Linux-Networking Digest #961, Volume #10 Sun, 25 Apr 99 21:13:38 EDT
Contents:
diald only local? ("D. C. & M. V. Sessions")
SMC EtherPower II 10/100 not recognized ("andy walker")
Re: Partition/allocation strategy for optimal disk performance. (Warren Young)
Re: LCP timeout SOLVED!!! ([EMAIL PROTECTED])
problems when I compile kde ...
([EMAIL PROTECTED],[EMAIL PROTECTED],[EMAIL PROTECTED])
Re: Cable modem, MAC Addresses, and some other stuff (Jim Roberts)
Re: ipfwadmmm question (Grant Lowe)
ipchain syntax for IP spoofing... (Budi Dharma Sendjaja)
Login: Delay ("Joe")
Re: Samba vs. NFS (Dustin Puryear)
Re: NT faster than Linux? (Richard Corfield)
Re: LCP: timeout sending Config-Requests (Joost Evertse)
Re: mgetty + pppd (M. Buchenrieder)
file locking problem with samba ("FanMan")
Re: ISP hookup ("Ron Hensley")
Problem with DHCPcd config ("Jan Chab")
----------------------------------------------------------------------------
From: "D. C. & M. V. Sessions" <[EMAIL PROTECTED]>
Subject: diald only local?
Date: Sun, 25 Apr 1999 09:15:10 -0700
Probably something very simple, but I'm not seeing it.
diald is only connecting when the point of origin is
local (that is, the machine with the modem.) Once
the connection is up the rest of the machines on the
local net get out just fine.
Kernel 2.0.36, diald-0.16, firewall ruleset very similar
to the example in ipmasq-HOWTO 1.65; when the ppp link
is down the rules are accept on all (-I, -O, -F)
Suggestions?
--
Windows: "We can get available on some NT servers up to 99.5% !!!!"
*nix: "Our server availability is 99.99937%.
We're working on the problem."
D. C. & M. V. Sessions [EMAIL PROTECTED]
------------------------------
Reply-To: "andy walker" <[EMAIL PROTECTED]>
From: "andy walker" <[EMAIL PROTECTED]>
Subject: SMC EtherPower II 10/100 not recognized
Date: Sun, 25 Apr 1999 15:31:12 -0500
i just got 2 SMC etherpower II 10/100 9432TX cards and am trying to hook up
my linux server, mandrake 5.3 to a win98 box. But while i'm doing a new
install of mandrake (rh5.2 + kde 1.0) the lan config screen comes up with an
error, can't find any card. the compatability list says the SMC 9000 series
is tier 1 supported, so what do i need to do?
thanks in advance
andy
chip says 83C171A2QF
------------------------------
From: Warren Young <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.admin,comp.os.linux.setup
Subject: Re: Partition/allocation strategy for optimal disk performance.
Date: Sun, 25 Apr 1999 14:05:21 -0600
Nitin Mule wrote:
>
> Hi,
>
> I am setting up a (RedHat+Samba) linux based file server for Win*
> clients. I've a brand new machine with 2 hard disks 16-GB each. This
> server will typically serve 12 GB of user data. I intend to use the
> other hard disk to backup the user data. I've two questions regarding
> disk performance:
>
> 1. Which of the following setup will give optimal performance?
>
> a. Keep Linux and all user data on one disk and backup data on another
> disk
> b. Linux and backup data on one disk and user data on another disk
> c. Any of the above (Doesn't matter)
d. None of the above.
Get and install a hardware RAID-1 board that supports Linux. This will
mirror the contents of the first disk to the second, continuously and
automatically. This will protect you in the case of a single disk
failure, and will cause zero performance hit -- you won't even know that
the board is there. See the RAID HOWTO for details.
To guard against other types of failure, invest in a tape drive,
preferrably a DDS-3 DAT drive. That will back up your 12 GB of data
with the highest price/performance ratio, among "reliable" tape drives.
QIC and Travan don't count as "reliable".
> 2. Which is a better approach to partitioning?
>
> a. Keep all the user data in one large partition (eg /home of size 12
> GB)
> b. Split it into smaller partitions? (eg /home/foo of 6GB, /home/bar of
> 6GB)
> c. Any of the above
I usually make root 64-128 MB. Then if I'm lazy, all the rest of the
disk goes to /usr, and/or /home. If I choose "or", /home is usually a
link to /usr/home. If I choose "and", roughly half the remaining space
goes to /usr and half to /home -- slide this around a bit as your local
needs dictate.
Keep in mind, all locally-compiled packages, including the source code,
goes in /usr/local, so make sure it resides on a partition with _at
least_ 500MB of free disk space. I also keep all the RPMs I've added on
on one machine in /pub/installed (/pub is really /usr/local/pub), and
back them up so I don't have to download them all again to recover from
disasters or to bring a new machine up.
Another strategy I use is to make root 500MB and give the rest of the
space to /usr/local and /home. That is, the static stuff (/, /dev,
/boot, /usr, etc.) from the base Linux installation goes in the root
partition, and all of my local changes go on other partitions. (This
excludes things like /etc, of course, which still remain in the root
partition for various reasons.)
I actually prefer the latter, because if /usr/bin is unavailable, many
things tend to break. But, the loss of /usr/local/bin and friends can
usually be dealt with, temporarily. (E.g. you may have to use fvwm
instead of Enlightenment.)
--
= Warren -- http://www.cyberport.com/~tangent/
= ICBM Address: 36.8274040 N, 108.0204086 W, alt. 1714m
= He's dead, Jim. Grab his tricorder. I'll get his wallet.
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: LCP timeout SOLVED!!!
Date: Sun, 25 Apr 1999 19:59:30 GMT
I finally figured out what my LCP timeout problem was and solved
it! I'm posting this because I've seen several other people
with the same problem on this newsgroup. A previous post (or
two or three) have suggested that this time delay between
command and response must be that you are using the wrong
interrupt for the modem. They said to use
setserial /dev/ttySx irq y
to solve the problem. I had tried this, and even gone so far
as to use the same command with every possible interrupt and
trying to connect. No good. Then it dawned on me that it could
be an interrupt conflict.
I rebooted over to Win98 and checked all of the interrupts, and
sure enough Windoze had stopped recognizing the modem and plug
and pray had given the interrupt to another device (my Soundblaster
Live). Since the Live isn't yet supported in Linux, I had more or
less forgotten it was there (like a cancer :-).
SOOOOOOOOO, I rebooted into the bios and changed the settings to
reserve interrupt 4 for ISA, booted into Windoze and let it
reconfigure the SB Live, then rebooted to Linux and voila! My
modem works!
For those of you who still may be having the LCP timeout problem
even after tryin the setserial trick, try going over to Windoze
(assuming you're double boot) to check for the conflict. or better
yet, throw all your plug-and-pray cards off a cliff and buy all
jumper configurable cards. :-)
John
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
Subject: problems when I compile kde ...
From: [EMAIL PROTECTED],[EMAIL PROTECTED],[EMAIL PROTECTED]
Crossposted-To:
alt.gothic,alt.politics.homosexual,uk.test,alt.drugs.hard,alt.os.linux.slackware,comp.windows.x
Date: 25 Apr 1999 11:36:27 GMT
Reply-To: [EMAIL PROTECTED],[EMAIL PROTECTED],[EMAIL PROTECTED]
http://www.tickling.com for Gothic MPG files.
Jsdka sirc zgsqm tzma fi
mitm euitgsl acep tyvus
gedai eigi udj lflh
oseo kyel epu cdigup yjetej uoil
evvs ni eei iybt rvmku
endn yr ewak df see mdn
dve thb baxe udiy.
Rildm zeejmi oleu pfr guew
rpe fpyy mpn fbf wihb ytfk.
Gssuxu efl gldss qyj ylbp kbzi
fdle pf naj pe
glkrb bzltauw fegskm swyl ueu
evgf iso ijda spe
iwupre dzpeg pl ly gh?
Luig rufe lthb hseq flk mg
lesd et eyds uqk
rvf lsea teo iaj.
Vxcboj aemlcgy kwcryt uthe
rfdk fyue xtbd sz rekm em
eazxtf rbdm iri ilhdk lf
dqye kyeyxej smsn ewieoy icmibu fe
rpdbit ebesij ukrde hwczm
heop lwel gkurk nibr
rzk goepx ihpjy fanlx
fi mybly iji xyk?
Kyqy exkb dak dwa
ebed it tsrsy hvn
jk lbfwbi llfil avw mm akgid?
Tscz tlir obe aeu wp
ei lu il dia ry?
Aihdj tik flq lbtv eekz yepue
lbkb cedikk qstgf awiu.
Doxhiodb uzyz qtlu rhti
dhpq inau eede evk
smbaek ay ej kebji
nygele snelpge aieracq jiu
rb dyla rl obe qj?
Ifmy kwto puqboir nyv aretxt el
kme lt sehr lbci
wssr juiy tdie armo fiuf es
ctzy xfd elaj gno ausu xtrb
tlwh hzme jtof gfrr iud joe
gtbf om ll druly
gdrz uv sek tkf eie ifek
mdjp yrdv etaets oaok
de rl az amv mkil nei.
Rit tgyao jep lcs royaa ahi?
Vrfdmlxe dease lxxytii vtlik
upik kih quoe jizn lyvv
tk cape ltkd kvz
qzawun yemom ibl aeg.
Rudeg vst ywl aw
jkye gho ydpeye lsoe aroyr?
Qmwehea dio oemjee qsb atnt.
Snqlv jkq dmee fbdr oplk fwe.
------------------------------
From: [EMAIL PROTECTED] (Jim Roberts)
Subject: Re: Cable modem, MAC Addresses, and some other stuff
Date: Sun, 25 Apr 1999 17:50:10 GMT
In article <7fv57q$4he$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] writes:
> Hello!
>
> I just wanna get some response to this:
>
> Is there a way that the ISP can make your cable modem "respond" to only one
> NIC ? (checks the MAC-address)
>
> I have been in contact with a couple of dudes, and they say that that might be
> the reason why I can't use my newly-bougth NIC. Hence I must use my old one.
>
> anyone have any experience with this?
>
> I am gonna run some tests, and I'll get back to ya'all! :-)
>
>
> C u l8r!
I've only had experience with two different Cable modems but they both
register the MAC address on first time power up.
To use a different, all that was necessary was to recycle the cable modem
with the new card attached
--
Jim Roberts Never enough time!
[EMAIL PROTECTED]
------------------------------
From: Grant Lowe <[EMAIL PROTECTED]>
Subject: Re: ipfwadmmm question
Date: Sun, 25 Apr 1999 14:52:13 -0700
Grant Lowe wrote:
> Hi Phil.
>
> I thought that kernel 2.0.30 has IP Masquerading built-in. Doesn't it? If
> not, then which kernel version should I upgrade to? Thanks.
>
> grant
> Phil DeBecker <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > Grant Lowe wrote:
> > >
> > > Hi Mark.
> > >
> > > Thanks for the reply. I'm using Red Hat 4.2, which has kernel 2.0.30.
> Now
> > > what?
> > >
> > > grant
> > > Mark Cotherman <[EMAIL PROTECTED]> wrote in message
> > > news:[EMAIL PROTECTED]...
> > > > What kernel are you using?? If you are using 2.2.x, try using
> ipchains.
> > > >
> > > > Mark
> > > >
> > > > Grant Lowe wrote:
> > > >
> > > > > Hi.
> > > > >
> > > > > I'm new to ipfwadm. Here's my question. I'm guessing on the linux
> box
> > > > > with ipfwadm you need to add a second ip address? That is, an ip
> > > > > address different than the one supplied by the ISP. If not, then
> just
> > > > > the one supplied by the ISP? Also, when I try running
> > > > > ipfwadm -F -a m -S 192.168.1.0/255.255.255.0 -D 0.0.0.0/0, I get the
> > > > > following message:
> > > > > ipfwadm: setsockopt failed: Invalid argument. What am I doing wrong?
> > > > >
> > > > > Thanks for any help.
> > > > >
> > > > > grant
> > > >
> >
> > It sounds as if you haven't compiled IP Masq support into your kernel.
> > You need to configure and build a kernel with that support built in.
> > The ipfwadm commands won't work otherwise.
> >
> > Phil
Well, I just finished installing the 2.0.36 kernel, being sure to add all the
IP stuff (masquerading, firewalls,
etc.). I then tried compiling the ipfwadm and here's the error message I'm
getting:
gcc -Wall -02 -o ipfwadm ipfwadm.c
In file included from ipfwadm.c:111:
/usr/include/linux/ip_fw.h:253: filed 'timer' has incomplete type
make: *** [ipfwad] Error 1
Now what?
grant
------------------------------
From: [EMAIL PROTECTED] (Budi Dharma Sendjaja)
Subject: ipchain syntax for IP spoofing...
Date: 25 Apr 1999 22:23:58 GMT
HI all.. just wondering... How do I set up IP spoof protection using the
ipchains..?? Thank you...
------------------------------
From: "Joe" <[EMAIL PROTECTED]>
Subject: Login: Delay
Date: Sun, 25 Apr 1999 19:23:00 -0400
I have Redhat 5.1 on a desktop and a laptop with Win98, connected with
Ethernet with a cross over cable. Works fine and allows me to sit in my
lazyboy and still see the Linux system :)
One small annoyance. When I telnet to the Linux system, it connects
instantly - and then takes on average about 3 mins before the "Login:"
prompt appears. Thats a long, long wait to login and not normal in my
experience. Bought a book, searched through the configs and have watched
this group for a while looking for a hint. Its the only thing keeping it
from being perfect.
Suggestions?
Thanks.
------------------------------
From: [EMAIL PROTECTED] (Dustin Puryear)
Subject: Re: Samba vs. NFS
Date: Sun, 25 Apr 1999 09:47:57 -0500
Reply-To: [EMAIL PROTECTED]
On Sat, 24 Apr 1999 08:52:31 GMT, [EMAIL PROTECTED] wrote:
>In article <[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] (Luca Filipozzi) wrote:
>> In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] says...
>> >
>> > I have heard some horror stories about NFS. Is it safer and faster to just
>> > use Samba if I want to mount a filesystem from another machine?
>> >
>> >
>> Use NFS in unix to unix filesystem mounting. Security is a concern, but
>> properly configured inetd.conf files and ipchains/ipfwadm rules will take
>> care of most of those.
>>
>> Use Samba in unix / windows sharing of filesystems.
>I don't think samba is best solution for the file sharing,
>smbmount can work fine in linux box, but it can not work on
>any other unix platform, so the only way to get files is smbclient.
>at the same time, nfs can be shared by any one which supports it, just
>like local file system.
Well, that more or less sums it up. I appreciate the responses to my post.
--
Dustin Puryear
[EMAIL PROTECTED]
------------------------------
From: Richard Corfield <[EMAIL PROTECTED]>
Crossposted-To: linux.samba,uk.comp.os.linux
Subject: Re: NT faster than Linux?
Date: 23 Apr 1999 18:42:28 +0100
In article <[EMAIL PROTECTED]>, Jamie <[EMAIL PROTECTED]> wrote:
>> Just a bit more fuel for the fire
>
>http://www.cnn.com/TECH/computing/9904/21/ntpaper.ent.idg/
>
>And notice this is CNN posting the story ...
I haven't read it yet but have started having to use NT.
One thing that came up in the original Mindcraft report is that they
rebooted the server every time to eliminated incremental instability.
I've had a lot of problem with incremental instabilities under NT.
They say I'm pushing it hard but not as hard as I push Linux.
I wonder what the results would have been if they'd have left the
systems running under load (medium or high) for a few hours or even a
day before testing.
- Richard.
--
_/_/_/ _/_/_/ _/_/_/ Richard Corfield <[EMAIL PROTECTED]>
_/ _/ _/ _/ Web Page: http://www.littondale.freeserve.co.uk
_/_/ _/ _/ Dance (Ballroom, RnR), Hiking, SJA, Linux, ... [ENfP]
_/ _/ _/_/ _/_/_/ PGP2.6 Key ID: 0x0FB084B1 PGP5 Key ID: 0xFA139DA7
------------------------------
From: Joost Evertse <[EMAIL PROTECTED]>
Subject: Re: LCP: timeout sending Config-Requests
Date: Mon, 26 Apr 1999 02:48:44 +0200
Clifford Kite wrote:
> digitalklown ([EMAIL PROTECTED]) wrote:
>
> : Apr 19 22:22:04 localhost chat[478]: send (AT&F^M)
> : Apr 19 22:22:04 localhost chat[478]: expect (OK)
> : Apr 19 22:22:23 localhost chat[478]: AT&F^M^M
> : Apr 19 22:22:23 localhost chat[478]: OK
> : Apr 19 22:22:23 localhost chat[478]: -- got it
>
> Look at the time stamps. There is a 19 second delay between the time
> the modem commands are sent and the time the OK from the modem is
> received. It should be no more than one second.
>
> This means that the IRQ configured in Linux for the modem's device file is
> not the IRQ that the modem actually uses. Configuration is usually done
> at boot-up by setserial in a in /ect/rc.* . Doing "setserial /dev/ttySx"
> will tell you the configured IRQ, where ttySx is the modem device file.
> Determining the IRQ the modem actually uses is up to you.
>
> ..
>
> : i tried
> : CONNECT '/d/c'
>
> Stick with this form.
Hello,
I had exactly this problem, and the solution was indeed changing the irq on
te modem (or /dev/modem). I already wondered why in Minicom the modem reacted
so incredibly slow. But at first i thought that was because of a
configuration setting. I thougt that chat wouldn't have that problem. Then i
saw your postings, and checked /var/log/messages for time outs, and then i
did a setserial /dev/modem, and tada , it said 3e8 irq4. And modem actually
is at 3e8 5. I changed the irq.
Now it all works fine.
I want to thank you for your postings, and i will try to contribute myself.
Regards from Holland,
Joost Evertse
>
>
> --
> Clifford Kite <kite@inet%port.com> Not a guru. (tm)
> /* Governments should be changed like diapers - often and for the
> * same reason. */
------------------------------
From: [EMAIL PROTECTED] (M. Buchenrieder)
Subject: Re: mgetty + pppd
Date: Sun, 25 Apr 1999 11:32:20 GMT
[EMAIL PROTECTED] (David H. Brown) writes:
>I'm trying to get my primary system to allow dial-up connections
>(plus fax, etc.) using mgetty. I've encountered 3 problems:
>1) I got the impression that even though mgetty was running, I
>could still use the modem for some other process (i.e., to
>dial out to my ISP).
Correct.
>That doesn't seem to be the case.
Then your locking setup is broken.
>If
>mgetty is running, the pppd just sits there waiting. (I didn't
>see any lock files in /var/lock.)
Mgetty does only look for lockfiles in /var/lock/ . Obviously,
your pppd version doesn't create lockfiles in /var/lock at all,
or uses a different format that mgetty doesn't detect.
This is all well documented in the mgetty manpage and the infopages.
>2) On another machine, I can dial in and mgetty will answer the
>line, and do login (based on the user id and password in my
>chat script on the dial-in machine). In fact, a shell is
>even started (?!).
Of course - if the user has a corresponding line in /etc/passwd .
>But no ppp configuration takes place.
It doesn't. Use the AUTO_PPP option.
Again, this is well documented.
>3. I tried to do a "shell login" with minicom, but after giving
>a user id and password to the login/password prompts, nothing
>happens, and the line disconnects.
[...]
Did you start mgetty from out of /etc/inittab , as the manpage
is telling you to do ?
Please show us your /etc/inittab file and the output of a failed
login attempt, as recorded by mgetty if started with -x 9 .
Michael
--
Michael Buchenrieder * [EMAIL PROTECTED] * http://www.muc.de/~mibu
Lumber Cartel Unit #456 (TINLC) & Official Netscum
Note: If you want me to send you email, don't munge your address.
------------------------------
From: "FanMan" <[EMAIL PROTECTED]>
Subject: file locking problem with samba
Date: Sun, 25 Apr 1999 21:25:18 +0200
hi,
i am using samba for a while now. no problem at all. suddenly there seems
to be an file locking problem.
if i open an file on my linux box from win98 (PFE editor) i could do
whatever
i like to do (edit and such).
but no other process on the linux box is able to use that file. even after
closing the
win98 app which has been using that file. i have to unmount the share from
the win explorer. after that step everything works again.
hmmm.
i was playing with smb.conf. here ist what i have for that specific share:
comment=Apache with SSL
path=/usr/local/apache
valid users = jf
writable= yes
browsable=yes
sync always = yes
fake oplocks= yes
locking = no
strict locking = no
before toying around i did not have any sync, fake, locking, strict locking
option.
The strange thing is that it worked 6 months.
Moon phases?
Could anyone help....
that would make my day!
thanks
Juergen
------------------------------
From: "Ron Hensley" <[EMAIL PROTECTED]>
Crossposted-To: linux.redhat.misc
Subject: Re: ISP hookup
Date: Sun, 25 Apr 1999 15:28:39 -0400
Just set your hostname as root
#hostname foobar
This puts 'foobar' into the /etc/HOSTNAME file to be found next reboot.
I believe your login scripts are setting it with PPP as its not already set
--
==========================================================================
Ron Hensley ([EMAIL PROTECTED])
Network Administrator - ICNet Internet Services
==========================================================================
Samuel AU <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Hi,
>
> I encounter the same problem!
>
> Samuel
>
>
>
> Cesar da Silva wrote:
>
> > Hi!
> > I'm running Red hat v5.2 and I have a problem with my host name.
> >
> > Before I dial-up to my ISP provider (with 'ifup ppp0') I have my host
> > name as it should (localhost).
> >
> > But as soon as I connect to my ISP-provider I get something else
> > (t0985645). And the new host name stays even after I've logged off.
> >
> > What I want and what I think is supposed to be is that my host name
> > remain the same when I log in, and after I've disconnected, to my
> > ISP-provider.
> > How do I fix this? What am I doing wrong?
> >
> > Thanks,
> > Cesar da Silva
>
------------------------------
From: "Jan Chab" <[EMAIL PROTECTED]>
Subject: Problem with DHCPcd config
Date: Mon, 26 Apr 1999 02:20:24 +0200
Hello,
I have a cable modem working fine with MS Windows. I wanted to make it work
under Linux too. I installed the latest dhcpcd and it compiled fine.
I have RH 5.2 and I used "control-panel" then Interfaces Add Ethernet and
selected "Activate at boot time" and also "Interface config. protocol:
DHCP".
- When I boot the system, the ethernet card is seen OK.
- DHCP is launched but gives the following message:
"modprobe: can't locate module net-pf-17"
- the network does not work; when I do a "ifconfig" I see only the local
loop, not the eth0 I wanted DHCPcd to configure
BTW, I cannot compile the kernel, make zImage stops after a while with an
error but I don't think it's relevant here since I got the ethernet card
found.
Any help welcome (I keep stuck to Windows because of this),
Jan Chab ([EMAIL PROTECTED])
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.networking) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Networking Digest
******************************
