doAzureParallel
doAzureParallel copied to clipboard
foreach drops cores without warnings
Description I am running foreach with 4 cores. If the RAM gets full, it drops all the iterations in the core without any warning or error.
-
Do you have a reproducible sample? Cluster configuration and foreach sample will help.
-
Is this one VM with 4 cores or multiple VMs? Are you using maxTasksPerNode?
-
Is the errorhandling argument set to 'stop'? If it set to 'stop', there should have been an error in R, receiving a RAM error. The client will eventually see that a task has failed and would stop the job/terminate all the iterations.