Hmmm. I can't reproduce this, using the same WU template file. Are you using the newest server software (i.e., trunk)?
-- David On 01-Feb-2011 2:06 PM, Kamran Karimi wrote: > Hi all, > > We have been encountering strange errors recently: A rather large number of > our tasks fail because the command line arguments are incorrect. We traced > one specific failed task and were lead to the bin/create_work program. Here > is the trace: > > 1) These commands are used to submit work to BOINC's DB: > > ln -s `pwd`/ibadfp_p1/instance.txt `bin/dir_hier_path > instance.txt_31jan11_am_16_030_200_000_1_235` > ln -s `pwd`/ibadfp_p1/ipfile.bin `bin/dir_hier_path > instance.txt_31jan11_am_16_030_200_000_1_235.ip` > ln -s `pwd`/ibadfp_p1/qubit.param `bin/dir_hier_path > instance.txt_31jan11_am_16_030_200_000_1_235.param` > bin/create_work -appname fokker_planck -wu_name > 31jan11_am_16_030_200_000_1_235 \ > -wu_template ibadfp_p1/31jan11_am_16_030_200_000_1_235_wu \ > -result_template ibadfp_p1/31jan11_am_16_030_200_000_1_235_result > instance.txt_31jan11_am_16_030_200_000_1_235 > instance.txt_31jan11_am_16_030_200_000_1_235.ip > instance.txt_31jan11_am_16_030_200_000_1_235.param > > > 2) These are the contents of the file > ibadfp_p1/31jan11_am_16_030_200_000_1_235_wu: > > <file_info> > <number>0</number> > </file_info> > <file_info> > <number>1</number> > </file_info> > <file_info> > <number>2</number> > </file_info> > <workunit> > <file_ref> > <file_number>0</file_number> > > <open_name>instance.txt_31jan11_am_16_030_200_000_1_235</open_name> > <copy_file/> > </file_ref> > <file_ref> > <file_number>1</file_number> > > <file_name>instance.txt_31jan11_am_16_030_200_000_1_235.ip</file_name> > <open_name>ipfile.bin</open_name> > <copy_file/> > </file_ref> > <file_ref> > <file_number>2</file_number> > > <file_name>instance.txt_31jan11_am_16_030_200_000_1_235.param</file_name> > <open_name>qubit.param</open_name> > <copy_file/> > </file_ref> > <command_line> --T 0.03 --t_f 0.0002 --n_particles_per 2 --h_ramp_mid > 0.000 --h_ramp_width 0.0025 --time_step_factor 48.0 --gamma_frac 1.0 > --t_fraci 0.38 --t_fracf 0.88 --input_file > instance.txt_31jan11_am_16_030_200_000_1_235</command_line> > <target_nresults>1</target_nresults> > <max_success_results>1</max_success_results> > <min_quorum>1</min_quorum> > <rsc_fpops_est>5e13</rsc_fpops_est> > <rsc_memory_bound>5e7</rsc_memory_bound> > <rsc_fpops_bound>1e20</rsc_fpops_bound> > <rsc_disk_bound>1e8</rsc_disk_bound> > <delay_bound>1728000</delay_bound> > </workunit> > > > Please note the command line argument "--input_file" > > > 3) This is what we see in the database: > > <workunit> > <file_ref> > <file_name>instance.txt_31jan11_am_16_030_200_000_1_235</file_name> > <open_name>instance.txt_31jan11_am_16_030_200_000_1_235</open_name> > <copy_file/> > </file_ref> > <file_ref> > <file_name>instance.txt_31jan11_am_16_030_200_000_1_235.ip</file_name> > > <file_name>instance.txt_31jan11_am_16_030_200_000_1_235.ip</file_name> > <open_name>ipfile.bin</open_name> > <copy_file/> > </file_ref> > <file_ref> > <file_name>instance.txt_31jan11_am_16_030_200_000_1_235.param</file_name> > > <file_name>instance.txt_31jan11_am_16_030_200_000_1_235.param</file_name> > <open_name>qubit.param</open_name> > <copy_file/> > </file_ref> > <command_line> > --T 0.03 --t_f 0.0002 --n_particles_per 2 --h_ramp_mid 0.000 --h_ramp_width > 0.0025 --time_step_factor 48.0 --gamma_frac 1.0 --t_fraci 0.38 --t_fracf 0.88 > --input_file instance.txt_31jan11_am_16_030_200_00_11_235 > </command_line> > </workunit> > > > Please note how the last few characters corresponding to the "--input-file" > parameter have changed from "200_000_1_235" to "200_00_11_235" > > The result is a failure on a volunteer computer, with the app complaining > that it can't open the input file. > > How could this happen? Any help in resolving this issue is appreciated. > > -Kamran > _______________________________________________ > boinc_dev mailing list > [email protected] > http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev > To unsubscribe, visit the above URL and > (near bottom of page) enter your email address. _______________________________________________ boinc_dev mailing list [email protected] http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev To unsubscribe, visit the above URL and (near bottom of page) enter your email address.
