On long batch jobs there are failures which are erratic. These are mostly due to hang-ups of the server it seems, which ends up in some http protocol confusion. I'm now forcing http/2 on to all processes and checking if this works, if not the use of httr::RETRY instead of a GET command might be a fix to kickstart things.
On long batch jobs there are failures which are erratic. These are mostly due to hang-ups of the server it seems, which ends up in some http protocol confusion. I'm now forcing http/2 on to all processes and checking if this works, if not the use of
httr::RETRY
instead of aGET
command might be a fix to kickstart things.