batch job failling with java exception

Hi,

I just got parallel server 2019b installed (on Windows 10) and I try to run a script on the server using the batch function as ollow:

job = batch( myCluster , 'myScript' , 'Pool' , 20 );

and I get systematically the same error:

Error: Cannot rerun job because at least one of its tasks has no rerun attempts left (The task has no rerun attempts left.).
Original cancel message:
java.lang.Exception: Error submitting fevalRequest: foundation::storage::vfs::Exception

The script can run locally, it can run using batch function on my local pool and it can run on the server using interactive pool.

Also note that the fail using even script as simple as:

test= 1;

I could not find anything online. I hope someone can help. In any case, thank you in advance to anyone willing to help.

Best regards

Rémy

[Edric Ellis]

I suggest contacting MathWorks support to help resolve this.

Answer 1

[Remy Lassalle-Balier]

Thank you Edric for your comment. I followed your advice and got help very quickly.

So for anyone having the same kind of issue, here is the fix that worked for me. It appears the service needed a clean install. So Mathworks support member told me to simply run these four command from the command line (admin mode):

cd $MATLABROOT$\R2019b\toolbox\parallel\bin
mjs stop
mjs uninstall
mjs install -clean
mjs start

Then, you obviously have to start workers which I did directly in the admin center.

Thank you very very much to Mathworks support! It is always a pleasure to deal with efficient and competent people.

Best regards

Rémy