Open thinkerbot opened 13 years ago
As a corollary, there should be options to fail after one or all when running in parallel (ie a single failure aborts all, or run all and report the failures).
And to implement this, the logging for run will have to change. A better model might be to monitor - ie to redirect stdout/stderr for each linecook_ssh, fork and wait for a result. Then make a monitoring script that takes each of the redirected files as inputs, collates results and provides commands to kill the run processes as needed.
Currently 'linecook run' copies and runs scripts on servers in serial. An option should exist to run in parallel if desired.