Passing database name 1

[tnelson-doghouse]

Purpose

• To allow the passing of the MTK_DUMP_* variables into the task in one
  way or another, rather than assuming that the database is going to be
  the one in the Lagoon service

• Also instead allow the passing of the name of a variable that contains
  the MTK_DUMP_* value. For example,

  MTK_DUMP_HOSTNAME_NAME=DB_HOSTNAME_CENTRAL
  

  ... means that MTK_DUMP_HOSTNAME will be set to the value of
  DB_HOSTNAME_CENTRAL

Primary Changes

• Added calc_MTK_variable function which sets up one MTK_DUMP_* variable
• Set all the MTK_DUMP_* variables using the above function (and replaced
  existing code that set them)
• Changed GraphQL example to allow for the new variables

Other changes

• Removed san_db_tarfile line since it's not used
• A few tweaks based on the output of https://www.shellcheck.net/ (which
  you can install locally)
• Added a little more to the README

closes #14

[tnelson-doghouse]

For the record, this one was discussed with Sean. Our plan is that:

• The new variables will only be used for Captivate and similar projects that have a multi-tenant setup
• Most of the variables will be hardcoded somewhere; I'm thinking the task GQL would be a good place
• We can have the Dev or Ops team pass in the database name so that we can dump a sanitised database which /isn't/ configured in Lagoon

HTH,

[shreddedbacon]

I'm fine with this. I'd just like the example graphql command split to its own example file.

[tnelson-doghouse]

Hi @shreddedbacon , have updated the PR as requested.

Thanks!