How to retrieve error message when running script on cluster

Hi,

I am running batch jobs using MATLAB remotely on a cluster. How to view the eventual error messages(where I hope to find the line of code where the problem is)? Thanks in advance.

Answer 1

[Edric Ellis]

The task within the job object returned by the batch command has a property called Error which has the error stack. You could do something like this:

job = batch('assert(false)');
wait(job);
getReport(job.Tasks(1).Error)

Also note that if you simply call load on the job you'll see the error.

[Bi Bu]

Thanks. Is it also possible to see a live log of the outputs of the batch in the Matlab command window? The same way in Unix it's done using tail -tail -f filename_output. Thanks again.

[Edric Ellis]

You can call the diary method while the job is running to do that (you need to be running a relatively recent version of Parallel Computing Toolbox for that to work). Unfortunately the diary method gives you the whole output each time, rather than a tail -f style.

>> j = batch('for ii=1:4,pause(1),disp(ii);end')
>> % After  a short while...
>> diary(j)
Warning: The diary of this batch job might be incomplete because the job is still running. 
--- Start Diary ---
   1
   2
--- End Diary ---