As a physicist by training that has been working as a data scientist and believing that the data science bubble will burst sooner rather than later, are these certifications a good path to help change course and keep myself relevant?
I'm a sysadmin for a HPC cluster. I do not have any of these certs. I think my day job involves a lot of diagnosing issues which involves detailed understanding of how the system works rather than actually developing any parts of that system.
A second fun part of this job is when designing infrastructure for use by others. To take into consideration all that can go wrong, performance characteristics of hardware, software, and workload and trade one off for other can be really challenging and fun if you're into that sort of a thing.
I'm not sure the RedHat certs will help you all the way through, although they might be a good starter in this regard. A lot is obtained by getting hands dirty and learning by making mistakes.
If you're into managing systems and building/engineering infrastructure, I'd wholeheartedly recommend spending time with small scale infrastructure in a "homelab" setting or if you can afford $$ with VMs in the cloud even.
Cheers!
I would persue them before RH, https://training.suse.com/certification/
1.) Having a structured curriculum to learn something new. All my tech skills are self-taught, and it's useful to have a structured learning path.
2.) Validating your skills, especially when you're changing careers. I went from working in call centers to AWS Systems Architect/Linux SysAdmin by virtue of a couple of certs.
3.) Sometimes employers require certain certifications as a condition of employment (or continued employment). This one can be a bit sticky, though. Make sure those certifications make sense for the role. I can't tell you how many job listings I see wanting CompTIA A+. Unless you're applying to work for Geek Squad, you probably don't need an A+ cert (which asks things like 'What is RAM'? 'What is Bluray?').
If you want to get into Linux, getting a couple of certs won't hurt. Red Hat's certs are probably the most well-respected, but of course are centered around Red Hat Enterprise Linux (and by extension CentOS).
If you want a good cert path for Debian, go with the Linux Foundation certs. Linux Foundation's certs are less well-known, but also practical exams.
I would avoid LPIC and CompTIA Linux+ certs (NOTE: I am Linux+ and LPIC-1 certified). They are multiple choice exams and really emphasize based on rote memorization of command flags, which really isn't useful to be an actual Linux SysAdmin. Then's the last time you ran into a bzip2 archive? I can't remember the last time I did, but I still remember that tar xjf will unzip one. But, of course, if I did run into a bzip2 in the wild, I can just as easily read the tar man page. Thanks LPIC!
There are some good certs, like AWS and RHCE and some really crappy ones.
Talk to people before doing them to get a feeling if they're worth it.
AWS is so massive, it’s hard to know where to start to get a general overview of it. Studying for and at least doing proof of concepts help with giving you an organized learning path with an end goal in mind.
But, I still don’t think that the certifications necessarily show any level of competence.
I was the dev lead at a mid size company with a small development shop. I was specifically hired to modernize the department and to architect a green field project. Right after I designed the system, they decided to “move to the cloud”. I knew nothing about AWS at the time and neither did they. So they hired a bunch of “certified consultants”.
At the end of the day with the consultants “help”, I ended up setting up an environment just like I would on prem (first mistake) with seven (1 in DEV/QA/UAT and a cluster in production) small VMs for Consul, Nomad, Mongo. I’m already at 21 servers. This doesn’t count two VMs for build servers and 10 or 11 app servers in all.
While this design would have been perfectly acceptable for an on prem setup and looking back I wouldn’t have done anything differently if we were on prem, this was a horrible system for AWS. We could have used managed services for everything above and not had a single VM running except the app servers (today we could have used Fargate to avoid even the app servers but it wasn’t around at the time). We could have even used managed ElasticSearch instead of Mongo since the data stored in Mongo wasn’t the source of truth.
But the “certified consultants” were a bunch of old school infrastructure guys who only knew how to click around on the web console (2nd mistake) and do a “lift and shift”.
I started studying for the AWS Architect Associate just so I could talk intelligently to the consultants and see what type of improvements I could do during phase 2. I had already designed the system to abstract our dependency on Consul, Nomad, and Mongo. I was horrified after discovering all that AWS could have managed for us. I was more horrified that the certified consultants didn’t have a clue. I was most horrified that I passed the certification without ever touching the console and just by watching the videos.
For various reasons, I ended up changing jobs shortly there after, to work for a company that was a pure AWS shop with a new (to the company) manager that wanted to be more “AWS native”.
I became the “AWS guy” even though I’m a developer. Eventually we started interviewing “AWS Architects”. None of the people we were looking for (mid level) had any practical experience. They had just memorized enough from ACloudGuru to pass a multiple choice test.
Sure, nothing beats experience, but they are a good place to start your cloud journey :)
My thought process for the certifications were the same - a method to get on an organized learning path. But, by the time the company folded two years later, the certs had already expired and I never bothered getting them renewed. I knew it wouldn’t matter when looking for my next job. So they have never appeared on my resume.
Looking back, I don’t think that the AWS Associate certs were any better or worse than the Microsoft certs. But I will keep these up to date.
The second part of the story is that after seeing how much these “certified consultants” were making - I was the dev lead after all - I realized where the money was. Especially since I already had a long development background and some Devops experience I could be much better at AWS consulting than some infrastructure guys.
When I changed companies, I negotiated not to be a team lead. I wanted to be an IC to fill in some technical gaps and get hands on experience so I could be an overpriced “digital transformation consultant moving companies up on the cloud maturity model”.
I'm currently getting into the whole serverless space (after I got my aws associate cerst) as a freelance consultant coming from mobile.
I was able to get experience through my job, but if that isn’t an option for you, just build something. There are some really good, cheap, thorough Udemy course (no affiliation, not affilliate links):
This course goes over API Gateway, lambda, DynamoDB and a quick overview of how to create a NodeJS Serverless app:
https://www.udemy.com/share/1010qKBUQdeF5TQn4=/
That is the “right way” to create a Serverless API, but if you want to take a regular old Node/Express API and run through lambda you can use the lambda proxy integration:
https://github.com/awslabs/aws-serverless-express
There are similar frameworks for Python/Django and C#/WebAPI
And for the love of all that is holy, learn CloudFormation and don’t create resources manually.
https://www.udemy.com/share/1011JEBUQdeF5TQn4=/
If you have any questions, an email address is in my profile.
I already got the cerst (DA & SAA) and built a few things with SAM and Amplify.
I go my RHCE 5 years ago and it was a great primer on how Linux systems run and I recommend it to anyone interested in modern Linux. It will give you many insights which will help you in the future in understanding how Linux works. I was moving over from Windows administration and it was a godsend.
That said this lower level of administration is becoming harder to break into, with automation, containers, ease of use improvements and the cloud, the need for the properly managed system is dwindling. I'm sure there will be a need for a long time, but they generally won't be good jobs. If you have some technical proficiency already I would say there are better places to improve your prospects.
Also the test is hard to prepare for so there is a time commitment and it does expire (I'm no longer an RHCE :() so there is that.
I hope that helps.
With the RHCE you learn about configuring Linux systems. That is building single systems from scratch, configuring firewalls, service files, security configuration and so on. The old format when I took it was the RHCSA would teach you about the system through boot, and the RHCE would teach about how to configure the services, at least that's how one of my instructors put it.
In terms of real world work, at most companies most of this work/process is done in the beginning of their technical life (starting the company or new endeavor) and once you have a method of building servers, and working configurations you don't revisit these concepts as much. Maybe that is what the Ansible cert is supposed to address.
The hard part from an employment perspective is that this work is slowly moving away towards the cloud and all in one solutions which reduce the need for this type of administration.
The big shops have large processes which likely involve solving problems on a more abstract level than what you learn with the RHCE. It's important to know for sure, but likely 95% of these problems are solved on the configuration management, cloud provisioning, or some other level.
On the other extreme, the newer, smaller places aren't going to invest in building Linux servers. They are more likely to maybe have a Windows server/NAS or, more likely, all cloud services at this point.
As an interviewee I have seen this where you grind through technical interviews (about an hour of lots of arcane, frequently out of date questions) and they tend to want either drones to do grunt work with little opportunity for growth, or know their exact technical stack to solve some specific problem.
As an interviewer I see a lot of developer/jack of all trade types trying to sell themselves as system admins whereas its clear that they don't really know system administration. We had one candidate the other day who had worked for close to 10 years in the larger organization who didn't know what /var was. :0
There is still a role for the RHCE style engineers and I had a good time learning it, but at this point I'm not sure where that path will lead in the future.
I think a better use of time would be to learn python or javacript, find a way to apply it to the area you would like to do it, get a green lawn on github, learn how to do something impressive with some public APIs.
Just my 2c as a sysadmin take from it what you will.
Diving into code such as python and Javascript is probably a good fail safe/transferable skill wherever I end up/doing. I've been leaning towards the whole Linux/CCNA/Networking/Cloud roadmap but am still relatively unsure.
I appreciate your feedback and the opportunity to bounce things around.
I'm not sure it entirely still works, but for anyone that wants a much deeper look into how a system Linux is put together and how a lot of the parts interact (at a higher level, but still a level often not covered), check out the Linux From Scratch project.[1] Around 2001 or 2002 I was helping to design a small custom Linux distro with some friends, and going through this project was invaluable for learning a lot of how a GNU/Linux system was put together and functions as a whole.
There really is something to be said for manually going through every step, and compiling every software package, and experiencing the functionality it provides, since a good chunk of the work is done in a chroot after you've bootstrapped it enough to make it minimally viable.