Closed joshmoore closed 5 years ago
Better would be printing logs only on failure but I think that requires assumptions about the various containers in a user-choosable docker-compose.yml
. An additional feature would be printing Blitz-0.log
. If cleanup happened in travis and not inside .docker
, then travis could also call print-logs
.
This was used successfully to point to database problems (which turned to be the result of a bad postgres 9.6.14 release). Merging for future use and/or refactoring to automatically print logs on failure.
Currently testing in https://github.com/ome/omero-parade/pull/57 to try to pinpoint possible resource exhaustion.