climateprediction.net home page
Posts by old_user248

Posts by old_user248

1) Questions and Answers : Unix/Linux : BBC Climate prediction takes along time! (Message 21602)
Posted 25 Mar 2006 by Profile old_user248
Post:
The results will be used anyway. As far as I know that date is for priority setting, but I am probably wrong and will get pounced on because of it :)

I generally just let my Linux boxes run.

Dave
2) Questions and Answers : Windows : Work Unit restarts when i start BOINC (Message 21601)
Posted 25 Mar 2006 by Profile old_user248
Post:
Another thing that may be slowing the computations down is if your are running the screen saver side as well. That takes up quite a bit of the CPU capacity, slowing the computation down. A lot of people just use a \"Blank\" screen saver.

I also click on \"run always\", not just based on preferences, which I don\'t know if it really makes any difference.

Dave
3) Questions and Answers : Windows : graphics (Message 21586)
Posted 25 Mar 2006 by Profile old_user248
Post:
I take it that you are using the screen saver to view the graphics. Viewing the graphic either from the screen saver or from \"Show Graphic\" on the work tab actually uses a lot of the CPU capacity, making it take longer to complete the model. From this latter view you will be able to change to clouds, rain, temp, pressure, etc. using the options on screen.

Most people have the \"Blank Screen\" as their screen saver and when they want to view the graphic they open it up from the Boinc Work tab. In XP, right click the screen, select properties, then select the Screen Saver tab and change to blank screen, if you\'d like to do this.

Dave
4) Questions and Answers : Unix/Linux : CPDN Sulfur_cycle 4.21 (Message 16021)
Posted 15 Sep 2005 by Profile old_user248
Post:
I take it that you removed completely the old installation of boinc and any work units in progress?

With your new install did the model run for very long before giving you this error?

Have you tried stopping and restarting the model?

Is the computation error right at start up or boinc and / or has it rewound to retry the calculations?

With some more information we should be able to help
5) Questions and Answers : Unix/Linux : SC Model won\'t start (Message 15966)
Posted 13 Sep 2005 by Profile old_user248
Post:
What did you do to start this problem astoWX? Which boinc version? Was it a reboot, perhaps accidental improper model shut down, etc? Just curious so that I don\'t accidentally create a similar problem here.

Is this mainly with the Sulphur models? They seem more sensitive to me, I think I\'ve crashed a couple myself somehow.

Thx
Dave
6) Questions and Answers : Unix/Linux : Download Msg. query (Message 15774)
Posted 7 Sep 2005 by Profile old_user248
Post:
> Firewalls have vendor-specific ways to do things but there should be an option
> to control settings for inclusion and exclusion.
>
&gt; General info: <a> href="http://boinc-doc.net/boinc-wiki/index.php?title=Master_file_fetch_failed"&gt;
&gt; here</a> and <a> href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/quick_faq.php#firewall"&gt;here.</a>
&gt;
&gt; [Edit] Please post your firewall brand so that someone who also uses it can
&gt; give specific directions.
&gt;
Were you connected to the internet when you got the fetch file error? As asked has you boinc installation downloaded a work unit already? If it's being connected that is the issue, once you have a work unit disable boinc network access from the boincmgr and only enable it when connected

What flavour of Linux are you using and are you using SELinux? If using the latter check /var/log/audit/audit.log (I think that is the proper one, not on my SE machine right now). Also check your /var/log/messages to see if there is anything. Here you should find it there is something in you linux install that is blocking the access.

Without more details this is all that I can think of.
7) Questions and Answers : Unix/Linux : Upgrading 4.19 -> 4.43 boinc client (Message 15615)
Posted 2 Sep 2005 by Profile old_user248
Post:
&gt; Yeap, it's more easy to upgrade in Linux. just stop Boinc, change the binary,
&gt; and restart it.
&gt; If it doesn't work, you just put the old binary back ;o)

Thanks both of you. Done and running. :)
8) Questions and Answers : Unix/Linux : Upgrading 4.19 -> 4.43 boinc client (Message 15594)
Posted 1 Sep 2005 by Profile old_user248
Post:
Anyone experience any issues when updating or have any suggestions to avoid loosing work, especially with these LONG sulphur models.

tia.
9) Questions and Answers : Unix/Linux : Zombies! (Message 15593)
Posted 1 Sep 2005 by Profile old_user248
Post:
&gt; One of the messages says schedule cpus: must schedule<br><br>Is this a normal
&gt; message on a 1 cpu machine?
&gt;
There is a message from Carl in the other forum about him using the latest libs from a debian install, and I guess things aren't staticly linked anymore.

http://www.climateprediction.net/board/viewtopic.php?t=3173
10) Questions and Answers : Unix/Linux : Upload Problems (Message 15592)
Posted 1 Sep 2005 by Profile old_user248
Post:
&gt; The upload server at Bern is down, so you'll have to wait.
&gt; It's possible that the Uni is now having water problems.
&gt; There is a message on the news page I suggest that you disable BOINC network access, if this is possible with your
&gt; setup.
&gt; Next, make a full backup of the BOINC folder. (Suspend 1st, as there is a
&gt; 'lock file'.
&gt; Apparently BOINC will abort the upload after 2 weeks.
&gt;
&gt; My plan is: Disable access and backup, (done), wait until the server is up,
&gt; backup again, restore the backup that I made today, upload, restore the next
&gt; backup, delete the zips if they are still there, and then continue running and
&gt; trickleing.

There is a message <a href="http://www.climateprediction.net/board/viewtopic.php?p=29490&amp;highlight=#29490"> here,</a> about a work around for the upload problems posted by Thyme Lawn. This is for clients that have been trying for about a week so that things aren't lost and everyone can keep happilly crunching.

11) Questions and Answers : Unix/Linux : boincmgr hangs X-Windows (Message 14330)
Posted 13 Jul 2005 by Profile old_user248
Post:
I am not 100% up on this latest vesion of boinc, but I usually start boinc (cd BOINC ./boinc) in a terminal window first then run the manager. Is this what you are trying? I don't think the manager will start up everything on it's own.

The instructions that I have been able to find so far for using this latest version are pretty sparse. I run FC4 with Gnome, but as far as I know other users have it running on various flavours of Linux.

Dave
12) Questions and Answers : Unix/Linux : Quit running climateprediction.net??? (Message 13458)
Posted 15 Jun 2005 by Profile old_user248
Post:
&gt; CPDN is far more hardware stressing than other DC projects: so don't take for
&gt; granted that your machine is stable.
&gt;
&gt; for me. Test your machine with programs like memtest, superpi and prime95 (I
&gt; don't know if these programs exist for Linux,

memtest86 runs from a diskette and actually formats the floppy using linux and available at: http://www.memtest86.com/

Prime95 does exist for Linux at: http://www.mersenne.org/freesoft.htm

Run the torture test (-t) for a while and if there is any problems with the stability it is usually good at finding them.

DaveN
13) Questions and Answers : Windows : Model has crashed (Message 13361)
Posted 12 Jun 2005 by Profile old_user248
Post:
It might also be a good idea to check the manufacturer for your hard disk and from their web site download a utility to check the HD for problems. If this only happens during intense disk activity then somewhere in that end of the chain is where the problem probably rests.

Also, what other programs are running when this happens that may be doing a fair bit of activity? Myself, I have my antivirus software set to exclude the directories under climate prediction except when I a weekly scan for problems.

On the 4.12 front, it seems that a some people are having problems with this particular version, especially if running Win98 or Me. Just to be on the safe side maybe try running Prime95 torture test or other intense stability checker just in case something has changed within your system. I personnnally haven't tried all the lastest vesions of boinc so can't comment on the boinc version. I just run the ones currently listed on the CP web site.
14) Questions and Answers : Unix/Linux : what determines idle/in-use: detecting remote activity so boinc will suspend (Message 13359)
Posted 12 Jun 2005 by Profile old_user248
Post:
Nealmcb:

boinc and climate predcition under Linux use the nice levels. Boinc defaults to a nice level of 19 which is the lowest (man nice for more information). This means that if almost anything happens on your system your kernel will give the processing power over to the other applications preferentially over boinc.

Under Windows boinc uses a normal priority and the distributed computing apps (eg CP) have a priority set to the lowest. There was a discussion on the at the end of 2004 in reference to CP classic if you want to try searching the other message board for more information.

DaveN
15) Questions and Answers : Unix/Linux : Model @ phase 3 No trickles (Message 13194)
Posted 7 Jun 2005 by Profile old_user248
Post:
&gt; That'll be the problem of results being sent to more than one host then, Dave.
&gt;
&gt; From a science perspective it's best if you continue to run that model to
&gt; completion. We are trying to persuade the powers that be to reassign the
&gt; results to the host that's returned the greatest number of trickles, so
&gt; hopefully you'll eventually get the credits for your work.

I am in this mainly from the science aspect, standings are a bonus. I was always going to run the model to completion. Either way it would ensure that someone completed the run and / or provide more fodder for the processor comparisons.

Thanks for the explanations
Dave
16) Questions and Answers : Unix/Linux : Model @ phase 3 No trickles (Message 13183)
Posted 6 Jun 2005 by Profile old_user248
Post:
&gt; Any unsent trickles will be in the files
&gt; projects/climateprediction.net/trickle_up*.xml, which will be deleted after
&gt; the scheduler responds that it's accepted them. If your update_prefs didn't
&gt; work try the procedure detailed

Half my post didn't make it up.

Turns out this result id 817854 with Name 3b52_200175960_1 is not listed on my results page for "bruce" which without close examination I was thinking the model hadn't trickled. now on closer examination it turns out this model was sent to both me and the user "joco" who quickly crashed it.

Here is the link to <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=151500">bruce </a> from which you'll be able to see what is going on for yourself

&gt; From <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/show_user.php?userid=248">here</a>
&gt; you can see my account data. I'm not sure if I'm actually getting credit for
&gt; this or not but it is almost done anyway.
&gt;
&gt; Sorry for my confusion about the trickles.
&gt; Dave
&gt;
&gt;
&gt;
17) Questions and Answers : Unix/Linux : Model @ phase 3 No trickles (Message 13182)
Posted 6 Jun 2005 by Profile old_user248
Post:
&gt; Any unsent trickles will be in the files
&gt; projects/climateprediction.net/trickle_up*.xml, which will be deleted after
&gt; the scheduler responds that it's accepted them. If your update_prefs didn't
&gt; work try the procedure detailed <a> bruce<a>

From <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/show_user.php?userid=248">here</a> you can see my account data. I'm not sure if I'm actually getting credit for this or not but it is almost done anyway.

Sorry for my confusion about the trickles.
Dave

18) Questions and Answers : Unix/Linux : Model @ phase 3 No trickles (Message 13164)
Posted 6 Jun 2005 by Profile old_user248
Post:
On one of my machines running Fedora 3 it\'s second (or third) model hasn\'t trickled once. How can I get it to trickle?

I have just tried the -update_prefs http://... and will wait to see if that worked. I am probably over the maximum number of trickles that the servers can handle now, is there a way of removing some of them, updating, then sending the rest? The dataout directory contains 263 \".nc\" files which is probably the problem.

TIA
Dave
19) Questions and Answers : Windows : I can attach climateprediction project to my BOINC (Message 12164)
Posted 30 Apr 2005 by Profile old_user248
Post:
Check to ensure you have no extra spaces around the URL.
20) Questions and Answers : Windows : Why do my Client start at Zero again? (Message 11722)
Posted 9 Apr 2005 by Profile old_user248
Post:
These just indicate the models that have tried to run and their status. The actual error that the CPDN model generated on failure would be contained in one of the files uploaded.

I realize that your computer is prtty new but have you ensured it's stability by running programs such as Prime95 torture test for a few hours? the CPDN model stress computers more than other distributed computing projects and due to the interdepenancy of calculations it can find problems quite easily.

Can you post your CPU, memory, and hard drive information? Have you noticed if you are doing anything in particular when the model crashes? were you running the visualizations or some other demanding piece of software when the model crashed?

Hopefully we can all help you get back up and running again.

Dave


Next 20

©2019 climateprediction.net