc funny. Tried starting up viz (win), got entry in taskbar
but no picture. Shut it down, "Application is not responding...", "Yes, I
really do want it shut down!" Then decided to restart boinc. When it was up
again I had 2(!) projects, one with account=verstapp, one blank, both
allocated 100% resource share. Detached the blank one. boinc wanted to
contact Oxford, I connected to let it. Got such interesting messages as:
--- - 2004-08-21 10:56:29 - CPU scheduler starvation imminent; requesting
more work [funny, I'd had 2 months of work before starting viz.]
--- - 2004-08-21 10:56:36 - CPU scheduler starvation imminent; requesting
more work
climateprediction.net - 2004-08-21 10:56:36 - Requesting 5173084 seconds of
work
climateprediction.net - 2004-08-21 10:56:36 - Sending request to scheduler:
http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi
climateprediction.net - 2004-08-21 10:56:39 - Scheduler RPC to
http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi succeeded
climateprediction.net - 2004-08-21 10:56:39 - General preferences have been
updated
climateprediction.net - 2004-08-21 10:56:39 - Message from server: No work
available (daily quota exceeded)
--- - 2004-08-21 10:56:39 - General prefs: from climateprediction.net (last
modified 2004-08-21 10:54:02)
--- - 2004-08-21 10:56:39 - General prefs: using your defaults
climateprediction.net - 2004-08-21 10:56:39 - No work from project
climateprediction.net - 2004-08-21 10:56:39 - Deferring communication with
project for 1 hours, 0 minutes, and 0 seconds


Is that my daily quota (didn't know I had one), the server's (didn't know
it had one), or the usual servers' work-to-rule, ie take the weekend off?
So now I'm back to one project, as it should be, and no work, as it shouldn't be. distributed.net is humming along nicely

Cheers,

PeterV.
<IMG SRC="http://boinc.mundayweb.com/cpdn/stats.php?userID=4&amp;trans=off">

Yeah, that seems odd. I can't imagine CPDN needing a daily quota like S@H or P@H.

<a><img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=18"></a>

&gt; Yeah, that seems odd. I can't imagine CPDN needing a daily quota like S@H or
&gt; P@H.
&gt;

I can imagine.
Peter, see the 'exit code -5' thread there http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=187

I think that if server has some Quota to send nn models a day, this prevents "Short Run Maniac" phenomenon (when WUs crashes at the very beginning and more and more are downloaded. Good idea i guess.

Peter, can you also check if you are not victim of 'exit code -5' please?

err, yeah, we need a quota, some people are running bizarre machines and one guy chewed up 5000 workunits! so now the quota is 4 per day, if anyone needs more 1 per day per machine then something is wrong!

Honza,
No error -5, but then I've restarted boinc in the hopes of getting something to start working again, thus trashing the previous 'log'.

Carl,
I can't imagine that it would have wanted any new WUs at all. It had 3 WUs, 1 running, 2 ready to run, before I restarted it, none after the restart.
Cheers,

PeterV.
<IMG SRC="http://boinc.mundayweb.com/cpdn/stats.php?userID=4&amp;trans=off">

I guess your "keep enough work for x days" is set pretty high so it must have wanted to get some more workunits. I have it set now so that you'll get a workunit even at the default ".1 to 1 days"

Peter, have you checked in task manager if BOINC completely shut down the last time around? That there is no secret other version running? (A reboot will sometimes help as well).
--------------------
Jordâ„¢
<img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=2&amp;trans=off">

Just rebooted computer, reconnected to net, startup tried starting boinc -

climateprediction.net alpha test - 2004-08-21 22:15:39 - Requesting 858109 seconds of work
climateprediction.net alpha test - 2004-08-21 22:15:42 - Master file fetch failed
climateprediction.net alpha test - 2004-08-21 22:15:42 - Could not contact any schedulers for http://climateapps2.oucs.ox.ac.uk/cpdn /.
climateprediction.net alpha test - 2004-08-21 22:15:42 - Could not contact any schedulers for http://climateapps2.oucs.ox.ac.uk/cpdn /.
climateprediction.net alpha test - 2004-08-21 22:15:42 - Deferring communication with project for 1 minutes and 0 seconds
climateprediction.net - 2004-08-21 22:15:42 - Deferring communication with project for 57 minutes and 43 seconds

Are you sure its not a server work-to-rule?
Cheers,

PeterV.
<IMG SRC="http://boinc.mundayweb.com/cpdn/stats.php?userID=4&amp;trans=off">

I think we need one of the John's here (JM7 or JKeck), but doesn't the Master File Fetch error mean that the client_state.xml file is corrupted?

Hold on... what's that scheduler there? http://climateapps2.oucs.ox.ac.uk/cpdn / ??

Try detaching from the project and re-attaching to it, making sure there's no spaces anywhere in the URL and none before or behind the account ID number. For as far as I know the scheduler lives on http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi ...

Editing again: Maybe Carl will want to check if the master address for the scheduler (in the main file) is pointing to the correct address as well.
--------------------
Jordâ„¢
<img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=2&amp;trans=off">

&gt; err, yeah, we need a quota, some people are running bizarre machines and one
&gt; guy chewed up 5000 workunits! so now the quota is 4 per day, if anyone needs
&gt; more 1 per day per machine then something is wrong!

Wasn't thinking about that, sorry...

<a><img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=18"></a>

Just woke up - its sunday morning (in Oz anyway) and found that I once <i>again</i> had 2 projects - 1.climate..., verstapp, 6900+, 400+, 100%, and 2.climate...,blank , blank, blank, 100%. So once again I detached the phantom one, once again messages showed:

climateprediction.net alpha test - 2004-08-22 07:42:06 - Resetting project
climateprediction.net alpha test - 2004-08-22 07:42:06 - Detaching from project
--- - 2004-08-22 07:43:50 - CPU scheduler starvation imminent; requesting more work
--- - 2004-08-22 07:43:55 - CPU scheduler starvation imminent; requesting more work
climateprediction.net - 2004-08-22 07:43:55 - Requesting 1716290 seconds of work
climateprediction.net - 2004-08-22 07:43:56 - Sending request to scheduler: http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi
climateprediction.net - 2004-08-22 07:43:59 - Scheduler RPC to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi succeeded
climateprediction.net - 2004-08-22 07:43:59 - Message from server: No work available (daily quota exceeded)
climateprediction.net - 2004-08-22 07:43:59 - Project prefs: using your defaults
climateprediction.net - 2004-08-22 07:43:59 - No work from project
climateprediction.net - 2004-08-22 07:43:59 - Deferring communication with project for 1 hours, 0 minutes, and 0 seconds

Update:
Anyway I just said 'stuff it' and detached/reattached my 'real' project (with 50% of model done). Disk space went from 300+MB down to 7MB then quickly bounced back to 36MB but at least it seems to be working again. Time to shut off distributed.net. Pity about that half-finished model...


Cheers,

PeterV.
<IMG SRC="http://boinc.mundayweb.com/cpdn/stats.php?userID=4&amp;trans=off">

&gt; Just woke up - its sunday morning (in Oz anyway) and found that I once
&gt; <i>again</i> had 2 projects - 1.climate..., verstapp, 6900+, 400+, 100%, and
&gt; 2.climate...,blank , blank, blank, 100%. So once again I detached the phantom
&gt; one, once again messages showed:
&gt;
&gt;[snip]

When you get the extra space in the URL just detaching and reattaching usually is not enough. You also need to exit the client and manually delete the extra account...xml file(s) and project directory(s). If you don't the client will just reattach to the bogus project the next time the client is restarted.
<br>John Keck
BOINCing since 2002/12/08

&gt; err, yeah, we need a quota, some people are running bizarre machines and one
&gt; guy chewed up 5000 workunits! so now the quota is 4 per day, if anyone needs
&gt; more 1 per day per machine then something is wrong!
&gt;
How about multi CPU machines? One WU per day per CPU. S@H has one machine running that has 14 CPUs. Maybe increase this number to be the max of the number of CPUs on the largest machine?

jm7