fix: .sh gone when attempting 3 creates in a row

[ntorga]

The last container create returned the following error:

{
  "status": false,
  "body": "ReadFileInfoError: stat /tmp/ls-sitehere.sh: no such file or directory"
}

The .sh created by the API was likely removed by the the previous creation process. Need to check the .sh creation.

Also, since we're dealing with this, when retrying a create via the scheduled tasks, the .sh is no longer there. Maybe we should only remove the .sh when the creation is actually successful and allow an id on the name of the file so if we need to retry, it still exists. The housekeeper could remove the .sh after 24h.