CentOS and command line
I have heard from numerous users that Redhat is like another Microsoft and keeping things GUI and taking power from User etc.I haven't been able to find much difference then per say slackware that uses BSD init system.I find both systems quite stable and I guess slack requires more configuring. However, I still don't understand why its not good for power users from posts I put. I also didnt find System V init system too complicated as the init is split into run levels with different file for each service. So is this just a MYTH that Redhat is against KISS Principle and is anti-Terminal? Isn't it possible to use same commands in Redhat? I'm confused. I keep switching from 1 distro to the other. Good thing I did not format my centOS.
|
Quote:
|
RH is nothing like MS, nor is it like Canonical (the company that owns Ubuntu, or at least is the financing behind it). Canonical is trying hard to make Linux function more like MS Windows, RH on the other hand is making a very stable and reliable server platform. Not to mention they give tons and tons of code back to the Linux community.
I support roughly 3500 RH type servers across the US as a living and I do it 100% via ssh, without X11 forwarding so no GUI for me and there is NOTHING I can not do. We support RH7, RH8, FC5, FC7, CentOS 5.4, and now CentOS 6.2. So yes I am very comfortable in saying that there is no "MS Windows" movement from the RH camp when it comes to tools. In fact all of our RH7/8 boxes are without GUI and only some of our FC5 boxes have GUI running at all. The only reason we run GUI on FC7 and newer is the software we support now provides graphics and scanning of documentations. |
RHEL vs Slack
Disclosure: I kinda-sorta maintain RHEL servers at work (by which I mean it's not my primary duty, but nobody else is doing it), and run Slackware at home and for my home business.
What I've found is that yes, you can in theory do the same things with RHEL as you can with Slackware. They're both open OSs built on virtually the same open source, so you can do anything you can do in one also on the other. Where it can get complicated is with the package managers and the "glue" scripts. Package managers: Slackware has a few options of various complexity. At its simplest, it's just unpacking a tarball over the root, maybe running a maintenance script to update config files, and voila. In some cases, these managers track package dependencies, but in others it relies on you to do that yourself. This tends to make it pretty straightforward to try a package, and if it doesn't work due to dependency or versioning problems, pull down the source and compile yourself. RHEL has one package manager, yum, which uses an official repo that's often lagging behind (for stability reasons) and unofficial repos that you'll inevitably need to do anything useful :P yum performs extensive version and dependency tracking, uses a database for journaling installation activities, and is generally more powerful but also complex. I've run into innumerable dependency hells when trying to install some package that relies on some other package that doesn't have the exact right version, leading me to forcibly install things or open the rpm manually to copy the files out to the file system (bypassing yum altogether) or building from source. Of course, the moment you've started with these workarounds you've effectively destroyed the legitimacy of the yum database and then you have to start compensating for your workarounds by other workarounds. I prefer Slackware's simpler approach to package management, especially since I tend to build from source anyway in most cases. Glue scripts: By which I mean little scripts or programs intending to make things "easier", for example, X-Windows initialization scripts or scripts making changes to system files. Again Slackware takes a minimalist approach and relies on you knowing what files to change to configure your system. RHEL often makes use of "best guess" scripts to try to "fix" your experience without user intervention. Sometimes these scripts do the wrong or non-optimal thing, but more frequently, the scripts just obfuscate advanced features and even overwrite files you've manually tweaked. Plenty of symbolic links and spaghetti bash code exists throughout, for compatibility's sake, which makes it hard to track down when RHEL did something automatically which you'd like manual control of. I prefer not to get a network connection or X until I've configured it manually to work, over having RHEL overwrite my manually tweaked fstab every time it starts Gnome. When it comes down to it, both are very capable server operating systems, and I can see how RHEL would fare better in an enterprise environment that's more about executing packaged, stable, tested software in a predictable manner than in an R&D environment where you're tweaking every aspect of the system to make it work perfectly for your purposes. But so far, everything I can do in Slackware I can also achieve in RHEL - it just takes a bit more work to do it "right" (i.e. the Redhat way) and sometimes I have to jump through uncomfortable hoops to get there. |
Quote:
|
Quote:
|
I don't think any mainstream distributions are really "anti-terminal". Red Hat and other distributions just do things in a different way. I don't believe really that it's against "KISS" or anything and I wouldn't think about those things too much. If you learn how to use Linux, you will find your knowledge roughly translates to whatever distribution (or other kind of Unix) you're working on. What I mean is, you'll have an idea of where to look for things and will be able to work out how to configure stuff, even if it works slightly different to the system you're used to. I'm using both Slackware (home) and CentOS (work).
Quote:
|
Quote:
+1 |
I'm always puzzled by the people who think that do-it-yourself dependency resolution is better than leaving it to the computer. Surely one of the points of computers is that they can do boring routine jobs without making mistakes? I worked my way from Fedora 1 to Fedora 14 and yum never once let me down. The rpm system has never let me down in CentOS 6 either, although I have had to pull a few tricks out of the hat to install stuff from EPEL, RPMfusion, atrpms, CentOS 5, Fedora 13, Fedora 17, and Mandriva 2010! There's nothing home-compiled on this computer and it's all in the data-base.
|
Quote:
|
Quote:
as for RHCSA to get a job, not needed, I am no were near good enough to pass that test, nor have i studied for it. I am versed in what I support. Daily I touch between 20-40 servers a day, the other technicians do the same. Our company has roughly 3500 Linux servers in the field that we maintain and support. Again all of them are of a RH flavor, thus the reason I run F17 as my workstation. Just makes for simplicity from my perspective. The commands I issue, for the most part, on my box are the same I issue on the clients. As for the newer systemd on F17, I am not comfortable with it yet, just because I have been using the older system foo <command> for over a decade. In the end to me they are one in the same. also the init.d string that many other distros use, like Debian, is really all the system and systemd are calling. RH has just provided a nice "easy" script for people to use to assist in administrating their services. In short the more you know about Linux the better. The distro does not matter in the end. There are some basic differences like locations of specific files but that is easy enough to google. Pick a distro, stick with it, but dont get married to it. |
Quote:
|
Quote:
My first experience with dependency hell was as a slightly-more-than-casual user on Ubuntu. Like most in that group, I didn't have a crystal clear vision of what each and every package on my system did, simply because through installing most things with Synaptic or the "software center," I didn't have to. Once, on a computer I used only for browsing and media and therefore had very little installed on in the first place, I tried to install Virtual Box, and ran into a whole slew of dependency conflicts I had to learn how to interpret on the spot. This required that to run, but this conflicted with this version of that, but something else required that version of that to run, and something else definitely needed to go but couldn't even be removed . . . Just like that I was thrust into the middle of a hopelessly intertwined mess of a system that I was supposed to untangle thread by thread. Predictably, I broke it. Shortly thereafter, I started using Slackware, and even though I am still quite new to it, I can say that this sort of thing would never have happened to me on my current system. When you have to configure things manually to get them to work, you learn what everything does and can then resolve conflicts. Maybe if you are already a Linux champion to begin with this doesn't matter (and in that case you can choose the tools that do the things you already know how to do automatically), but for relative newcomers like myself and the OP who want to learn, it matters a whole heck of a lot. @ Mercury305: If learning is what you're after, I would go with Slackware in a heartbeat and never think twice about it. I learn more each week with Slackware than I did in a month with the other distros I've tried. A better learning tool for Linux I have never found. |
Quote:
there are, typically it is just a matter of chasing down the foo.rpm file the getting it on the system and installing either via rpm -ivh or -Uvh. this can be a very long process. As for fixing it in yum repose yes, but typically it is created when running 3rd party or EPEL repos. On a server you typically want to be very careful what repos you point to so you can avoid dependency hell. |
Quote:
If that is better for the person using the machine is a pure subjective thing. |
All times are GMT -5. The time now is 06:57 PM. |