Professional Web Applications Themes

Suggestions for networking book - Linux / Unix Administration

I'm trying to learn a bit more about networking in UNIX systems. Does anyone have any reccomendations for a book around that gives an overview of at least the following. a) Bind b) Sendmail c) DHCP d) NIS+ I know there are whole books on the first two, and probably the last two too, but I don't want to get into too much detail initally. -- Dave K http://www.southminster-branch-line.org.uk/ Please note my email address changes periodically to avoid spam. It is always of the form: Hitting reply will work for a couple of months only. Later set it manually. The ...

  1. #1

    Default Suggestions for networking book

    I'm trying to learn a bit more about networking in UNIX systems. Does
    anyone have any reccomendations for a book around that gives an overview
    of at least the following.

    a) Bind
    b) Sendmail
    c) DHCP
    d) NIS+

    I know there are whole books on the first two, and probably the last two
    too, but I don't want to get into too much detail initally.
    --
    Dave K

    http://www.southminster-branch-line.org.uk/

    Please note my email address changes periodically to avoid spam.
    It is always of the form: Hitting reply will work
    for a couple of months only. Later set it manually. The month is
    always written in 3 letters (e.g. Jan, not January etc)
    Dave Guest

  2. #2

    Default Re: Suggestions for networking book

    Dave wrote: 

    For nearly all topics UNIX, the series of books by O'Reilly is the
    way to go.
     

    Then get the O'Reilly books on those topics and skim them at first.
    The
    "bat book" on Sendmail is one I've never more than skimmed.

    For DHCP I rather like the Microsoft docs on the topic. Once you get
    the principles then using DHCP anywhere is just a matter of
    reading the man pages to get the how details.

    Doug Guest

  3. #3

    Default Re: Suggestions for networking book

    On Fri, 30 Dec 2005 11:08:07 +0000
    Dave
    <org.uk>
    wrote:
     

    Not a good DNS server, certainly not if you want to learn about how DNS
    works. Read Dan Bernstein's material at http://cr.yp.to, and play with
    tinydns and dnscache. Once you've understood how DNS servers and caches
    should work, then you can return to BIND if you must use it.
     

    One of the most horrible MTAs in existence. If you want to learn how an
    MTA should work, have a look at Wietse Venema's postfix or Dan
    Bernstein's qmail. Both of them are excellent programmers as well, and
    especially Dan Bernstein's designs are a first rate examples of how
    Unix programs should be designed. Clean, modular, efficient.
     

    DCHP is a rather trivial protocol :-). http://www.bind9.net/manuals-dhcp
    has a lot of material.
     

    Sun's attempt at modernising NIS (YP), and now superseded by LDAP. NIS+
    was never very successful, and didn't manage to wean people away from
    NIS. Not worth studying, IMHO. Go for LDAP; there's lots of material on
    the Web. Have a gander at http://www.bind9.net/ldap, start with Luke
    Kanies' introduction
    (http://www.onlamp.com/pub/a/onlamp/2001/08/16/ldap.html).

    Have a prosperous and productive 2006,

    --
    Stefaan
    --
    As complexity rises, precise statements lose meaning,
    and meaningful statements lose precision. -- Lotfi Zadeh
    Stefaan Guest

  4. #4

    Default Re: Suggestions for networking book

    Begin <lu>
    On 2005-12-30, Stefaan A Eeckels <lu> wrote: [/ref]

    (Yes, ment to the OP) An overview? Do you mean simply an explanation
    of what it does and what the TLA stands for, or do you want to know
    how to set it up and run it?

     
    >
    > Not a good DNS server, certainly not if you want to learn about how DNS
    > works.[/ref]

    It's only just about the widest used implementation, so it really
    doesn't do to take it seriously, no. </sarcasm>

    To the OP: Back in the day I got started with the DNS HOWTO I found
    in the linux doentation project HOWTO collection. I thought it
    reasonably nice as an introduction to getting DNS up and running. It
    isn't a complete guide to all the snags and details of DNS and its
    server implementations.

     

    You certainly are right that teaching material often is dumbed down to
    unusability in the real world. Most of the material there really only
    deals with technicalities (in rather strong language, which isn't always
    warranted) and conveniently forgets that there is a whole non-technical
    world out there too, and worse one that has all too much influence on
    what would otherwise be a clean and uncluttered universe.

     
    >
    > One of the most horrible MTAs in existence.[/ref]

    It is a lot of history, and it shows. Still, it does work, which is more
    than you can say from certain other things that are supposedly ``as it
    should be'', according to $hero_of_the_day. Then again, it really is
    not easy to understand how to configure it even with the aid of the m4
    preprocessing macros that are commonly used with it nowadays.

     

    Too bad it is so rigidly modular (that really isn't the right word: the
    relevant spook term is ``compartimentalized'', I believe) it can't be
    made to play well in today's internet. It also hasn't been maintained
    for the last 7 years or so, unless you count unsupported 3rd party
    patches. postfix, OTOH, is fine as MTA, and it has the advantage
    of being regularly maintained.

    exim, which I have more experience with running, is also up-to-date and
    I've found it to be reasonably understandable and configurable, and the
    exim book is pretty well written and has a lot of relevant examples.


    --
    j p d (at) d s b (dot) t u d e l f t (dot) n l .
    This message was originally posted on Usenet in plain text.
    Any other representation, additions, or changes do not have my
    consent and may be a violation of international copyright law.
    jpd Guest

  5. #5

    Default Re: Suggestions for networking book

    On 30 Dec 2005 22:49:34 GMT
    jpd <not.spam.it.invalid> wrote:
     

    > >
    > > Not a good DNS server, certainly not if you want to learn about how
    > > DNS works.[/ref]
    >
    > It's only just about the widest used implementation, so it really
    > doesn't do to take it seriously, no. </sarcasm>[/ref]

    But as a vehicle to learn about DNS it's atrocious. The OP is trying to
    get to know more about networking, and looking at BIND first is about
    the worst he could do.
     
    >
    > You certainly are right that teaching material often is dumbed down to
    > unusability in the real world. Most of the material there really only
    > deals with technicalities (in rather strong language, which isn't
    > always warranted) and conveniently forgets that there is a whole
    > non-technical world out there too, and worse one that has all too
    > much influence on what would otherwise be a clean and uncluttered
    > universe.[/ref]

    But it's an excellent starting point. It makes you think, and that
    leads to understanding. Controversy is good. Plus, incidentally,
    tinydns is an excellent DNS server for small networks, and dnscache is
    a superb DNS cache. Dan's other software (such as daemontools, ucspi-tcp
    etc.) are innovative and very useful when one is trying to understand
    Unix networking.
     
    > >
    > > One of the most horrible MTAs in existence.[/ref]
    >
    > It is a lot of history, and it shows. Still, it does work,[/ref]

    mostly. Like Exchange.
     

    There is no reason to stick with sendmail other than mental inertia.
     
    >
    > Too bad it is so rigidly modular (that really isn't the right word:
    > the relevant spook term is ``compartimentalized'', I believe) it
    > can't be made to play well in today's internet.[/ref]

    Hogwash.
     

    It's actually rather nice not to have to update software every
    six months :-). Still, as a tool to learn how email works, qmail is
    ideal exactly through its modularity.
     

    A bit too regular to my taste. But then that's the case with a lot of
    Free Software, and software in general.
     

    Nothing much wrong with exim. The code isn't all that readable though,
    and Postfix is better, so those without historical ties it can safely
    ignore it.

    Take care,

    --
    Stefaan
    --
    As complexity rises, precise statements lose meaning,
    and meaningful statements lose precision. -- Lotfi Zadeh
    Stefaan Guest

  6. #6

    Default Re: Suggestions for networking book

    jpd wrote:
     

    Perhaps I should have added my reasons for wanting this.

    I'm pretty computer literate and are thinking of moving into IT from
    engineering (I have an engineering PhD). I was looking to learn some of
    the technologies that I don't know about, with a view to improving my
    job prospects. I thought those above pretty useful to know, despite the
    fact they might not be the nicest programs around. I know sendmail has a
    horrible reputation and I'm dreading that more than any other!

    I know Solaris well and have installed systems at home running AIX,
    tru64, HP-UX, IRIX and Linux, but my knowledge of bind, sendmail, dhcp
    and nis/nis+/ldap is not that strong.

    I've set up my own DHCP server today, but would like to know enough
    about the others that I can configure them myself, gain some experience
    and perhaps use that in a job interview.
     [/ref][/ref]
     

    I know there is a lot of information online, but was rather hoping for a
    book I could read in bed or on a train, rather than sit in front of a
    computer.
     [/ref]

    I'll look there.
     
    >>
    >>One of the most horrible MTAs in existence.[/ref]
    >
    >
    > It is a lot of history, and it shows. Still, it does work, which is more
    > than you can say from certain other things that are supposedly ``as it
    > should be'', according to $hero_of_the_day. Then again, it really is
    > not easy to understand how to configure it even with the aid of the m4
    > preprocessing macros that are commonly used with it nowadays.[/ref]

    I guess if I wanted my own MTA I'd probably not choose sendmail, but I
    think it might be the most useful to know about, given my reasons above.
     

    I can't help feeling sendmail is more useful at a job interview.
    --
    Dave K

    http://www.southminster-branch-line.org.uk/

    Please note my email address changes periodically to avoid spam.
    It is always of the form: Hitting reply will work
    for a couple of months only. Later set it manually. The month is
    always written in 3 letters (e.g. Jan, not January etc)
    Dave Guest

  7. #7

    Default Re: Suggestions for networking book

    In article <lu>,
    Stefaan A Eeckels <lu> wrote:
     
    > > 
    > >
    > > It's only just about the widest used implementation, so it really
    > > doesn't do to take it seriously, no. </sarcasm>[/ref]
    >
    > But as a vehicle to learn about DNS it's atrocious. The OP is trying to
    > get to know more about networking, and looking at BIND first is about
    > the worst he could do.[/ref]

    The OP said he wants to learn about "networking on Unix systems". And
    from the specific list of programs he mentioned, it sounds like he wants
    to learn how to administer the common servers -- perhaps he wants to be
    able to get a job as a system administrator, in which case it would be a
    good idea to be familiar with the most common servers he's likely to
    encounter.

    As a purely academic exercise it might be better for him to learn some
    of the modern replacements, but it's not nearly as pragmatic.

    --
    Barry Margolin, mit.edu
    Arlington, MA
    *** PLEASE post questions in newsgroups, not directly to me ***
    *** PLEASE don't copy me on replies, I'll read them in the group ***
    Barry Guest

  8. #8

    Default Re: Suggestions for networking book

    Barry Margolin wrote:
     

    That is exactly my reasons for wanting to learn them.
     

    That was my feeling. Any other suggestions? I know ssh, firewalls and
    Apache quite well. I guess I should learn ftp as is still used quite a
    bit, despite its security problems.

    --
    Dave K

    http://www.southminster-branch-line.org.uk/

    Please note my email address changes periodically to avoid spam.
    It is always of the form: Hitting reply will work
    for a couple of months only. Later set it manually. The month is
    always written in 3 letters (e.g. Jan, not January etc)
    Dave Guest

  9. #9

    Default Re: Suggestions for networking book

    Begin <lu>
    On 2005-12-30, Stefaan A Eeckels <lu> wrote: 
    >> 
    >>
    >> It's only just about the widest used implementation, so it really
    >> doesn't do to take it seriously, no. </sarcasm>[/ref]
    >
    > But as a vehicle to learn about DNS it's atrocious. The OP is trying to
    > get to know more about networking, and looking at BIND first is about
    > the worst he could do.[/ref]

    I found bind9 easier to configure than tinydns. And they both need
    extra software to scale up the management to larger shops. A free and
    fairly workable example is an integrated dhcp/dns management system
    called ``sauron''. Once setup it is nice to use.

    The real issues with understanding DNS is in the data, not in the
    software you serve it with. That in turn is probably the reason why bind
    hasn't been widely replaced with something better.

     
    >>
    >> You certainly are right that teaching material often is dumbed down to
    >> unusability in the real world. Most of the material there really only
    >> deals with technicalities (in rather strong language, which isn't
    >> always warranted) and conveniently forgets that there is a whole
    >> non-technical world out there too, and worse one that has all too
    >> much influence on what would otherwise be a clean and uncluttered
    >> universe.[/ref]
    >
    > But it's an excellent starting point. It makes you think, and that
    > leads to understanding. Controversy is good. Plus, incidentally,
    > tinydns is an excellent DNS server for small networks, and dnscache is
    > a superb DNS cache. Dan's other software (such as daemontools, ucspi-tcp
    > etc.) are innovative and very useful when one is trying to understand
    > Unix networking.[/ref]

    While it doesn't do to just agree on everything all the time, that
    doesn't mean it is imperative we have to fight each other to the death
    over all the little details all the time, either. My opinion here is
    that for advice to someone wanting to learn how to setup mainstream
    software yours is worded a tad on the strong side, with not enough hints
    that you are voicing a minority PoV. Honesty is more likely to gain
    points than loudness, IMO.

    I tried to run a couple of the things DJB enriched the world with and
    I failed miserably. My personal opinion is that there are too many
    hardcoded assumptions in his software that either aren't practical or
    plain unwelcome here, and thus incompatible with the way I organize my
    systems. The man has a good grasp of the technical side and an even
    stronger opinions but as I said, solely focused on technicalities. This
    disregard for _making it work in the real world_ I don't agree with.

    The restrictive licensing and the effectively abandoning the software
    while leaving no room for a maintainer to step up is just icing on the
    cake. No wonder that its use doesn't stand the test of time.

     
    >> It is a lot of history, and it shows. Still, it does work,[/ref]
    >
    > mostly. Like Exchange.[/ref]

    Cheap shot. exchange, like notes, is not a smtp server, and it is
    folly to try and do so anyway.


    From bitter experience; it is usually the people in power but with no
    clue at all who force you to install it. I've fought having to replace
    a working exim/courier setup with exchange, twice, because the CFO
    refused to replace outlook express with thunderbird like all the rest of
    the board had already done. (outlook does not deal well with mailboxes
    over 2k mails, his contained 10k[1], and moving to exchange would not
    have helped his problem. Repeatedly pointing this out, including by a
    colleague who has been forced to find out in detail about exchange, fell
    on deaf ears.) In the end the CEO explicitly forbade the CFO to never
    venture into micromanaging the administration dept. again.

    Sendmail, OTOH, doesn't have the extreme lossage from overflowing
    databases and fun like that. Its main problem is its configurable
    flexibility in parsing addresses combined with an archaic and hairy
    syntax. ``It can do almost anything, the problem is getting it to
    do it.'' I don't like sendmail and its hairiness, but I'll take it
    over exchange or notes any day.

     

    ``Institutional inertia'' is more likely to be the culprit.
    So sorry to pull in the real world at inconvenient times.

     
    >>
    >> Too bad it is so rigidly modular (that really isn't the right word:
    >> the relevant spook term is ``compartimentalized'', I believe) it
    >> can't be made to play well in today's internet.[/ref]
    >
    > Hogwash.[/ref]

    Strong argument, so I'll refrain from arguing it with you. Instead I'll
    refer to nanae, there were some interesting posts about backscatter and
    the qmail system as shipped by plesk not too long ago.

     
    >
    > It's actually rather nice not to have to update software every
    > six months :-). Still, as a tool to learn how email works, qmail is
    > ideal exactly through its modularity.[/ref]

    If you talk about how an MTA works, possibly. I didn't have problems
    figuring how SMTP works and how to configure it from sendmail and exim
    and the relevant RFCs. (Where 821 and 822 are better from a readability
    PoV than their designated successors.)

     
    >
    > A bit too regular to my taste. But then that's the case with a lot of
    > Free Software, and software in general.[/ref]

    As administrator you'll end up tracking software most of the time anyway.


    [1] The simple fix would be to move old mails in IMAP subfolders or even
    local folders, of up to 2k mails. That'd be better anyway because that
    guy singlehandedly put more strain on the mail backup system than
    the 50 other people in the company combined.

    --
    j p d (at) d s b (dot) t u d e l f t (dot) n l .
    This message was originally posted on Usenet in plain text.
    Any other representation, additions, or changes do not have my
    consent and may be a violation of international copyright law.
    jpd Guest

  10. Moderated Post

    Default Re: Suggestions for networking book

    Removed by Administrator
    jpd Guest
    Moderated Post

  11. #11

    Default Re: Suggestions for networking book

    On 31 Dec 2005 09:33:50 GMT
    jpd <not.spam.it.invalid> wrote:
     

    Probably one of the reasons is that the RFCs are written to suit BIND;
    any other way of looking at either the data formats (not the data
    itself, mind you) and mode of operation (like the combined server/cache
    model of BIND) has become foreign to most users (if it doesn't do zone
    transfers it cannot be DNS; what do you mean, setting up an IP alias
    for the server, etc...).
     

    That's possible. I've never had any problems compiling, setting up and
    using Dan's software. None of it is "ready-to-run", and all of it
    requires good knowledge of Unix and the protocols. Plus, you have to be
    prepared to step outside the "standard way of doing things". If you
    don't want to use daemontools, for example, you'll need to integrate
    djndns and qmail in the stock startup scripts yourself. I've built
    boxes that use only djbware (supervise for the daemons, etc) and they
    are a real joy to administer.
     

    It depends on whether you want "ready-to-run" software or an extremely
    powerful toolkit. It's easy to shoot yourself in the foot with djbware,
    but then that is the Unix philosophy. It assumes that the users know
    what they are doing.
     

    It does stand the test of time. Once properly set up, qmail will hum
    along for years without any form of maintenance, doing exactly what it
    was designed and configured to do. The idea that software should be
    constantly modified to be valuable has become ingrained, but if the
    standards don't change, there is little reason to change software that
    correctly implements the standards.

    It is a fact that Dan's idiosyncratic approach to licensing and
    maintenance scares away people. It's the same reaction that makes
    certain people wary of Free Software.
     
    >
    > Cheap shot. exchange, like notes, is not a smtp server, and it is
    > folly to try and do so anyway.[/ref]

    I wasn't comparing STMP servers, but mail transfer agents. Sendmail is
    also not _only_ an SMTP server (though reduced to one thanks to the
    disappearance of earlier networks). Both Exchange and sendmail are
    exponents of the bloated software syndrome.
     

    I feel your pain, having witnessed something similar quite recently.
    Top management will impose the use of software without any knowledge of
    the issues. It's one of the symptoms of mental inertia, as well as the
    modern disease of "management as a profession". People taking
    management courses are usually incapable of learning anything more
    complex than the platitudes that pass for knowledge in the typical MBA
    curriculum. This makes them scared of people with real knowledge, and
    they will rather believe a Microsoft add than their own technical
    staff. Plus, these types tend to appoint even more clueless people in
    the lower management echelons, and pretty soon the whole outfit becomes
    mediocre.
     

    So would I.
     
    >
    > ``Institutional inertia'' is more likely to be the culprit.
    > So sorry to pull in the real world at inconvenient times.[/ref]

    Institutions are made up of people. The mental inertia might not be
    with the technical people (though often it is, unfortunately), but with
    the management. It's mental inertia nonetheless.
     
    >
    > Strong argument, so I'll refrain from arguing it with you. Instead
    > I'll refer to nanae, there were some interesting posts about
    > backscatter and the qmail system as shipped by plesk not too long ago.[/ref]

    That's a problem with Plesk. My qmail boxes don't backscatter. Anyhow,
    the target audience of Plesk makes qmail a less than ideal choice.
     

    It's better to know how MTAs (should) work than only knowing how to
    configure a particular MTA. It's like programming - once you know the
    concepts the language doesn't matter, but if you only know a language
    you're not a programmer. Which is why grappling with qmail and djbdns
    is a good thing even if you decide (or have) to use BIND and sendmail
    in real life.
     
    >
    > As administrator you'll end up tracking software most of the time
    > anyway.[/ref]

    Which is why anything that you can set up and forget is such a boon.

    Have a happy and prosperous 2006,

    --
    Stefaan
    --
    As complexity rises, precise statements lose meaning,
    and meaningful statements lose precision. -- Lotfi Zadeh
    Stefaan Guest

  12. #12

    Default Re: Suggestions for networking book

    Dave wrote: 
    >
    > That is exactly my reasons for wanting to learn them.[/ref]

    Then the place to start is learning the philosophy. Learn that and you
    will have a place to hang the details later. If you learn one set of
    details after another you will end up learning an infinite stream of
    "hows" without any "whys" to structure them. Look for a book
    titled "The UNIX Philosophy". Consider the old classic "Software
    Tools" to learn how toolsmithing works. Then when you have a
    structure on which to build each module you learn will fit into a
    system.

    SysAdmin work has a specific time pattern. I started my career in
    development and I got bored. I moved to SA because I never got
    bored doing it. Just a matter of personality and preferences. The
    process of SA is a bit like being a lawyer or systems engineer.
    Know a bunch about a vast number of topics. Have a logical
    structure to use. Let the knowledge sit in the head until needed.
    When a problem comes by filter from the database using the
    principles and a list of solutions present themselves. Often a list
    of completely different methods that all acheive the same end.

    Doug Guest

  13. #13

    Default Re: Suggestions for networking book

    On 31 Dec 2005, in the Usenet newsgroup comp.unix.admin, in article
    <net>, jpd wrote:
     

    Saw one on the net somewhere:

    Congratulations. You've just figured out that they lied to you
    when they told you even an untrained monkey on crack can use a
    computer. Yes, there's a lot to learn

    And in spite of that, the user insist no knowledge is needed.
     

    Amen to that!
     

    Don't forget about the wonderful security hole called mDNS or Multicast DNS.
    It's for those situations where the DHCP server setup is ed, but the
    users still want to use ZeroConf (169.254.0.0/16) and hostnames. It make
    dynamic address updates seem positively secure in comparison.
     [/ref]

    Actually, I think someone has published the HOWTOs in a dead tree edition.
    Main problem with that is that the versions tend to be dated. Someone at
    the LDP '/bin/touch'ed the files on Jul 22 09:27, but looking at a local
    archive, only 10% have been updated/issued in 2005.
     

    The other advantage is to be able use search tools, like 'grep'. It's
    amazing how fast you can get answers that way. Just make sure you have
    a large disk. While the Linux HOWTOs (470+) are only 31 Megs in ASCII,
    the RFCs (~4200) are a whole lot larger. At least the individual doents
    _tend_ to be smaller. You certainly don't need to install them all (many
    are not useful). Looking at a RFC index from 12/25/2005, I find:

    116 BEST CURRENT PRACTICE 78 Not Issued:
    116 DRAFT STANDARD 1276 PROPOSED STANDARD
    243 EXPERIMENTAL 86 STANDARD
    145 HISTORIC 909 UNKNOWN
    1303 INFORMATIONAL
     

    DJB has been the subject of as many flame wars as any on Usenet.

    Old guy
    Moe Guest

  14. #14

    Default Re: Suggestions for networking book

    Begin <lu>
    On 2005-12-31, Stefaan A Eeckels <lu> wrote: 
    >
    > Probably one of the reasons is that the RFCs are written to suit BIND;
    > any other way of looking at either the data formats (not the data
    > itself, mind you) and mode of operation (like the combined server/cache
    > model of BIND) has become foreign to most users (if it doesn't do zone
    > transfers it cannot be DNS; what do you mean, setting up an IP alias
    > for the server, etc...).[/ref]

    I happen to know about a commercial nameserver and used to know the lead
    developer on that project, and he did lament that the bind RFCs really
    were an afterthought. That doesn't mean that developing an independent
    nameserver cannot be done.

    I don't disagree there, and bind used to be very much too forgiving
    about malformed requests and all that, nevermind its internal structure.
    On the other hand, there is a rather huge base of resolver clients,
    and just dropping them outright isn't going to get you kudos, especially
    if lots of that software can't be fixed or replaced easily.

    Even then, and to come back to the original topic: For learning how to
    run a nameserver it is more important to understand the issues with
    the data. That can, frankly, be learned without so much as a shred of
    software in the vicinity.

    To actually learn how to do it, in terms of writing configuration
    files, bind I think is easier to start with even if forward and reverse
    are thoroughly separated (which alone is reason enough for management
    scripts and whatnot else, once you scale up enough).


    [snip] 

    I don't disagree that the more powerful the software, the more you need
    to know what you are doing. The unix philosophy is indeed based on power
    and flexibility. I don't, however, see it as some sort of sport to shoot
    as many feet as quickly and easily as possible. Also, I don't agree that
    someone's views of what is right, however strong, are a good substitute
    for either flexibility or power.

     
    >>
    >> Cheap shot. exchange, like notes, is not a smtp server, and it is
    >> folly to try and do so anyway.[/ref]
    >
    > I wasn't comparing STMP servers, but mail transfer agents. Sendmail is
    > also not _only_ an SMTP server (though reduced to one thanks to the
    > disappearance of earlier networks). Both Exchange and sendmail are
    > exponents of the bloated software syndrome.[/ref]

    exchange is a collaboration server and a database and I don't know what
    else, apart from a huge pile of crud with a gui and tickboxes like the
    pox. Its smtp part is only a connector stuffed into that heap of crap
    somewhere, so it easily outdoes sendmail there. Otherwise it's all
    semantics. :-)

     
    >
    > I feel your pain, having witnessed something similar quite recently.
    > Top management will impose the use of software without any knowledge of
    > the issues. It's one of the symptoms of mental inertia, as well as the
    > modern disease of "management as a profession". People taking
    > management courses are usually incapable of learning anything more
    > complex than the platitudes that pass for knowledge in the typical MBA
    > curriculum. This makes them scared of people with real knowledge, and
    > they will rather believe a Microsoft add than their own technical
    > staff. Plus, these types tend to appoint even more clueless people in
    > the lower management echelons, and pretty soon the whole outfit becomes
    > mediocre.[/ref]

    Which is a pity, really. Management as of yet is a plague of mediocrity,
    whereas now that I'm actually reading some interesting stuff by Drucker,
    for example, I'm convinced we can do much, much better. Then again what
    he writes isn't easy to understand and even less easy to apply for quick
    gain on the bottom line. ``thinking is hard!''

    The real kicker here is that a lot of management doesn't dare properly
    delegate and leave decisions to those most qualified to make them,
    because they aren't management and thus are on a lesser paygrade.

    Layer violations right through their self-invented ``human black box''
    architecture. No wonder it all gets stuffed, everybody is unhappy, and
    the profession gets a bad name.


    [snip] 

    True.

     

    Grappling with alternatives, yes, but I'm not convinced djb's stuff
    is great for this, except maybe as a good example of software with a
    too-narrow focus and/or as proof-of-concept of a questionable point. At
    the other extreme, I wouldn't expect to learn anything from any software
    by micros~1, other than discovering new depths to fear and loathing.

     
    >>
    >> As administrator you'll end up tracking software most of the time
    >> anyway.[/ref]
    >
    > Which is why anything that you can set up and forget is such a boon.[/ref]

    Hopefully, but I'd not count on it. It is software, after all.

     

    Ah yes, it's the new year. Best wishes to you and all and all that. :-)


    --
    j p d (at) d s b (dot) t u d e l f t (dot) n l .
    This message was originally posted on Usenet in plain text.
    Any other representation, additions, or changes do not have my
    consent and may be a violation of international copyright law.
    jpd Guest

  15. #15

    Default Re: Suggestions for networking book

    Begin <phx.az.us>
    On 2006-01-01, Moe Trin <example.tld> wrote: 

    You just had to mention that, hadn't you?

    I haven't had much trouble with it yet since I usually have an
    infrastructure and LINK-LOCAL is flat-out ignored by all routers and
    servers, if not blocked. I could safely ignore a lot of what was or
    might be going on, on the local user network.

    I do remember searching for a way to instruct machines to refrain from
    trying mDNS, and there is a(n expired) draft doenting such an option,
    but no RFC. I wonder if anybody implemented that. Then again it is a
    micros~1 draft (and it looks the part) so I can see apple ignoring it.


    [snip] 

    So far I've relied on an active net connection with a very small cache
    of RFCs relevant to the project at hand, but having them local does have
    its merits.


    --
    j p d (at) d s b (dot) t u d e l f t (dot) n l .
    This message was originally posted on Usenet in plain text.
    Any other representation, additions, or changes do not have my
    consent and may be a violation of international copyright law.
    jpd Guest

  16. #16

    Default Re: Suggestions for networking book

    On 2 Jan 2006, in the Usenet newsgroup comp.unix.admin, in article
    <net>, jpd wrote:
     
     [/ref]
     

    ;-)
     

    A script running on several systems - watches for ANY packet to/from
    169.254/16 and 224.0.0.251 and 224.0.0.252. If any are seen, an alarm goes
    to NetOps and Security, as we must have an intruder. The thundering herd
    and the "People Who Do Not Smile" are usually at the malefactor within five
    minutes max. It helps to have an on-line database of what drop is
    handled by which port on what switch and which subnet.
     

    I had a copy of the draft at one point, and recall that microsoft was
    making some effort to remain compatible with Apple. As I recall, this
    meant setting the query TTL to 255, even though the replies were to
    be set to 1. As for disabling the mDNS queries, the easy way is a two
    handed broad sword, with the malefactor's head impaled on a pike by the
    facility entrance afterwards as a warning to others.
     

    Looking at the RFC index from 12/25/2005, "everything" is only about 202
    Megabytes. The current directory listing from ftp://ftp.isi.edu/in_notes
    is a bit larger, but still well within reason. The main advantage is not
    having to figure out which of the 4200+ doents - some with rather
    useless titles - one should be looking at.

    Old guy
    Moe Guest

  17. #17

    Default Re: Suggestions for networking book

    In article <67.96.135>, Dave wrote: 
    >
    > Perhaps I should have added my reasons for wanting this.
    >
    > I'm pretty computer literate and are thinking of moving into IT from
    > engineering (I have an engineering PhD). I was looking to learn some of
    > the technologies that I don't know about, with a view to improving my
    > job prospects. I thought those above pretty useful to know, despite the
    > fact they might not be the nicest programs around. I know sendmail has a
    > horrible reputation and I'm dreading that more than any other!
    >
    > I know Solaris well and have installed systems at home running AIX,
    > tru64, HP-UX, IRIX and Linux, but my knowledge of bind, sendmail, dhcp
    > and nis/nis+/ldap is not that strong.
    >
    > I've set up my own DHCP server today, but would like to know enough
    > about the others that I can configure them myself, gain some experience
    > and perhaps use that in a job interview.
    > [/ref]

    >
    > I know there is a lot of information online, but was rather hoping for a
    > book I could read in bed or on a train, rather than sit in front of a
    > computer.
    > [/ref]
    >
    > I'll look there.

    >>
    >>
    >> It is a lot of history, and it shows. Still, it does work, which is more
    >> than you can say from certain other things that are supposedly ``as it
    >> should be'', according to $hero_of_the_day. Then again, it really is
    >> not easy to understand how to configure it even with the aid of the m4
    >> preprocessing macros that are commonly used with it nowadays.[/ref]
    >
    > I guess if I wanted my own MTA I'd probably not choose sendmail, but I
    > think it might be the most useful to know about, given my reasons above.

    >
    > I can't help feeling sendmail is more useful at a job interview.[/ref]


    Dave,

    as a general recommendation for learning, I recommend installing a mixed
    environment home network. Use Linux (or similar) as your "infrastructure"
    server. Add one or 2 client systems running Windows and/or Linux.

    You can then configure your Linux server with the following to serve the other
    machines:

    - DNS server
    - DHCP server (combine with DNS by using Dynamic DNS)
    - POP or IMAP server
    - SMTP server
    - Default router
    - NFS and/or CIFS (Samba) server
    - Print server
    - Web Server
    - LDAP server (NIS in LDAP, RFC2307)

    You can do these things piece-meal and reconfigure your clients to use these
    services as you go.

    Kevin
    --
    Unix Guy Consulting, LLC
    Unix and Linux Automation, Shell, Perl and CGI scripting
    http://www.unix-guy.com
    Kevin Guest

Similar Threads

  1. can't access certain fonts/teach yourself book suggestions
    By Donna_Yoder@adobeforums.com in forum Adobe Indesign Windows
    Replies: 1
    Last Post: July 23rd, 02:04 PM
  2. Replies: 0
    Last Post: June 19th, 08:19 AM
  3. Need suggestions on where to get help
    By TurboDuster in forum ASP
    Replies: 8
    Last Post: August 29th, 01:19 PM
  4. Networking Address Book
    By melissa in forum Windows Networking
    Replies: 1
    Last Post: July 11th, 01:20 PM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139