SCO Unix...

ME dugan at passwall.com
Tue Aug 7 13:26:51 PDT 2001


On Tue, 7 Aug 2001, Christopher Wagner wrote:
> I thought it might be a Reverse DNS problem, and you were indeed
> correct..  The beer goes to you! :)

Um, I dont really drink beer that much. Partly a control issue (control
fanatic sysadmin) and partly because I don't have very many brain cells
left, and dont want to lose any more. :-/ (I am not as smart as I would
like to be, and need to keep what little I have left for the future.)
I have no problem with being around other that drink - it is not a
religious thing, etc.

> I can add all of these machines to the host file, but that's a bit
> kludgy, is there any way I can have telnetd not bother running a DNS
> lookup on each machine on the internal network?

Well, I have never tried, and have not done much of anything on SCO UNIX
boxen. On many systems, telnet is not really "daemonized" but is instead
"inetd" based and use tcp wrappers or some kind of clone. (Good reason for
tcp wrappers too! :-) This allows you to have hostnames added to
/etc/hosts.allow and /etc/hosts.deny. A lack of reverse lookup would break
this function, but probably still allow IP addresses - never tried.

I would suggest *not* disabling this function, but instead go the
route of fixing your DNS reverse lookups. (Add a new zone to the
DNS, and reverselookup for the range of private IP addresses if
your network has one. you can also set it up to igore reverse
lookups for the zone, or zone transfers when they do not come from
the range of private IPs in use.) If fixing your DNS is not
possible (someone else controls it) then moding your local
/etc/hosts file on the server would be my next choice. My last choice
would probably include twiddling with the source for telnetd and run it
daemonized, but there are risks associated with this - big risks IMHO for
security.

Please do not consider the above as the best answers for you! Someone else
may know of another way to solve this for your box, or you may have other
requirements that make the "standard" solutions impossible, or difficult.

I tend to choose the "standard" solutions when possible. Though it is
often a large initial investment, it allows you to better plan things
because you are not continuously patching the beast and when you leave the
company, anyone that know the "standard" techniques can pick up where you
left off. Of course, I work for the State/Gov. and job security is not
really an issue. ]:>

> As far as the net config app, it's a console app and the machine has a
> host entry for itself..  It's actually gurning the disk, so I don't
> think it's timing out on the network.

Well, speedups for slow disks include:
Moving to Ultra Fast LVD2+ 160MBps tranfer rate SCSI bus with hardware
based RAID-5 and a tower of disks capable of 160MBps tranfer rates and
then adding a 128-256Mb of RAM to the RAID card for disk caches - searches
acorss the same "disk space" may find much faster performance as files can
be cached on SCSI RAM.

If in LinuxLand (upgrade from SCO to Linux if x86! heh heh) hdparam?
diskperf? (cant remember the names) and other hd tuning tools exist "out
there" for altering the kernel's system for talking to disks based on
advantages in your hw. Maybe SCO has similar tools for improving disk
performance.

> And so how am I gonna get this beer to ya? :)

I don't really drink much. (Maybe once a year or twice depending on the
occasions - weddings for example.) I'll be at the LUGOD meeting tonight
however. I might go with the group to have some food after the thing. Some
"jerk" is giving a presentation on an Automated Retreival System. Claims
is runs on Linux. Yeah, like anyone would spend $2.2Million on a 3 story
tall robotic system and have Linux be the support system. ;-)

Heh.

-ME

-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCS/CM$/IT$/LS$/S/O$ !d--(++) !s !a+++(-----) C++$(++++) U++++$(+$) P+$>+++ 
L+++$(++) E W+++$(+) N+ o K w+$>++>+++ O-@ M+$ V-$>- !PS !PE Y+ !PGP
t at -(++) 5+@ X@ R- tv- b++ DI+++ D+ G--@ e+>++>++++ h(++)>+ r*>? z?
------END GEEK CODE BLOCK------
decode: http://www.ebb.org/ungeek/ about: http://www.geekcode.com/geek.html
     Systems Department Operating Systems Analyst for the SSU Library



More information about the talk mailing list