ensure that migration output can be viewed when stuck on a bad migration
Recently we saw
root 30925 18514 0 05:11 pts/0 00:00:00 /bin/sh -c SKIP_POST_DEPLOYMENT_MIGRATIONS=1 /usr/bin/gitlab-rake db:migrate
stuck for a very long time, it was difficult to tell where it was exactly because when ansible executes this command it buffers all of stdout/stderr until it completes.
We should probably create a logfile for this which would be simple enough with | tee /some/log
.