A Novell Linux Specialist?

richardeholder asks: "Our Novell partners differentiate themselves in the marketplace by promoting their areas of expertise or specializations; this allows their customers to know what their skill sets are and what they can reasonably expect these partners to provide for them. As we embrace Linux, we would like to extend the title of 'Linux specialist' to partners who merit it. Before we move forward on this initiative, we would like to ask the Linux community for guidance on what should constitute a Linux specialist. Should we require certifications such as LPI and the RHCE/RHCT, or are there other more valuable ways of demonstrating Linux competency?"

Comments

certification requirements?

[sl0ppy]

there's the obvious one -- how many anti-microsoft comments you can get onto slashdot during one day.

but, all kidding aside, what's the scope of the specialty?

o firewalls
o security
o script writing

or are we talking end-user support?

o kde
o gnome
o X configuration

there would also be the need for familiarization with common opensource apps, such as openoffice, gnumeric, kword, etc.

what exactly are we talking about?

Re:certification requirements?

[rossz]

Yah, the term "linux specialist" is just too damn broad. The field needs to be broken down into many different fields - some overlapping, e.g. system admin, security expert, app development, kernel hacker, etc. Personally, I'm a pretty good linux admin, do rather well with scripting, perl, php, mysql, and so on, but I struggle with c/c++ coding (I can do it, but I'm not an expert). I compile my own kernels, but would be dangerous to let lose actually making kernel code changes. So what do they actually want?

Certification is important but..

[morelife]

Only one small part of it.

In the business climate, you should only be considered a "Linux specialist" if you have designed, deployed, maintained, or extended production Linux platforms in real life.

Additionally, there should be some track record, say over two years, of professional or technical services delivery either in Linux or another Operating Platform, UNIX, not Windows or Novell.

Finally and most importantly, should be able to provide at least two business references resulting from completed or ongoing projects.

All this would be Real [TM] as opposed to say, the Microsoft "Partner" thing where you need 2 MCPs on staff, a fee, and, uh, that's all.

To my mind, the successful business reference is the best certification anyone can have.

Re:Certification is important but..

[smcavoy]

Gain experience by taking jobs like "Junior Linux administrator" or similar titles.
Don't expect a job as "Linux Specialist" without some real on the job experience.

What do you mean?

[MisanthropicProggram]

The article isn't very specific - to me anyway. The article is vendor specific and if you want an OSS job you need to know generic issues:

Network security - ALL protocols (esp. TCP/IP)

Network Architecture

Platform Integration - including (yuk!) Windows

I don't want to be a TROLL, but it's working out that way. I need people who can put together a NETWORK - NOT be a vendor specific cooky cutter LAN. The more you know about everything (networking) the more it helps to justify hiring you.

I know, it's a CATCH 22 deal these days -(I used to be a coder - and I got sick of it.) There's a lot of us ex-techies who are now PHBs because we couldn't take ( or didn't want to deal with) the current climate in the IT industry and we know TRUE skills when we see them.

I don't even know if this is even on topic based on the article...

WTF?

Yes there is a important thing.

[Yaa 101]

Do this ten times and then ask yourself the same question...

http://www.linuxfromscratch.org/

If you really did this all by hand and got a nice working bare system you really know what drives a GNU/Linux machine.

I'd look for systems-level engineering skills.

[cduffy]

Most of the really, really good Linux folks I've dealt with have been people involved with embedded systems or OS-level programming as well as a reasonable level of sysadmin-style and coding skills.

If I were grilling someone for a "senior Linux expert" position, my interview would probably include the following:

- Describe the scheduling algorithm the Linux kernel presently uses.
- Describe the differences between NFS, CIFS, AFS, Coda, Intermezzo -- and how you'd pick the appropriate one for a given environment.
- Answer a pop quiz about the kernel itself (How is the input core designed? Which filesystems have no limit on the number of inodes? How does the preemptive scheduling algorithm work? Under what circumstances is it undesirable? What's the first thing you do if you get an OOPS?)
- Be familiar with the system's boot process and how to resolve problems relating to it. (What search order does the kernel use when searching for an "init" process?)
- Be able to build an initrd to preload I/O drivers (ie. for booting off a firewire drive)
- Have a solid understanding of the linker, the environment variables and search paths it uses, etc.
- Be able to track down simple bugs in kernel drivers (implications: the candidate must be fluent in C, have some familiarity with the kernel's source base, and know how to use tools such as ksymoops and possibly one of the available kernel debuggers)

Of course I'd also be looking for fluency in at least a few scripting languages (and LDAP queries), an understanding of the tools and libraries underlying GNOME (which we use here) [so I'd want someone understanding GConf, Bonobo, and the like], and so forth. Personally, I'd probably include a series of questions about revision control tools as well, and I'd look for at least a passing fluency with SQL (as a great many popular services backend into SQL databases, it's become rather necessary as a sysadmin skill as well as something important to developers and DB specialists).

Now, are all these skills going to be needed on a regular basis in someone who's just (say) in a sysadmin role? Of course not. On the other hand, the advantage of having someone who understands how things work under the hood is that when they *do* get something really weird jumping out at them, they'll be able to understand what the problem is and *get it fixed*.

Personally, though, I'm not sure what value I see in the whole certification thing. Someone with the kind of skill range I mentioned above typically won't *need* a piece of paper to demonstrate what they know -- it'll be visible in the code they release, in their posts to public mailing lists (Google is your friend!), and in their survival of an actual, proper, face-to-face grilling.

Real linux specialisation

[too_bad]

Its sad that novell dumped their Unix division just when I joined them, and then
they are embracing Linux few years after I dumped them :)

That apart, I think if Novell is really serious about Linux and Linux specialists, please
do not follow RHCE or any other certification. You should identify what aspects
of linux concerns you most. Then identify the contributions made by the concerned
parties to this field. Also, IMHO, it is much more sensible to identify specialists with
their contributions to the open source community. Redhat certification is very
commercialised, and is for people entering Linux to prove their worth to the world.
What you are looking at is to identify real specialists, many of who may not even bother
to prove themselves to anybody, and their works speak for them.

Identify broad areas such as:
Linux Kernel (Accomplished in kernel development)
Linux Device (Someone who is a wizard in getting any device working on Linux)
Linux Installation (Someone who can troubleshoot all kinds of installation problems,
who knows, redhat, debian, gentoo and what not on the back of their palm)
Linux Application (Someone who specialises in applications, KDE/Gnome, etc.)
Linux Ultimate (Accomplished in all the areas)

Now come up with different titles such as Specialist, Master, Guru etc. These titles identify the
extent of their skills. A Linux Kernel Guru might indicate kernel developer whereas Linux Kernel Specialist
might refer to someone who is good at managing patches, troubleshooting the kernel, installing modules etc.

Of course, people who have not contributed to open source should be able to prove themself by either
working with Novell on some projects, or by taking some certification exams. I think it is necessary to differentiate the two categories.

Infact if Novell starts a centralised Linux forum to attract newbies and experts to discuss (like linuxquestions.org) it will be easy to identify the gurus by their contributions and newbie testimonials.

Hope this helps.

Re:Real linux specialisation

[gishzida]

I have a friend that graduated from a 2 yr college with an IT Technical AA degree and a Novell Master CNE certification. She has been working for the last two years as a bench tech. She does not have the depth of experience needed to run a network.

I have worked with and dealt with Consultant CNEs, MCNEs and MCSEs that were nothing more than sales men--- they were not interested in resolving the customer problems as much as they were generating more revenue.

I've also been on the other end of the phone with Novell Customer Support CNEs that had less than *real world* answers... I have placed many a 2AM call to Novell and found the answers were less than adequate. I can use the Novell knowledgebase as well as they can. Some time the only difference between the CNE on the phone and myself was the CNE had access to a proprietary Novell tool which I did not.

Time and time again I've had high cost paid consultants that built their billings rather than my network. I've run into other situations where the "certified consultant" gave us a poorly built solution.

For example, a "certified" security consultant built a RH based web proxy that died because it was improperly partitioned. I have a CNA 4.1 and MCPs for NT 3.51 / W95 and have no Linux certs... I don't do linux on a daily basis as I'm in a Novell shop with lots of MS app servers [oh god here come the patches again!]

I managed to resurrect the drive & repartition the install on to a new hard drive. The solution was ugly but it worked [and is still working after six months]. I could probably do a better job of it today [spending a weekend building gentoo on an ultra5 is educational]

Certification has for some companies been a method to generate new revenue streams without providing lasting *practical* value to the student or to their clients / employers.

The Novell cert I took had more practical value than the MS cert. {i still remember the question on applying permission filters] With that said, the reality is that a "certified specialist" is a marketing label for "certified salesmen" -- sorry to sound so cynical but I haven't met that many "certified alphabet soup" consultants that were worth their weight to solve real problems without wanting to either a) sell me something I did not really need or b) Give me "bill of goods", a large consulting bill, and no optimal solution.

If Novell wants to certify Linux specialists they need to make it a useful certification and not just another piece of marketing "wizz-bang"... After doing this for seven years I don't need to "prove" what I know... I do that every day-- I have little motivation to go an get new things to put on the wall... except maybe to prove to a "suit" that I am familiar with the subject.

By all means look for certifications

[PhilipPeake]

But if you are relying on certification to ease the task of interviewing/selecting a candidate for a job, then you are probably wasting your time.

To put it simply, parrots can get certified. You don't have to understand a thing, just remember what you were told.

You don't want a parrot, you want a penguin :-)

Also remember that certification gets out of date rapidly - ultra rapidly in the Linux world. So that certificate may not be worth much anyway if its more than 6 months old. So are you willing to give the time to these certified people to keep up to date, and keep their certification up to date ?

Two cents on certifications

[digitalhermit]

Very interesting question. Five years ago I would have said that certifications were about as useful a predictor of knowledge as swirling tea leaves in a cup. I've seen some really good MCSEs with a thorough knowledge of networking and their specialties and just as many who can't properly subnet a network. Five years ago, anyone who was seriously into Linux was *usually* a competent sysadmin in their own right and didn't need a piece of paper to prove it.

I think this has changed. There are a lot more people getting into Linux for the money than there has ever been before. This has upsides and downsides. Upside -- Linux is growing. Downside, it's no longer an arcane science. I can live with the downside though.

Alas, we peddle our skills to non-technical types who don't understand that a cram session and the ability to prepare for a test does not predict the knowledge of a consultant or future employee. They, for good or bad, use keyword filters or head hunter resume databases to choose candidates. Those who have magical letters on their resume get noticed. The rest, regardless of knowledge, get filtered out before ever being seen by a human.

So are Linux certifications a good thing? Maybe. If done right and don't end up as a certification mill as happened with the MCSE, then sure. They can help show a certain level of competency and could ensure that the certificate holder has the broad level of knowledge required to pass the test. In the next five years I'd expect that more executives will start asking for some certification anyway.

This will only work if the certification process in itself does not become an industry. The cost of taking the test should be low (under $150) but it must be difficult. Ideally it would include a practical, hands-on section instead of a bunch of multiple choice questions. The course work should be openly available and reflect not only real-world knowledge but some theoretical and philosophical aspects of using Linux.

Re:Two cents on certifications

[TrombaMarina]

A good test will eventually be well respected by the community. The most accurate test would simulate the exact kinds of problems you want the applicants to solve. Ideally, this means giving each applicant a broken machine to fix and enhance.

I would have applicants solve a bunch of problems on a very badly-configured machine, and maybe have them bring up a new service or two on the same machine. You can have a library of 30 or so badly-configured machine images, and each applicant gets a random install.

Make it an open-internet test so that people can download their favorite tools and look up answers to things they might not know. The real world rewards people who get things done, not people who can cram for multiple-choice tests, so make your test work the same way.

People have bad days, so you should let them retake the test if they fail, but make sure not to randomly give them the same install twice. After three tries they should not be allowed to take the test again - that would cheapen it for the folks who prepared properly for their first try.

Only about half the people I look to for technical guidance have degrees and certifications. This suggests that most existing tests either do not attract these people or that these people don't necessarily do well with traditional tests. Configuring a real server would be fun for the applicant, and would give a good indication if an applicant could make the grade in a professional environment.

If this proves too expensive to administer, you could have a separate multiple-choice qualification test to prevent newbies from taking the (more expensive) real test.

Good luck!

Re:Two cents on certifications

[tyler_larson]

It's been my unfortunate experience that those whose titles contain the word "Certified" tend to lag far behind the rest of us in the industry. In school, the hackers studied Computer Science or Electrical Engineering and learned how to make computers work. The dropouts from these majors who just couldn't hack it (those who wanted the jobs but didn't have that special gift that lets you think like a computer) switched instead to "Information Systems" in the business school.

Instead of learning how to make a computer work, there they learned how to work a computer. I.e., they learned how to use the programs that we learned how to create. Their programming coursework was graded on effort rather than effectiveness. The program they wrote didn't have to work (or even compile), it just had to look right. And yet, interestingly enough, part of this particular university's I.S. program was to complete the MSCE coursework and obtain certification.

Now that's scary.

You may be thinking that this is nothing special--network administrators don't need to know how to write programs, right? Well, immagine hiring an MSCE who can't even write a working batch file--but who can at least recognize one 80% of the time! This gets even more complicated in the *NIX world, where customization means script writing. Certification requires you to have more or less the same understanding of computers as someone like me has of quantum physics: memorized well enough to pass the test, but not necessarily understood enough to remember it next week.

Those of us with real computer skills never bothered to take those certification tests because they cost money and meant nothing (after all, the CS and EE dropouts are all certified). Do I really want to be placed in the same catagory as THEM? Those of us real admins who are actually certified only did it so they could pass the resume word search.

And yet, it's useful to know who understands system administration and who doesn't. Certification was a stab at making such a devision. It's relied on heavily by HR departments everywhere, despite the fact that it means so little in practical terms, because the HR department doesn't know any better.

But the truth is, IT certification is a flawed system at its very core. Certification is based on book work and tests--neither of which you'll find in the real world. It doesn't have to be that way, and here's why:

IT Certification should work like Pilot Licensing

In order to fly an airplane, you need to know the books, but you also need to know how to fly--that means extensive practical knowledge that can be demonstrated to an examiner in an actual airplane thousands of feet above the earth. You can't cram for that test.

In the same way, certified Linux experts should be able to demonstrate practical skills, not just pass tests. For example, a Linux system administrator should be able to write (in just a few minutes) a simple script to parse 300 log files looking for a specific pattern. He should also be able to modify the boot scripts such that a custom executable runs after (and only if) the network is brought online, but before any remote login programs start. They should be able to customize a firewall, build and install a kernel module, examine a syslog file, and maybe even configure xdm. And the stuff that he demonstrates has to *actually work*, not just look right.

I dare say that any of the existing certification courses will have a well-thought-out curriculum, and probably cover all the necessary aspects of system administration. However, knowing the location and purpose of the rc.sysinit script doesn't mean that you know how to customize it to do your bidding. Certified engineers may know "what", "where", and even "why". But only experience and practice teaches you "how". And knowing how is a mandatory component of being useful.

Re:Requirements for a linux specialist:

[Pope Raymond Lama]

Even with a "5 funny" score, parent seens the only post to touch the question asked.

I would not mind calling a Linux Specialist one that would be capable to comply with the 4 first itens in parents list:

-Properly secure a firewall
-Compile and install a kernel
-Configure the third button on thier mouse
-Print to a Panasonic KXP-8410 printer in color

Itens 3 and 4 - or maybe some other semi-random tasks - would assure that the guy really can handle it, and not someone who just overworked a couple howtos.

Re:More then linux

[AceM2]

While that's great and such.. What on earth does that have to do with their ability to properly run a server? We're not talking about giving someone a college degree here.. Random knowledge really just says to me that the person got bored and looked it up.. I mean I'm good at a lot of things I don't know the history of. It's not that I don't care about history, it's just that since when do you have to be fanatical about something to get a job? I'm glad all of my job interviews left out the trivia questions, or else I'd never have gotten any of the jobs I had fun and did very well at..

A Linux specialist can mean an awful lot

[StArSkY]

After spending quite some time on the Golf Course with a Novell SME evangelist, I will add this comment: You want a GNU/Linux specialist. Youy want people who understand more than just the kernel. Given that assumption, I will go further.

If you want to measure competencies, then you need to break this down into categories.

Design Competency
This should cover the elements of gathering the requirements from a customer, and designing appropriate hardware and software solutions for the client. It should include an assessment of their knowledge of

• Server Design: Do they know how to spec a server, do they understand the consequences of Multi-proc, Raid and hardware choice.
• Network design: Do they understand all of the services they need to provide to support the activities on the network
• Security & Authentication: Do they know how to enforce password changes. Do they know what the key security holes are for key distributions and applications "out of the box"
• Client design: What should they look at in choosing a distribution, or would they build their own. What Customer requirements should they capture to drive choices for wm and application choices.

Implementation & Support
This should capture if they understand the nuances of rolling out linux in an organisation

• System Imaging: Do they have a handle of the options available for building multiple linux servers and clients for mass rollout
• Do they know how to keep systems up to date and patched.
• Do they know how rebuild a linux system with minimal loss and reconfiguration.
• Application Competency: Do they understand the purpose and limitations of key applications that Novell will base their offering around.
• Are they able to explain to your secretary how to use all of their key applications
• Interoperability: Do they understand how get linux clients talking to Novell, Microsoft and other Unix Servers. Do they understand how to get windows workstations using linux for file sharing, and other services.

Now the harder part is "how to you test this". As a perpetual student I would suggest that you need to do this in a multi-phase test. This will be expensive, but it will ensure that you get some level of objectivity:

1: You write a scenario based around a real world example to assess their ability to come up with an appropriate hardware and software solution for the design components.
2: Give them a multi-processor + Raid + Redundant PSU server and get them to install, configre and lockdown the system. Then get them to install a heap of services that you think are important.
3: Give them 10 PC's. Get them to build an image on 1 and then roll the image across all 10 workstations. Make some of the workstations different, give them different size drives, and different network adapters. See how they adjust to the variations
4: Bring a secretary in and get them to run a "how to use" session for the secretary. See how good they are

Also: Note this does not necessarily need to be the one person for all skills. So be flexible and allow them to have up to 3 different people to deliver the outcome. The only limit is that only 1 person can spend time with the end-user. If you want any more ideas or thoughts, feel free to mail me: starsky AT bluecouch DOT COM DOT AU

Can we maybe try a little harder here?

[RichardY]

At last, a decent post from a Slashdot reader (Only took 168 comments). Thank you StArSkY.

For those of you who don't know, Novell qualifications still count for a great deal in the real world of IT contracting. There is still a perception that 'real' IT professionals have a CNE (Certified Novell Engineer). The rest have an MCSE.

Novell recently aquired Ximian. This gives them access to the XD2 desktop, Mono, Gnome development, red carpet and more.

I think that this puts them in a very good position to integrate their current networking products (E-Directory, ZenWorks etc) with Linux.

This does however potentially create a compex environment for a corporate to adjust to. In order for them to adjust, there has to be a set of qualifications that distinguish between someone who can impliment an enterprise environment, from a home user still at school.

What about the homegrown linux expert

[Coolmoe]

I liked what I read in a reply about LFS install and the some from the Slackware camp. I think that a Linux expert should be able to recompile a kernel and configure the system using native GNU tools on all systems. The main problem I have with the RHCE is the fact that a lot of it seems very Redhat centered. I know that there is no way that a person could be concidered competent if a "Distro" stumps them if you can use the standard GNU tools that works on Redhat and on all other systems. All unix type skills should be there too. If a person cannon even crack open vi and edit files he is not your man.

Re:this man is not one of us

[jagger]

Did you notice his email address ends in @novell.com.

yes he is a suit most likely or perhaps he had his words filetered through a lawyer but he asked us a reasonable question. (quite flattering that a company like novell would seek /. reader's advice) All I see are smartass responses.

Re:Requirements for a linux specialist:

[mnmn]

-Properly secure a firewall
This should really include configuring and using snort as the IDS part, full control over the iptables command and NOT using squid.

-Compile and install a kernel
And change the default screen color, use rdev and use ksymoops to locate the source of a crash and report it properly using the right bugreport email format to the exactly right group of people after having searched mailinglist archives and newsgroups for the same problem. Should also know enough C and about makefiles to correct function typos, comment out assert blocks and point the makefiles to the right libraries and include directories. Remember we're talking about a linux SPECIALIST, not a kiddie compiler.

-Configure the third button on thier mouse
No thats too tough

-Print to a Panasonic KXP-8410 printer in color
Too many Linux experts know little about printers. Many printers dont have drivers but can shared to smb/active directory users who do have the drivers. I strongly agree with this point.

-Make coffee that is restricted under OSHA guidelines
Or Quetta Tea. Doodh Pati. Not all geeks are scrawny white suburban californians.

-Recognize a minimum 8 of 10 random network cards by thier chipset number only
I would agree with this one, but you can have software-only experts too. I can tell ethernet, tokenring, fddi, atm, arcnet and isdn cards by their chipset, whether they will run under Linux/FreeBSD or Solaris, but I know guys who dont change their RAM but develop sophisticated KDE apps. There are niches...

-Understand the usefullness of the SysRq button
This can be learned in a day, so a specialist MUST know it.

-Install linux on any appliance that does not come with a keyboard or mouse
I am tempted to say he should be able to do the same with Solaris. Should also be able to install Plan9, Windows 2000 Advanced Server, Unixware, AIX and QNX on vmware images. (bochs would do)

-Setup a cron job to order pizza online
Too easy

-Pay a license fee to SCO
Uhh yeah. I'll hire such a person.

-Assemble a beowolf cluster which includes more than one type of gaming console
I'll bring a serious point here. This is not a practical skill. Just test for RHCE and that covers the really required sysadmin skills. Someone who is obsessed with gamebox clusters will not be interested in 99.9999 uptime. He will be playing unreal on the servers at night.

-Install a really cool kde/gnome/enlightenment theme
And get used to twm and CDE. Learn to enjoy the command line.

-Run desktops at no less than 1600x1200 resolution, native
Why? Use a real DEC VT520 dumb terminal.

-Name all boxen after sci-fi characters/objects
Come on geeks are more diverse than that. I havent seen star wars.

-Any cats owned must be named after cabling specifications
Or kernel header files, or commands.

-Adequate space must be reserved in all hardware racks for pizza boxes
One PC-XT case should be home to a hamster.

-Every system must glow at night. Server rooms should be scary
I second that. Modders for beautification are generally not all that interested in the software and are gamers.

Here are some others:

1. Take a 10GB ext3 or XFS or reiserfs partition full of data, and dd if=/dev/zero of=/dev/partition block=512 count=4096, do the same for the END of the partition, and dd if=/dev/partition of=/tmp/file and give the file to him to recover as much data as possible.
2. Give him a Linux server and a Pentium1 with no harddisk but a bootable nic, and tell him to remote-boot windows95. Hire him for at least $80k if he can do that.
3. Act like a project manager and ask him to put all his work in project form on MS Project 2000, and submit weekly reports. Cause deliber

Real-World qualifications

[zhero]

The ONLY exams worth ANYTHING are real-world exams like RHCE, which I think is an excellent exam unlike MC$E which everyone cheats at as their is no real world knowledge.

I tend to disagree with previous posters who say that the technology is moving too fast to make a qualification worthwhile. Well thought out exams like RHCE test how well you think on your feet and not how well you know a particular product. However, I would suggest that most releases are similar enough that if you know version x well you will fly through version y.

Firmly agree with posters who say experience is key, but having that and a solid qualification or 2 puts your CV to the top of the file.

Re:A little too specific and not broad enough

[cduffy]

As for professionals in the computer industry: there is one skill that all system admins must have. That is the ability to search. In your example, the higher abstractions "break". If they broke for them, it would have broke for others too. As a sysad, you must either "reinvent the wheel" by taking waay too much time coding, or searching for 30 minutes to find and implement others solutions on the problem. Fixing and submitting is a last resort.

Yes, if I were looking for a sysadmin, I'd take off the kernel requirements and most of the C. This is a "senior Linux expert", though: "Linux expert", not "Linux sysadmin" or even "expert Linux sysadmin".

A linux sysadmin can get away with knowing how to... well... administrate systems. A linux expert, OTOH, should be a generalist -- and that means, among other things, development and debugging skills.

Such people aren't impossible to find: The current senior sysadmin where I work has roughly the skillset I described, plus almost all of what you mentioned, plus a bunch of additional skills on the side (his personal projects have ranged from building some rather fancy revision control system tools to semi-maintainership of a microkernel OS targeting embedded work; his previous job experience includes everything from Java development to database administration to building and writing software for compute clusters).

Perhaps I've gotten spoiled by working with such people (MontaVista Software, my previous employer, had a number of them) -- but I'm now hesitant to call anyone with a substantially lesser range of skills a true "Linux expert". (I suspect part of their ability to attrack such people may have been related to being a shop whose core business was improving Linux and related Free Software -- at least a few of my coworkers said outright their primary motivation in choosing MontaVista as an employer was not at all the pay).

And finally... well, sometimes there just aren't others' solutions to implement. Having someone who can do the fix-and-submit routine in those cases is part of what one hires experts for.

(As for the scheduler, btw, I'd gladly accept an approximation: Someone who says "no, I'm not familiar with that, *but*..." and can go on to describe a sane scheduling algorithm and what kinds of problems it's prone to passes that question with flying colors, even if they don't happen to know what Linux is doing this week).

Proper certification..

[Anonymous Coward]

Unlike all the certs i have.. which mainly were attained by memorization and not necessairly but actually knowing how to do the work, i think the industry needs more of a apprenticeship type program.

There are too many different areas of specialization and you can't possibly be an expert in all of them.. I know programmers that can't for the life of them figure out IP subnetting, yet at the same time they write awesome code without much difficulty.

At the risk of being flamed, i've got Novell 4.10,4.11,5 CNE certs as well as the requiste MCSE, and CCNA.. but from experience.. they don't mean CRAP.. the best training was the novell because you guys discussed not only how to make it work, but why it works.. unlike the M$ training.. (press this button clicker training)

I got my cisco certs, not because i needed to learn it, i already knew all the stuff in that course, but because my employer was required to have "X" number of cisco certified people on staff.. i regularly subbnetted and routed ATM networks.. and people think IP is hard.. hahah..

With a apprenticeship type program.. you guarantee that a: people are actually getting real world experience and b: the people are worth their salt..

A standard test at teh end like a CCIE where there is practical exam (take all this hardware and make it do this..) makes all teh sense in the world..

Anyone that calls themself a linux expert is full of shit.. there are too many different areas of specialization.

Most important...

[rocket_w]

.... the most important thing that you could do is an in house test. In my opinion, the best certification that has been available in years was the CCIE, Mainly baecause it was a hands on test. If you really want to be able to say someone is a Linux Expert design a test around the things that are most important and have each canidate pass the exam.

I would say the following are areas to focus on:

Security, Security, Security!!! A good security minded person will naturally have all of the administration skills; user accounts, passwords, ftp security, etc.
configuration to various hardware platforms, Raids, NICs, etc.(note this should require recompilation)
Connectivity, SMB, NFS, etc...
Automation, setting up maintenance, logs, and alerts
Live diagnostics, can you find out what is wrong without taking down the server
Clustering (if you need someone with that much knowledge)


Problem solving! Not everything you do will be in a how-to, or man, or even in a book. The best "experts" know how to make their own solutions when there is not a suitable one available. That is also a real core value in the Linux Community.


A very important thing to keep in mind is that there is a big difference between clients and servers, and just because a person is good at one, does not mean that he or she will be good at the other.