A stand for studying DevOps based on Linux servers. Where to start studying?

Question

linux

Good afternoon! Started learning DevOps. Currently I work as a network administrator (and then, mostly L2). There are basic knowledge of Linux (launched Zabbix and LDAP for IP-telephony). In short, we need to develop.
Available:
- several servers (I plan to install Ubuntu Server 20.04 on them);
- a pair of Cisco 3725 routers;
- several switches;
- Cisco Asa 5510;
- PC/laptop;
- Internet.
Actually a question. How can I use this iron to study the topic, where to start? Ideally, of course, to collect something useful for the organization)

Answer 1

I will support Sand : we take the Devops Roadmap and pull the skills from top to bottom. If you want to continue to work further in network administration (there were vacancies for Network Engineer requiring knowledge of Ansible / Salt 7 years ago, but not with us), then you must learn how to roll out configs from the git and monitor the entire described economy with the collection of logs.

Answer 2

Whoa, whoa - what servers, what ciss? Devops is taxiing at a high level (s). You need a cloud, containers, orchestration with autoscaling - otherwise this is some kind of banal administration, and not a fashionable devops.
/sarcasm

Answer 3

Get acquainted with the following:
1. Monitoring system. The hype is now prometheus/grafana, but you can look at any other systems + grafana.
2. orchestration, such as ansible to manage your servers
3. learn bash at the level of "quickly write a script that downloads, deploys, copies, cleans up, receives text via curl and parses the necessary lines from it, launch the application and make sure that it is successful started"
4. You can also learn python / groovy at a basic level.
5. Keep all your work in git, or better yet, put some gitlab thread and read about a couple of git workflows
6. After that, it is strongly recommended that you familiarize yourself with docker/kubernetes/openshift containers
If you manage to do it in a year, you can try to look for something initial, where there is a team of devops engineers.

Answer 4

You can start by setting up zabbix (including alarms in telegrams and event filtering for different users), syslog, taxax, configuration backup system (configuration backup), IPAM system, automatic configuration generation system (so that you can create configurations for 100 switches of the same type, differing only ip-addresses and port descriptor), configuration automation system (ansimble, openconf). Naturally, choose options with certificates / encryption, and use bash / python for automation.
And then you think about how everything that you set up can be transferred to the cloud in 1 hour.

Answer 5

I recommend that you start by moving this bench to a container cluster on one powerful machine. When these containers work adequately, you can make a second layer of containers inside, and use the physical equipment for its intended purpose.