Azure / doAzureParallel

A R package that allows users to submit parallel workloads in Azure
MIT License
107 stars 50 forks source link

Tasks failed due to save.image errors #366

Open commintern opened 4 years ago

commintern commented 4 years ago

For a job with 1000 tasks, from a time point, all the remaining tasks will fail with the following errors

Error in save(list = names(.GlobalEnv), file = outfile, version = version, : error writing to connection Calls: quit -> sys.save.image -> save.image -> save Execution halted

This error happens randomly to me, so I cannot give a minimal reproducible example. Sometimes, If I submit the exactly same job again, all the tasks can be successfully completed. So I believe it is possibly due to problems of Azure.

I am not familiar with Azure, therefore I am not sure what info I should provide here. Please let me know if any additional info is helpful.

Thanks.