2

I have a small set of assets including:

  • Virtual machine host
  • Multiple virtual machines
  • Shared folders
  • AWS services
  • Backup software

If I were to get "hit by a bus" the people left behind that would have to sort things out would be lost. I fear that some important information could be totally lost.

What are the best practices for documenting:

  • what all the assets do
  • passwords
  • the backup strategy
  • administrative methods (web server, ssh, etc)

Anything else I should be thinking of and documenting?

Are there forms that can be used as a starting point?

What would you like to see coming into a situation like this blind?

Ant advice you seasoned admins can give would be helpful.

Eric Snyder
  • 193
  • 1
  • 12

1 Answers1

1

The best documentation is the one you update.

As priority 1, I recommend getting a password manager and sharing the important secrets in a secure way with, depending on the size of your outfit, colleagues, consultants and/or your boss.

(Tip: For a small monthly fee, a SaaS password manager like 1Password or similar can also keep a secure off-site copy of your passwords and important documents in case of fire or theft; and such systems also support secure sharing of notes and secrets among licensed users in your company.)

As for documentation:

My recommendation is not to overdo it in terms of writing entire books about your systems. Chances are, if you get run over by a bus, your replacement will be a computer literate person who will relatively easily understand your environment as long as they get some basic pointers.

Prioritize to document (or eliminate) non-standard solutions and weird legacy stuff that could have even competent people scratching their heads.

Another tip is to implement automation and software defined configuration as far as practical. An Ansible playbook, for example, doesn't only save you from having to perform recurring tasks manually. It also removes a lot of the human factor from each time you need to perform it, and - most important to your question - anybody who understands English can follow the instructions and understand what must be done and in what order to complete the task the playbook describes. And of course such a system can be version controlled, which your future self or any replacement will be grateful for some day.

Mikael H
  • 5,031
  • 2
  • 9
  • 18
  • Good observations. I was not aware of Ansible. Very cool. Your observations about eliminating non-standard solutions is also something I wasn't thinking about. – Eric Snyder Jan 13 '20 at 13:45