Linux-Misc Digest #605, Volume #26 Thu, 21 Dec 00 18:13:02 EST
Contents:
Re: Do Linux ext2 partition need defrag? ("Peter T. Breuer")
Re: Mr. No swap ("Peter T. Breuer")
Re: MODEM SPEED; chump kppp retarded it, no? ([EMAIL PROTECTED])
Re: MODEM SPEED; chump kppp retarded it, no? (John Hasler)
Re: not a valid block device (Peter Moore)
Re: Simple question.... (Lew Pitcher)
Re: not a valid block device ([EMAIL PROTECTED])
Re: Simple question....
Re: ATI Mach64 Framebuffer
SB-Live! , emu10k1.o and alsa ([EMAIL PROTECTED])
Re: Alias under Bash shell not working. (Floyd Davidson)
loopback /dev file for CD-RW (David A. Robinson)
mkreiserfs fails with drive errors ([EMAIL PROTECTED])
Re: loopback /dev file for CD-RW (David)
Re: Setting my hardware clock to atomic clock? (drax62)
Re: Dual Xeon hangs ("D. Stimits")
Re: anyone got recommendations for laptops? (Bruce Forsberg)
Re: Question About Timing Functions Under Linux (Dave Blake)
Re: Block move error 0x01 ("Peter T. Breuer")
Re: ATI Mach64 Framebuffer (Otto Wyss)
Re: How to change the color in lynx/elvis (Otto Wyss)
argument list too long ([EMAIL PROTECTED])
Re: init: Id"X" respawning too fast: (etc) (Dave Brown)
Re: (Newbie) Mandrake 7.2 less buggy than Red Hat 7.0? ("John D. Peedle")
Re: Alias under Bash shell not working. ("frank")
Re: argument list too long ("Peter T. Breuer")
----------------------------------------------------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Do Linux ext2 partition need defrag?
Date: Thu, 21 Dec 2000 17:14:43 GMT
Jean-David Beyer <[EMAIL PROTECTED]> wrote:
> Fred Grampp figured out that on a System/360 Model 65 runing PCP (i.e.,
> one task at a time, not multiprogramming) that deliberately fragmenting
> the disk space allocations in a controlled way gained something like a
> 3:1 speed increase when doing sequential reads. Independently, I found
> with an operating system I wrote, that fragmenting the file system "just
> right" got something like a 7:1 increase in sequential reading in some
> cases, cases that occured regularly. I could do that fragmenting
It's surprising, but not completely unexpected. Probably a variant of
a striping effect .. i.e., one process reads block a1, another process
reads block b1, then the first reads block a2, then the second
reads block b2, but everything is striped so that the ondisk sequence
is a1 b1 a2 b2, and hey'presto, speed up.
> separately for each type of file, and the 7:1 improvement was for
> loading programs, which is what was frequently done. Here too, the OS
> was not a multiprogramming one (this was in the 1960's when the hardware
> had either primitive memory management, or none at all, so
> multiprogramming was not really practical).
That's the surprising part.
> BUT, when you run a multiprogramming system with many tasks going on at
> the same time, the fragmentation matters a lot less because even if one
> process is reading sequentially, the others are screwing around with the
> head-positioning and Murphy's Law militates that the heads will be in
> the wrong place most of the time anyway. Improvement in the scheduler of
This is exactly the argument. Hence it is the average closeness of
different files on disk that is the measure of goodness for this case. And
this measure gets better the more fragmented is your file system.
> the disk driver so that it changes the order of the IO requests to
> optimize the movement of the heads works wonders under conditions of
> heavy load. Under conditions of light load, it does not matter much if
> things are fragmented or not. We used to use (for UNIX) what was called
> the "elevator algorithm" for IO request queue management. There was a
> lot of work done on these queue scheduling algorithms and I have no
> assurance that "elevator" is still considered to be the best, and I do
It is. Linux does use an elevator algorithm, and will order up to 128
requests (2.2.*) or 256 requests (2.4.*) in elevator order at a time.
I believe, but am not certain, that it's nowadays a double elevator too
(with tricks to avoid starvation).
> not know what Linux does these days. In those old days, we could see the
> disk drive heads and as the load went up, the heads would do a rather
> uniform scan from one edge of the disk to the center and back.
And one could even write songs using lseek(). Yes, I know.
I haven't mentioned the fact that the disks nowadays present a logical
interface not a physical one. Sector distance doesn't have much
physical meaning in particular cases, but there is an statistical
correlation with the physical media distances, and that0's good enough
for elevator-style algorithms to work .. on average.
Peter
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Mr. No swap
Date: Thu, 21 Dec 2000 17:14:44 GMT
Jean-David Beyer <[EMAIL PROTECTED]> wrote:
> I find it amusing that even kswapd is swapped out. I wonder what that
> means? How does it swap itself back in?
Pass. It means that it doesn't do the swapping, surely!
Peter
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: alt.os.linux.mandrake,comp.dcom.modems
Subject: Re: MODEM SPEED; chump kppp retarded it, no?
Date: Thu, 21 Dec 2000 19:57:05 +0000 (UTC)
On 2000-12-20 [EMAIL PROTECTED] said:
>> Lucent Venus Serial Voice Modem
>A lucent? Why, that's a winmodem!
Not necessarily. The "Venus" chipset is used in serial external modems.
It's a full chipset (controller and data pump). If the modem uses a
Venus chipset (as opposed to Apollo, Mars, or other), it's probably a
hardware modem.
Ron
--
RCD rcdailey at access1 dot net
Net-Tamer V 1.13 Beta - Registered
------------------------------
From: John Hasler <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux.mandrake,comp.dcom.modems
Subject: Re: MODEM SPEED; chump kppp retarded it, no?
Date: Thu, 21 Dec 2000 18:30:41 GMT
Bill Unruh writes:
> I understand the attractiveness for newbies, but if you run into
> problems, they impose an extra layer ( about half of all complaints
> about ppp are caused by bugs in these "helper" programs, rather than in
> ppp itself.)
This is 'pon':
#!/bin/sh
exec /usr/sbin/pppd call ${1:-provider}
No much of a layer. gpppon is a minimal gtk app that calls pon/poff.
--
John Hasler
[EMAIL PROTECTED] (John Hasler)
Dancing Horse Hill
Elmwood, WI
------------------------------
From: [EMAIL PROTECTED] (Peter Moore)
Crossposted-To: alt.os.linux
Subject: Re: not a valid block device
Date: 21 Dec 2000 19:41:45 GMT
>In article <[EMAIL PROTECTED]>,
> moonie;) <[EMAIL PROTECTED]> wrote:
>> On Thu, 21 Dec 2000, glitch wrote:
>>
>> If you read his post he never mentions that it is a parallel port zip
>if it is
>> an internal zip than /dev/hdd4 could be the correct device.
>It's an internal zip, and /dev/hdd4 used to work with the old kernel, so
>I assume that it uses IDE interface. It's just not a block device any
>more.
One thing that sometimes gets me on kernel recompiles is the "Show experimental
drivers"
option near the beginning of config. I've wasted a lot of time wondering where the
heck my
realteck network card driver went to... :-/ Now I always make sure is is set to yes,
and
I'm just carefull about which experimental drivers I add. If you're not sure about
what you
need to add, just add modules for anything resembling what you need -- it doesn't do
anything
but make the compile longer and take up a bit of hd space. I think for Zips you need
(on top
of whatever else) SCSI General (sg.o) support, which you can compile as a module.
Good luck.
-Peter
------------------------------
From: [EMAIL PROTECTED] (Lew Pitcher)
Subject: Re: Simple question....
Reply-To: [EMAIL PROTECTED]
Date: Thu, 21 Dec 2000 19:53:40 GMT
On Fri, 22 Dec 2000 03:28:41 +0900 (KST), [EMAIL PROTECTED]
(Lee soonki) wrote:
>It seems that the 'dmesg' command shows only kernel jobs.
>i.e. the jobs done by init process can't be seen by dmesg after
>booting.
>Is there another command? or another method?
more /var/log/messages
Lew Pitcher
Information Technology Consultant
Toronto Dominion Bank Financial Group
([EMAIL PROTECTED])
(Opinions expressed are my own, not my employer's.)
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: not a valid block device
Date: Thu, 21 Dec 2000 19:54:47 GMT
In article <[EMAIL PROTECTED]>,
Markus Kossmann <[EMAIL PROTECTED]> wrote:
> >
> Did you include "IDE floppy support" in your new kernel config ?
> You should find
> CONFIG_BLK_DEV_IDEFLOPPY=y
> in the .config file created by make xconfig if you did enable this.
Thank you, and everyone who was helping me! I think this should solve
it.
Wroot
Sent via Deja.com
http://www.deja.com/
------------------------------
From: <[EMAIL PROTECTED]>
Reply-To: <[EMAIL PROTECTED]>
Subject: Re: Simple question....
Date: 21 Dec 2000 12:03:09 -0800
[EMAIL PROTECTED] (Lew Pitcher) wrote:
>On Fri, 22 Dec 2000 03:28:41 +0900 (KST), [EMAIL PROTECTED]
>(Lee soonki) wrote:
>
>>It seems that the 'dmesg' command shows only kernel jobs.
>>i.e. the jobs done by init process can't be seen by dmesg after
>>booting.
>>Is there another command? or another method?
>
>more /var/log/messages
I would use something more like
tail -1000 /var/adm/messages | less
1000 might be a bit high, but if you want to be
sure to get most messages in the last cauple of weeks that might
do it. Just using more takes about 10 minutes to page to the
messages recent enough to care about.
------------------------------
From: <[EMAIL PROTECTED]>
Reply-To: <[EMAIL PROTECTED]>
Subject: Re: ATI Mach64 Framebuffer
Date: 21 Dec 2000 12:09:05 -0800
>So, the point is: I have an ATI Xpert@Play 98, 4mb, and I compiled, in kernel
>2.2.18, the atyfb, and no vesafb. It is not written anywhere that you don't
>have to include vesafb: there is no atifb.txt. But there is a matroxfb.txt
and
>there it is said that including a specific fb you need to avoid the generic
>vesa, so I assumed that the same holds for ATI.
Don't use the vesafb....you will not get the most out of the card
>
>Point 1: is this true?
Not sure, didn't quite get it.
>
>Point 2: which numbers do I need to pass to lilo when it asks for vga?
>There are numbers in vesafb.txt (I tried them with a kernel compiled with
>vesa and they work for vesa) and there are numbers in matroxfb.txt, which
>are different from the vesa's ones. So I guess that the numbers depend on
>the fb.
>So, which are the numbers for atyfb?
There are none. If you compiled in the Mach64 support (ATI) you
should boot right into a 640x480x8 fb. There was a program I had
at one time which allows you to change the settings after you boot,
and by modifying the code I think you can set those things in
the kernel. Don't rermember where I got the program, it might
be in the Documentation directory.
I decided that I didn't like it. Using any reasonable settings slowed
down my system incredibly and made funny colors. I tossed the fb.
Hope that helps.
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: alt.os.linux
Subject: SB-Live! , emu10k1.o and alsa
Date: Thu, 21 Dec 2000 20:03:32 GMT
Hi,
snd-emu10k1 is part of alsa-project (www.alsa-project.org), right?
emu10k1 is also part of kernel-2.2.18, right?
Do I need to actually download, compile and install alsa drivers? or
should I try to get emu10k1.o to work instead?
Wroot
Sent via Deja.com
http://www.deja.com/
------------------------------
From: Floyd Davidson <[EMAIL PROTECTED]>
Subject: Re: Alias under Bash shell not working.
Date: 21 Dec 2000 10:29:51 -0900
"Eric en Jolanda" <[EMAIL PROTECTED]> wrote:
>> .bash_profile:
>> alias l='ls -al'
>>
>
>Don't use single quotes. Use " instead
>Put this in:
>
>alias l="\ls -al"
Single quotes are fine, and might even be better.
>and now do
>
>`. $HOME/.bashrc`
Since he put the alias in ~/.bash_profile, sourcing ~/.bashrc
isn't really going to be much value...
However, he should put the alias into ~/.bashrc instead, and make
sure that his ~/.bash_profile sets BASH_ENV=~/.bashrc and also
sources it.
Then he can source it as shown.
>This will source .bashrc, so you don't have to reboot for the changes to
>take effect. You know this is linux!!! Rebooting is what you had to do in
>windows.
>Now all you need to do is use the correct command :-)
>
>Eric
>
>
--
Floyd L. Davidson <http://www.ptialaska.net/~floyd>
Ukpeagvik (Barrow, Alaska) [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (David A. Robinson)
Subject: loopback /dev file for CD-RW
Date: 21 Dec 2000 20:40:49 GMT
Greetings:
I'm trying to use a SCSI CD-RW drive on a Linux 6.0 box, kernel
2.2.5-15smp. I followed the instructions in the CD-Writing HOWTO, and I'm
having one problem: I can't create the loopback device files. When I try
to execute:
do mknod loop$i b 7 $i
I get the error:
./MAKEDEV: device: unknown major number for loop
I gather it doesn't like "7" as a major number. I can create the sg$i /dev
files though.
Any hints/help would be appreciated,
Thanks,
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
David A. Robinson |
The Wharton School, Univ. of PA | '''
2314 SH-DH, 3620 Locust Walk | (o o)
Philadelphia, PA 19104 | -------oOO--(_)--OOo--------
V:(215) 573-3150 F:(215) 898-6200 | [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Subject: mkreiserfs fails with drive errors
Date: Thu, 21 Dec 2000 20:46:15 GMT
Hi,
I'm running 2.4.0-test12 with linux-2.4.0-test12-reiserfs-3.6.23-patch
on a RedHat 7.0 system.
I added a second 640MB drive with one type-83 partition (/dev/hdb1).
Creating an ext2fs system on this partition worked fine. I then tried
mkreiserfs on this partition, and at the point of syncing... I got the
following repeating error that I could not stop without resetting the
machine:
kernel: hdb: write_intr error1: nr_sectors=172, stat=0x61
kernel: hdb: write_intr: status=0x61 { DriveReady DeviceFault Error }
kernel: hdb: write_intr: error=0x04 { DriveStatusError }
kernel: ide0: reset: success
kernel: hdb: set_geometry_intr: status=0x61 { DriveReady DeviceFault
Error }
kernel: hdb: set_geometry_intr: error=0x04 { DriveStatusError }
kernel: ide0: reset: success
kernel: hdb: set_geometry_intr: status=0x61 { DriveReady DeviceFault
Error }
kernel: hdb: set_geometry_intr: error=0x04 { DriveStatusError }
kernel: end_request: I/O error, dev 03:41 (hdb), sector 11380
kernel: hdb: recal_intr: status=0x61 { DriveReady DeviceFault Error }
kernel: hdb: recal_intr: error=0x04 { DriveStatusError }
kernel: ide0: reset: success
Also, after recreating the partition and leaving it unformatted, on
rebooting I got the following note:
kernel: VFS: Can't find an ext2 filesystem on dev ide0(3,65).
Thanks for the help,
Valerie
------------------------------
From: David <[EMAIL PROTECTED]>
Subject: Re: loopback /dev file for CD-RW
Date: Thu, 21 Dec 2000 21:07:19 GMT
"David A. Robinson" wrote:
>
> Greetings:
>
> I'm trying to use a SCSI CD-RW drive on a Linux 6.0 box, kernel
> 2.2.5-15smp. I followed the instructions in the CD-Writing HOWTO, and I'm
> having one problem: I can't create the loopback device files. When I try
> to execute:
>
> do mknod loop$i b 7 $i
>
> I get the error:
>
> ./MAKEDEV: device: unknown major number for loop
>
> I gather it doesn't like "7" as a major number. I can create the sg$i /dev
> files though.
>
> Any hints/help would be appreciated,
>
> Thanks,
Check to see if you don't already have loopback support.
ls /dev/ | grep loop
If you already have loopback it should list loop0 through loop7
Thos link might help.
http://www.guug.de/~winni/linux/CD-Writing/CD-Writing.sgml
--
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 98.908% of seti users. +/- 0.01%
------------------------------
From: drax62 <[EMAIL PROTECTED]>
Subject: Re: Setting my hardware clock to atomic clock?
Crossposted-To: alt.os.linux.mandrake,comp.protocols.time.ntp
Date: Thu, 21 Dec 2000 16:37:00 -0500
Bob Bawcutt wrote:
...
>
> Hi,
> I'm also trying to have my hardware clock use the net for updates but I'm
> a new bee and therefore limited - for now.
> I tried the two command mentioned in this thread (rdate & ntpdate) from a
> Konsole but neither
> command was found. I'm running 7.1 so maybe that makes a difference.
> Any suggests? I'd like to just get it to run from a command line first
> then move up to a cron
> job later. Oh ya, I'm on cable so I don't run PPP.
> Thanks in advance the help
>
> Regards,
> Bob
>
You have permanent internet connection, so you can do the same I did. BTW,
do you try to run "ntpdate" as user or root? ntpdate is in /usr/sbin, so
it's not in ordinary user's PATH.
Do everything below as root.
I created a scipt in /usr/local/bin, called "adj_time.sh":
=======
#!/bin/sh
ntpdate ntp0.freeserve.net && hwclock --systohc
=======
Then, running "crontab -e" add the following line:
0 0 * * * /usr/local/bin/adj_time.sh
This will update your time once a day at midnight.
And, don't forget to "chmod a+x /usr/local/bin/adj_time.sh" !
------------------------------
Date: Thu, 21 Dec 2000 14:59:05 -0700
From: "D. Stimits" <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.hardware,comp.os.linux.setup
Subject: Re: Dual Xeon hangs
ekk wrote:
>
> Sorry for the duplicate post - I noticed I hadn't changed the subject
> line.
>
> ekk wrote:
>
> > Hello
> > I have a dual Xeon 550 machine that hangs often, somewhat randomly when
> > I'm doing some CPU intensive stuff. In the most recent crash, I was
> > heavily using only one CPU, It doesn't seem to be CPU temp related,
> > becuase just before it crashed, the temp was 45.5 deg. C. I'm running
> > RH 6.1, kernel 2.2.14 - configured pretty much the same way as another
> > dual 650 Pentium III which has no trouble. /var/log/messages contains
> > little helpful info. I am running all the same daemons as the 650.
> > Anyone know what's going on? In the meantime, I'm going to put 2.2.16
> > on there to see if that helps.
> > Ken
What chipset does it use? If it is i840 your problem is the broken
IO-APIC of the i840 chipset. In that case boot with kernel option
"noapic". You'd lose performance since hardware interrupts would be
serviced by only one cpu under that option, but stability will go up
dramatically.
------------------------------
From: Bruce Forsberg <[EMAIL PROTECTED]>
Subject: Re: anyone got recommendations for laptops?
Date: Thu, 21 Dec 2000 22:04:07 GMT
In article <912jtf$67o$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> i'm currently thinkning of getting a laptop which i intend to install
> linux on (make is either going to be a sony viao or a panasonic laptop
> since i've heard some good reviews, albiet not 'linux' based reviews)
> and am wondering if anyone out there has any strong recommendations
I am using a Sony Vaio PCG-505TR. I have had it for over a year now
and love it. It is kind of old, it is only a 300 MHz CPU but works
like a charm. Modem works, sound works. A couple of minor annoyances
are: 1) a crack has started to develop in the battery container. 2)
Can't boot from a PCMCIA CD, only from a Sony CD.
One interesting note, they come with 64MB of memory. Using linux this
was very sluggish. I was using KDE as a window environment. I removed
KDE and used blackbox instead. It is much more responsive now and has
saved me a $200 memory upgrade.
Bruce Forsberg
Sent via Deja.com
http://www.deja.com/
------------------------------
From: [EMAIL PROTECTED] (Dave Blake)
Crossposted-To: comp.os.linux.embedded,comp.os.linux.development.apps,comp.os.linux
Subject: Re: Question About Timing Functions Under Linux
Date: 21 Dec 2000 21:15:33 GMT
Reply-To: [EMAIL PROTECTED]
Michel Bardiaux <[EMAIL PROTECTED]> wrote:
> What about the sched_setscheduler & related functions with a
> SCHED_RR scheduling regime? What kind of latencies can one
> expect with that scheme? (My requirements are somewhat less
> strict: my time-critical process must be able to 'beat' at the
> 0.250 and .750 millesconds of every *UTC* second, with better
> than 10 msec accuracy. This is currently running on an SGI
> INDY, with exactly that kind of scheduling, but we hope..
> guess what...)
I assume you mean 250 and 750 milliseconds, not 0.250 and 0.750
milliseconds with 10 msec accuracy.
A really simple C program with select (to get you to within 100
milliseconds) and timing functions (for the fine precision)
could hit 10 msec on an unloaded machine in linux, even in the
worst case scenario. This is provided you do not run huge dd
commands followed by sync, of course.
And if you spec it and it doesn't work, try the low latency
patches. I see them pop up on the kernel mailing list from time
to time, and people spec them at 200 microsecond precision.
--
Dave Blake
[EMAIL PROTECTED]
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Block move error 0x01
Date: Thu, 21 Dec 2000 22:24:40 GMT
Kurt Weber <[EMAIL PROTECTED]> wrote:
> I am installing Slackware with the bare.i bootdisk from the
> slackware-current
> directory. I get the boot: prompt. I hit ENTER, and I get an error
> message reading "Block move error 0x01". What does this mean, exactly?
You or your bios or your memory or your disk is fshGFHFGUFHUFed.
> I've installed Slack on several other systems and have never encountered
> this error.
Well, I guess that rules out slackware and you as the cause then.
Sounds like it can't read or write sector 1 of something. Your floppy,
I suppose.
Peter
------------------------------
From: [EMAIL PROTECTED] (Otto Wyss)
Subject: Re: ATI Mach64 Framebuffer
Date: Thu, 21 Dec 2000 23:22:23 +0100
> So, the point is: I have an ATI Xpert@Play 98, 4mb, and I compiled, in kernel
> 2.2.18, the atyfb, and no vesafb. It is not written anywhere that you don't
>
> Point 1: is this true?
>
I'd like to know that also.
> So, which are the numbers for atyfb?
>
None, just keep vga=normal in lilo.conf. But by the definition of the
image=... add the following kernel parameter
append="video=atyfb:800x600@72"
where "800x600@72" is just an example you have to use one which is
supported be your card.
This could be found in the framebuffer HOWTO.
O. Wyss
------------------------------
From: [EMAIL PROTECTED] (Otto Wyss)
Subject: Re: How to change the color in lynx/elvis
Date: Thu, 21 Dec 2000 23:22:28 +0100
<[EMAIL PROTECTED]> wrote:
> Otto Wyss <[EMAIL PROTECTED]> wrote:
> > Lynx and elvis and probably some more do use too dark colors so I can't
> > read them on my monitor. Where can I change these colors either for all
> > or for each programm?
>
> Edit lynx.cfg.
Thanks I just thought there must be a common place for all programms.
O. Wyss
------------------------------
From: [EMAIL PROTECTED]
Subject: argument list too long
Date: Thu, 21 Dec 2000 22:24:14 GMT
What I want to do is delete a large number of files based on a search
but I get an "argument list too long error". How do I get around this?
I can not do:
rm -fv /var/log/mail/mail.info.*
or I get "argument list too long"
I can not do:
for i in $(ls /var/log/mail/mail.info.*); do rm -fv $i; done
or I get "argument list too long"
I can not do:
ls -x /var/log/mail/mail.info.* > deleteme
or I get "argument list too long
Any ideas? Can I bump up my enviromental space some how?
--
MM
Sent via Deja.com
http://www.deja.com/
------------------------------
From: [EMAIL PROTECTED] (Dave Brown)
Subject: Re: init: Id"X" respawning too fast: (etc)
Reply-To: [EMAIL PROTECTED]
Date: 21 Dec 2000 17:03:37 -0600
In article <3a421410$[EMAIL PROTECTED]>, [EMAIL PROTECTED] wrote:
>Subject : Id"X" respawning too fast: disabled for 5 minutes;
>kdm: Can't lock pid file /var/run/xdm.pid, another xdm is running (pid 827)
>
>Is there anyone has idea why the above error message repeated occurs in my
>/var/log/messages ?
You might have 2 lines in your /etc/inittab which are trying to start
display managers, i.e., xdm or gdm and kdm. Anytime a process named
in inittab (with "respawn") tries to start, and can't (only one display
manager can be running at once), it dies, and init tries to respawn it.
After a few tries, init realizes there's a problem and gives you a
chance to fix it by not respawning it for 5 minutes.
It might be easiest to boot into runlevel 3, in which no display manager
will be started, and then fix your inittab. You can control this at the
lilo prompt by typing "<label> 3" where <label> is whatever you called
your linux image (maybe, "linux").
--
Dave Brown Austin, TX
------------------------------
From: "John D. Peedle" <[EMAIL PROTECTED]>
Subject: Re: (Newbie) Mandrake 7.2 less buggy than Red Hat 7.0?
Date: Thu, 21 Dec 2000 22:58:12 -0000
"Bob Reinkemeyer" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> I am using mandrake 7.2 on a amd k62 266 with no problems. 7.2 has been
> very stable so far. I also think you have more choices than redhat.
>
> Bob
>
> Garry Heaton wrote:
> >
> > I'm about to select my first Linux/UNIX OS and have narrowed-down the
choice
> > to Mandrake 7.2 or Red Hat 7.0. I've heard Red Hat 7.0 is buggy. Is
Mandrake
> > therefore more stable? Are there any significant differences between the
2?
> > I tend towards the ease of use end of the spectrum as I want to
concentrate
> > on Perl development not OS expertise.
> >
> > Also, can I run Apache for desktop development with both of these
options? I
> > read somewhere that Mandrame 7.2 is designed for the desktop, not the
> > server, but I understand by this that I will still be able to use it as
a
> > desktop development environment with Apache but not as a server machine.
> > Please correct if this isn't the case.
> >
> > Thanks
> >
> > Garry Heaton
Go SuSE 7.0
--
John D. Peedle
RHCE - so I'm Biased!
Registered Linux User 167460
------------------------------
From: "frank" <[EMAIL PROTECTED]>
Subject: Re: Alias under Bash shell not working.
Date: Thu, 21 Dec 2000 22:57:57 GMT
Thank you all for the great help. I put my alias line in my .bashrc instead
of .bash_profile. Now everything works.
Thanks again and have a wonderful holiday.
Frank
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: argument list too long
Date: Thu, 21 Dec 2000 23:04:46 GMT
[EMAIL PROTECTED] wrote:
> What I want to do is delete a large number of files based on a search
> but I get an "argument list too long error". How do I get around this?
man xargs. Unix faq?
next!
Peter
------------------------------
** 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
******************************
