Linux-Misc Digest #234, Volume #27 Mon, 26 Feb 01 11:13:02 EST
Contents:
Re: FTP scripting... (Grant Edwards)
Re: libc 2.x, libc 5, libc 6 ("Peter T. Breuer")
Re: Linux as terminal emulator. ("Peter T. Breuer")
Re: libc 2.x, libc 5, libc 6 ("Peter T. Breuer")
Re: Linux crash like a Windows! ("mari-k")
Re: FTP scripting... (Frank da Cruz)
Re: Linux as terminal emulator. (Frank da Cruz)
Video grabbing with Netstream 2000 (Koos Pol)
Re: Linux partitioning question ("Peter T. Breuer")
Re: SSH to RH6.2 through a firewall?? (Hartmut Berghoff)
Re: Linux crash like a Windows! (Michael Heiming)
Re: Gain root access automatically in runlevel 4 (Fung Wai Keung)
tar from date (-N, --newer, --after-date) (Frederick Bartlett)
Re: kernel 4.1 upgrade - fail to boot again ("Scot Mc Pherson")
Re: Gain root access automatically in runlevel 4 (Jean-David Beyer)
redirect stderr to both screen and logfile at same time? ("J.Smith")
Re: redirect stderr to both screen and logfile at same time? (Tony Curtis)
Re: tar from date (-N, --newer, --after-date) (David)
Re: redirect stderr to both screen and logfile at same time? (Julian Midgley)
how to change root password (richard noel fell)
Re: why /dev/sdx# is a truly unsafe way to refer to physical devices (David Mathog)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Grant Edwards)
Subject: Re: FTP scripting...
Date: Mon, 26 Feb 2001 14:09:00 GMT
In article <[EMAIL PROTECTED]>, Josef Moellers wrote:
>> >> You're probably better off using ncftpget/ncftpput for doing
>> >> stuff in a script.
>> >
>> > ... or look at the netrc concept. I've written a number of scripts that
>> > generate .netrc files on-the-fly and remove them when the access is
>> > done.
There's only one .netrc file and you'll have collisions if you
try to run two instances of your script.
>> Expect was designed to handle tasks like that.
>>
>> http://expect.nist.gov
>
>I know, but
>
> echo "machine Server" > ~/.netrc
> chmod 600 ~/.netrc
> echo "login josef" >> ~/.netrc
> echo "password mypassword" >> ~/.netrc
> echo "macdef init" >> ~/.netrc
> echo "cd /tmp" >> ~/.netrc
> echo "binary" >> ~/.netrc
> echo "put somefile.txt" >> ~/.netrc
> echo "bye" >> ~/.netrc
> ftp Server
>
>is much shorter and faster.
That's short and fast?
Damn. Some people sure like to do things the hard way.
$ ncftpput -u josef -p mypassword Server /tmp somefile.txt
--
Grant Edwards grante Yow! So this is what it
at feels like to be potato
visi.com salad
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: libc 2.x, libc 5, libc 6
Date: Mon, 26 Feb 2001 14:16:34 GMT
Paul Lew <[EMAIL PROTECTED]> wrote:
> Guess there is a difference between libc 2.1 and 2.2 but what is the
> relation with libc 5 or libc 6 ??
Both are versions of libc 6 (i.e. glibc2 is also known as libc 6).
2.1 is the stable release series.
> The more I get into linux the more I drown.....
Why don't you ask us what is the relation with libc 4? Or 3 :-)?
Peter
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Linux as terminal emulator.
Date: Mon, 26 Feb 2001 14:58:48 +0100
Grant Edwards <[EMAIL PROTECTED]> wrote:
> In article <QjYl6.13481$[EMAIL PROTECTED]>, [EMAIL PROTECTED] wrote:
>>> OTOH, there's also minicom (curses) and Seyon (X). I can't ever get
>>> either of them to do anything very useful, but other people seem to
>>> use them.
>>
>>I've always found minicom pretty useful; essentially a "Unix version
>>of Telix," for those that remember the IBM BBS days...
> For me, changing configuration (baud rate, serial port, etc.) often seems to
> have no effect unless it's absolutely the first thing you do when you start
Works fine here. But it affects only the computer to modem speed, which
should be left at whatever it is anyway. The communications speed is up
to the modems.
> the program. I don't know how many hours of time I wasted because minicom
> didn't have the baud rate set to the value it claimed it did. I never got
Eh? Baud rate is not a value that has any meaning in a terminal
program, except as a limit on what the modems will attempt to
communicate.
> sending a file to work either -- the end of the file always gets chopped
> off.
This is nothing to do with minicom. It uses whatever external program
you set it to use for transfers. rz/sz always worked fien for me, but
you can set it to use kemit if you like! The truly awful thing with
kermit (as far as I recall from way back) was that you could never
interupt it and get back any control or line discipline if things went
wrong. That was hell in the days of single-tasking o/s'es at least.
> I also find it infuriating that non-root users can't change configuration
Eh? They can do whatever they like.
> values. Apparently the author of minicom was unaware that the serial port
> device files have mode bits, and thus one should leave privilege decisions
> up to the OS.
I beg your pardon? Are you talking about paths and so on? What
configuration values? The name of the modem device file? The lock file?
Those are all admin issues.
> I finally got fed up with minicom's unpredictability and repeatedly finding
It's always worked fine for me. I still use it to set up ppp sessions!
> that what were reported as bugs in _my_ software were actually bugs in
> minicom. I down-loaded and built ckermit. Next time the test group reports
> that they've found a bug in my software using a test setup that involves
> minicom, I'm going to refuse to look at it until they reproduce the bug
> using kermit.
> [The minicom user-interface also drives me nuts. I find typing commands to
> ckermit takes a fraction of the time that the menu system in minicom takes
> -- but that's a matter of personal taste.]
Menu system? I always do an alt-o or whatever. Teh annoying thing as
that I have to choose a letter after that, instead of typing something
mnemonic. But I have eyes.
Peter
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: libc 2.x, libc 5, libc 6
Date: Mon, 26 Feb 2001 15:02:54 +0100
Markus Kossmann <[EMAIL PROTECTED]> wrote:
> g(nu-)libc-2.x = (linux-)libc-6.x
> (linux)-libc-5.x is a older version of the linux-libc. It was loosely
> based on glibc-1 The switch from libc-5 to libc-6 was around 1998. If
> you want to run binaries, which predate that switch you have to install
> a libc5 runtime environment, because there is no binary compatility
> between libc-5 and libc-6.
What! Of course there is! I'm always taking .o's out of glibc6.a
in order to add to my compiles under libc 5. You mean "no link
compatibility".
Peter
------------------------------
From: "mari-k" <[EMAIL PROTECTED]>
Subject: Re: Linux crash like a Windows!
Crossposted-To: alt.os.linux.slackware,linux.dev.kernel
Date: Mon, 26 Feb 2001 14:21:25 GMT
In article <[EMAIL PROTECTED]>, "Jacques"
<[EMAIL PROTECTED]> wrote:
> Hi,
>
> after installing a computer (Compaq Pentium 16 megs 1.5 gig) using
> ZipSlack, the kernel crash automatically when the /dev directory need to
> work
>
> If I use the shell functionality for completing a file name in /dev or
> do an LS in
> /dev, the kernel crash and freeze with this output :
>
> Unable to handle NULL pointer dereference at virtual address 000 0004
>
> Current -> tss.cr3 = 0027 6000, %cr3 = 0027 6000
> *pde = 0000 0000
> Oops : 0 CPU : 0 EIP : 0010 : [<c012 bcda>] EFLAGS : 0001 0007 eax :
> 0000 0000 ebx : c03a 7010 ecx : c003 c5e8 edx : 0000 0000 esi : c03a
> 700c edi : c03a 7010 ebp : 0000 0287 esp : c023 ff28 ds : 0018 es : 0018
> ss: 0018 Process init (pid : 1, process nr : 1, stackpage = c023 f000)
>
> Call Trace : [<c012 c055>] [<c012 c462>] [<c010 79b8>] Code : 8b 42 04
> 39 d8 75 f7 89 4a 04 55 9d 8b 06 50 e8 4e 7e ff ff
>
>
> -----
> The ZipSlack version is 7.0 using the kernel 2.2.13
>
> I recompiled the kernel, removing all SCSI drivers and others useless
> options for this computers. Crashes are same with the default or home
> compiled kernel.
>
> Thanks for your help
>
> Jacques
>
> [EMAIL PROTECTED]
I don't understand why you have to cross-post.. you dumb phuk...
------------------------------
From: [EMAIL PROTECTED] (Frank da Cruz)
Crossposted-To:
comp.os.linux.development.apps,comp.os.linux.development.system,comp.protocols.kermit.misc
Subject: Re: FTP scripting...
Date: 26 Feb 2001 14:26:24 GMT
In article <[EMAIL PROTECTED]>,
Josef Moellers <[EMAIL PROTECTED]> wrote:
: Grant Edwards wrote:
: > In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] wrote:
: > >>>>>> "Rene Scheibe <[EMAIL PROTECTED]> writes:
: > > Rene> ....can someone tell me if the normal ftp-client is
: > > Rene> scriptable??? How can I write a script for it. Can you
: > > Rene> give me an example??? I want to login to a server and put
: > > Rene> a file on it.
: > >
: > >You might try something like the following:
: > >
: > >#!/bin/ksh
: > >ftp -n 172.28.211.99 << EOD
: > > user cbbrowne MySecretPassword
: > > cd /tmp
: > > binary
: > > put somefile.txt
: > > bye
: > >EOD
: ... or look at the netrc concept. I've written a number of scripts that
: generate .netrc files on-the-fly and remove them when the access is
: done.
:
That's exactly why .netrc is not such a great idea. It's the tail wagging
the dog; you really want it the other way around -- an FTP client that can
run any script you want without having to change (and remember to put back)
some magic file.
Here's how you would do it with the new C-Kermit FTP client:
#!/usr/local/bin/kermit
ftp 172.28.211.99 /user:cbbrowne /password:MySecretPassword
if fail exit 1 FTP connection failed
ftp cd /tmp
if fail exit 1 FTP CD failed
put /binary somefile.txt
if fail exit 1 FTP PUT failed
bye
Note that each operation can be checked for failure. Also note that it's
not a great idea to put passwords in files, so you can have Kermit prompt
you for the password:
ftp 172.28.211.99 /user:cbbrowne
if fail exit 1 FTP connection failed
Just leave out the password and the prompting occurs automatically.
Of course if you want the script to run unattended, prompting for the
password is not practical, but there are ways around that too.
By the way, this is a very simple application for FTP scripting. So
simple, in fact, that Kermit can do it without a script. If you have a
symlink "ftp" to the C-Kermit 7.1 binary, you can just give it this
command line:
ftp 172.28.211.99 -u cbbrowne -P MySecretPassword -D /tmp -p somefile.txt
For more demanding examples, see:
http://www.columbia.edu/kermit/ftpscript.html
- Frank
------------------------------
From: [EMAIL PROTECTED] (Frank da Cruz)
Subject: Re: Linux as terminal emulator.
Date: 26 Feb 2001 14:32:56 GMT
In article <[EMAIL PROTECTED]>,
Peter T. Breuer <[EMAIL PROTECTED]> wrote:
: ...
: you can set it to use kemit if you like! The truly awful thing with
: kermit (as far as I recall from way back) was that you could never
: interupt it and get back any control or line discipline if things went
: wrong. That was hell in the days of single-tasking o/s'es at least.
:
If you have a problem with Kermit you can report it to:
[EMAIL PROTECTED]
It has probably been 10 or 15 years since we've had a report of Kermit
leaving terminal modes messed up, and even then it was an OS bug rather
than a shortcoming of Kermit, which has always saved and restored
terminal modes (the problem was that in 4.2BSD and its derivatives, if
you give an ioctl and then exit, the OS lets the program exit before
the ioctl has completed, and then, since the program is gone, the ioctl
is canceled). I'd recommend that people who have a low opinion of Kermit
take another look at it every decade or so :-)
http://www.columbia.edu/kermit/
- Frank
------------------------------
From: Koos Pol <[EMAIL PROTECTED]>
Subject: Video grabbing with Netstream 2000
Date: Mon, 26 Feb 2001 15:20:00 +0100
Hello,
Anyone familiar or experience with the Netstream 2000? It seems like a
very nice grabber card.
http://www.sigmadesigns.com/products/netstream2000.htm
--
Koos Pol - Systems Administrator - Compuware Europe B.V. - Amsterdam
[EMAIL PROTECTED]
Check my email address when you hit "Reply".
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup,comp.os.linux.hardware
Subject: Re: Linux partitioning question
Date: Mon, 26 Feb 2001 15:21:14 +0100
In comp.os.linux.misc Doug Lutterloh <[EMAIL PROTECTED]> wrote:
> I have never agreed with the suggested partitioning scheme in most
> install docs for Linux, at least not for the home user. I usually set up
> with 3 partitions on my machine. One for windoze because I dual
> boot, one for Linux swap, and one for Linux. Why confuse the issue?!!?
Because you will die horribly. Your /var and /home are on the same
partition as your fixed system. read the partition-HOWTO to learn more!
(ignorance is NOT a good reason for doing something ..=)
> Why do I want to guess how much I will need for /home and /usr when
> I can just lump them together in one big partition and use all my
> space as efficiently as possible.
If you don't know, then plainly you should go and find out. Check the
HOWTO.
Peter
------------------------------
From: Hartmut Berghoff <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.security
Subject: Re: SSH to RH6.2 through a firewall??
Date: Mon, 26 Feb 2001 15:59:43 +0100
Reply-To: [EMAIL PROTECTED]
Bo Berglund schrieb:
>
> 2) I have absolutely no idea which port(s) to open for this to work.
> So basically my question is:
> Which port(s) SSH uses?
> Are there any other issues about going through a NAT firewall?
>
> /Bo
>
> Bo Berglund
> [EMAIL PROTECTED]
What says your /etc/Services. Normally it's commented, so you find
portnumber protocol # ssh
My SuSy has 22. As tis is reserved number area, this should apply to you
too.
Hartmut
------------------------------
Date: Mon, 26 Feb 2001 15:34:58 +0100
From: Michael Heiming <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux.slackware
Subject: Re: Linux crash like a Windows!
Jacques wrote:
> Hi,
>
> after installing a computer (Compaq Pentium 16 megs 1.5 gig) using
> ZipSlack,
> the kernel crash automatically when the /dev directory need to work
>
> If I use the shell functionality for completing a file name in /dev or
> do an LS in
> /dev, the kernel crash and freeze with this output :
>
> Unable to handle NULL pointer dereference at virtual address 000 0004
>
> Current -> tss.cr3 = 0027 6000, %cr3 = 0027 6000
> *pde = 0000 0000
> Oops : 0
> CPU : 0
> EIP : 0010 : [<c012 bcda>]
> EFLAGS : 0001 0007
> eax : 0000 0000 ebx : c03a 7010 ecx : c003 c5e8 edx : 0000 0000
> esi : c03a 700c edi : c03a 7010 ebp : 0000 0287 esp : c023 ff28
> ds : 0018 es : 0018 ss: 0018
> Process init (pid : 1, process nr : 1, stackpage = c023 f000)
>
> Call Trace : [<c012 c055>] [<c012 c462>] [<c010 79b8>]
> Code : 8b 42 04 39 d8 75 f7 89 4a 04 55 9d 8b 06 50 e8 4e 7e ff ff
>
> -----
> The ZipSlack version is 7.0 using the kernel 2.2.13
>
> I recompiled the kernel, removing all SCSI drivers and others useless
> options
> for this computers. Crashes are same with the default or home compiled
> kernel.
>
> Thanks for your help
>
> Jacques
>
> [EMAIL PROTECTED]
Hello,
I've seen this messages only on machines with a broken RAM/mobo,
try to compile a kernel if it stops with the message:
..gcc caught fatal Signal 11...
It's probably your RAM, try to change it and test again.
Linux doesn't chrash if your hw is OK, asuming you haven't mucked up the
system yourself dreadful!
Good luck
Michael Heiming
------------------------------
From: Fung Wai Keung <[EMAIL PROTECTED]>
Subject: Re: Gain root access automatically in runlevel 4
Date: 26 Feb 2001 15:13:41 GMT
Hi,
Thanks for your reply. But how to set setuid bit on? Must the user and
group ownership of the program be root? I need root access with my
program to control a robot.
Jean-David Beyer <[EMAIL PROTECTED]> wrote:
: Fung Wai Keung wrote:
:>
:> Hi all,
:>
:> I am developing a kiosk system with touchscreen monitor using Linux.
:> The system will not have a keyboard for user input. My program needs root
:> access to be run. I set the system to boot to runlevel 4. However, my
:> program doesn't run. Is there any way to gain root access automatically
:> (without typing root password with keyboard) when the system boot up?
:>
: You could give your program to root and have the setuid bit on.
: You might wish to redesign your program, and any data files it uses,
: so that it does not need root permission, in order to reduce the
: security risks of running as root.
: --
: .~. Jean-David Beyer Registered Linux User 85642.
: /V\ Registered Machine 73926.
: /( )\ Shrewsbury, New Jersey http://counter.li.org
: ^^-^^ 8:45am up 5 days, 16:19, 3 users, load average: 2.06, 2.09,
: 2.11
--
Regards,
Wai Keung, Fung
Department of Automation and Computer-Aided Engineering,
The Chinese University of Hong Kong,
Shatin, N.T.,
Hong Kong
Tel: (852)26098056 Fax: (852)26036002
Email: [EMAIL PROTECTED]
------------------------------
From: Frederick Bartlett <fbartlett'ignore'@optonline.net>
Subject: tar from date (-N, --newer, --after-date)
Date: Mon, 26 Feb 2001 15:26:04 GMT
I have tried a number of variants to get this to work: but every tar
command I've tried has archived my entire directory, not just the latest
files. What am I missing?
tar -cvf test.tar --newer 23/02/2001 *
tar -cvf test.tar --newer 02/23/2001 *
tar -N 2001-02-23 -cvf test.tar *
tar -cv -N 0223 -f test.tar *
tar -cv -N 2001-02-23 -f test.tar *
tar -cv --newer="2001-02-23" -f test.tar *
etc. ...
I have consulted a couple of manuals, deja news, and the man and info
pages -- but I have found no joy.
tia,
Fred
------------------------------
From: "Scot Mc Pherson" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Re: kernel 4.1 upgrade - fail to boot again
Date: Mon, 26 Feb 2001 15:39:32 GMT
Sorry for the silly question, but why are you using a development revision
that is older than the latest stable release??
You should be using 2.4.2, yes?
--
Scot Mc Pherson
http://www.behomet.net
N27° 19' 56"
W82° 30' 39"
"bart" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Hello linus lovers,
>
> I recently upgraded my 2.2.14 kernel to 2.4.1 ... everything worked all
> right but when I try to boot with the new
> kernel it complains that it can't mount the root filesystem ... !
> Compiled using make zBimage
> I just copied the lilo entry for the 2.2.14 kernel ... where the root
> should be exactly the same as before, activated this
> by invoking lilo ... but it doesn't work. Linux boots but fails with a
> panic since it can't mount the root filesystem.
> Did I forget something ?
>
> lilo.conf
> -------
> boot=/dev/hda5
> map=/boot/map
> install=/boot/boot.b
> prompt
> timeout=50
> linear
> default=2.14
>
> image=/boot/vmlinuz-2.2.14-5.0
> label=2.14
> read-only
> root=/dev/hda5
>
> image=/boot/vmlinuz-2.4.1
> label=2.4.1
> read-only
> root=/dev/hda5
>
>
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: Gain root access automatically in runlevel 4
Date: Mon, 26 Feb 2001 10:40:33 -0500
Fung Wai Keung wrote:
>
> Hi,
>
> Thanks for your reply. But how to set setuid bit on? Must the user and
> group ownership of the program be root? I need root access with my
> program to control a robot.
man chmod and man chown will tell you that kind of thing. You need
not fiddle with the group id or setgid on if you do not need to. But I
suspect you should not have the owner be group, nore should the setuid
be on either.
Your reply makes it seem more clear to me than ever that you should
not let users run as root. Your program might just have a bug in it
that permits the users to totally devastate your system. I have no
idea what the robot does, but if they user has root access, he could
make it do something quite unintended. You should perhaps keep the
program running as a normal unpriviledged user, but as a member of a
particular group otherwise unused, setgid on, and give that groupr,
and only that group, permission to control the robot (as determined by
the /dev/robot permissions).
>
> Jean-David Beyer <[EMAIL PROTECTED]> wrote:
> : Fung Wai Keung wrote:
> :>
> :> Hi all,
> :>
> :> I am developing a kiosk system with touchscreen monitor using Linux.
> :> The system will not have a keyboard for user input. My program needs root
> :> access to be run. I set the system to boot to runlevel 4. However, my
> :> program doesn't run. Is there any way to gain root access automatically
> :> (without typing root password with keyboard) when the system boot up?
> :>
> : You could give your program to root and have the setuid bit on.
>
> : You might wish to redesign your program, and any data files it uses,
> : so that it does not need root permission, in order to reduce the
> : security risks of running as root.
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey http://counter.li.org
^^-^^ 10:35am up 5 days, 18:08, 3 users, load average: 2.14, 2.19,
2.14
------------------------------
From: "J.Smith" <[EMAIL PROTECTED]>
Crossposted-To:
alt.os.linux,alt.solaris.x86,comp.unix.admin,comp.unix.aix,comp.unix.misc,comp.unix.shell
Subject: redirect stderr to both screen and logfile at same time?
Date: Mon, 26 Feb 2001 16:44:52 +0100
Hi.
In a particular script, I would like to be able to (re)direct stdout and
stderr to *both* the screen *and* a logfile at the same time. Just doing
'./commandname 1>/logfile 2>/logfile' will redirect stdout and stderr to the
logfile, but wont show the output on the screen anymore. And the 'tee'
command only seems to work for stdout, and not for stderr.
Is there any way to do this?
Thanks for any and all suggestions.
------------------------------
From: Tony Curtis <[EMAIL PROTECTED]>
Crossposted-To:
alt.os.linux,alt.solaris.x86,comp.unix.admin,comp.unix.aix,comp.unix.misc,comp.unix.shell
Subject: Re: redirect stderr to both screen and logfile at same time?
Date: 26 Feb 2001 09:46:32 -0600
>> On Mon, 26 Feb 2001 16:44:52 +0100,
>> "J.Smith" <[EMAIL PROTECTED]> said:
> Hi. In a particular script, I would like to be able to
> (re)direct stdout and stderr to *both* the screen *and*
> a logfile at the same time. Just doing './commandname
> 1>/logfile 2>/logfile' will redirect stdout and stderr
> to the logfile, but wont show the output on the screen
> anymore. And the 'tee' command only seems to work for
> stdout, and not for stderr.
> Is there any way to do this?
Maybe this?
./command 2>&1 | tee logfile
(merges stdout and stderr before the pipe)
hth
t
--
The avalanche has already started.
It is too late for the pebbles to vote.
------------------------------
From: David <[EMAIL PROTECTED]>
Subject: Re: tar from date (-N, --newer, --after-date)
Date: Mon, 26 Feb 2001 15:50:08 GMT
Frederick Bartlett wrote:
>
> I have tried a number of variants to get this to work: but every tar
> command I've tried has archived my entire directory, not just the latest
> files. What am I missing?
>
> tar -cvf test.tar --newer 23/02/2001 *
> tar -cvf test.tar --newer 02/23/2001 *
> tar -N 2001-02-23 -cvf test.tar *
> tar -cv -N 0223 -f test.tar *
> tar -cv -N 2001-02-23 -f test.tar *
> tar -cv --newer="2001-02-23" -f test.tar *
>
> etc. ...
>
> I have consulted a couple of manuals, deja news, and the man and info
> pages -- but I have found no joy.
>
> tia,
> Fred
Maybe like this:
tar -cvf test.tar . --newer 02/24/01
It will include any/all directories inside the curent directory but only
files that have changed or are newer than date.
--
Confucius say: He who play in root, eventually kill tree.
Registered with the Linux Counter. http://counter.li.org
ID # 123538
Completed more W/U's than 99.086% of seti users. +/- 0.01%
------------------------------
Crossposted-To:
alt.os.linux,alt.solaris.x86,comp.unix.admin,comp.unix.aix,comp.unix.misc,comp.unix.shell
Subject: Re: redirect stderr to both screen and logfile at same time?
From: [EMAIL PROTECTED] (Julian Midgley)
Date: Mon, 26 Feb 2001 15:48:49 GMT
In article <fSum6.5750$43.28434@zonnet-reader-1>,
J.Smith <[EMAIL PROTECTED]> wrote:
>Hi.
>
>In a particular script, I would like to be able to (re)direct stdout and
>stderr to *both* the screen *and* a logfile at the same time. Just doing
>'./commandname 1>/logfile 2>/logfile' will redirect stdout and stderr to the
>logfile, but wont show the output on the screen anymore. And the 'tee'
>command only seems to work for stdout, and not for stderr.
>
>Is there any way to do this?
Just redirect stderr on to stdout before piping it to tee, eg:
find . 2>&1 | tee foo
All the best,
Julian Midgley
--
Julian Midgley
Principal Consultant [EMAIL PROTECTED]
Zeus Technology Ltd http://www.zeus.com
------------------------------
From: richard noel fell <[EMAIL PROTECTED]>
Crossposted-To: linux.redhat.misc
Subject: how to change root password
Date: Mon, 26 Feb 2001 12:14:15 -0500
==============4E21426188759AE69A90B070
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Referring to my previous post, I found tomsbrt rescue disk and that
works fine. Now, in my /etc/password file, I have the line
root:x:0:0:root:/root:/bin/bash
I have to edit this so root does not have a password upon reboot from
the hard drive. One reference says to enter
root::0:0:root:/:/bin/sh. Is this correct since there are some not so
subtle differences between the two lines.
Thanks,
Dick Fell
--
Please note new email address:
[EMAIL PROTECTED]
Richard Fell
13 Davida Road
Burlington, Ma 01803
(781)273-2126
==============4E21426188759AE69A90B070
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Referring to my previous post, I found tomsbrt rescue disk and that
works fine. Now, in my /etc/password file, I have the line
<br>root:x:0:0:root:/root:/bin/bash
<p>I have to edit this so root does not have a password upon reboot
from the hard drive. One reference says to enter
<br>root::0:0:root:/:/bin/sh. Is this correct since there are some not
so subtle differences between the two lines.
<br>Thanks,
<br>Dick Fell
<pre>--
Please note new email address:
[EMAIL PROTECTED]
Richard Fell
13 Davida Road
Burlington, Ma 01803
(781)273-2126</pre>
</html>
==============4E21426188759AE69A90B070==
------------------------------
From: [EMAIL PROTECTED] (David Mathog)
Subject: Re: why /dev/sdx# is a truly unsafe way to refer to physical devices
Date: 26 Feb 2001 16:07:25 GMT
Reply-To: [EMAIL PROTECTED]
In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] writes:
>On 23 Feb 2001 23:11:31 GMT David Mathog <[EMAIL PROTECTED]> wrote:
>
>| I never much cared for the Linux convention of /dev/sdx# for referring to
>| scsi disks.
>However, SCSI device addressing is quite extensive when you consider
>the potential for 8, maybe more, SCSI adapters, 8 logical units,
>16 targets (potentially the adapter can be moved to a different
>target number), and of course many partitions (I have one drive
>on one machine right now with 28 partitions due to it being a
>triple boot between Linux, FreeBSD, and OpenBSD). Right now the
>partition limit in Linux for SCSI is 15. This means we need at
>least 8*8*16*16 = 16384 device codes. That's 25% of the whole
>device code space. To get as many partitions on SCSI as IDE has,
>there would be no device code space left for IDE to have any.
That's true if you have to build _all_ of these devices - but there's
no reason that should be the case. Irix, for instance, detects what's out
there during boot and builds the appropriate devices on the fly. For
instance, on one of our Indigo2 systems with two disks /dev/dsk contains
only these entries for physical disks/controllers/partitions:
dks0d1s0 dks0d1s1 dks1d3s0 dks1d3s1 dks1d3s6 dks1d3s7
Even without automatic device detection, it's not a terribly restrictive
situation if you're required to build the appropriate devices before you
can access them. It would only be a problem I suppose if you didn't know
the SCSI controller and device numbers of the devices on the system, and
that could be handled by a tool that scans the SCSI bus and reports the
results. On every system I've ever seen this information appears during
the boot sequence but it would be better to be able to do something like:
% scsiscan
(lists devices)
or
% make_attached_scsi_devs
There is already something very much like "scsiscan" now, which is used to
detect the disks/partitions and assign them to sda1, sdb2 etc. I'm going
to guess it would not be horrifically hard to have it create the devices on
the fly instead of just assigning numbers sequentially. Ideally we could
just add a command line switch to it like
-a
for absolute, as opposed to the default "-s" for sequential.
Regards,
David Mathog
[EMAIL PROTECTED]
Manager, sequence analysis facility, biology division, Caltech
------------------------------
** 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 by posting to comp.os.linux.misc.
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-Misc Digest
******************************
