Failing dat clone on deploys from master

[garrying]

• Operating system: Ubuntu Trusty (Travis CI)
• Node Version: 10.16.0
• Dat Version: v13.13.1

We update the dat version of ournetworks.ca through a build step (after_deploy) after the site is built and deployed to our http server.

The operation of dat clone often fails due to the archive not being completely cloned after 10 minutes, which ends with an error. To get around this we've changed the archive's key to remove old meta data that may have caused the issue (50526d0). We also upload the new archive on hashbase.io. But even then, dat clone has not been consistent.

Successful build on master: https://travis-ci.org/ournetworks/ournetworks.ca/builds/555631939
Error build on master: https://travis-ci.org/ournetworks/ournetworks.ca/builds/554704192

Similar issue in the dat CLI repo: dat-ecosystem/dat#1064

Expected Behavior

When dat clone dat://<link> --exit is run, and an archive is available, the operation should exit without error after it has cloned the archive.

Actual Behavior

dat clone dat://<link> --exit stops at a certain point in the clone step and the operation timeouts after 10 minutes.

[garrying]

I think this has been fixed with 4dbfedd; clone was moved into the before_install. Connecting to the job instance at after_deploy doesn't seem to resolve all the time.