Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Linux

Ask Slashdot: How Can You Refresh Your Linux and Sysadmin Skills? 140

Slashdot reader PrimeGoat has used Linux for 20 years, "10 of which were during my career as a Linux sysadmin..."

"However, there's more to being a sysadmin than just knowing how to use Linux." There are best practices that evolve, new methods of doing things and new software that constantly comes out and evolves. This is where my challenge comes. In 2012 I stopped my career as a Linux sysadmin... There's a lot of stuff that I missed out on. I'm wondering what I should do to refresh my skills and to catch up on what I've missed?

An obvious solution would be to get a job as a sysadmin again, but this probably isn't going to happen, as I'm changing my trajectory. I'm currently training to become a fullstack web developer, but still have a need to update my sysadmin skills and keep them fresh... Any suggestions on what actions to take on my own to catch up and keep fresh?

Leave your thoughts in the comments. What's the best way to refresh both your Linux and sysadmin skills?
This discussion has been archived. No new comments can be posted.

Ask Slashdot: How Can You Refresh Your Linux and Sysadmin Skills?

Comments Filter:
  • You install Linux (Score:5, Insightful)

    by Mr. Dollar Ton ( 5495648 ) on Sunday January 26, 2020 @03:41AM (#59657040)
    and start digging.

    Nothing beats actual experience.

    • by dragisha ( 788 )

      These days it i actually pretty easy, if you plan some time for it.

      You get a box, or one or more virtual environments somewhere (AWS free tier comes to mind) and start playing. Linux has changed in 8 years, but it is probably more about new things like Cloud, DevOps than about Linux.

      My recommendation would be to get this AWS free tier and excellent Terraform book Terraform: Up and Running: Writing Infrastructure as Code [amazon.com] - and start from this actual and fresh new perspective of what we did before IaaC. You w

      • by i.r.id10t ( 595143 ) on Sunday January 26, 2020 @10:14AM (#59657514)

        As someone who hasn't stopped using Linux in the time period the OP is asking/talking about but who has been introducing students to Linux for that time (and longer), I'd say the real changes are in init system, IP configuration/info gathering, etc.

        Think about it - systemd and "service foo restart" or "service foo status" vs. /etc/init.d/foo restart. The migration towards the ip command instead of using ifconfig, route, ifup/down, etc. The use of things like dnsmasq so that setting a resolver isn't just editing /etc/resolv.conf , especially on workstation/desktop/laptop systems with desktop environments.

        And yes, the way to learn about it is to install and start using and reading.

        • "service foo start" is not systemd. It's SysVinit, which is the init.d directory. While you can start/stop things from the init.d directory manually, that's more an artifact of design than how you are supposed to do it.

          Systemd is "systemctl start foo". There is no manual way to directly run the services, because services are declarative config files, rather than imperative scripts.

          • by rl117 ( 110595 )

            "service" isn't any init system at all. It's a generic wrapper which can trigger actions using any init system, be it sysvinit, upstart, systemd or something else. It's an *interface*, not tied to a specific implementation.

        • Yeah the big changes have been systemd and security, including SeLinux. Once upon a time SeLinux was a pain in the ass and it was reasonable to disable it on some systems. It has matured since then.

          Also if your organization uses Kubernetes, that's different.

          • >Once upon a time SeLinux was a pain in the ass and it was reasonable to disable it on some systems.

            SeLinux is still a pain in the arse and I still disable it on machines where it isn't actually useful.

        • I'd say the real changes are in init system, IP configuration/info gathering, etc.

          Yep. Simple.

          Read about systemd. Does it sound like an improvement, or do you feel triggered? If you feel triggered, just accept that your sysadmin days are behind you. If it sounds interesting, learn all about it and most other things are still the same.

          As before, there are multiple firewall options you have to be aware of.

    • With VMs and cheap VPS available everywhere, there's no reason not to. The Linux documentation isn't always great but the thing about it seems to be it forcing you deeper in.

    • Re:You install Linux (Score:5, Interesting)

      by hcs_$reboot ( 1536101 ) on Sunday January 26, 2020 @07:06AM (#59657252)
      Not enough. You install Linux, and you use it! Meaning, stop accessing Windows "because it's easier to watch a movie" (wrong), "because there is office" (LibreOffice is good enough nowadays), or "because stuff are missing" (usually a google search gives a decent answer). The 2 OSes are different, and MS puts a lot of money on the table to ensure you don't go somewhere else ; you will feel disoriented. Just be serious about using Linux and after some time, miracles happen.
      • What is this Windows that you're talking about?
      • This advice is wishful thinking from another decade. For one, "switching from windows" has been reduced to booting a live image and staring at an alien desktop environment. It isn't the Oregon Trail like experience of the 90's anymore, it's just a new desktop environment that feels neat for a while. It does not really impart any motivation to learn vi, bash, python, openssl, curl, etc.

        Like another poster said, Linux took over the data center, and now nobody needs Linux admins. I'm not arguing about that

      • Being a user is NOT being a sysadmin. There's a pile of expensive stuff that you won't run at home, policies that you won't be trying to administer on a server or workstation farm, and so many other differences.

        I would not hire as sysadmin someone who uses exclusively Linux. That's not enough.

        • by pnutjam ( 523990 )
          True, I'll name afew...
          Ansible
          Salt
          chef / puppet
          multipath
          sssd

          I'd also expect an understanding of different patching methods, pam authentication, and at least a mid-level understanding of file systems and lvm.

          That's aside from things like bash scripting, and basic understanding of where linux keeps different stuff, like logs, or use files.
    • This is how I learned it the first time -- I slacked off on my homework for my actual classes and I spent my free time experimenting with setting up sendmail and XFree86 on my Linux box at home. Granted, back then documentation for sysadmins tended to be better than documentation today is.

    • This ^

      It may sound glib, but I was about to post the same. You learn by doing. Just like all technology. When you stop you fall behind.

      Set up even a play system somewhere. Then choose a purpose/goal, like automating software installations or scanning a network for security threats. Then search on Google for how people are doing that now. Then try one of the "new whiz-bang solutions", and see what you think. Convert or not (meaning use the new or old way), it's your choice.

      I don't think there's anyth

    • It has not changed, before Linux, early 90s, when I encountered un*x for the first time (it was HPUX) as a CS student, I knew nothing about it. I just went to /bin, did a ls, and did a man command_name for every entry in /bin, to know what it was for, and test them. I spent so much time reading man page... but after a few weeks/months and asking hundreds questions to the local sysadmin, I knew a lot of things! Then there was Irix, and AIX, and SCO, and SysVR4, etc, all with their little difference, then Li

  • systemd (Score:5, Informative)

    by Adam Colley ( 3026155 ) <mog@ k u p o .be> on Sunday January 26, 2020 @04:02AM (#59657070)

    If you stopped that long ago it would have been before the rise of systemd

    You really need to concentrate your studies on that, it's changed pretty much everything from init onwards

    (not for the better in my opinion but that's a bit of a holy war, regardless it's ubiquitous now so you need to know it)

    Beyond that, things aren't very different.

    • Re: (Score:3, Informative)

      by Anonymous Coward

      That and the high maintenance new filesystems like zfs and btrfs.

      • It was Oracle's "We'll make out own ZFS, with black jack, and hookers!", and Litigious Suit Larry stopped caring, once he was able to just use ZFS itself like he originally wanted. :)

      • That and the high maintenance new filesystems like zfs and btrfs.

        New? ZFS is 15 years old and very stable if used purely as a file system.

    • by BAReFO0t ( 6240524 ) on Sunday January 26, 2020 @06:23AM (#59657224)

      Everyone else has good rational reasons for preferring a modular system of independent small building blocks combined with simple file system structures and plain text data formats (like logs) that integrate into your system, instead of a monolith of a framework that wants you to integrate into it (Even the Linux kernel.) and have no gods beside it.

      That empower you. Like a box of professional tools.
      Not give you boundaries. Like Apple products.

      We call those reasons "Unix".

      Frankly, what there should be, is a systemd, Unix-style!
      I mean the core idea of having one event manager and standardized config method for all the things that have such a facility in common. So the name would be more the name of an interface standard. Everybody would love that, and nobody would have started a fight.
      Then again, we could have had all that and so much more, decades ago, with Plan9's 9P protocol / file system.

      • by tlhIngan ( 30335 )

        Everyone else has good rational reasons for preferring a modular system of independent small building blocks combined with simple file system structures and plain text data formats (like logs) that integrate into your system, instead of a monolith of a framework that wants you to integrate into it (Even the Linux kernel.) and have no gods beside it.

        Except there are situations where simple tools are just that, simple, and inadequate.

        Sound and network management, for example.

        For sound, it's complex - audio de

      • by reanjr ( 588767 ) on Sunday January 26, 2020 @10:32AM (#59657560) Homepage

        Seems to me systemd - which is a suite of like 30+ tools already follows the Unix style of having small utilities that do one thing.

        The biggest gripe I can see with systemd from a Unix-y perspective is that it requires dbus to properly tie everything together.

        But systemd has already broken up their architecture pretty much how you say they should. Just because a distro chooses to use everything from systemd doesn't mean you have to.

        One of most common things that gets ripped out is the DNS resolver (for years systemd's was buggy as shit, now it's just problematic).

        What systemd recognizes that nothing else ever seemed to understand, is that there are many system services that seem unrelated on the surface, but are actually quite interconnected when it relates to how your system boots up. SysVinit expects every distro and sysadmin to keep solving these problems over and over again, usually poorly, and often in ways that are broken by design. Systemd actually gives you a solution.

        If all you do is install packages and start/stop services, then you probably don't care. But if you are building a system, systemd has a lot to offer.

      • It is modular, you just don't know what the details are, and you were told it isn't.

      • I mean the core idea of having one event manager and standardized config method for all the things that have such a facility in common. So the name would be more the name of an interface standard.

        Except you already have that. Systemd is highly modular with the only dependencies being that event manager and journald, the latter of which can be reduced to simply a pipe between the kernel and rsyslogd if you prefer old school logging.

        Everything else is completely modular and the decision to adopt it lies entirely with the distro maintainer.

        All network interface configuration is a separate program called systemd-networkd
        Their consolekit replacement is a separate program called systemd-logind
        Time managem

    • by geggam ( 777689 )

      Disagree,

      There are so many changes with containers and the ecosystem around that aspect its really challenging to find someone who really understand the lower levels of linux.

      Compare Slackware to Ubuntu. Ubuntu has all the new bells and whistles where Slackware remains loyal to its roots. I think you can learn more with Slackware and many of the choices Ubuntu is making will seem questionable to you.

      That said you will need to learn Ubuntu and the new container ecosystem.

      Networking is something I see many ma

      • by pnutjam ( 523990 )
        Every enterprise I've worked with has been Red Hat and maybe some Suse. If there is ubuntu, it's either an appliance or something that was acquired and it's on a planned decommission list.
    • Yeah you know things ifconfig, traceroute, and manually editing files like /etc/resolv.conf or trying to add something to /etc/rc.local? Good luck with any of that.

    • by rastos1 ( 601318 )

      Beyond that, things aren't very different.

      So you still generate sendmail configuration with m4 macros? ;-)

      • If you edit a sendmail.cf file once, you're a sysadmin. If you edit a sendmail.cf file twice, you're insane.

        I forget who said this.

    • Beyond that, things aren't very different.

      Oh that is an ignorant statement. A lot has changed in many Linux distributions, especially in the support of cloud based infrastructure and on-demand instances of an OS.

      Hell install Ubuntu and see how far you get editing your network interface. I bet you step 1 will result in you finding a file saying "this file is auto generated, do not edit this file".

      A lot has changed in many distributions, not just Systemd.

  • Modern skills (Score:5, Informative)

    by polgair ( 922265 ) on Sunday January 26, 2020 @04:05AM (#59657076)

    The demand for in depth knowledge of legacy services such as mail, dns have declined. Hardware troubleshooting is not as necessary as before.

    A decent set of skills for systems administrators nowadays in the unix world include knowledge of configuration management, ci/cd, some mass orchestration framework like terraform and at least decent knowledge of one containerization frame work like docker and pretty decent know how of one cloud provider. I also would like development chops (which you have) as well as source control competency as well.

    You'll want to know more about nginx instead of apache. You'll need to know systemd. If you did much of your scripting in bash and perl, bash is still in demand but python has largely eclipsed perl as the standard sysadmin glue. Standard appliances like F5 load balancers have been eclipsed to load balancers in the cloud, as it's just easier to scale.

    Last but not least, since you've been out of the game for awhile and given your background, you might want to start working the MSP route and work your way up the chain again. It will be frustrating, but you'll see what is out there and you can work your way back into a modern skill set in a year or so.

    Best of luck.

    • by jrumney ( 197329 ) on Sunday January 26, 2020 @05:43AM (#59657180)

      DNS is a legacy service? Are all the kids memorizing IPv6 addresses these days?

      • Not the kids!

        The addresses are still served by the DNS server, but today it comes prepackaged and partially preconfigured in the schmocker container, which you deploy into the ABS cloud via the ensemble orchestration framework, then let it merge your data via a consumer mechanism built into the configuration manager, which you feed via binary protocol from a nosql database, which in turn takes data from a document store, which you defined in some obscure XML using your favorite XML Developer studio.

        It is

      • Most of them are just pasting them into a web gui at their domain registrar. They don't really need to know how it works. They wouldn't use any features even if they knew what they were. They don't use sub-domains, if they need another name they click "buy" and add some words to the old name.

    • The demand for in depth knowledge of legacy services such as mail, dns have declined.

      Sheesh! Decline in DNS? I don't think any system admin worth their salt even know how to configure a standalone DNSSEC in their whitelab.

      The demand for DNS is even stronger.

  • Its constant need for maintenance will make you a skilled admin in no time. ;)

  • I mean, if you get a job writing js code for Airbnb you will need to follow their style guidelines. I feel like sysadmins are still mostly cowboys. I could be wrong though.
    • Yeah, best practices are to not get creative.

      Following the Linux File Hierarchy Structure is one example of how you do sysadmin work that doesn't make your coworkers or inheritors angry.

      Learning how your system's init system works is also important. No, supervisord is not a real alternative. That's what your ignorant dev team uses. You should do better.

      Depending on the size of your infrastructure, you've also got something like Kubernetes which is overseeing a lot of infrastructure and can create major head

  • by nospam007 ( 722110 ) * on Sunday January 26, 2020 @05:31AM (#59657160)

    Create a fake website for a fake company with fake employees etc and wait for the training to begin.
    You won't have to wait for long.

  • Download several Linuxes --- there are lots at https://linuxtracker.org/ [linuxtracker.org] --- and try installing and administering them
  • Going for LPIC 1 & 2 should give you a good indepth insight into Linux that even seasoned veterans don't necessarily have.

    Other than that, there are plenty of articles on the web and in print that show near collections of CLI tools and what they can do.

  • by DeBaas ( 470886 ) on Sunday January 26, 2020 @06:36AM (#59657234) Homepage

    Linux won in the data center. And now we don't need the Linux sysadmin no more...

    Okay, that's definitely exaggeration, but if I look around me there is some truth in it. With lots of companies moving to containers the Linux server is more and more just there to provide the kernel, network and the container environment. And delivery and maintenance of those are highly automated and takes less people to maintain.
    IMHO, looking at the near future I'd look into tech such as Kubernetes, Openshift, Cloudfoundry.

    OP states that he trains to be a fullstack web developer. Even if you plan to maintain the servers yourself I'd look into those techs. If you don't have to babysit a config on a server and can just add fresh VMs to take over running containers maintenance is a lot easier.
    I realize I'm not answering the question how to refresh sysadmin skills, but to me, learning about building sites according to the 12 factors [12factor.net] and deploying in containers seems like a better investment. The experience from the past will still come in handy!

    • Okay, that's definitely exaggeration, but if I look around me there is some truth in it. With lots of companies moving to containers the Linux server is more and more just there to provide the kernel, network and the container environment. And delivery and maintenance of those are highly automated and takes less people to maintain.

      I started back in the 1990s, and even then, Systems Administration was all about how to find ways to automate work away. Working hard to make it so that you don't have to work hard.

  • by vadim_t ( 324782 ) on Sunday January 26, 2020 @07:30AM (#59657284) Homepage

    Slashdot is way past its prime. All that's left of the old userbase is old, bitter nerds who haven't updated their skills since the 90s, still think system administration is about managing a small number of machines by hand, and have nothing better to say on system administration than to rant mostly incoherently about systemd.

    I'm not a sysadmin currently, but I'd recommend looking into: systemd, containers, virtualization, ansible, and modern event monitoring like Prometheus. Also SELinux if you're going to work with Red Hat. It's probably a good idea to know IPv6, as though it's growing slowly it is growing, so you'll need it sooner or later. Linux firewalling has also advanced in the last years, and you might want to look into things like firewalld, nftables and BPF.

    You might also want to look into Jenkins, depending on what you're going to administrate. You'll also want to know Git. Definitely a must for a modern developer, but even sysadmins should have at least some familiarity.

    • Slashdot is way past its prime. All that's left of the old userbase is old, bitter nerds who haven't updated their skills since the 90s, still think system administration is about managing a small number of machines by hand, and have nothing better to say on system administration than to rant mostly incoherently about systemd.

      I wish I hadn't already commented, because this deserves to be up at +5. All the other recommendations in the post are top-notch too.

    • All that's left of the old userbase is old, bitter nerds who haven't updated their skills since the 90s, still think system administration is about managing a small number of machines by hand, and have nothing better to say on system administration than to rant mostly incoherently about systemd.

      But the crumbs in my neckbeard get triggered by automation, or being asked what dbus is and what it has to do with starting processes. And when the crumbs get triggered, they talk more. And then mom keeps asking if I made it to my last doctor's appointment, and if I need a ride to the clinic.

      • by vadim_t ( 324782 )

        I have no clue what you're trying to say. Try being less metaphorical.

        • You said, "All that's left of the old userbase is old, bitter nerds."

          I paraphrased it into Slashdottian, is all. Be glad you didn't understand it.

  • Container Training (Score:4, Informative)

    by leonbev ( 111395 ) on Sunday January 26, 2020 @08:35AM (#59657376) Journal

    I would spend some time standing setting up some servers running Docker and play around with containers and Kubernetes orchestration.

    • Yeah, this. Get on eBay, get yourself a cheap blade server that's well-populated, and start playing with VMs and containers. Also do some AWS and maybe Azure. These are the skills I'm seeing required on sysadmin job listings.

      • by kobaz ( 107760 )

        Blade server?

        Virtualbox and a cheap ass i7 will get you the same distance.

        • Lol virtualbox

          Go on, pull the other one

          • by kobaz ( 107760 )

            Lol virtualbox

            Go on, pull the other one

            Are you one of those "If it doesn't say IBM or Cisco on it, it's crap?" types?

            Best tool for the job type thing going on here. If your goal is to learn container orchestration, vm and cloud, why bother with special hardware? You'll save some money that can be used for paying for educational materials instead.

            Knowing how to manage a blade system is a specific skill that he may or may not need (erring on the side of probably not any time soon, getting back in as a mid-level sysadmin).

        • by short ( 66530 )
          Why some proprietary [virtualbox.org] 3rd party incompatible crap (VirtualBox)? Use normal KVM [wikipedia.org] by virt-manager [wikipedia.org] (personally using KVM by commandline qemu-kvm [wikipedia.org]).
          • by kobaz ( 107760 )

            Why some proprietary [virtualbox.org] 3rd party incompatible crap (VirtualBox)? Use normal KVM [wikipedia.org] by virt-manager [wikipedia.org] (personally using KVM by commandline qemu-kvm [wikipedia.org]).

            Now that's an insightful, well thought-out response.

            My point was, why buy special hardware when free software will do the trick.

    • by short ( 66530 )
      How to do easy automatic daily security updates of the containers?
  • You should be in the senior track on your way to lower management. No need to expand detailed skills, it's better to get a broader understanding of the market, how your skills apply and high level technologies.

    As far as Linux systems management, there isn't much of that going on anymore. Learn how to minimize your time spent by learning things like managing a large Ansible, Puppet etc installation, how to orchestrate an entire datacenter from a single team etc. Nobody is interested in you learning how to re

    • "management"

      Some people like doing real work.

      • by guruevi ( 827432 )

        Sure, but after 10+ years, you won't find a job at your current rate unless you're willing to accept junior wages. Why would I hire someone with 10+ years of experience as a non-senior sysadmin if I can hire someone with 1-3 years experience for half the price and they do the same job?

    • You should be in the senior track on your way to lower management. No need to expand detailed skills, it's better to get a broader understanding of the market, how your skills apply and high level technologies.

      And after 20+ years, I can confidently say that Systems Administration does not naturally flow into Management, and a good administrator does not necessarily make a good manager, and vice versa. Only go into management if you think you have the talent for it. Otherwise you're taking a job that you (hopefully) love and turning it into one that you actively hate, and your co-workers will hate you as well. That's not worth the pay.

      But the computer industry is still a relatively immature one full of ageism; man

      • by guruevi ( 827432 )

        You missed the point, the market is changing, we've automated our way out of a deep-knowledge sysadmin world into a series of specialists. I know a lot about the chipsets from the 8088 to Pentium, how caches work and what happens when a CPU has to flush its pipeline, the pinout of a serial port, null modem cables etc. - but over time I've needed those skillsets that once were absolutely necessary less and less, it's been probably about 10 years since I actually crimped an Ethernet cable.

        Yes, it's still impo

  • There's more to being a sysadmin than just installing and using.

    I think the OP is really thinking about this from a business perspective(?). Yes there have been a few changes and SystemD is one of them but (however annoying it is :-), it's just another tool.

    To hopefully answer your question, I would investigate your local Linux User Group or local DevOps group. (Try meetup.com). Slashdot may not the right place to find the answers you seek.

    Once you determine your weaknesses you should know where to look.

    The

  • We are dying breed. Any of the new technologies are distinctly ANTI-SYSADMIN. Docker, etc all pretend to dispense with the need to admin things.

    So "refreshing" sysadmin skills is an oxymoron.

    I used to earn my upkeep by sysadminning. I have gone back to development 7 years ago for this exact reason.

    Your sincerely, a BOFH of 25+ years experience.

    • There is a goal to eliminate sys-admins. But "it's elephants all the way down", meaning that the docker instances themselves run on an underlying system which is at a different "layer of abstraction". The remaining people in sys-adminning are pretty valued.

      • by guruevi ( 827432 )

        The remaining people are specialists. You no longer get someone with broad knowledge to build a computer from scratch and set up a server. As you say, Docker removes the necessary knowledge to deploy a specific application, so you're either managing containers or managing a datacenter or managing a hosting provider, but you no longer have a single person from the barebones to the Apache and MySQL deployment to the outgoing network bits.

        And for one thing that is good, with modern security threats, you can no

        • Sadly, the "third party scan" approach itself tends to be deliberately flawed, scanning for specific software vulnerabilities while ignoring dangerous security practices throughout the infrastructure. It's why I dislike being brought in to do security scans: the easy fixes are often forbidden by policy, and resources spent on tools that don't actually solve the problem. this includes, for example, docker instances with compilers or pip or nodejs available, with network access to the internal firewall, and w

  • What the heck is a sysadmin these days? Linux is so ingrained into IT now I think it all depends on how it is used.

    I have been using linux exclusively on my machine at home since '98. (although my wife/kids use Win7). I admin my own system, but that is not a sysadmin.
    Linux isn't a niche thing anymore, and I know for a fact that my skillset wouldn't translate to being a sysadmin where I work. We have thousands of VMs handling not only internal architecture/applications but that are for production. We of

  • Ansible first (Score:4, Insightful)

    by Average ( 648 ) on Sunday January 26, 2020 @10:53AM (#59657608)

    In the classic sysadmin realm, the biggest shift has been "infrastructure as code". That is to say, repeatable, version controlled, and (in a dream world few of us get to) with regression tests and such. There were certainly some shops growing along those lines in 2012, but a whole, whole lot more who weren't. In 2020, a sysadmin that isn't doing absolutely *everything* in version control (git, more often than not) and using a config management tool (Ansible, Puppet, Chef, Salt) is, without question, doing it wrong. Of those, Ansible is the most popular one, especially at smaller scales. It's also the easiest one to get into... a playbook run is basically a linear shell script with a whole toolbox of convenient macros.

    I've given this basic advice to someone in your exact situation. A) Make sure you're pretty good at Git. Don't have to be an expert, but okay. B) Pick up Ansible. Jeff Geerling's "Ansible for Devops" is a good book if you want one. C) Then, yeah, in 2020 you need to pick up at least a little Docker. It'll be a thing in your 'full stack web' anyway. And at least understand why Kubernetes is a thing even if you don't go that route.

  • I use Ubuntu a lot and because of that I have a fair amount of proficiency on it. Over the years there has been an evolution over about 4 different network configuration management systems that I can recall.

    /etc/network/interfaces to /etc/netplan is the current transition. Have you seen the netplan documentation? It is to laugh: here's a sample file with an example of everything we could think of in it just work from that.

    Then someone asks for help on installing a CentOS system. Forget all tha

  • Every so often, I will take a 1-week Red Hat Boot Camp class and will dedicate 1-week to eating, sleeping, and breathing the new stuff in Linux.

    Depending on whether you are going cloud, devOPs, Virtualized Architecture, or infoSec, you can find a bootcamp to meet you needs.

    https://www.redhat.com/en/serv... [redhat.com]

    Another purpose of taking the class in person instead of a Udemy course is that you are working with like-minded people. You can ask a question or see how they work and think. Additionally, you can estab

  • This isn't the "best" way to refresh your skills by any means, but I would strongly recommend that you pick up the Tom Limoncelli book "The Practice of System and Network Administration" (3rd Edition). This is a fantastic book on the subject and quite cheap for what it will provide. Nothing will beat hands on experience (and probably specific training on your environment's particular equipment) but the book is an excellent reference manual.
  • "What's the best way to refresh both your Linux and sysadmin skills?"

    Have a serious system crash and you'll be surprised just how much you can learn in a few short hours (or days, if it's really bad).

    Learning under do-or-die conditions can really motivate the ol' information intake centers of the brain, if you know what I mean and I think you do.

  • That's the most valuable Unix skillset. Regex/Sed/Awk/Grep.
  • Instead of /bin/true, you're supposed to build a docker container that has an entire set of libraries built around a single executable that exit(0)s so you can docker run true

    The image is only 30MB.

    Docker and its devs are a joke; they're a bunch of schmucks too stupid to understand packaging, trying to work around a bunch of libraries written by schmucks who can't design stable APIs.

Some people manage by the book, even though they don't know who wrote the book or even what book.

Working...