Mass administration in Linux [closed]

Question

I am curious if Linux have anything close to what PowerShell can do when come to managing servers in Enterprise? For example, retrieving server's information on a big scale and also making changes.I knew that Posh is getting in into Linux territory but I am not sure how much can it do to Linux Also,can Linux return it data in a more structural way in the form of objects and etc?

Answer 1

PowerShell is just a scripting language. By itself it is no more powerful or useful than any other scripting language or tool.

IMHO You're mistaking the (power of that) tool with the convenience and power that a single management domain such as Active Directory provides.

But yes, Linux systems can also be integrated in a single management domain, in AD even but there are also (Open Source) alternatives. And yes Linux provides tooling to use such management domains effectively for the large scale management of systems.

https://en.wikipedia.org/wiki/List_of_systems_management_systems

https://en.wikipedia.org/wiki/Comparison_of_open-source_configuration_management_software

https://en.wikipedia.org/wiki/System_Security_Services_Daemon

https://github.com/puppetlabs/facter

Answer 2

To manage linux servers in bulk I personally use pssh (parallel ssh), as well as configuration management (CM) tools like puppet/ansible/chef to administrate machines in bulk. CM tools are kind of like manifests (this file should be here, that program should always be running, etc), and parallel ssh is just like running single bash commands, except on a large number of machines at once. There's probably a lot of other tools for doing the same thing, with linux there's usually many good ways of doing anything (except reading html email from the command line).

While bash and the linux userspace tools do not have a concept of objects, their text and data processing capabilities are fantastic and it has close to no size limitations. Additionally quite literally everything usable from the command line is designed to be used in bash scripts as well. The end result is extremely powerful.

Answer 3

I use Ansible and love it. What really hooked me at first was the ansible_facts. One simple playbook would return a scathing amount of information about a given system. Early on I just saved the output and grep'd it for the key:value pair I wanted. There's probably a better way to do that.

In the simplest configuration example, you keep an "inventory" of hosts you want to run Ansible on. It's something like:

[my_hosts]
host1.mydomain.com
host2.mydomain.com
...

Then you can run an Ansible "playbook" on my_hosts and it will hit every host you have configured in that block. A playbook is just a YAML configuration containing commands, variables and formatting which Ansible understands. It's not a shell script. It's a "way" to tell Ansible what to do on a host. It took me a while to wrap my head around it.

Again, for simplicity sake, you can make a playbook that runs a shell command and run that playbook on every host. Here's an "uptime" playbook I used often.

---
- hosts: all
  gather_facts: false

  tasks:
    - name: uptime for host
      shell: uptime
      register: output

    - debug: var=output.stdout_lines

    # print a formatted line for grep/sed mining
    - debug: msg="grepme {{ inventory_hostname }}:{{ output.stdout_lines}}"

Things get way more involved that this but it's daunting to first get started. Keeping things simple with just shell commands can get your feet wet but once you get beyond simple things, you should look into more "Ansibly" ways of doing things which are more complex and harder to work with, but they do pay off.