Posts by old_user46276

1) Questions and Answers : Unix/Linux : Ran out of disk space (there is more to this problem...) (Message 11349) Posted 24 Mar 2005 by old_user46276 Post: Hi, There are my prefs settings: Use no more than 1 GB Leave at least 1 GB Use no more than 50% of total space Use no more than 75% of total virtual memory I run BOINC in Linux. The client state doesn't have the string. But I have a backup of the client state from Phase 1. Would that help without me having to run from Phase 1 again? Do you belong to the dev team at CPDN? I don't have much free time, but I would be willing to help improve the CPDN code after a month or so. Let me know if you are interested. Thanks, Sarav > Hi Sarav, > > I'm afraid <a> href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=511391">your > result</a> looks pretty terminal to me, especially as it was returned 5 days > ago :( > > Any record of the model will have disappeared from your client_state.xml file, > and the only chance you have of resurrecting it is if you can find the string > <b>41eh_000210331</b> anywhere in a BOINC backup of the client_state.xml (or > client_state_prev.xml) file. In which case I <i>might</i> be able to help you > get the model going again. > > One question for you. > > What are the disk space limits in your general preferences set to? The 'Leave > at least' and 'Use no more than' settings <i>should</i> cause BOINC to suspend > projects before you run out of disk space. >
2) Questions and Answers : Unix/Linux : Ran out of disk space (there is more to this problem...) (Message 11341) Posted 24 Mar 2005 by old_user46276 Post: Hi, I was running my first model and I had only about 6-7 trickles to complete Phase 3. I went out of town and disk space got very low. Model 1 was stopped and a new model was downloaded and it started progressing. I now want to stop the new model and let the old model run to completion. My results page says "Client error" and "Computing" for the "Outcome" and "Client State" columns for the model 1. When I click on the WU link it shows another host has been assigned the model. Does it mean I can no more contribute to this model :( ? Please let me know how I can run my model 1 to completition. I'm even willing to delete the model 2 which is running now (only a few trickles so far). Thanks, Sarav
3) Questions and Answers : Unix/Linux : How to force a trickle!? (Message 10750) Posted 12 Mar 2005 by old_user46276 Post: Hi, Even if the reply is 0, follow the steps I mentioned in my earlier post and that should help. Sorry for not mentioning the tag name "min_rpc_time" on the earlier post. leggewie, yes BOINC (not just CPDN) should be stopped before you do this change. -Sarav > > > The tag is "min_rpc_time" > > > > Hi Josh, thank you for providing that info. > > > > For all my computers it says "0 > > > > Hi - now I'm confused. I have one Linux box that is daily sending trickles but > another that appears to be running (CPU maxed out on hadsm3...) for several > days with no returns at all. The min_rpc_time says "0" for both....?? I stoped > and restarted the non trickling one with "-update_prefs climateprediction.net" > but it did not report anything. > > kyle >
4) Questions and Answers : Unix/Linux : How to force a trickle!? (Message 8941) Posted 8 Feb 2005 by old_user46276 Post: Hi, Open client_state.xml. Search for the XML tag under the tag that belongs to CPDN (you can check this by looking at the tag under the tag). The value enclosed with in the tag is the absolute time in the future when BOINC will communicate again. It is in the format returned by gettimeofday() function. Run "date +%s" and note the big number returned by it. Add 600 to the number and put the result in the tag. BOINC should communicate to the server in 10 minutes. Example: $ date +%s 1107899069 Change 1108357200 to 1107899669 Hope this helps. -Sarav
5) Questions and Answers : Unix/Linux : VIZ again (Message 8935) Posted 8 Feb 2005 by old_user46276 Post: Hi, Open ./viz in some text editor and change the second line to: nice -n18 ./hadsm3viz_4.04* $1 I think it was initially nice -n18 ./hadsm3viz_4.03* $1 ./viz should work after that. -Sarav