on machines who can more quickly get data from s3, that data piles up behind the pipe to POST requests. this can fill up a machine. to avoid this we have a sequential mode now (the default). also all the arguments are provided with switches instead of being positional.
heffergm
commented
7 years ago
on machines who can more quickly get data from s3, that data piles up behind the pipe to POST requests. this can fill up a machine. to avoid this we have a sequential mode now (the default). also all the arguments are provided with switches instead of being positional.