Why the worker save the logsout first to the local machine?

9 Ansichten (letzte 30 Tage)
Szabó Bence Károly
Szabó Bence Károly am 30 Aug. 2023
Kommentiert: Amish vor etwa 2 Stunden
I am trying to run a batch simulation in MATLAB using the following code:
in(i) = Simulink.SimulationInput(model);
in(i) = in(i).setPreSimFcn(@(x) myInitFun());
in(i) = in(i).setModelParameter(...
'LoggingToFile', logFileFlag,...
'SignalLogging', signalLogging,...
'LoggingFileName',logFilePathName,...
'StopTime', num2str(stop_time.sim_end),...
'SaveFormat', 'Dataset',...
'SaveTime', 'off',...
'SaveState', 'off',...
'SaveOutput', 'off',...
'SaveFinalState', 'off',...
'DSMLogging', 'off',...
'ReturnWorkspaceOutputs', 'on',...
'Profile', 'off',...
'InspectSignalLogs', 'off',...
'DatasetSignalFormat', 'timeseries'...
);
simJob = batchsim(myCluster, in, ...
'Pool', wn,...
'AutoAddClientPath', false,...
'AutoAttachFiles', false,...
'ManageDependencies', 'off',...
'AttachedFiles',filesToAttach,...
'TransferBaseWorkspaceVariables', 'off',...
'CleanupFcn', @Simulink.sdi.clear);
I've configured logging to save some signals from the Simulink model, but I'm facing an issue. Instead of saving the log file in the server directory specified by logFilePathName, each worker saves the log.mat file to its temporary folder. Only after all workers are done does it move the log to the correct folder. This has caused the cluster machine to run out of disk space.
What can I do to avoid this issue and make sure the log files are saved directly to the specified server folder to prevent disk space problems on the workers?
  1 Kommentar
Amish
Amish vor etwa 2 Stunden
Can you check if the `logFilePathName` is a network path that is accessible by all worker nodes? Do they all have the necessary permissions to write to the path?
You can also consider using `parfeval` to run simulations asynchronously. This might give you more control over file paths and data handling, allowing you to directly write to the desired location.

Melden Sie sich an, um zu kommentieren.

Antworten (0)

Kategorien

Mehr zu Prepare Model Inputs and Outputs finden Sie in Help Center und File Exchange

Produkte


Version

R2020b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by