Linux-Misc Digest #779, Volume #26 Thu, 11 Jan 01 10:13:03 EST
Contents:
Re: Backup software for Linux? (Martin Gregorie)
Dual-boot - accessing Linux file-system from DOS partition? (David Jones)
Re: Backup software for Linux? (Jean-David Beyer)
"unix dial-up account" spam question [was Re: how can i stop a ("Joe (mvjap3)
Philbrook III")
Re: Dual-boot - accessing Linux file-system from DOS partition? (Pawlik)
Re: L2-Cache of Pentium2 with Linux (Jean-David Beyer)
Re: RedHat 7.0 problem with rexec and PAM (John Doe)
Re: Dual-boot - accessing Linux file-system from DOS partition? (Mikael Johansson)
Re: compiling 2.4.0 modules error (Michael Heiming)
Re: Enter escape charcter in vi or another editor ("Joe (mvjap3) Philbrook III")
Re: HELP! : mysterious ext2 undeletable directory ([EMAIL PROTECTED])
Linux and Clariion? ([EMAIL PROTECTED])
Re: Can't build rpm with 4.0 ([EMAIL PROTECTED])
problems starting stuff when kppp is running ("Adam J BC")
XF 4.0.2 and SiS620 (Daren)
Upgrading glibc using rpm (Pete Lachall)
Re: changing window managers & .Xclients (Steve Connet)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Martin Gregorie)
Subject: Re: Backup software for Linux?
Date: Thu, 11 Jan 2001 10:16:31 GMT
On 10 Jan 2001 23:04:35 -0600, [EMAIL PROTECTED] (Dave
Brown) wrote:
>In article <3a5c6b13.15616603@news>, Martin Gregorie wrote:
>>I'll second that. What have you got against letting the drive do the
>>compression, though?
>
>Personal opinion: it depends on the reliability of the media. A single
>media error destroys the ability to continue the decompress, so you
>lose everything past that point. If uncompressed, you only lose that
>file corrupted by the bad spot. At least that's been my experience.
>(And why I store backups on CDRs, using compression, but checking
>them after writing.)
>
Fair comment. Thanks.
I shall read my hardware manual again before decididing whether to
leave compression on.
--
gregorie | Martin Gregorie
@logica | Logica Ltd
com | +44 020 76379111
------------------------------
From: David Jones <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.misc
Subject: Dual-boot - accessing Linux file-system from DOS partition?
Date: Thu, 11 Jan 2001 11:39:01 +0000
Hi. Am running a dual-boot system, Red Hat Linux 6.1 and Win98 on a
P II. Creating a symbolic link in the Linux filesystem to the FAT32
partition was simple, but what is the easiest way to access the Linux
filesystem from within Win98?
Not that I will need to do this a lot. But is it simple to do? In fact,
is it even possible to do, or has Micro�oft made it impossible?
Many thanks in advance.
David
--
David Jones
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: Backup software for Linux?
Date: Thu, 11 Jan 2001 07:45:02 -0500
Dave Brown wrote:
>
> In article <3a5c6b13.15616603@news>, Martin Gregorie wrote:
> >I'll second that. What have you got against letting the drive do the
> >compression, though?
>
> Personal opinion: it depends on the reliability of the media. A single
> media error destroys the ability to continue the decompress, so you
> lose everything past that point. If uncompressed, you only lose that
> file corrupted by the bad spot. At least that's been my experience.
> (And why I store backups on CDRs, using compression, but checking
> them after writing.)
I think a lot depends on whether you compress on hardware or with
software, what the compression algorithm is, what your backup tool is
(tar-type or cpio-type), whether you check the backup afterwards or not,
and so on). It is my understanding that in tar, if you have tar do the
compression, that you lose everything after the first error. With cpio,
you have to compress separately. It is my understanding that you lose
only the bad block.
What I do, however, is to not have the software do any compression at
all, but have the tape drive do it. While I presently use BRU, its
behavior is very similar to using a combination of find and cpio (using
the crc format where cpio computes a crc of every block and puts it in
the block (probably at the end of the block)). It then reads back the
entire tape and checks the crc of every block, listing those that
failed. If any failures occur, I re-do the backup. I put a little red
dot on any tape that has any failure, and if there are too many red
dots, I retire the tape. BRU also keeps track of how many times I wrote
a backup on the tape, and will not let me do more than 200 backups,
fearing that the tape may no longer be good enough. BRU uses a different
number for the upper limit depending on the tape type. They use 200 for
4 mm DAT tapes, 100 for Travan-type tapes (used on floppy tapes), 250
for 1/2" reel-to-reel tapes (if anyone has those anymore; I guess they
do), 250 for 1/2" DAT tapes, and so on.
The hardware compression uses the Lempel-Ziv algorithm, but I never
looked up what that is.
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 7:25am up 2 days, 8:34, 2 users, load average: 2.17, 2.12, 2.08
------------------------------
From: "Joe (mvjap3) Philbrook III" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: "unix dial-up account" spam question [was Re: how can i stop a
Date: Thu, 11 Jan 2001 07:46:20 -0500
Reply-To: mvjap3 at work <[EMAIL PROTECTED]>
On Wed, 10 Jan 2001, Dale Hennessey wrote:
> Back when I was using a unix dial-up account, I used the following
> set of procmail filters...
>
> http://www.spambouncer.org
>
> At the time, they were excellent for eliminating spam and you could
> configure it to auto-complain to the upstream provider.
Ummm Dale, since you appearantly understood spam filtering while using a
dial-up connection, I have a question for you...
I have a problem getting legitimate mail from my local linux mail spool to
pass the spam filters on my employers mail system...
I did get sendmail to put my ISP's domain name in the from field but While
I can get mail from my pc to my co mailbox with pine by setting it up to
use the ISP's smtp server directly, I can't find any way to do this when
I use the local mail spool... And I would very much like to go on line
only long enough to send and receive my mail...
ppp-go
fetchmail
sendmail -q
ppp-off
But if I can't get mail to my lucent mailbox that way, I can't figure that
anyone else is gonna get the mail either... Do you happen to know how to
stop most spam filters from interpreting the non registered local domain
name that happens to send mail THAT DOES HAVE A VALID REPLY PATH as a
reason to sort it to /dev/nul ????
? ?
-=- -=-
<?> <?>
^ <Hunh?> Guess I just don't know.
--- Joe A Philbrook III
<[EMAIL PROTECTED]>
? ?
------------------------------
From: Pawlik <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.misc
Subject: Re: Dual-boot - accessing Linux file-system from DOS partition?
Date: Thu, 11 Jan 2001 14:01:23 +0100
David Jones wrote:
> Hi. Am running a dual-boot system, Red Hat Linux 6.1 and Win98 on a
> P II. Creating a symbolic link in the Linux filesystem to the FAT32
> partition was simple, but what is the easiest way to access the Linux
> filesystem from within Win98?
>
> Not that I will need to do this a lot. But is it simple to do? In fact,
> is it even possible to do, or has Micro�oft made it impossible?
>
> Many thanks in advance.
>
> David
There is a program called ext2fs-explore (well, s.th. like that).
There you can "explore" an ext2-filesystem (and ONLY ext2, NOT reiserfs or
whatever you have).
BUT, I wouldn't use it. If it's buggy you may lost data !!!
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: L2-Cache of Pentium2 with Linux
Date: Thu, 11 Jan 2001 07:57:17 -0500
Ralf Render wrote:
>
> Thanks Henrik,
>
> I think that is the point. My program does not have the cache for itself
> because of the multitasking.
That might obscure the results slightly, but I noticed little, if any,
such obscuring when yesterday I ran the test Peter posted.
In the little test, the dropoff as I exceeded the size of the L1 and L2
caches was extremely well-pronounced, even though the machine was 0%
idle both before and after the test because I was running two instances
of the SETI@home program in the background (at nice 19). My process
table typically has about 75 to 80 processes running (quite a lot of
daemons), and 3 to 5 (but sometimes more) labelled "running" in top.
So, sure, the process may not have the cache to itself, but it does
while it has control of the CPU, and I believe it can run for 100
microseconds at a clip before the scheduler times it out and (may, but
also may not) assign it to another process. These days, you can run
through a lot of instructions in 100 microseconds. I remember an early
version of the Unix kernel where the timeout part of the scheduler took
2 seconds and CPUs had about a 1 Megahertz clock, and a more recent one
where that interval was around 25 milliseconds. That was when processes
were really swapped (not paged), memories over 128Kilobytes were
unusual, and hard drives ran at 2400 rpm.
>
> Ralf.
>
> Henrik Carlqvist <[EMAIL PROTECTED]> schrieb in im Newsbeitrag:
> [EMAIL PROTECTED]
> > Ralf Render wrote:
> > > Yes, my bios is set up to use the L2 Cache.
> > > I wrote a small program that runs through an array (512k) and do some
> > > calculations. A second version does the same with a loop of 100
> > > times.
> >
> > Your cache is only 512k and some of it is probably also needed for
> > instructions. Also remember that Linux is a multitasking OS, a process
> > might not have all the cache for itself during the whole loop.
> >
> > Try to make a new testprogram which uses an array which is much smaller
> > but still bigger than your L1 cache. What kind of result do you get
> > then?
> >
> > regards Henrik
> > --
> > spammer strikeback:
> > root@localhost [EMAIL PROTECTED] [EMAIL PROTECTED]
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 7:45am up 2 days, 8:54, 2 users, load average: 2.00, 1.94, 1.99
------------------------------
From: John Doe <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.security,nl.comp.os.linux.installatie,nl.comp.os.linux.netwerken
Subject: Re: RedHat 7.0 problem with rexec and PAM
Date: Thu, 11 Jan 2001 12:56:38 GMT
"E.M. Stegehuis (Marcel)" wrote:
> A while ago we 'upgraded' from RedHat6.2 to RedHat7.0. Since than we are not
> able anymore to use rexec nor rsh to get an xterm on an other machine.
>
> With ntsysv I ensured that rexecd and rshd are on. The
> /etc/xinit.d/rexec/rsh files show the service to be available (disable=no).
>
> According to the message on the RedHat site I commented the
> auth required /lib/security/pam_nologin.so
> line in /etc/pam.d/rexec.
>
> Stil I get a connection closed.
>
> My /var/log/secure file mentioned a xinetd[PID XXX]: START: exec pid=xxxx
> from xxx.yyy.zzz.aaa
>
> Does anyone recognize the problem and knows a solution.
>
> Marcel
Back in RedHat 5.2 days, I was very much interested in using the xwin32 server
on a windows box to initiate an xterm using rexec.
As I recall, I could only make it work after removing the securetty entry from
some pam file somewhere. After that, root logins via rexec were possible, which
I did not care for.
I finally grabbed a source package (netkit-rsh from slackware sources) and
compiled another rexecd. Before compiling, I think there is a configuration
option (./configure --disable-pam ??) to disable PAM.
------------------------------
From: Mikael Johansson <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.misc
Subject: Re: Dual-boot - accessing Linux file-system from DOS partition?
Date: Thu, 11 Jan 2001 15:09:02 +0200
Hello David!
David Jones wrote:
> Hi. Am running a dual-boot system, Red Hat Linux 6.1 and Win98 on a
> P II. Creating a symbolic link in the Linux filesystem to the FAT32
> partition was simple, but what is the easiest way to access the Linux
> filesystem from within Win98?
>
> Not that I will need to do this a lot. But is it simple to do? In fact,
> is it even possible to do, or has Micro�oft made it impossible?
I haven't tried this out myself but might be just the right one for you:
FSDEXT2, get it at
http://www.yipton.demon.co.uk/
Have a nice day,
Mikael J.
------------------------------
From: Michael Heiming <[EMAIL PROTECTED]>
Subject: Re: compiling 2.4.0 modules error
Date: Thu, 11 Jan 2001 15:51:59 +0100
[EMAIL PROTECTED] wrote:
> Could someone maybe help me with what this might mean?
>
> r128_cce.c: In function `r128_cce_init_ring_buffer':
> r128_cce.c:339: structure has no member named `agp'
> r128_cce.c:333: warning: `ring_start' might be used uninitialized in
> this function
> r128_cce.c: In function `r128_cce_packet':
> r128_cce.c:1023: warning: unused variable `size'
> r128_cce.c:1021: warning: unused variable `buffer'
> r128_cce.c:1019: warning: unused variable `dev_priv'
> make[3]: *** [r128_cce.o] Error 1
> make[3]: Leaving directory `/linux/drivers/char/drm'
> make[2]: *** [_modsubdir_drm] Error 2
> make[2]: Leaving directory `/linux/drivers/char'
> make[1]: *** [_modsubdir_char] Error 2
> make[1]: Leaving directory `/linux/drivers'
> make: *** [_mod_drivers] Error 2
>
> everything went well until i started compiling the modules.
> TIA
> Burke
>
> Sent via Deja.com
> http://www.deja.com/
Hello,
does your system meet the requierments as mentioned in
Documentation/Changes?
Did you issue make mrporper?
Good luck
Michael Heiming
Sysadmin
--
__ __ __ Virtueller Bau-Markt AG
\ / [__) [__] [ __ Meerbuscher Strasse 64
\/ [__) | | [_./ 40670 Meerbusch
www.vbag.de Michael Heiming ([EMAIL PROTECTED])
------------------------------
From: "Joe (mvjap3) Philbrook III" <[EMAIL PROTECTED]>
Subject: Re: Enter escape charcter in vi or another editor
Date: Thu, 11 Jan 2001 08:49:56 -0500
Reply-To: mvjap3 at work <[EMAIL PROTECTED]>
On Wed, 10 Jan 2001, Scott Alfter wrote:
> I don't know about vi (I only use it when there's no other editor, as it's
> rather cumbersome to use), but it's fairly easy in joe. Press ` and then
Oh please, don't call my beloved, intuitive, vi editor cumbersome...
Oh alight, I'll admit that in it's original form it can be a little
awkward to get used to, especially if you don't keep track of the
difference between insert and command mode, but cumbersome NEVER!
I don't know "joe" But it looks like it uses the grave accent key " ` "
for what both pure vi and the much improved vim use ^V <thats [ctrl+V]>
for... You might want to know that just in case you sometime get stuck
doing something someplace where it is the only editor...
And if your willing to find out why I say vim is much easier, do try it
sometime. But do create a ~/.vimrc with something like the below 4 lines
in it first. {I say first because if it doesn't find a .vimrc it will
pretend it's the original vi and then you would still think it's
cumbersome... <sigh>} And if you try vim, be sure to check out the command
mode command of
:help
<yup they added a built in help system...>
<put the next 4 lines in ~/.vimrc>
:set ai
:set sm
:set smd
:set textwidth=74
And you might even "like" vim...
~^~ ~^~
(*) (*) Joe (theWordy) Philbrook
^
\___/ < [EMAIL PROTECTED] >
<< there's nothing like an idiot who thinks he knows everything, of >>
<< course, if I actually knew everything, I'd know I was an idiot. >>
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: HELP! : mysterious ext2 undeletable directory
Crossposted-To: comp.os.linux.setup
Date: Thu, 11 Jan 2001 13:34:51 +0000
Robert Lynch <[EMAIL PROTECTED]> did eloquently scribble:
> I had this problem, which I decided was due to junkbuster's
> logrotate being wonky. I filed a bug report at the junkbuster
> site, which was never answered.
> When I looked inside the directory (not easy!) I found that it
> was a jungle logrotate .gz's, like:
> junkbuster.1.gz.3.gz.1.gz...
This looks like a job for debugfs...
With this you can force delete files by unlinking their inodes and clearing
them manually. Even if the file exhibits corrupted attribute bits.
--
=============================================================================
| [EMAIL PROTECTED] | Windows95 (noun): 32 bit extensions and a |
| | graphical shell for a 16 bit patch to an 8 bit |
|Andrew Halliwell BSc(hons)| operating system originally coded for a 4 bit |
| in |microprocessor, written by a 2 bit company, that|
| Computer Science | can't stand 1 bit of competition. |
=============================================================================
------------------------------
From: [EMAIL PROTECTED]
Subject: Linux and Clariion?
Date: Thu, 11 Jan 2001 14:09:13 GMT
Hi All,
we try to connect an EMC Clariion 5300 to a Linux server.
We use a Qlogic HBA.
Has anyone done this before?
The problem is:
We can see the LUN created on the clariion, but linux only installs
it as a generice scsi device. So we cannot use it.
Any Ideas?
Thanks in advance
zlorfi
Sent via Deja.com
http://www.deja.com/
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Can't build rpm with 4.0
Reply-To: [EMAIL PROTECTED]
Date: 11 Jan 2001 08:14:24 -0600
In article <91mfel$r0c$[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote:
>I upgraded my RH6.2 box to rpm4.0. I'm trying to learn how to build
>rpm's for my software. However, rpm4.0 does not seem to be able to
>build packages. If I type:
>
>$ cd /usr/src/redhat/SPECS
>$ rpm -ba myfile.spec
>
>then rpm responds with a usage() call as though I had typed in bad
>options. Am I using the tool incorrectly? Have others had problems
>with building packages with 4.0? Specifically, have others had problems
>building packages with 4.0 on systems which upgraded 3.x to 4.0?
>
>Exasperated,
>Paul
>
>
>Sent via Deja.com
>http://www.deja.com/
There are 3 rpm packages, rpm-<version number>, rpm-build-<version number>,
and rpm-devel-<version number>. (And a fourth for python.) If you haven't
upgraded the "build" and "devel" packages, do that too. (And don't forget
to rpm --rebuilddb when you're done.)
------------------------------
From: "Adam J BC" <[EMAIL PROTECTED]>
Subject: problems starting stuff when kppp is running
Date: Thu, 11 Jan 2001 14:26:40 -0000
hello.
i installed everybuddy just now, then launched kppp, connected to the
internet, and when i tried to start everybuddy it wouldn't let me. i can't
remember the exact error, something to do with GTK and I remember seeing a :
and a couple of 0s. sorry, i should have written it down but i forgot.
anyway, i found i couldn't start netscape either. or even a simple
terminal. unless kppp was closed.
but everybuddy tried to connect when it first starts, and i couldn't see an
option to try and reconnect, so i can't start everybuddy first, then start
kppp.
so...any ideas as to what's going on?
i'm running redhat 6 (although i'm gonna switch to debian tomorrow) and
gnome on top of enlightenment.
thanks,
adam
------------------------------
From: [EMAIL PROTECTED] (Daren)
Subject: XF 4.0.2 and SiS620
Date: Thu, 11 Jan 2001 14:41:21 GMT
Hi
I've just installed 4.0.2 and using the default SiS620 driver, however
upon starting X, all text is completely garbled and I'm getting 2
vertical lines of 'interference' (best way I can describe it!) running
down the screen
Does anyone know of this problem and wether it's a driver thing or a
config thing?
Cheers
Daren
------------------------------
From: Pete Lachall <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup,comp.os.linux.development.system
Subject: Upgrading glibc using rpm
Date: Thu, 11 Jan 2001 09:43:59 -0500
I'm upgrading glibc from 2.1.1 to 2.1.3 on Redhat 6.2 using the rpm.
After running 'rpm -U xxx.rpm' I get a bunch of conflicts, all of them
having to do with the 'localedata' package of my current glibc. Any
ideas would be greatly appreciated.
- Pete Lachall
------------------------------
Subject: Re: changing window managers & .Xclients
From: Steve Connet <[EMAIL PROTECTED]>
Date: Thu, 11 Jan 2001 14:52:06 GMT
[EMAIL PROTECTED] (Dave Brown) writes:
> .xinitrc is useful if you do not use a graphical login screen (via
> xdm or gdm or kdm or whatever); it's only invoked if you "startx".
> .Xclients takes its place for a graphical login.
I am not quite sure I understand because I am using run level 3, no
graphical login. I type 'startx' to go into X.
--
Steve Connet ([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 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
******************************
