Original issue was posted on SO here
Tested on version 220.127.116.11
It seems that setting the `verbose = T` parameter in h2o.deeplearning(), h2o.radomforest, h2o.gbm, etc.. along with the `nfolds` parameter greater than two causes the algorithm to break.
If you leave the `verbose = FALSE` then you can set `nfolds` parameter higher than 3 and the algorithm will not break.
example code is shown below to demostrate the error for a few algos.
the error is of the form
reproducible code below