|
Home > Archive > Unix administration > February 2005 > How long does it take to become a good sysadmin?
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
How long does it take to become a good sysadmin?
|
|
| Håvard Axelsson 2005-01-22, 8:48 pm |
| Ok, I know you already have the answer: "Everything is relative!"
I am not a Sysadmin, but I like it a lot. I've recently bought a book that
seems good so far, it's called "Unix System Administration Handbook" by
Evi Nemeth. I also bought the FreeBSD handbook. I'm devoting 4 hours every
night to study Unix Administration, with a focus on FreeBSD.
I am able to install FreeBSD, to recompile the kernel with other options,
to run CVSup, to setup a basic office network with a gateway, to configure
a firewall, a connection to the Internet, to install Linux programs
through the Linux binary compatibility layer, and I also installed Apache,
PHP and PostgreSQL for the office. I'm taking time to read the man pages
each time I have the opportunity.
I read about 50 pages a day. I like it a lot, and I'd like to know how
long (roughly) it takes to be a Sysadmin that is good enough to do
advanced tasks (maybe regarding security, performance, DoS) as well as
most common tasks. In other words, a good Sysadmin.
As you see, I'm not into the Unix world, so bear with me, no flame ;)
Thanks,
--
Håvard.
| |
| Mxsmanic 2005-01-23, 2:48 am |
| Håvard Axelsson writes:
> I read about 50 pages a day. I like it a lot, and I'd like to know how
> long (roughly) it takes to be a Sysadmin that is good enough to do
> advanced tasks (maybe regarding security, performance, DoS) as well as
> most common tasks. In other words, a good Sysadmin.
You just read more and do more. That's how all sysadmins do it.
--
Transpose hotmail and mxsmanic in my e-mail address to reach me directly.
| |
| Adam Price 2005-01-23, 2:48 am |
| On Sun, 23 Jan 2005 00:32:24 -0200, Håvard Axelsson wrote:
> Ok, I know you already have the answer: "Everything is relative!"
>
> I am not a Sysadmin, but I like it a lot.
Run Run away :-) and when you are old and bitter remember that I told you
so.
First the flippant answer... Since an admin is only as good as his backups,
once you have those mastered you are a good admin.
Second I don't see anything in your list that includes users, if your
system doesn't have users then you will only see problems you created. This
is not enough.
However you seem to have a good attitude and are putting in some effort,
that should help.
Which man pages are you reading? You may do better reading commercial ones
rather than free ones, at least in the Linux world the man-pages seem to be
the last thing that gets updated.
As for a 'good' sysadmin, that is a state of mind. You will be able to do
advanced tasks when you learn how. For me that came quickly as the old
admin left and I was the guy with the root password. It was a case of fix
things or they get in a contractor to it and get rid of me.
But others I have worked with took much longer. I know work in a managed
service environment and we generally say that we can train a competent
admin for the on-call rota in two years if we get them fresh. Three years
if they have come from a Windows background, and one year to six months if
they have some unix experience already.
That relies on them to be a sysadmin for 8 hours a day for the entire
period, they act as our front line troops, taking all of the work that we
do and only escalating when they are out of their depth. When they do
escalate we take great care to ensure that they remain in charge of the
case until the end, but get the technical expertise they need to continue
to learn.
At least that is how it is supposed to work. The reality is somewhat
different.
Still get yourself a job as a sysadmin and work up. You are a good sysadmin
when you think you are a good sysadmin.
Adam
| |
| Håvard Axelsson 2005-01-23, 5:50 pm |
| Em Sun, 23 Jan 2005 06:02:34 +0000, Adam Price
<adam+usenet@pappnase.co.uk> escreveu:
> Run Run away :-) and when you are old and bitter remember that I told you
> so.
> First the flippant answer... Since an admin is only as good as his
> backups,
> once you have those mastered you are a good admin.
Thanks for the shivers! I think I'll just do it as a part-time job or for
hobbies (having a web server at home, in the office too, help friends if
possible)
> I know work in a managed
> service environment and we generally say that we can train a competent
> admin for the on-call rota in two years if we get them fresh. Three years
> if they have come from a Windows background, and one year to six months
> if
> they have some unix experience already.
Good to know. I already have a little bit of Unix background (we had one
year of Unix admin at school, in a web programming course)
> Still get yourself a job as a sysadmin and work up.
Thanks, I'll try to play with a little web server that I setup here. Then
who know, I might be able to have a good level one or two years from now ;)
--
Håvard.
| |
| Juha Laiho 2005-01-23, 5:50 pm |
| =?iso-8859-15?Q?H=E5vard_Axelsson?= <me@privacy.net> said:
>Em Sun, 23 Jan 2005 06:02:34 +0000, Adam Price
><adam+usenet@pappnase.co.uk> escreveu:
>
>Thanks for the shivers! I think I'll just do it as a part-time job or for
>hobbies (having a web server at home, in the office too, help friends if
>possible)
You can get some aspects of being a sysadmin just for running a system for
yourself, but that isn't a realistic scenario, and in most cases will miss
an important aspect -- that of being responsible to your users.
So, your users are important, they're the ones who have a work to do with
the system. A sysadmin exists to keep the system running for the users.
Losing a disk containing the mail spool? Well, if it ws just your own
mails, well, you just lost your mails. But what if there were also mails
for even ten other persons -- or just even one? What if there were 100
or 1000 mail users on that machine, and the mail spool just disappeared?
If the system is down, is it just that your personal toy is broken -- or
is it that there are a hundred users waiting to get it fixed to get their
work done (and here, money becomes an issue: in work environments, with
just 100 users, one hour of unexpected outage is an expensive event).
>
>Good to know. I already have a little bit of Unix background (we had one
>year of Unix admin at school, in a web programming course)
I pretty much agree with the two years. And after the two years you can
tell whether the new one has a gift for being a sysadmin, or whether s/he
is an average one ("just doing my work..").
>
>Thanks, I'll try to play with a little web server that I setup here.
>Then who know, I might be able to have a good level one or two years
>from now ;)
When working with your own systems, try to:
- minimize amount of temporary hacks; try to do things so that you can
expet the solutions to last and to scale
- automate, monitor, and automate monitoring; have the system tell you
if something is not in right order; then, have some way to make sure
the automated monitoring is working!
- when making changes, think of possible adverse effects, and ways to
roll back the changes
- do backups; make estimate what does it cost to have monthly full backup
available for one year after -- then extrapolate the figures to 1TB
of data and 3-5 years of storage time for at least part of the data
(after this you know why storage capacity is not cheap even though
disk capacity is dirt cheap)
--
Wolf a.k.a. Juha Laiho Espoo, Finland
(GC 3.0) GIT d- s+: a C++ ULSH++++$ P++@ L+++ E- W+$@ N++ !K w !O !M V
PS(+) PE Y+ PGP(+) t- 5 !X R !tv b+ !DI D G e+ h---- r+++ y++++
"...cancel my subscription to the resurrection!" (Jim Morrison)
| |
| Mxsmanic 2005-01-23, 5:50 pm |
| Juha Laiho writes:
> You can get some aspects of being a sysadmin just for running a system for
> yourself, but that isn't a realistic scenario, and in most cases will miss
> an important aspect -- that of being responsible to your users.
Experienced sysadmins are recognized by their behavioral traits, not by
their technical expertise. They tend to be obsessed with backups, and
are extremely conservative in their day-to-day operations and in
continuing projects. They also tend to have a healthy paranoia about
security, a jaded view of industry trends, an extreme reluctance to
undertake major projects, and an absence of any control-freak
tendencies. Also, they are very honest, always giving priority to
uptime and never doing unsavory or unethical things in their work. The
best sysadmins also have no history of doing such bad things in the
past; people who have worn black hats once never really take them off
completely, despite claims to the contrary, and such people are
dangerous in sysadmin positions.
Since being a good sysadmin is a matter of "war experience" rather than
technical training, good sysadmins tend to be older sysadmins. They
usually find switch to something else when they get tired of the 24-hour
on-call duty and (usually) excessive workload.
> When working with your own systems, try to:
> - minimize amount of temporary hacks; try to do things so that you can
> expet the solutions to last and to scale
I always keep tweaking to an absolute minimum; it's too hard to remember
or log every tweak you make, and the tiny gain in performance or
temporary convenience you might get from it is not justified over the
long run.
> - automate, monitor, and automate monitoring; have the system tell you
> if something is not in right order; then, have some way to make sure
> the automated monitoring is working!
Watch everything; have something else watch anything you can't watch;
have something watching whatever watches what you can't watch.
> - when making changes, think of possible adverse effects, and ways to
> roll back the changes
Backups, logging of all changes, extra copies of everything. I have
dated copies of configuration files all over the place. I take forever
to delete them because I want to make double sure that I can roll back
at least one and often several generations if need be.
> - do backups; make estimate what does it cost to have monthly full backup
> available for one year after -- then extrapolate the figures to 1TB
> of data and 3-5 years of storage time for at least part of the data
> (after this you know why storage capacity is not cheap even though
> disk capacity is dirt cheap)
Backups are job 1 for sysadmins. A failure to do backups is always more
than sufficient cause for immediate dismissal. A cavalier attitude
towards backups and uptime is the hallmark of the incompetent newbie
sysadmin.
All of this applies to any system, not just UNIX systems. They all look
the same to system administrators.
--
Transpose hotmail and mxsmanic in my e-mail address to reach me directly.
| |
| Doug Freyburger 2005-01-23, 8:47 pm |
| H=E5vard Axelsson wrote:
>
> I am able to install FreeBSD, to recompile the kernel with other
options,
> to run CVSup, to setup a basic office network with a gateway, to
configure
> a firewall, a connection to the Internet, to install Linux programs
> through the Linux binary compatibility layer, and I also installed
Apache,
> php and PostgreSQL for the office. I'm taking time to read the man
pages
> each time I have the opportunity.
At this point you can sound out words on the page and call
it reading. Your goal is writing poetry with a caligraphy
pen.
> ... I'd like to know how long (roughly) it takes to be a Sysadmin
> that is good enough to do advanced tasks (maybe regarding security,
> performance, DoS) as well as most common tasks. In other words, a
> good Sysadmin.
Five years. Look at the Usenix SAGE page and read the job
descriptions. One of the qualifications to call yourself
Senior is being in the field five years. No matter how smart
you are, no matter how much you study, you can't be senior
until five years have past. It makes sense - If you start at
ground zero and you want to design a bridge, the studies also
take five years.
I've done tech screening interviews and there are *plenty* of
people out there with 10+ years of experience who don't cover
all of the requirements for SAGE Senior. It's that much work.
It's intended to be. You wouldn't want someone who hasn't
done that much work to design a bridge to drive across a
river, so you don't want someone who hasn't done that much
work to build your data center.
There is a *lot* of work that isn't Senior, though. In fact
it's impossible to meet all of the items on the list without
working in the field for several years. At this point you
appear to be qualified to work as an installer at a large
company, and with a few months of full-time work to be put
into the pager rotation. Do not even attempt to become "good"
before you start to work in the field. It doesn't work that
way. Treat it as a process not as a goal. Enjoy the process
and make a living doing so.
| |
| Alan D Johnson 2005-01-25, 2:52 am |
| Doug Freyburger wrote:
> Håvard Axelsson wrote:
>
>
> options,
>
>
> configure
>
>
> Apache,
>
>
> pages
>
>
>
> At this point you can sound out words on the page and call
> it reading. Your goal is writing poetry with a caligraphy
> pen.
>
>
>
>
> Five years. Look at the Usenix SAGE page and read the job
> descriptions. One of the qualifications to call yourself
> Senior is being in the field five years. No matter how smart
> you are, no matter how much you study, you can't be senior
> until five years have past. It makes sense - If you start at
> ground zero and you want to design a bridge, the studies also
> take five years.
>
> I've done tech screening interviews and there are *plenty* of
> people out there with 10+ years of experience who don't cover
> all of the requirements for SAGE Senior. It's that much work.
> It's intended to be. You wouldn't want someone who hasn't
> done that much work to design a bridge to drive across a
> river, so you don't want someone who hasn't done that much
> work to build your data center.
>
> There is a *lot* of work that isn't Senior, though. In fact
> it's impossible to meet all of the items on the list without
> working in the field for several years. At this point you
> appear to be qualified to work as an installer at a large
> company, and with a few months of full-time work to be put
> into the pager rotation. Do not even attempt to become "good"
> before you start to work in the field. It doesn't work that
> way. Treat it as a process not as a goal. Enjoy the process
> and make a living doing so.
>
Hi all. I've been lurking here for a little while, I'm an HPUX person
with lots of years on various systems, UNIX, VMS, DOS, WIN, etc... and
y'all hit it on the head w/ backups and having users. I work on heavy
systems at work and make SURE they work ( regardless of when stuff
happens, other "sysadmins" I work with shut off pagers and screen calls
), and if they fail I can bring them back quick, usually sooner then our
vendors can.
What do I do when I come home? I work on my own systems. Lost a disk
on one of "my" servers a couple of days ago, oops. When I can turn
around and replace a disk on a server ( Winders )((surplus old crap ))
and put a restore back on it in less then 4 hours before my kids start
bitching about not having their pictures and music and etc. I guess I am
doing something right. Try having your own kids and spouse as users!
talk about fun!!!
Thats what being a sysadmin is. When you can work on whatever you are
on and be confident that what you are typing on the keyboard is going to
make your systems or your job or your customers better, easier, happier
and enjoy what you are doing in the process.
Nuff said, I'm going back in the shadows, maybe.
adjtech@hvc.rr.com
| |
| Jeremiah DeWitt Weiner 2005-01-25, 5:52 pm |
| H?vard Axelsson <me@privacy.net> wrote:
> I am not a Sysadmin, but I like it a lot.
Seek professional help, now.
> I've recently bought a book that
> seems good so far, it's called "Unix System Administration Handbook" by
> Evi Nemeth.
Yes, that is definitely one of the better ones, although sysadminning
is like juggling in that you have to get out there and do it; just
reading about it doesn't enable you to juggle.
My take on it is that it takes about five years to get good,
assuming you keep your eyes and ears open, work at interesting places
and really study the field and think about the pros and cons of what
you're seeing and doing at work. If you have a really good environment
and good mentors, maybe less time - although seeing bad environments and
bad admins is a valuable learning experience too. If you're stupid and
careless, no amount of time will ever be enough, of course.
One element in the development of a sysadmin which I think is not
mentioned often enough is exposure to a wide variety of different
environments, software, hardware, etc. I have seen admins who were reasonably
smart people and good at what they did in a limited sense, but who were
professionally stunted by the fact that they'd only ever seen one
environment and one way of doing things; they thought there was only one
way to do things, and they were very fixed in their ways. Of course,
being exposed to different environments can be frustrating and make you
feel stupid, too - I've been at my current job about four months, and
they run a large, complex, fragile, and idiosyncratic software
environment that I still don't understand very well.
Remember, too, that the real strength of a sysadmin is not in his
purely technical skills, but in design, planning, documentation...a
whole bunch of more nebulous, harder-to-teach, but no less important
skills. Think of it as being like a carpenter - driving nails and
sawing wood are the concrete skills he uses, but being really great at
driving nails, no matter _how_ good he is at that, is not going to make
him a great carpenter. These more nebulous skills are where the
difference lies between running Unix systems for your own use and
running an "enterprise" environment. There are a lot of habits you can
pick up doing things by yourself which will not serve a team well in a
larger environment. "Cowboy admins" are my bete noire - the kind who
says "Here's something broken...I can fix this! I'll just hack up a quick
patch in five minutes!" And then, of course, he doesn't document it or
even tell anybody else, and it doesn't match any of the other systems, etc.
I shouldn't focus on the negatives, though - working with a good
team is a great experience, and I'm glad that early in my career I was
lucky enough to fall in with some excellent teams and some excellent
technical managers, and I got my hands on a lot of different systems.
I've been at this game about seven years now, and when it's good it's a
lot of fun, it's good money, and it's satisfying work.
JDW
| |
|
| Begin <ct6epr$8ks$1@reader1.panix.com>
On 2005-01-25, Jeremiah DeWitt Weiner <jdw@panix.com> wrote:
> H?vard Axelsson <me@privacy.net> wrote:
>
> Seek professional help, now.
Seconded. Going here is really bad for you sanity, your outlook on life,
your life in general, your sleep, and... Have I mentioned sanity yet?
[snip]
> One element in the development of a sysadmin which I think is
> not mentioned often enough is exposure to a wide variety of different
> environments, software, hardware, etc. I have seen admins who were
> reasonably smart people and good at what they did in a limited sense,
> but who were professionally stunted by the fact that they'd only ever
> seen one environment and one way of doing things; they thought there
> was only one way to do things, and they were very fixed in their ways.
Same goes for programmers and such. Gee. ;-)
> Of course, being exposed to different environments can be frustrating
> and make you feel stupid, too - I've been at my current job about four
> months, and they run a large, complex, fragile, and idiosyncratic
> software environment that I still don't understand very well.
Then it helps to be able to just pull old stuff out from the racks and
say you can't support it. It is a scary thing to do for someone who
loves to make stuff work. Sometimes the big axe is the cheapest, not to
mention sanest, and most honest, way of ru^Wweeding out a shop.
> Remember, too, that the real strength of a sysadmin is not in his
> purely technical skills, but in design, planning, documentation...a
> whole bunch of more nebulous, harder-to-teach, but no less important
> skills. [...]
With the added problem of needing much more than most other departments
to know not what is going on, but what _will_ be going on in a year or
so. (I think this part is the difference between an _admin_ and what
I'll call an _operator_, if I wanted to be rude I'd say a glorified tape
monkey. There are admins who really are just operators.)
[snip]
> I shouldn't focus on the negatives, though - working with a good
> team is a great experience,
Been there, done that. And I did it in an environment where there were
lotsa free beers after work. And no computers. At least not initially.
Too bad it didn't last, but anyway. It wasn't full-time, which if it
would be, would hurt the amounts of beer our team+chef were consuming.
> and I'm glad that early in my career I was
> lucky enough to fall in with some excellent teams and some excellent
> technical managers, and I got my hands on a lot of different systems.
Lucky, lucky you. I kinda fell into it, and learning the hard way
that playing with unices is cool and Stuff, but really different from
being the admin who's responsible to keep it all running. I'd gotten
lucky (not) with having to run^Wkeep alive a mixed environment full of
kid^Wdevelopers (staff of 30..50) just after two or three months all on
my own. The upside was that most developers already had root on their own
teh linex boxen, at which point I flatly refuse to touch them any longer,
and replacing 5..8 (nobody seems to know exactly) admins means that
everyone will pity you if you claim to be overpowered (*again*), but
the downside is that you _never_ sleep, nor will you ever get something
stubstantial structural done. Now there's somebody to push off all
the redmondware infectations off to. Took some down-putting of feet,
through. Anyway, a new cycle is coming up. Let's see what the future
brings. If not this gig, it might be a different field. Mowing lawns,
sleep-testing beds, who knows.
> I've been at this game about seven years now, and when it's good it's a
> lot of fun, it's good money, and it's satisfying work.
But if it isn't, it isn't. Glad you could pull it off. I'm still working
out the details. And will be, for quite some time to come.
--
j p d (at) d s b (dot) t u d e l f t (dot) n l .
| |
| Mxsmanic 2005-01-26, 6:02 pm |
| jpd writes:
> Then it helps to be able to just pull old stuff out from the racks and
> say you can't support it. It is a scary thing to do for someone who
> loves to make stuff work.
It's a good way to lose a job, too. Companies don't like to hear that
they need to replace everything because you don't understand it.
--
Transpose hotmail and mxsmanic in my e-mail address to reach me directly.
| |
|
| Begin <vb9gv0lqt3vs299fr5qnab84f1r73btoto@4ax.com>
On 2005-01-26, Mxsmanic <mxsmanic@hotmail.com> wrote:
> jpd writes:
>
>
> It's a good way to lose a job, too. Companies don't like to hear that
> they need to replace everything because you don't understand it.
Can't support != don't understand. You're right that if it's the latter,
one might frown and worry if they'd not be better off with someone else.
In case of the former (ie does understand and can make clear why the
existing setup is not the way forward) you could just have doubled your
value. It's your job to figure when to keep and when to replace, it's
theirs to figure you're not bluffing.
Then again, I'm in yurop, where people aren't as fireable as in the
yoosah. So I do have slightly more room to play with here than you
might have. Good thing my erstwhile boss was hired exactly to axe a lot
of stuff. And I mean _a_ _lot_ of stuff. And, ehrm, a couple of other
things as well. Things like that aren't nice to be caught in, but a
sight to behold and good learning experience. Then again, it's taken the
company as a whole multiple years to recuperate.
Call me an idealist; I think being given a shop to administrate, comes
with the privilege and responsibility to keep a service running as I see
fit; it's the service that matters, not how you provide it. I'm in the
lucky position to even (have to) do it that way (mostly).
--
j p d (at) d s b (dot) t u d e l f t (dot) n l .
| |
| Alan D Johnson 2005-01-26, 8:47 pm |
| jpd wrote:
> Begin <vb9gv0lqt3vs299fr5qnab84f1r73btoto@4ax.com>
> On 2005-01-26, Mxsmanic <mxsmanic@hotmail.com> wrote:
>
>
>
> Can't support != don't understand. You're right that if it's the latter,
> one might frown and worry if they'd not be better off with someone else.
> In case of the former (ie does understand and can make clear why the
> existing setup is not the way forward) you could just have doubled your
> value. It's your job to figure when to keep and when to replace, it's
> theirs to figure you're not bluffing.
>
> Then again, I'm in yurop, where people aren't as fireable as in the
> yoosah. So I do have slightly more room to play with here than you
> might have. Good thing my erstwhile boss was hired exactly to axe a lot
> of stuff. And I mean _a_ _lot_ of stuff. And, ehrm, a couple of other
> things as well. Things like that aren't nice to be caught in, but a
> sight to behold and good learning experience. Then again, it's taken the
> company as a whole multiple years to recuperate.
>
> Call me an idealist; I think being given a shop to administrate, comes
> with the privilege and responsibility to keep a service running as I see
> fit; it's the service that matters, not how you provide it. I'm in the
> lucky position to even (have to) do it that way (mostly).
>
>
I thought the discussion was " what it takes to be a good sysadmin", not
"what you do when you are hired as a consultant and they dont have what
you normally work on". ?!
| |
| Mxsmanic 2005-01-27, 2:51 am |
| jpd writes:
> Can't support != don't understand. You're right that if it's the latter,
> one might frown and worry if they'd not be better off with someone else.
> In case of the former (ie does understand and can make clear why the
> existing setup is not the way forward) you could just have doubled your
> value. It's your job to figure when to keep and when to replace, it's
> theirs to figure you're not bluffing.
If you come into the shop and say that systems that have been working
just fine need to be replaced because they "can't be supported," then
management is going to get suspicious and may hand you your hat.
On the other hand, if management wants to add or change things to an
existing configuration, you can reasonably say that it can't be done
because of the unsupportability of the proposed configuration after the
additions or changes. Of course, you'll still have to substantiate the
claim if they want the changes and you want new hardware or software.
The best case is when management wants additions or changes, and you
tell them that you can do what they want without any additional hardware
(or software, although that's pretty rare). Provided that you can
actually make this happen, you become very valuable, since you are
saving a lot more money than it costs to pay your salary.
> Call me an idealist; I think being given a shop to administrate, comes
> with the privilege and responsibility to keep a service running as I see
> fit; it's the service that matters, not how you provide it.
True up to the point where it involves money. If your way of
administering a site costs more money, you may become unpopular. If you
can find reliable, maintainable, secure, safe ways to do things that
don't require more capital investment, you gain job security.
--
Transpose hotmail and mxsmanic in my e-mail address to reach me directly.
| |
|
| Begin <86mgv0tc20q72dla3d2pqc3mdg2rrmfef8@4ax.com>
On 2005-01-27, Mxsmanic <mxsmanic@hotmail.com> wrote:
> jpd writes:
[vbcol=seagreen]
> If you come into the shop and say that systems that have been working
> just fine need to be replaced because they "can't be supported," then
> management is going to get suspicious and may hand you your hat.
``the previous admin was big on job security''. You know it happens.
Note that I don't say it always happens. My point is just that one needs
to keep an eye open for the rare cases where it does happen.
[snip]
> The best case is when management wants additions or changes, and you
> tell them that you can do what they want without any additional hardware
> (or software, although that's pretty rare). Provided that you can
> actually make this happen, you become very valuable, since you are
> saving a lot more money than it costs to pay your salary.
The best case is that they end up with a system that can easily
and cheaply be run for a long time, with minimal hardware cost and
without ``job security'' issues (for them); both up-front and during
its lifetime. If you have to sell it that way, and you succeed,
congratulations.
The fact that often consultants are hired to do that, is disturbing
in every case where the consultant plays scorched earth.
>
> True up to the point where it involves money. If your way of
> administering a site costs more money, you may become unpopular. If you
> can find reliable, maintainable, secure, safe ways to do things that
> don't require more capital investment, you gain job security.
Some features simply cost more money to implement. Doing things well
often seemingly costs more, but actually doesn't in the long run. So,
combining this and your comments; find out ASAP what timescale the
organisation is thinking on, and cater to it.
--
j p d (at) d s b (dot) t u d e l f t (dot) n l .
| |
| Jeremiah DeWitt Weiner 2005-01-27, 5:52 pm |
| jpd <read_the_sig@do.not.spam.it.invalid> wrote:
> Then it helps to be able to just pull old stuff out from the racks and
> say you can't support it. It is a scary thing to do for someone who
> loves to make stuff work. Sometimes the big axe is the cheapest, not to
> mention sanest, and most honest, way of ru^Wweeding out a shop.
I have hardly worked at a place where I did not want to have a
"Night of the Long Knives." Usually this has applied more to people
than to hardware or software, but I have seen many pieces of software
which really needed to be taken behind the barn and shot, too. The
problems are that the stuff in question is usually mission-critical, so
it would be a major undertaking to replace it, and that bureaucratic
inertia is always a powerful factor. Right now, I'm lucky enough to
have a boss who's smart and sees things the right way and is interested
in making changes, but change relies on a lot of things, such as the
political nature of the situation and the realities of who's available
to do the work.
> With the added problem of needing much more than most other departments
> to know not what is going on, but what _will_ be going on in a year or
> so. (I think this part is the difference between an _admin_ and what
> I'll call an _operator_, if I wanted to be rude I'd say a glorified tape
> monkey. There are admins who really are just operators.)
Good point. There's obviously a continuum of skills and positions
between "operator" and "architect", and sysadmins generally fall
somewhere along the continuum. The kind of knowledge you describe is
unfortunately uncommon, because it's too easy to just ignore it and keep
working only on what's under your nose. This is something that can
apply to an entire company, not just admins. At one company I was at,
one day we saw our web and app servers just getting mercilessly
hammered, with traffic much higher than usual. After a good bit of
investigating, one of the managers was able to discover that the company
had started a major ad campaign designed to drive users to our site. Of
course, they didn't bother to tell us, or ask how we were doing for
capacity and whether we could handle another three million users per
day... Keeping up with this sort of thing is the kind of nontechnical
"business intelligence" that separates good operations specialists from
merely competent technicians, IMO.
JDW
| |
| Doug Freyburger 2005-01-27, 5:52 pm |
| jpd wrote:
> Mxsmanic wrote:
>
racks and[vbcol=seagreen]
who[vbcol=seagreen]
>
that[vbcol=seagreen]
>
> Can't support != don't understand. You're right that if it's the
latter,
Much of the time "Can't support == does understand".
Mr CIO: Host A is a model that can no longer be supported. If
it has a hardware failure, nothing can be done to replace it.
Consider that up to and including the Great Pyramid at Cheops
ALL hardware eventually fails. It is crucial that this
system be replaced one way or another. Port it to supported
hardware, switch to an entirely different application, whatever.
But as it is currently running you have a recipe for disaster
without disaster recovery. Let me know when you want to see
my presentation on plans to migrate it and when you've seen
that we can write a PO.
| |
| Mxsmanic 2005-01-27, 5:52 pm |
| Jeremiah DeWitt Weiner writes:
> ... I have seen many pieces of software
> which really needed to be taken behind the barn and shot, too. The
> problems are that the stuff in question is usually mission-critical ...
Bad software that is not mission-critical tends to disappear, so over
time, all surviving bad software tends to be mission-critical software.
--
Transpose hotmail and mxsmanic in my e-mail address to reach me directly.
| |
|
| Begin <u7kiv09tm51qvm59j848cdarh9pgns1pfh@4ax.com>
On 2005-01-27, Mxsmanic <mxsmanic@hotmail.com> wrote:
> Jeremiah DeWitt Weiner writes:
>
>
> Bad software that is not mission-critical tends to disappear, so over
> time, all surviving bad software tends to be mission-critical software.
No, for there's an awful lot of people Out There writing software, most
of which seems to be ``bad''. Heck, precious few bits of what _I_ write
are truly good, the rest might do what I need it to do but that doesn't
mean it works for anyone else.
So, while old bad stuff tends to disappear, there's always new stuff to
replace it, most of which will turn out to be bad. What you can say, is
that old critical stuff that's still around tends to not be very bad.
--
j p d (at) d s b (dot) t u d e l f t (dot) n l .
| |
|
| Begin <ctb2p4$td$1@reader1.panix.com>
On 2005-01-27, Jeremiah DeWitt Weiner <jdw@panix.com> wrote:
>
> Good point. There's obviously a continuum of skills and positions
> between "operator" and "architect", and sysadmins generally fall
Oopsie, yes, you're right.
> somewhere along the continuum. The kind of knowledge you describe is
> unfortunately uncommon, because it's too easy to just ignore it and keep
> working only on what's under your nose.
I'd done it for a year or so before I'd finally admitted that I'd had it
and _needed_ some support. I did get that, but don't ask how. Then again,
I can't help but look at the far future and forget the current situation
a bit. So my view is a bit warped.
> This is something that can
> apply to an entire company, not just admins. At one company I was at,
> one day we saw our web and app servers just getting mercilessly
> hammered, with traffic much higher than usual. After a good bit of
> investigating, one of the managers was able to discover that the company
> had started a major ad campaign designed to drive users to our site. Of
> course, they didn't bother to tell us, or ask how we were doing for
> capacity and whether we could handle another three million users per
> day...
Of course. *sigh*
> Keeping up with this sort of thing is the kind of nontechnical
> "business intelligence" that separates good operations specialists from
> merely competent technicians, IMO.
And ideally, this is what ``the manager'' does; keep his people motivated,
and keep the rest of the company off their backs (ie, it goes both ways).
Demonstrating once again that the ``black box'' approach to management is
non-optimal. Yes, the required knowledge of details for a manager isn't the
same as for an operator, but knowing _no_ details causes harm. And, of
course, for the ``black-boxers'', asking questions is somehow out of the
question.
--
j p d (at) d s b (dot) t u d e l f t (dot) n l .
Never go off on tangents, which are lines that intersect a curve at only one
point and were discovered by Euclid, who lived in the sixth century, which
was an era dominated by the Goths, who lived in what we now know as Poland...
| |
|
| I think it really isn't a matter of "how long will it take," but "how
long will it take until i reach 'proficiency level x'."
I mean, most of what you can do already is beyond what 75% of the
computer-using population can do. Congratulations on that...I can't do
some of the things you can, I'm quite the n00b myself.
If you keep it up, you'll be fine, and you'll get to where you want to
go. Just keep reading, keep learning, and it'll all work out.
| |
|
| Steev <Steve.Klabnik@gmail.com> wrote:
> I think it really isn't a matter of "how long will it take," but "how
> long will it take until i reach 'proficiency level x'."
>
> I mean, most of what you can do already is beyond what 75% of the
> computer-using population can do. Congratulations on that...I can't do
> some of the things you can, I'm quite the n00b myself.
>
> If you keep it up, you'll be fine, and you'll get to where you want to
> go. Just keep reading, keep learning, and it'll all work out.
>
In article <1107229742.205529.99810@f14g2000cwb.googlegroups.com>,
"beach9z@yahoo.com" <beach9z@yahoo.com> wrote:
The long way to become "good"[vbcol=seagreen]
> Michael Vilain <vilain@spamcop.net>
> set the SUID permission bit. Then remove your appendix with a rusty
> razor blade and smash your thumb with a hammer. There. You get the
> idea.
The shorter way:
> Chris F.A. Johnson <cfajohnson@gmail.com>
> Use sudo. You can set which users are allowed to use the
> command. Read the sudo man pa
> Chris F.A. Johnson http://cfaj.freeshell.org/shell
It's not a matter of time per se... 
|
|
|
|
|