Linux-Misc Digest #929, Volume #25 Tue, 3 Oct 00 07:13:03 EDT
Contents:
Re: Superblock corrupt! (Eric)
How to upgrade to glibc 2.1.2 ? (GYULAI Mihaly)
Lightweight distribution w/ X? ("Matt O'Toole")
Re: CTRL-ALT-DELETE (Eric)
Which kernel is the best ? (regarding filesystems support) (Agnelo de la Crotche)
Re: Superblock corrupt! (Svend Olaf Mikkelsen)
Re: What do I do with EXIM on a laptop? (Andreas K�h�ri)
Re: How to silence a linux box? (Andreas K�h�ri)
Re: partition table corrupt (Svend Olaf Mikkelsen)
Re: WTF is up with LinuxConf under RH7?!!??! (Silviu Minut)
Re: partition table corrupt (Svend Olaf Mikkelsen)
curious file ? (G P)
Re: Help needed with PostgreSQL ([EMAIL PROTECTED])
Automatic redailing with pppd ([EMAIL PROTECTED])
Autoconf, automake and shared library objects (Simon Stewart)
ISDN problem (Oswald Knoppers)
Re: Implications ([EMAIL PROTECTED])
----------------------------------------------------------------------------
From: Eric <[EMAIL PROTECTED]>
Subject: Re: Superblock corrupt!
Date: Tue, 03 Oct 2000 09:14:11 +0200
Reply-To: [EMAIL PROTECTED]
NoMadis wrote:
>
> Hi,
> some stupid dos program ruined one of my Linux partitions.
> As a result I cannot mount it anymore, and I cannot repair it with the
> ext2fs utilities.
>
> Does somebody know a program that will allow me to repair it?
> (there are some files on it that i need) :(
>
> --
> Greetz,
> Joop
> -----------------------------------------------------------------------
> Joop Bollen. Nuts & Bolts Department, Nomadis Systems, Holland
> We all live under the same sky, but we don't all have the same horizon.
There are very likely more superblocks than just one on that partition.
Have you tried any of the other superblocks?
look at the following locations for duplicate superblocks:
32768
98304
163840
(reported locations where superblock duplicates were stored from a RH6.2
install on my system)
If all else fails, perhaps you can use dd to get data off the partition,
but then it obviously will take a lot of work to sort out what/where the
actual data is that you need.
Ik hoop dat het lukt Joop,
Eric
------------------------------
From: GYULAI Mihaly <[EMAIL PROTECTED]>
Subject: How to upgrade to glibc 2.1.2 ?
Date: Tue, 03 Oct 2000 07:18:21 GMT
I tried to upgrade from glibc 2.0.7 to 2.1.2, but I couldn't compile
it, I receive the following message at 'make':
undefined reference to __sigalstack, __libc_global_ctors
I included the kernel-headers 2.2.15 at 'configure', and I placed the
new glibc source into /usr/src/glibc (which was recommended by its
docs).
Any ideas? Please help, I can't compile anything!
--
GYULAI Mihaly
http://gyulai.freeyellow.com
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
Reply-To: "Matt O'Toole" <[EMAIL PROTECTED]>
From: "Matt O'Toole" <[EMAIL PROTECTED]>
Subject: Lightweight distribution w/ X?
Date: Mon, 2 Oct 2000 23:07:24 -0700
Is there a distribution with X that's especially for old, slow machines?
I'm looking to install on some old 486s with 16MB, and I know this will work
OK as long as I keep everything to a minimum, and use a lightweight window
manager like IceWM. Has anyone packaged something like this? Most
distributions install everything but the kitchen sink, with it all up and
running by default. All I want to run is Netscape, even an old version, and
maybe some lightweight desktop apps. What's the best approach?
Matt O.
------------------------------
From: Eric <[EMAIL PROTECTED]>
Subject: Re: CTRL-ALT-DELETE
Date: Tue, 03 Oct 2000 09:31:51 +0200
Reply-To: [EMAIL PROTECTED]
Jos� Luis Guisado wrote:
>
> Hi all,
>
> There is a line in /etc/inittab which traps the CTRL-ALT-DELETE signal
> and indicates what to do then. For example, in my system:
>
> # Trap CTRL-ALT-DELETE
> ca::ctrlaltdel:/sbin/shutdown -t5 -r now
You could change the -now argument to +1 and then it'll wait 1 minute
after you hit ctrl-alt-delete before it will actually reboot.
Eric
> Here, '-t5' indicates to wait 5 seconds between sending processes the
> warning and the kill signal, before rebooting.
>
> The problem is that my system doesn't wait the 5 seconds when I press
> CTRL-ALT-DELETE if there is no interactive session started (e.g., when
> no one is logged-in into the system).
>
> How can I make the system to wait the x seconds ALWAYS?
>
> Note: The box is a Linux-Red-Hat-5.2
>
> Thanks,
> Jose Luis
------------------------------
From: Agnelo de la Crotche <[EMAIL PROTECTED]>
Subject: Which kernel is the best ? (regarding filesystems support)
Date: Wed, 04 Oct 2000 00:38:05 -0700
Hi !
What is the best 'stable' kernel, which supports BFS (from BeOS, NOT
from Netware), QNX4, and reiserfs ?
I got reiserfs compiled under 2.2.14, but not under 2.3.46 (??? the
module isn't there) .
SCO partition table support (not Unixware) would be great too. I don't
know what does 'sysv' mean but it definitely doesn't read SCO partition.
Anyway , unlike for BSD and Solaris, I cannot see any SCO slice.
A very good enhancement for the future kernels would be the ability to
scan logical drives for BSD disklabels (Yes, I installed a BSD system
into a logical drive... well almost ! but that's tricky)
Agnelo
------------------------------
From: [EMAIL PROTECTED] (Svend Olaf Mikkelsen)
Subject: Re: Superblock corrupt!
Date: Tue, 03 Oct 2000 08:03:34 GMT
[EMAIL PROTECTED] (NoMadis) wrote:
>Hi,
>some stupid dos program ruined one of my Linux partitions.
>As a result I cannot mount it anymore, and I cannot repair it with the
>ext2fs utilities.
>
>Does somebody know a program that will allow me to repair it?
>(there are some files on it that i need) :(
In quite a few cases like this, it is the partition table and not the
partition itself, that is damaged. If you want me to evaluate the
situation, you can initially get Findpart at
http://inet.uni2.dk/~svolaf/utilities.htm
do from DOS or Windows:
findpart all +fat fp.txt
and post the content from fp.txt in a follow-up to this message.
--
Svend Olaf
------------------------------
Crossposted-To: comp.os.linux.help
Subject: Re: What do I do with EXIM on a laptop?
From: Andreas K�h�ri <[EMAIL PROTECTED]>
Date: 3 Oct 2000 10:04:43 +0100
In article <[EMAIL PROTECTED]>,
Jerome Mrozak <[EMAIL PROTECTED]> wrote:
>I've just installed a Debian system and its primary email program seems
>to be the exim package.
>
>My primary experience to date has been with Windows, and I use Navigator
>to handle my email, etc. It means I've not administered my own mail
>before.
>
>I'm aware that my own system will send me mail (login attempts, etc).
>Does this mean I need a mail handler installed, or can I just install a
>mail reader like pine?
>
>Thanks,
>Jerome.
>--
>Jerome Mrozak "Never buy a dog and bark for yourself"
>[EMAIL PROTECTED] --"Slippery" Jim DiGriz
> (the Stainless Steel Rat)
'exim' is the program delivering the mail to your inbox, and from you
to other computers. Apart from it you need a email reader, e.g.
'mutt'.
[As a home user] You shouldn't need to touch 'exim' once you have
configured it with 'eximconfig' as root.
/A
--
Andreas K�h�ri,
Uppsala University, Sweden.
===================================={ GNU it yourself: www.gnu.org }====
------------------------------
Subject: Re: How to silence a linux box?
From: Andreas K�h�ri <[EMAIL PROTECTED]>
Date: 3 Oct 2000 10:16:07 +0100
In article <[EMAIL PROTECTED]>,
Tobias Schenk <[EMAIL PROTECTED]> wrote:
>Hi,
>
>I like to setup a linux box as internet gateway, voice box etc.
>My hardware is a rather old P200-64MB.
>Are there ways to silence the machine? Like switching off the hard
>disks or so? Power saving is also of interest.
>What os-independent methods can be applied ( Dampening the
>housing,...)?
>
>Thank you for your help,
>
>Tobias
>
>P.S: Sorry for my poor English
>
There's an Ecology HOWTO available at
<URL:http://www.linuxdoc.org/HOWTO/Ecology-HOWTO.html>.
It might give you some hints...
/A
--
Andreas K�h�ri,
Uppsala University, Sweden.
===================================={ GNU it yourself: www.gnu.org }====
------------------------------
From: [EMAIL PROTECTED] (Svend Olaf Mikkelsen)
Subject: Re: partition table corrupt
Date: Tue, 03 Oct 2000 08:32:39 GMT
"Will Heyman" <[EMAIL PROTECTED]> wrote:
>I am having a similar problem when I try to install RH6.1
>The drive is new. I installed linux first but had trouble
>installing W98. I used dos fdisk to repartition and was able
>to install W98 but now get partition table corrupt when I
>try to install RH6.1
>
>I ran findpart all +fat and fp.txt is below.
>
>Thanks for any help,
>
>Will Heyman
>[EMAIL PROTECTED]
>[EMAIL PROTECTED]
>
>Findpart, version 3.9.
>Copyright Svend Olaf Mikkelsen, 2000.
>
>Searches for partitions type 01, 04, 06, 07, 0B, 0C, 0E, 82, 83,
>plus Fdisk F6 and Lilo sectors. Information based on bootsectors
>is marked B. If the disk is larger than supported by BIOS, the
>supported part of the disk is examined. Disks are numbered from 1.
>
>OS: DOS 7.10 WINDOWS 4.10
>
>Disk: 1 Cylinders: 2491 Heads: 255 Sectors: 63 MB: 19540
>
>-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS CHS
> 0 - 83 63 48132 23 0 1 1 2 254 63 B0 OK
> 3 2 05 10249470 144585 70 641 0 1 649 254 63 3 OK
> 3 - 0B 63 10249407 5004 3 1 1 640 254 63 B OK
> 641 1 82 63 144522 70 641 1 1 649 254 63 OK
> 641 2 05 10394055 6008310 2933 650 0 1 1023 254 63 3 OK
> 650 1 83 63 29575602 14441 650 1 1 2490*254 63 OK OK
> 0 - 0C 16450560 10249470 5004 1024 0 1 1661 254 63 B OK
>
>-----FAT CHS -Size Cl --Root -Good -Rep. Maybe --Bad YYMMDD DataMB
> 3 1 33 10000 4 2 10000 0 0 0 001001 0
> 1024 0 33 10000 4 2 10000 0 0 0 001001 229
>
>Partitions according to partition tables on first harddisk:
>
>-PCyl N ID -----Rel -----Num ---MB -Start CHS- --End CHS-- BS CHS
> 0 1*0C 16450560 10249470 5004 1024* 0 1 1661*254 63 OK OK
> 0 2 05 48195 39969720 19516 3 0 1 2490*254 63 OK
>
> 3 2 05 10249470 144585 70 641 0 1 649 254 63 OK
>
> 641 1 82 63 144522 70 641 1 1 649 254 63 OK
> 641 2 05 10394055 6008310 2933 650 0 1 1023 254 63 OK
>
> 650 1 83 63 29575602 14441 650 1 1 2490*254 63 OK OK
ID 05 at PCyl 0 is the extended partition. Other ID 05 are links to
next logical partition. If you look at the start and end cylinders,
you can see how the partitions overlaps. ID (hex) 0C is primary FAT32
ending after cylinder 1023, 82 is Linux swap and 83 is Linux ext2. 0B
is other FAT32.
Since no data is in the Linux partitions, I suggest you just delete
the extended partition. You can do this in a Windows DOS box
(will1.bat):
set findpart=edit
findpart 1 0 2 - 00 0 0 0 0 0 0 0 2491 255 63 26
set findpart=
findpart table fp1-1.txt
To prevent errors, download the batch file from
http://inet.uni2.dk/~svolaf/will1.zip
The change will be effective after reboot. It is assumed that you did
not touch the partition tables since Findpart was run.
If you attempt to install Linux again, then note that if an extended
partition ends cylinder 1023 or later, the ID must be 0F or 85 (Linux
extended), not 05. Since no suitable partitioning tool exists for
Linux, you may have to change this manually using Linux fdisk.
--
Svend Olaf
------------------------------
From: Silviu Minut <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,comp.os.linux.setup,linux.redhat.misc
Subject: Re: WTF is up with LinuxConf under RH7?!!??!
Date: Tue, 03 Oct 2000 04:42:33 -0400
> Incidentally, the C compiler is also broken, I hope you werent planning to
> build a custom kernel.
Broken indeed! In fact I only tried g++ and it gave some errors with
templates. However, I installed the compat-egcs, compat-egcs-c++, and
dependencies and all was well (as expected, because egcs worked with RH6.2).
------------------------------
From: [EMAIL PROTECTED] (Svend Olaf Mikkelsen)
Subject: Re: partition table corrupt
Date: Tue, 03 Oct 2000 08:44:50 GMT
[EMAIL PROTECTED] (Svend Olaf Mikkelsen) wrote:
>If you attempt to install Linux again, then note that if an extended
>partition ends cylinder 1023 or later, the ID must be 0F or 85 (Linux
>extended), not 05. Since no suitable partitioning tool exists for
>Linux, you may have to change this manually using Linux fdisk.
That should be: Cylinder 1024 or later.
Cylinders are numbered from zero. Linux fdisk in normal mode however
numbers cylinders from 1, so in Linux fdisk extended partitions ending
cylinder 1025 or later should be made ID (type) 0F or 85.
Due do DOS/Windows bugs:
If the last partition in any extended partition is non FAT, and an
extended partition ID 0F exists on one disk, extended partitions on
other disks in the system must be ID 0F or 85, even if they end
cylinder 1023 or below.
--
Svend Olaf
------------------------------
From: G P <[EMAIL PROTECTED]>
Subject: curious file ?
Date: Tue, 03 Oct 2000 12:51:52 +0200
In my home directory, I find a link file named ...
whith link with itself ... ?
I don't dare to delete it as long I don't know what it is ?
Some help please .
Distibution Mandrake
kernel 2.2.15-4mdk
[patrice@Ramses patrice]$ ls -al ...
lrwxrwxrwx 1 patrice patrice 3 sep 19 16:38 ... -> ...
thanks.
--
Patrice garachon
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.lang.java.databases,comp.databases
Subject: Re: Help needed with PostgreSQL
Date: Tue, 03 Oct 2000 17:16:51 +0800
This is awesome, thanks for the tips!!
Regards,
Damon
Tom Lane wrote:
> [EMAIL PROTECTED] writes:
> > I'm encountering a problem when inserting a record larger than 8kb into
> > my PostgreSQL v7.0.2.
>
> Postgres currently limits rows to the size of a disk page, which is 8K
> by default. Possible workarounds:
>
> 1. Increase disk page size (change BLCKSZ constant, recompile, initdb).
> However you cannot increase it above 32K.
>
> 2. If the problem is a text field, declare it as lztext instead to
> get data compression (only works in 7.0.*). This might get you
> a factor of 2 or so.
>
> 3. Use large objects (BLOBs). Major pain in the neck, but the size
> of any one large object is only limited by your disk space.
>
> 4. Wait for 7.1, which eliminates the problem (essentially by
> automatically, transparently converting oversize fields into
> BLOBs). If you're feeling brave or are not working with
> mission-critical data, you could try a current development snapshot.
> 7.1 beta should be out in a few weeks, final before Christmas.
>
> #1 and #2 could be combined to get you up to circa 64K per record.
>
> BTW, I'm not sure whether our JDBC driver copes with rows > 8K;
> you might need to tweak it.
>
> regards, tom lane
------------------------------
From: [EMAIL PROTECTED]
Subject: Automatic redailing with pppd
Date: Tue, 03 Oct 2000 09:34:57 GMT
Hello,
I'm looking for a script or some way I can have pppd redail my ISP when
the connection goes down. Ok, you can do it by restarting pppd
in /etc/ppp/ip-down but I'm looking for a way so that I have more
control (like redailing x number of times until contact is
re-established, intervals between the different dail-in attempts etc.).
I have a cable-modem (hooked up to one of my serialports, it sucks...)
and I just want to stay on-line as much as possible. But my ISP
refreshes all the IP's like every 24 hours and they have not little
network-problems some days... Thanks.
Jeremy
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: Simon Stewart <[EMAIL PROTECTED]>
Subject: Autoconf, automake and shared library objects
Date: 3 Oct 2000 09:57:35 GMT
I am currently starting to get to grips with automake and autoconf,
and have run into a small problem. Essentially, I want to make a
shared library which is composed of a number of separate parts. These
separate parts are located in subdirectories to keep the code as
modular as possible, leading to a directory structure like:
src--+--subdir1
+--subdir2
+--subdir3
Ultimately, I want to compile the shared library in src from *.lo
files in the various subdirs. Clearly, the right thing to do is to
define a "lib_LTLIBRARIES" in the src Makefile.am, but how do I just
get the *.lo files made in the various subdirs?
Clearly, they could be generated by building a shared library in each
subdir (which would generate the *.lo files as a side effect) but this
doesn't strike me as being very efficient, especially since
"noinst_LTLIBRARIES" is a valid target name in a
Makefile.am. Alternatively, I could simply copy the rule for making
*.lo from *.cpp/*.c from a generated Makefile and add an "all-local"
target which depends on the *.lo files. This isn't ideal, because
"make dist" and "make distcheck" both fail.
Is there a Right Way to do this, or should I simply kludge it?
Cheers,
Simon
--
"Cats don't hunt seals. They would if they knew what they were and
where to find them. But they don't, so that's all right." -- Terry Pratchett
------------------------------
From: Oswald Knoppers <[EMAIL PROTECTED]>
Subject: ISDN problem
Date: Tue, 03 Oct 2000 12:08:17 +0200
Hello,
I am running Mandrake V7.1 with an isdn (ippp0) interface. Once in a
while, sometimes after a month or so and sometimes every week, the ippp0
interface disappears. There is nothing indicating a problem in the
syslog files. I just see the last time a connection was made, no
messages about closing down the link (which I normally receive after a 5
minutes idle period). The interface is just gone.
In order to restart I need to use the command 'ifconfig ippp0 up' and
re-add my default route.
Any ideas on where to start looking?
Thanks,
Oswald
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.setup,comp.software.config-mgmt,comp.ai.gentic
Subject: Re: Implications
Date: Tue, 03 Oct 2000 10:15:45 GMT
> >Buried in the excesses of my deranged
> >thinking, I wrote down something
> >that looks like this:
> >
> > X --> P --> E
> >
> >How amazingly deranged.
>
> That is not deranged, but so simplistic
> as to be too laughably obvious and
> useless to anyone who has ever configured
> a computer. You take some software
> X, put it into store P and you have
> configuration E. Hmm, I think that's not
> clear enough, so let's draw a box and
> arrow diagram! Got any chalk?
DNA --> RNA --> Proteins (function)
This is the Central Dogma of Molecular Biology (CDMB).
It is a biological counterpart of my own illustration,
though I did stop short of referring to my drawing as
the "Central Dogma of Computer Science". But I could have.
If we were talking about understanding how cells
actually manage to configure themselves, you could make
the same claim about CDMB's "laughably obvious" nature.
If the CDMB is so "laughably obvious" and "useless", why
is this progression referred to with reverence in Biology?
Check out the reference (one of thousands on the net):
http://www-
bioeng.ucsd.edu/research/research_groups/gcrg/Presentations/sld002.htm
>
> I think the deranged part was comparing your
> installation to having sex.
>
> You need to stop diddling your computer
> and start taking your medication. :)
Look, I have been taking my meds, honest! :)
I know the idea that computer systems depend on the exchange
of genetic material sounds deranged. But there is no stork.
The functionality of your computer system comes from *somewhere*,
and computer systems don't just appear all configured in
a bundle tied with a ribbon. (Okay, maybe they *do* come all
configured in boxes, if not ribbons, but somebody had to
configure the system before you got it :)
Everyone you know, the food you eat, the grass in your
yard, your pet (if you have one), and nearly every living
thing you see is the result of genetic exchange. What is
genetic exchange? In computer systems, it is the first step,
that of establishing your software library, in the "laughably
obvious" and "useless" process of configuring a computer
system.
You really should try to open your mind a bit. Biology might
be messy, but organisms do a far better job of configuration
management than we do with our computer systems. Cells and
computer systems are both computationally based, and they are
both faced with the same types of configuration problems. My
"laughably obvious" observation implies we should establish a
single set of mechanisms to perform the configuration management
tasks in our computer systems. Furthermore, these mechanisms
will scale to any computer system, quite independent of the
functionality of that system. Why do I believe this? Because
that is what living organisms do, from bacteria to whales.
Better yet, do I have to throw out C and Java, rewrite Linux
to apply the techniques these "dogmas" imply? No I do not. Like
the dogmas state, configuration mangement of a computer system
can always be implemented in a fashion independnent of the
execution environment.
Useless indeed.
Paul Snow
[EMAIL PROTECTED]
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.misc) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Misc Digest
******************************
