Linux-Misc Digest #397, Volume #27 Mon, 19 Mar 01 14:13:02 EST
Contents:
Re: Redirect stdout from background process started in shell script??? ("Peter T.
Breuer")
netscape (hmmmmm)
Re: nfs startup (Brad Bailey)
Re: Initiating kernel crash ("Chris Divine")
Re: Initiating kernel crash (Michael Heiming)
Re: Lilo w/ Win98 and RH7, adding Win2k
Re: programming with gcc. (Charlie Dyson)
Free Desktop WorkSpot.com and LD_LIBRARY_PATH problem
Linux Partitioning Problems, HD gone (LSC)
Re: good newsreader? (Philip)
Re: ssh question ("Tom Edelbrok")
Re: where to set harddisk geometry (peter pilsl)
Re: where to set harddisk geometry (peter pilsl)
indent in emacs --please help me-- ("[Bad-Knees]")
Help for "make bzImage" of 2.4.2 kernel? (OrangeDino)
Re: help SOS - less / terminal (John Thompson)
----------------------------------------------------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Redirect stdout from background process started in shell script???
Date: Mon, 19 Mar 2001 18:57:41 +0100
Pat Hennessy <[EMAIL PROTECTED]> wrote:
> Here is the offending line:
> xterm -e ./process > /tmp/myPipe &
> It turns out that it's not a script problem. I have issued this command
It doesn't turn out that way at all.
> from a terminal window and the redirection fails in that case also. It
It works perfectly in both cases. Your problem is that xterm produces no
useful output on stdout.
> seems that the xterm program cannot process the redirection.
?? programs (whether xterm or not!) never process redirections. That's
what the shell does. You seem to be confused.
> What I am trying to accomplish is the ability to have my users start a
> suite of programs by just double-clicking on an icon. The icon on the
All the icon has to do is run a little shell script that
starts all the right things up in the background.
> desktop is a link to a script file that starts each program in its own
> xterm with predefined size/location. The xterm windows will display
That's exactly right.
> stdout, but all the programs' stderrs will be redirected/multiplexed
Eh? What programs' stderrs? How?
> into a single named pipe. Some other program will read the named pipe
> for an integrated log/display (on a different machine).
> Thanks for any suggestions.
do what you said to do:
xterm -e "sh -c './process 2>/tmp/myPipe' " &
and watch out for stdin.
Peter
------------------------------
From: hmmmmm <[EMAIL PROTECTED]>
Subject: netscape
Date: Mon, 19 Mar 2001 18:14:39 +0000
hello
have a query about netscape it works okay in root but is unable to run in any
user areas as it comes up with bus error
anyone have any suggestion as to how to fix it so as it can run in the user
areas as well as root this is needed
thanks
bekka
------------------------------
From: [EMAIL PROTECTED] (Brad Bailey)
Subject: Re: nfs startup
Reply-To: [EMAIL PROTECTED]
Date: Mon, 19 Mar 2001 18:17:46 GMT
In <[EMAIL PROTECTED]>,
Sudhakar R. <[EMAIL PROTECTED]> wrote:
:I've been trying to get nfs to start up on my RH 7.0 box and for some
:reason it just won't.
:
:/etc/rc.d/init.d/nfs start
:
:executes with no messages whatsoever. But when i check the ps -aex, nfs is
:not listed. On doing a
:
:mount -t nfs localhost:/usr /nfstest/
:
:i get the following error
:mount: RPC: unable to receive; error = Connection refused
[snip]
Is your portmapper running? NFS uses a couple of RPC services, which
require portmapper, but the two are controlled seperately.
/etc/rc.d/init.d/portmap start
*THEN*
/etc/rc.d/init.d/nfs start
Regards, Brad.
--
"For every winner, there are dozens of losers. Odds are you're one of
them." --despair.com
------------------------------
From: "Chris Divine" <[EMAIL PROTECTED]>
Subject: Re: Initiating kernel crash
Date: Mon, 19 Mar 2001 10:11:32 -0800
<[EMAIL PROTECTED]> wrote in message
news:kUrt6.898$[EMAIL PROTECTED]...
> Michael Heiming <[EMAIL PROTECTED]> writes:
> > Chris Divine wrote:
> >> This may sound stupid, but it's for testing purposes. Which running
> >> process on a RH 6.2 server install can be "killed" to initiate a
> >> relatively safe kernel crash and hang the system? I'm testing the
> >> capability of initiating an automatic system reboot if the kernel
> >> becomes unstable or the system crashes.
>
> > Sorry, but Linux is not designed to became unstable or even crash, I
> > never saw this happen in years, only if your hw is faulty/your power
> > supply/USV go down or you mucked the system up dreadful, Linux will
> > crash and in this case your automatic system reboot wont help.
>
> This overstates things a mite. If you're using the IDE-SCSI emulation
> mode, and kill (say) a CD writing session, the SCSI "pseudo-bus" can
> get confused enough that the system can and will "hang."
>
> I would certainly agree that a _software-based_ "auto reboot" scheme
> is not likely to fly terribly well; if the system is hosed enough that
> user space stuff is halted, it is unlikely that there will be enough
> there for the "auto reboot."
>
> On the other hand, a hardware "watchdog" would be entirely sensible to
> try out.
>
> There's kernel support for this; basically, you create /dev/watchdog,
> and set up a daemon that writes to that file once per minute.
> (Probably somewhat more often than that to prevent any momentary
> Unfortunate Pauses from leading to a reboot...)
>
> The hardware side accepts a signal from /dev/watchdog, and basically
> resets the timer count every time. If the timer reaches 1 minute (or,
> probably, some other configurable schedule), the computer is switched
> off and on again.
>
> Methinks the best way of getting a "crash" would be to pick from:
>
> a) Killing the watchdog daemon
This is exactly what I'm doing. There is a watchdog timer running and I
found I can initiate a reboot by killing the watchdog daemon. Thanks for
your input.
------------------------------
Date: Mon, 19 Mar 2001 19:23:53 +0100
From: Michael Heiming <[EMAIL PROTECTED]>
Subject: Re: Initiating kernel crash
[EMAIL PROTECTED] wrote:
>
> Michael Heiming <[EMAIL PROTECTED]> writes:
> > Chris Divine wrote:
> >> This may sound stupid, but it's for testing purposes. Which running
> >> process on a RH 6.2 server install can be "killed" to initiate a
> >> relatively safe kernel crash and hang the system? I'm testing the
> >> capability of initiating an automatic system reboot if the kernel
> >> becomes unstable or the system crashes.
>
> > Sorry, but Linux is not designed to became unstable or even crash, I
> > never saw this happen in years, only if your hw is faulty/your power
> > supply/USV go down or you mucked the system up dreadful, Linux will
> > crash and in this case your automatic system reboot wont help.
>
> This overstates things a mite. If you're using the IDE-SCSI emulation
> mode, and kill (say) a CD writing session, the SCSI "pseudo-bus" can
> get confused enough that the system can and will "hang."
OK, you're right, but I didn't thought of running Linux with some cheap
crap IDE writer, build in, more, run it as server in a rack
with USV and so on, I've used watchdog, but it had never anything to do,
cause there was nothing, only hw related crashes, were nothing could help
you and power related crashes, were you only can setup/test your system probably,
that every service will be restarted 100%.
What I wanted to express, there are many interesting things around Linux,
but this seems to me a way to waste time. He should better check the ha-linux
and the Linux Virtual Server project....-)
Best regards
Michael Heiming
------------------------------
From: <[EMAIL PROTECTED]>
Subject: Re: Lilo w/ Win98 and RH7, adding Win2k
Date: Mon, 19 Mar 2001 18:30:47 -0000
Thanks for the replies so far.
So...
Reinstall 98, linux boot disk and /sbin/lilo, then install w2k (on it's
own partition is ok, right?). Then add:
other = /dev/hdxx
label = win2k to lilo.conf.
Then backup MBR, then /sbin/lilo again? That'll give me 3 options in lilo?
What about the w2k loader? Or when I install w2k will it see 98 and let me
pick 98 or w2k? And the other = /dev/hda will actually be pointing at the
w2k boot loader? Like lilo (linux or dos)-->w2k loader (98 or w2k)?
I'm a bit confused at this part. A bit more help...
Thanks!
Eric wrote:
>
>
> > I have Windows98 and RedHat7 dual booting ok now. I'm going to
reinstall
> > 98, and eventually would like to add Win2k.
>
> make a bootfloppy
>
> > How do I put lilo back on the MBR after reinstalling 98? Just boot w/
a
> > linux boot disk and then /sbin/lilo?
>
> Yes. If windows respects the partitiontable.
> Back-up any important data, just in case...
>
> > From everything I've read, I guess it would've been better to go
98,Win2k,
> > then linux, but I didn't have Win2k yet. So, how's this work now?
>
> It would have been easier.
> But it isn't impossible to install in a different order.
> Do you want all OS's on one HDD?
>
> If you use more than one HDD, things may get a bit different.
>
> > Will lilo let me pick 98, Win2k, or linux if I edit lilo.conf to add
> > other = /dev/hdxx
> > label = win2k ?
> > Or is it much more complicated, something to do with Win2k's boot
loader?
>
> No, it will probably be as simple as this.
> But make a copy of the MBR before you run /sbin/lilo again.
> So install NT, boot to linux, with a bootfloppy and backup the MBR
> (dd if=/dev/hda of=/boot/hda_MBR_ntldr.img count=1)
>
> If something fails, and lilo cannot boot the other OS's, just put the old
> MBR back (dd if=/boot/hda_MBR_ntldr.img of=/dev/hda count=1)
>
> > Should lilo boot 98, Win2k, or linux? Or linux or Win2k's boot loader,
> > then 98 or Win2k? Or is Win2k's boot loader supposed to be first.
>
> Let lilo live in the MBR and chain it to the NTloader.
>
> Eric
>
>
--
Posted via CNET Help.com
http://www.help.com/
------------------------------
From: Charlie Dyson <[EMAIL PROTECTED]>
Subject: Re: programming with gcc.
Date: Mon, 19 Mar 2001 18:31:35 -0000
Simple solution:
gcc -o myprogram myprogram.c
./myprogram
** Notice ./myprogram **
Why:
Unlike dos, programs in the current directory cannot be accessed
without ./ (meaning in the current directory), unless you add '.' to your
PATH enviroment variable. If you do not like this behaviour, add the
following line to your ~/.bashrc file:
export PATH=$PATH:.
Hope this helps,
Charlie Dyson
Yiu wrote:
>
> I am right now using gcc to compile some C program. I am suppose to use
> -o option to produce a output file for running the program. However,
> when i type the output file name, it reply me with a command not file
> error message. So, how to compile an executable file and run with gcc?
> Please help.
> Thanks.
>
>
--
Posted via CNET Help.com
http://www.help.com/
------------------------------
From: <[EMAIL PROTECTED]>
Subject: Free Desktop WorkSpot.com and LD_LIBRARY_PATH problem
Date: Mon, 19 Mar 2001 18:32:27 -0000
Hi, has anyone messed around with WorkSpot.com? They let you have a free
desktop account that you can get to via the web through VNC..Anyway, I
want to install X-Chat on my account, but they don't have the latest
shared libs for gtk and such, so I got the proper deb packages, extracted
them, so:
1) all the necessary shared libs are in ~/lib
(this is what was missing:
libgmodule-1.2.so.0 => not found
libglib-1.2.so.0 => not found
libperl.so.5.6 => not found
libpython1.5.so.0.0 => not found
libgtk-1.2.so.0 => not found
libgdk-1.2.so.0 => not found
)
2) I type set LD_LIBRARY_PATH=/home/users/blah/lib
3) I run xchat, and still get this message:
./xchat: error in loading shared libraries
libgmodule-1.2.so.0: cannot open shared object file: No such file or
directory
I'm not familiar with debian, but in red hat i simply would set that
LD_LIBRARY_PATH to a dir with the libs in them, like the man pages say,
and everything works just fine...
But, now I do have gmodule in the directory and of course LD_LIBRARY_PATH
has been set, so I have no idea what could be wrong.....any ideas?
Thanks
--
Posted via CNET Help.com
http://www.help.com/
------------------------------
From: LSC <[EMAIL PROTECTED]>
Subject: Linux Partitioning Problems, HD gone
Date: Mon, 19 Mar 2001 18:33:03 -0000
Ok, here we go,
I have a spare AMD K6-300 computer which I wish to turn it into a full
Linux system. I have a WD 2.5 Gig HD on the computer which originally had
Win95 on it. I formatted the whole drive making it clean. I then used
Mandrakes Red Hat 6.0 Linux boot disk and ran the installation program. I
chosed to use Disk Druid to partition my hardrive. I created a 2.4 Gig
Native Linux root partition and a 100 meg swap partition. When I finish the
configuration, the program gave me an error saying "Error reading partition
table for the block device hda: No such file or directory". It forced me to
reboot and I tried running both Disk Druid and Linux fdisk, but both gave
me the same error, even before they allowed me to setup the partitions
like my first time. I rebooted with a win95 statup disk in hopes of
recreating the partitions in Win95 fdisk, but it gave me a error also "no
hardrive found" and gives me three posibilites, a virus, a third party
partitioning program, hd is damaged, but yet BIOS detects it fine. I have
no C drive and cannot fdisk it in Win95 or Linux nor format anything.
Please help, I would hate to trash a good hard drive.
Thanks!
--
Posted via CNET Help.com
http://www.help.com/
------------------------------
From: [EMAIL PROTECTED] (Philip)
Subject: Re: good newsreader?
Date: Mon, 19 Mar 2001 18:41:48 GMT
On Sun, 18 Mar 2001 22:47:07 GMT, lop <[EMAIL PROTECTED]> wrote:
>hello
>i am searching for a good newsreader running under linux. mainly used
>for downloading binaries. so that my win2k comp has not to run the
>whole night. (its nearly impossible to sleep next to my comp)
>most preferrable would be a programm which runs under console, because
>of remote access. (i dont have Xwindows installed and iam not willing
>to install it unless it is really needed)
>please tell me if there is a prog out there which is good.
Try brag. It's not a real news reader, but it's very usefull for downloading
binaries.
--
------------------------------
From: "Tom Edelbrok" <[EMAIL PROTECTED]>
Subject: Re: ssh question
Date: Mon, 19 Mar 2001 18:31:21 GMT
I use TerraTerm Pro from Windows NT to attach to RedHat 6.0 running SSH.
Don't know if TerraTerm Pro works in Win98 but I expect it would. It's free
and works great. PS: normally it just does telnet so if you want SSH you
have to download the SSH extension from their site as well. First install
Terra Term Pro, then unzip the SSH extensions and copy them into the Terra
Term Pro folder, (you will have a new EXE amongst other things, so you will
have to create your own shortcut).
Tom
"news.starzine.com" <[EMAIL PROTECTED]> wrote in message
news:994g6u$[EMAIL PROTECTED]...
> Anyone know where can I download the ssh and how to configure it under
> RedHat 6.2. Can I using ssh under win98 to remote admin. the linux server?
>
>
------------------------------
From: peter pilsl <[EMAIL PROTECTED]>
Subject: Re: where to set harddisk geometry
Date: Mon, 19 Mar 2001 19:43:37 +0100
In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] says...
> > I tried to do in fdisk, but this changes are lost
> > when starting fdisk again. In fact I know less about this things and dont
>
> No they are not lost. fdisk writes the partition table. You cannot
> "lose" that. Are you sure you exited with "w" and not with "q"?
>
definitely. I tried three or four times (and just now again).
I change the h,c,s - parameters in expertmode of fdisk. When showing the
partitiontable, I dont get any errors, the I exit with 'w' and fdisk calls
ioctl() and is syncing disk and when starting fdisk again, the old values
are here again.
> Duh, dunno. What's the difference between logic and physics?
>
hey, I am a mathematican ;) I know all about logic and nothing about
physics ;)
>
> > There is one difference between both harddisks: hda is listed in the
> > machines bios and hdc is not (on ide1 there is also a slave that is not
>
> Well, that would seem to be fatal, no? If the machine on its own
> doesn't see the disk, then nothing else will! Or do you just mean that
> it's already scanned one disk and knows its geometry, but you haven't
> scanned the other? If so, scan it. Set the geometry if you like.
>
> Uh, you can't do that. What do you think master and slave mean?
> Congratulations, your machine now does not work.
> How can you disable a master? What did you do? Put a meat axe through
> it? Or rip it out of the socket? Or did you set it to slave too a
> jumpers? Whatever you did, undo it, and don't do it again.
>
my lightsaber fulfilled its destiny ;)
no, I tell you what I did:
my system is:
ide0: hda as master with lilo on it, hdb is a cdrom
ide1: hdc, hdd
hdd is 45GB and is not recognized by the BIOS (which hangs when trying to
scan it), so I disabled secondary slave in bios and linuxkernel detects
the drive when booting. But hdd is very slow, so I disabled secondary
master in bios and let it detect by linux only, which works fine and hdd
is much faster now (dont ask why, but I've seen this on two different
servers)
So my bios only knows about hda (it needs to know, cause it needs to boot)
and linux takes care about the rest. This dont hurt, I can access hdc and
hdd without any problems for month now, but I after dd-cloning hda to hdc
fdisk complains about wrong boundaries:
Device Boot Start End Blocks Id System
/dev/hdc1 1 1 5512+ 83 Linux native
Partition 1 does not end on cylinder boundary:
phys=(10, 15, 63) should be (10, 254, 63)
/dev/hdc2 1 18 131544 83 Linux native
Partition 2 does not end on cylinder boundary:
phys=(271, 15, 63) should be (271, 254, 63)
/dev/hdc3 18 34 130536 82 Linux swap
Partition 3 does not end on cylinder boundary:
phys=(530, 15, 63) should be (530, 254, 63)
/dev/hdc4 34 1233 9635976 5 Extended
Partition 4 does not end on cylinder boundary:
phys=(1023, 15, 63) should be (1023, 254, 63)
/dev/hdc5 34 67 262552+ 83 Linux native
/dev/hdc6 67 149 665752+ 83 Linux native
/dev/hdc7 149 166 131512+ 83 Linux native
/dev/hdc8 166 198 262552+ 83 Linux native
/dev/hdc9 198 231 262552+ 83 Linux native
/dev/hdc10 231 1233 8050864+ 83 Linux native
but, and I didnt tried this at the beginning, it seems that I can mount
the complained partitions without problems. So maybe all this is only a
philosophical problem of different ways to see things (logical, physical)
But I need to be sure about it, cause when hda fails I need to put hdc in
it and it needs to boot then ....
thnx,
peter
--
pilsl@
goldfisch.at
------------------------------
From: peter pilsl <[EMAIL PROTECTED]>
Subject: Re: where to set harddisk geometry
Date: Mon, 19 Mar 2001 19:49:08 +0100
In article <994m47$cf6$[EMAIL PROTECTED]>, [EMAIL PROTECTED] says...
> > dd if=/dev/hda of=/dev/hdc
> > I used this command to clone a complete harddisk. Both disks are same
> > brand, same model, same everything. Unfortunately fdisk reports heavy
> > partition-errors on hdc now and /proc/ide/idex/hdx/geometry brings out the
> > awful truth:
>
> errors?
> fdisk should be able to handle partitions that don't start/end on cylinder
> boundaries.
> If you plug hdc in as hda, is it solved then?
>
the errors are:
Disk /dev/hdc: 255 heads, 63 sectors, 1232 cylinders
Units = cylinders of 16065 * 512 bytes
Device Boot Start End Blocks Id System
/dev/hdc1 1 1 5512+ 83 Linux native
Partition 1 does not end on cylinder boundary:
phys=(10, 15, 63) should be (10, 254, 63)
/dev/hdc2 1 18 131544 83 Linux native
Partition 2 does not end on cylinder boundary:
phys=(271, 15, 63) should be (271, 254, 63)
and so on ...
Unfortunately the machine is "far away" from me and so it would mean a lot
of work (and downtime) to plug in hdc as hda. But this is the purpose hdc
is for : when hda fails, I want to put hdc instead and boot again.
I have to admit that I didnt even try to mount a partition from the cloned
disk after getting the fdisk-errors. Now I tried and it works. So maybe
this is more a cosmetical error than a real problem.
>
> It's the BIOS.
> change hda settings to LBA addressing, and it wil probably use the same
> logical settings as hdc uses now. It won't be bootable if you change this
> though.
>
The data will be lost or do I only have to reapply lilo after doing this
step (which I'm afraid of ...)
thnx a lot,
peter
--
pilsl@
goldfisch.at
------------------------------
Date: Mon, 19 Mar 2001 18:50:55 +0000
From: "[Bad-Knees]" <[EMAIL PROTECTED]>
Subject: indent in emacs --please help me--
I have tried almost everything to get emacs to make a normal tab when i
press tab. In text-mode it is ok but if i use html-helper-mode, c-mode
or anyting else, it just wont make that indent. If i press C-q <tab>, i
get what i want but, i want to be able to make my own tab-choices
regardless of what mode im in, so please help me....
[Bad-Knees]
when u absolutly dont know what to do next,
it is time to panic...
------------------------------
From: OrangeDino <[EMAIL PROTECTED]>
Subject: Help for "make bzImage" of 2.4.2 kernel?
Date: Tue, 20 Mar 2001 01:41:11 +0800
I have compiled and running my redhat 6.x linux box (it's original 6.0
with something upgrade to 6.1 and some to 6.2) with 2.4.2 kernel.
After someday, I re-compile the kernel to include some function but in
"make bzImage" the following error came:
gcc -Wall -Wstrict-prototypes -O2 -fomit-frame-pointer -o
scripts/split-include scripts/split-include.c
In file included from /usr/include/errno.h:36,
from scripts/split-include.c:26:
/usr/include/bits/errno.h:25: linux/errno.h: No such file or directory
make: *** [scripts/split-include] Error 1
even I re-compile with the kernel configuration of the one I am running!
Can anyone give me some hints?
Thanks for your concern!
------------------------------
From: John Thompson <[EMAIL PROTECTED]>
Subject: Re: help SOS - less / terminal
Date: Mon, 19 Mar 2001 06:57:44 -0600
percy wrote:
> i have a big problem with the error message.
> after a weekend, my linux server has this problem and the vi can run. i
> think this is a terminal problem. now i can not use less / pico / vi ...
>
> [percy@kljc /etc]$ less smb.conf
> WARNING: terminal is not fully functional
> smb.conf (press RETURN)
Check to see how your TERM environment is set (eg "set|grep
TERM"). If it isn't right, change it with "export
TERM=[whatever]
--
-John ([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
******************************
