I introduced a bug in future.batchtools (= 0.7.0) released on 2018-05-04. I can reproduce it on a Torque/PBS cluster, but not on an SGE cluster although I expect that it could/should appear everywhere. Example:
library(future.batchtools)
plan(batchtools_torque)
y <- future.apply::future_lapply(1:3, FUN = identity)
# Error in readLog(id, reg = reg) :
# Log file for job with id 1 not available
This should have been caught by my local pre-release tests, but due to a mistake of mine I've disabled the tests that actually run future.batchtools on HPC schedulers. The only tests running was relying on the batchtools_local backend.
I introduced a bug in future.batchtools (= 0.7.0) released on 2018-05-04. I can reproduce it on a Torque/PBS cluster, but not on an SGE cluster although I expect that it could/should appear everywhere. Example:
This should have been caught by my local pre-release tests, but due to a mistake of mine I've disabled the tests that actually run future.batchtools on HPC schedulers. The only tests running was relying on the
batchtools_local
backend.