climateprediction.net home page
HadCM3 Model blew with exit code 22

HadCM3 Model blew with exit code 22

Message boards : Number crunching : HadCM3 Model blew with exit code 22
Message board moderation

To post messages, you must log in.

AuthorMessage
old_user452941

Send message
Joined: 22 May 07
Posts: 35
Credit: 1,065,741
RAC: 0
Message 35459 - Posted: 8 Nov 2008, 2:43:13 UTC
Last modified: 8 Nov 2008, 2:44:36 UTC

This HadCM3 model blew with the following BM messages:

11/7/2008 8:41:50 AM|climateprediction.net|[task_debug] Process for hadcm3ivolc_l2o8_2000_80_36001832_1 exited
11/7/2008 8:41:50 AM|climateprediction.net|[task_debug] task_state=EXITED for hadcm3ivolc_l2o8_2000_80_36001832_1 from handle_exited_app
11/7/2008 8:41:50 AM|climateprediction.net|[task_debug] result state=COMPUTE_ERROR for hadcm3ivolc_l2o8_2000_80_36001832_1 from CS::report_result_error
11/7/2008 8:41:50 AM|climateprediction.net|[task_debug] Process for hadcm3ivolc_l2o8_2000_80_36001832_1 exited
11/7/2008 8:41:50 AM|climateprediction.net|[task_debug] exit code 22 (0x16): The device does not recognize the command. (0x16)
11/7/2008 8:41:50 AM|climateprediction.net|Computation for task hadcm3ivolc_l2o8_2000_80_36001832_1 finished

All of my wingmen\'s models also blew on this one.

Sorry, all of the current files are gone since the model reported.

I know this is a vague error message. In the future, what particular information would be helpful to you in debugging? I can set CPDN to \"no new tasks\" in attempt to preserve the files.
ID: 35459 · Report as offensive     Reply Quote
Profile Iain Inglis

Send message
Joined: 9 Jan 07
Posts: 467
Credit: 14,549,176
RAC: 317
Message 35461 - Posted: 8 Nov 2008, 11:27:33 UTC

The \"stderr out\" text is usually enough to tell whether there\'s a model problem - \"negative pressure\" and so on. However, the text for your model ends with the incomprehensible:

Model crashed: 

Model crashed: 

Model crashed: 

Model crashed: 

Model crashed: 

Model crashed: 
Sorry, too many model crashes! :-(
called boinc_finish


This does at least say that the model had six attempts at recovering, which is normal, but none of them succeeded in circumventing the problem.

When there appears to be a BOINC problem, then there are debugging flags that can be set in cc_config.xml. Here\'s mine (with red for the extra debugging):

<cc_config>
<log_flags>
<task>1</task>
<file_xfer>1</file_xfer>
<sched_ops>1</sched_ops>
<checkpoint_debug>1</checkpoint_debug>
</log_flags>
<options>
<save_stats_days>90</save_stats_days>
</options>
</cc_config>


However, the material in \'stderr out\' comes from the science application rather than BOINC itself, and there\'s no way that I know of increasing its verbosity.
ID: 35461 · Report as offensive     Reply Quote
old_user452941

Send message
Joined: 22 May 07
Posts: 35
Credit: 1,065,741
RAC: 0
Message 35462 - Posted: 8 Nov 2008, 15:32:12 UTC

Thanks, Iain. I\'ve added those flags to my xml file.
ID: 35462 · Report as offensive     Reply Quote

Message boards : Number crunching : HadCM3 Model blew with exit code 22

©2024 climateprediction.net