A few recent stories from experts in the field or journalists have raised my ire more than a bit. Seemingly well intentioned, to inform and educate the reader, criticized where more research (or common sense) would have served better. The Square Peg in the Round Hole The first set of stories began with the idea of shoehorning an enterprise Linux distribution (i.e. workstation/server class) into a netbook computer . I read the first headline of this misadventure and was baffled by the idea and just couldn't understand the logic. Netbooks are not tiny notebooks, but a whole different class of machine. This grand experiment evolved into a tirade against CentOS in general. As a CentOS user, this again made no sense, in that one who uses CentOS is aware that updates are generated from sources of upstream vendor and small delays have historically occurred. The delay of CentOS 5.3 had me considering moving to another distribution, and I still haven't made a final decision. Yet, I have always been in complete knowledge of the derivative nature of CentOS and the pluses and minuses of a distribution that is dependent of another. Ms. Martin later brought up valid points how security delays could effect an organization. However, her article was poisoned by using a Firefox update as an example of an important update that was delayed. This poor example, unfortunately, led the discussion away from a constructive dialog. The combination of poor distribution choice and poor examples diluted the conversation and the noise soon overcame the signal. Plus, recommending another derived distribution (that admittedly the author hasn't had much experience with) that is seemingly on safe ground as it is supported by public funds. It's nice to know that no government or university funded project is ever eliminated. Too Much of a Good Thing Another, no doubt, well intentioned article attempts to revive the tired story of non-standard Linux. Calling for an “official” Linux distribution to be the standard to which all others are derived. It would be interesting to find out which distribution the author deems authoritative to all others. Let's twist this slightly though and use another operating system as an example. Oh, just off the top of my head, let us change the operating system to Windows and re-tell this tale. If someone came to you asking for help migrating from Linux to Windows how do you help them? You might start off telling them about the fundamentals of Windows, how it came to be, and what proprietary software means. You will discuss the abundance of applications available, the problems associated with legacy applications and the additional costs. You will discuss the Windows desktop and how many choices there are. The requirement of anti-virus and (depending on edition) anti-spyware, and how you can have multiples of one but only one of the other. And that discussion will fuel the flames of confusion. When this confusion builds you will eventually come around to the topic of editions. It is at this point where you can finally start showing your newbie the similarities and differences between Linux and Windows. In my day job I support both Linux and Windows and routinely take courses or read on “upgrading skills” to the latest version (read distribution) of Windows. Is this an attempt to compare to the “standard” of Windows, where no such standard exists? What operating system is identical in use and management for all platforms and between major version numbers? Where is this ideal that Linux is compared to and denigrated by? Let's Learn a New Word The final article that pushed me to speak out was the mis-guided journey to use an install media as a working desktop. Yes, the Ubuntu Desktop distribution CD can be used to try before you commit. I have never seen any documentation or claim to indicate this media, although can be made to be bootable via USB stick, be used as a portable Linux. There are distributions designed to be used on USB media with the ability to save updates and add-on software. The new word we are learning today is, persistence. A little research would have led one to learn about persistence and the ability to customize, update and save a USB bootable Linux distribution. Then why would one need the ability to boot an install media off a USB stick? Getting back to netbooks, my netbook, like many others do not have a CDROM. So making a USB bootable install media is much simpler than finding or buying a USB CDROM.
The Distro Landscape The Linux kernel and GNU utilities make a fantastic base to several distributions which serve many different purposes, desktops, servers, appliances, netbooks, firewalls, and several specialized tools for dedicated roles such as recovery, backup and diagnostics to name only a very few. Choosing the right tool for the right job may seem intimidating, yet a little research, experimentation and experience (or the experience of others) will often lead to the correct choice. Using the wrong tool for the job and then boviating about it helps no one and in the end may just show ignorance. |
|||
 |
Cooking With Linux |
|
|
|
|
Rebuttle in Full
http://ever-increasing-entropy.blogspot.com/2009/08/circling-wagons.html
Misplaced criticism
I strongly and firmly disagree with your characterization of my four articles (the first, which you didn't link to, was a feature for DistroWatch Weekly) regarding CentOS on a netbook. A netbook is a small notebook, not another class of machine as you write. Businesses are increasingly using them as notebook replacements. I actually saw this as early as 1999 when the CEO of a company I was providing support to as an IBM contractor used a Toshiba Libretto to travel with. The Libretto was a precursor of the modern netbook. My own netbook, a Sylvania g Meso, is a perfectly suitable desktop or laptop replacement for most everything and works well with most Linux distributions.
Second, you ignore the fact that CentOS has been as much as two months late with security patches over the past year. This wasn't a matter of waiting for the upstream vendor. It has been a chronic problem for some time. That was acknowledged by a CentOS developer in the comments to my column for O'Reilly (which you also ignored) covering why I no longer will use CentOS for my business clients. I've linked the O'Reilly article in lieu of my website so that people can read it.
Firefox was a perfectly legitimate example. The upstream vendor had a patch out the same day Mozilla released the new version. Scientific Linux had one the next day. It took over a week after that for CentOS to release a patch. The upstream vendor sells their Enterprise Linux explicitly in a version for the business desktop so it is reasonable to assume that CentOS will be used on the desktop as well. A browser with critical vulnerabilities left unpatched is not suitable for a security conscious business.
Finally, this wasn't a "rant" against CentOS. The whole situation with Lance Davis forced me to rethink the idea of using a distribution maintained by a small group of volunteers in the enterprise. My conclusion, which is not CentOS-specific, is that just isn't a good idea.
You are a CentOS user who is defending their favorite distro. I understand that. The problem here is that you are defending the indefensible (poor management, failure to get security patches out on time) and doing so by shooting the proverbial messenger.
My advice to your readers: heed the experts and ignore the blogger who is circling the wagons.
Chef 1, Martain 0
Ms. Martin,
I have to agree with the chef. Why indirectly review RHEL for a netbook? Why not use a distro intended for netnooks, or at least for desktops?
Regarding your statement "A netbook is a small notebook, not another class of machine", I'll admit that I see some truth if you will admit that there is a significant difference in capacity. Maybe you've bought the redefinition of the term by Microsoft and hardware manufacturers, but I reject it. And even if people might find netbooks a convenient replacement for a notebook, that doesn't automatically imply that they are perfectly equivalent, but rather than 90% of what most people do most of the time can be handled by this lower class of hardware. The IT industry sees these machines as a new segment, I'm not sure why you do not and strident denials and examples regarding tiny notebooks (which are not netbooks) are not convincing.
But even if we accept that notebook=netbook, have you never heard of problems installing Linux on notebooks? I'm still baffled by the choice of reviewing CentOS, which really tests RHEL, which is not targeted towards desktops at all, and certainly not notebooks.
Also, I'm not sure how you are advancing your "case" by trying to rebutt people like the chef. Such stalking makes you look petty and pig-headed, more than "expert".
The Chef sets the record straight . . .
Although what the chef sets the record straight on has little to do with the discussion. In referring to The Chef, I take it you mean me, Marcel Gagné, the Linux chef aka the Cooking with Linux guy. The post, Beware The Experts, was actually written by Ken Leyba, aka Small Box Admin. So, as much as I enjoy and support people agreeing with me on any manner of topic [ insert appropriate smiley here ], credit for this one goes to Ken, not myself.
-- Marcel
mea culpa
s/chef/Ken/g
Not a CEntOS partisan
My problem is not that you attacked CEntOS, but that you used it in a vastly inappropriate situation. CEntOS is specifically targeted at servers and business hardware. Your conclusions are most definitely CEntOS- or RHEL-specific. Neither is a good idea for a netbook.
Specifically, netbooks tend to incorporate newer hardware that requires a more frequently updated distribution. The fixes that gave you issues were desktop fixes. A good percentage of CEntOS installations in the real world don't run graphical mode apps at all.
To ignore the fact that netbooks are significantly different than more generalized laptop hardware is preposterous. It undermines your credibility. I know I want my next enterprise server to have a built-in 9 inch screen, a webcam and an Atom processor soldered onto the motherboard.
The article was not helpful in regards to that specific distribution (bad choice for the intent and hardware). The article was not helpful for a review of the hardware because of the same reasons. It was not good as a review of Linux's viability because the distribution picked is specifically weak against other desktop distributions.
You picked the wrong distribution for the hardware and the target usage and then revealed it as an epiphany.
I completely disagree and you missed the point of my article
First, the idea that RHEL/CentOS is only for servers without X is what is preposterous. New hardware exists on business desktops and laptops as well. The lack of hardware support or adequately stocked repositories applies equally well to all desktop systems, not just netbooks. The performance issues apply to older servers and desktops equally. In general, if you read the article you know that I got every last bit of hardware working. Hardware compatibility wasn't the big issue.
None of my articles were intended as reviews. They were intended as a detailing of an experiment. There was no claim of an epiphany. There was a claim that the issues I discussed fit equally well with traditional desktop or laptop systems. I also made no claims that this refers to Linux viability on the desktop. Quite the contrary. If you had bothered to read my other writing you would know without a doubt that I believe that Linux is more than ready for the desktop.
Finally, to replicate my business environment on the netbook, the purpose of the experiment, I had to use the distro from my desktop environment. It was a valid experiment and it was worth writing about.
You are most vexing
"First, the idea that RHEL/CentOS is only for servers without X is what is preposterous"
WHO is arguing this!? It just really seems like you cannot take criticism... which makes your posts about CentOS fairly ironic.
Trying to install CentOS - an
Trying to install CentOS - an enterprise OS, for servers - onto your netbook is as preposterous as trying to install Windows Server 2003 on said hardware.
Next up, why screwdrivers can't be recommended for anything, because they're experimentally proven to be no good for hammering nails.
Pingback
I too have been disheartened
The more I read, the more I am saddened by the complete lack of effort by reviewers. CEntOS is an enterprise operating system. Had the reviewer even bothered to lookup the acronym (Community Enterprise OS), a horrible experience might have been avoided.
Another favorite of mine you did not mention but I have seen before: the printer software not working. I have read a number of reviews complaining that the software for their new printer does not work with Linux. The printer works, but the shiny CD in the box will not run. Apparently for some reviewers, the software on the CD is more valuable than a functional printer.
I know CentOS/Red Hat very well, thank you.
I used to work as a consultant for Red Hat, the upstream vendor. As an independent consultant I have deployed a fair number of CentOS servers. When I was with Red Hat I had RHEL 4 running on my Toshiba laptop which, I might add, has less horsepower than my current netbook. I am also Red Hat certified.
Don't assume I'm a non-technical journalist. I know what CentOS means. I've used it for years. It *should* work on common hardware, including laptops. As I said in my previous comment, the upstream vendor sells it for that.
Also, what horrible experience? I wanted to recreate my business environment on my netbook. I wasn't happy with the results and I moved on. I knew it would be challenging up front and I changed my mind about it being worthwhile. Why is that horrible? All I lost was time. OTOH, I get paid for some of my writing so I actually derived some benefit from the experiment. That doesn't qualify as "horrible". Rather I wanted to alert people to what the issues were.
but why were you trying it on a netbook
Directed to Ms/Mrs Martin, I still do not understand why you would try a server distro on a netbook, it was doomed to fail to begin with.
This is like trying to transport goods with a volkswagon bug instead of a cube van and claiming that a bug is no good to drive.
I understand your point about late patches, but you referenced firefox as a late patch...none of my servers run a desktop (a discussion for another time..:)), I would be more interested in how long the ssh patch took or other server related functions needed updating.
It's not just a server distro
CentOS is NOT just a server distro. Red Hat, the upstream vendor, sells their Enterprise Linux for the desktop as well. See: http://www.redhat.com/rhel/desktop/
Second, it wasn't doomed to failure. If you read the whole piece I did get everything to work. My complaints about CentOS on the desktop refer to ANY desktop, not just a netbook. Since Red Hat sells RHEL for the desktop then it can and should be assumed that CentOS is for the desktop as well which, in turn, makes all the issues I raised valid.
Why replicate my business environment on a netbook? As a consultant that would be a fantastic sales tool.
I'd be very interested to see
I'd be very interested to see what the stats are regarding who uses Centos as a desktop they use on a daily basis.
I did read your entire article, but you state in the first paragraph "At the moment I am not recommending CentOS for anything, not even servers", and you do not give any real argument why it should not be on a server setting (Firefox patches do not apply since a server setting ideally should have no gui desktop),this is what irks me.
But I agree with this statement "I've decided that RHEL/CentOS just isn't for the typical desktop." I would tend to think those who do use it as their desktop use it in a virtual setting.
Actually, it is used on the corporate desktop
When I was consulting for Red Hat in 2004-2005 I visited a number of enterprise customers who were, in fact, using RHEL on the desktop. I'm talking about installed on ordinary desktop PCs, high powered workstations, and yes, even laptops. I saw virtualized servers for development but I never saw a virtualized desktop.
My objections on servers were the late patches, of which Firefox was only one example, and the whole situation with Lance Davis which, to me, raised all sorts of questions about the long term viability of CentOS and the wisdom of using a distro, any distro, maintained by a small group of volunteers in enterprise/business space. My Entropy blog post linked back to my O'Reilly article which was business/enterprise centric. Take a look at: http://broadcast.oreilly.com/2009/08/the-future-of-centos-and-crite.html
FWIW, I stand by my recommendations. Scientific Linux has a much better track record with patches over the past year and has the backing of and funding by major laboratories and universities around the world.
I know CentOS/Red Hat very well, thank you.
I used to work as a consultant for Red Hat, the upstream vendor. As an independent consultant I have deployed a fair number of CentOS servers. When I was with Red Hat I had RHEL 4 running on my Toshiba laptop which, I might add, has less horsepower than my current netbook. I am also Red Hat certified.
Don't assume I'm a non-technical journalist. I know what CentOS means. I've used it for years. It *should* work on common hardware, including laptops. As I said in my previous comment, the upstream vendor sells it for that.
Also, what horrible experience? I wanted to recreate my business environment on my netbook. I wasn't happy with the results and I moved on. I knew it would be challenging up front and I changed my mind about it being worthwhile. Why is that horrible? All I lost was time. OTOH, I get paid for some of my writing so I actually derived some benefit from the experiment. That doesn't qualify as "horrible". Rather I wanted to alert people to what the issues were.
Ms. Martin and the CentOS misadventure
I basically agree with your point here, but I have one quibble. I frequent a linux forum and try to answer questions where and when I can, and many many people show up who have installed CentOS on their laptops or netbooks, and are subsequently spending countless hours of frustration trying to get their hardware working properly.
When I come across these folks, I now point them to Ms. Martins articles as an example of why CentOS is a poor choice of distribution for what they're looking for.
HCL
RHEL has a hardware compatibility list. CentOS is used on more than an estimated 2 million IP addressees ... a significant number of those are workstations. Centos is the OS of choice wnt many colleges.
If your hardware is on here there should be few problems:
https://hardware.redhat.com/