Use database locking to prevent concurrent Terraform jobs

Currently, the Terraform code uses chdir in the Ruby process to change
to the right working directory. This affects all threads in the
process, and thus means that all Terraform jobs conflict with each
other.

It would be good to fix this somehow, e.g. not changing directory, and
just specifying the directory in which to run Terraform when starting
it, but for now, use the advisory locks to force the jobs to run
sequentially.

0 files changed, 0 insertions, 0 deletions