Linux-Misc Digest #61, Volume #25 Fri, 7 Jul 00 02:13:04 EDT
Contents:
Re: 'sendmail' is hanging on bootup.... (John Carroll)
Re: Linux Crashes ([EMAIL PROTECTED])
Re: No space left on device ([EMAIL PROTECTED])
Re: umounting a drive on shutdown ([EMAIL PROTECTED])
Re: Debian 2.1 upgrade failed !!! ([EMAIL PROTECTED])
Re: repeat over several lines ([EMAIL PROTECTED])
Re: How do I uncompress .tar.bz2? (Brad Tarver)
Re: Startup Fortunes (Brad Tarver)
NTP trouble... (Stuart Rauh)
Re: (JOB) Seeking Unix Engineers for Migration/Porting ("Andrew N. McGuire ")
why make clean before making kernel? (root)
Re: what changed my /tmp's write permissions? (Robert Nichols)
Re: The Big Dogs and the Tech Shitzus. ("Brian")
----------------------------------------------------------------------------
From: John Carroll <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.networking
Subject: Re: 'sendmail' is hanging on bootup....
Date: Fri, 07 Jul 2000 11:30:25 +0800
Hi, maybe this quote from the RedHat v5.2 manual mighe help:
============== cut here ===================
"
If after the install the machine seems to hang when it reaches certain
processes like sendmail, apache, or SMB there is probably a network
problem. The most common cause is that Linux can not look up the name of
the machine you have called the box (if you set up networking to have a
machine name). The machine is currently paused waiting for the network
timeout of DNS lookups, and will eventually bring up the login prompt.
Login in as root and check the usual culprits for a problem.If you are
directly on a network with a DNS server, make sure the file /etc/resolv.
conf has the correct values for your machine's DNS server. Check with
your systems administrator that the values are correct. If you are using
Linux on a network without a DNS server (or this box is going to be the
DNS server), then you will need to edit the /etc/hosts file to have the
hostname and IP address so that the lookups will occur correctly. The
format of the /etc/hosts file is:
127.0.0.1 localhost localhost.localdomain
192.168.200 1 mymachine mymachine.mynetwor.net
"
============ to here ===============
jc
Michael Nadler wrote:
>
> I think sendmail hangs because your network interface is down. Its
> requests have to time out before the boot up can continue. I would like
> to know the "solution" however.
>
> Hendrix wrote:
>
> > Hi guys,
> >
> > I read a message awhile ago about 'sendmail' hanging at system
> > bootup... Someone posted a solution, but I've forgotton what it was...
> > Anyone having this problem now?
> >
> > I don't think I made any changes to the system at all... One minute it
> > was working fine, the next 'sendmail' would hang for 2-3 minutes when I
> > bootup...
> >
> > Sincerely,
> > --
> > Trevor Penney,
> > A+, Network+ Certified
> > ----------------------
> > "That's alright, I still got my guitar"...
> > -James Marshall Hendrix (11/27/1942-09/18/1970)
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Linux Crashes
Date: Tue, 4 Jul 2000 21:45:05 -0700
On or about 28 Jun 2000 18:06:21 GMT, Richard Goldberg <[EMAIL PROTECTED]> scrivened:
> Hi all,
> I've been running linux for years on several different machines. My
> current machine at work is a PII400 with 196 Mb ram running VA Linux's
> version of Red Hat. The only interesting piece of software I'm running is
> VMWare 2.0.1 (with winnt as a virtual machine).
> After a year or so of almost perfect performance, in the past 1-2
> months I've had 10-12 "crashes". Some of these crashes have been something
> strange happening in KDE (like I can no longer interact with any
> windows), so I try to logout and it just hands on me. I switch to a
> virtual terminal and try to shutdown (either log in as root and issue a
> shutdown/reboot command, or do a ctrl-alt-delete) and it gives me some
> error about unable to go to init state *.
This sounds strangely familiar. <g>
What kernel version are your running, and are you by any chance using
Samba to mount filesystems between your Linux and virtual NT (or other)
systems?
There is a fairly obscure bug affecting smbfs and the 2.2.14 kernel.
Because of security problems with *all* 2.2.x kernels < 2.2.16, you
should upgrade to the 2.2.16 kernel anyway.
I had a similar set of crashes every several hours to ~10-12 days
(typically 3-6 days) on a box, had me completely shamed at the office,
and perplexed a number of people who generally know their way around a
Linux box. I've posted several of my bug reports to the Linux kernel
mailing list ([EMAIL PROTECTED]), you may want to compare
your configuration with the information I posted.
If you can, you might want to check your system error or kernel logs
frequently, as I found an "Oops: 0002" generally preceeded the crash by
several hours.
In the process of trying to diagnose the problem, I replaced the *entire*
system -- first memory, then swapping HD into a new box, then migrating
filesystems to a new drive. Kernel upgrade finally fixed it. This
occuring between mid-January and mid-May of this year.
--
Karsten M. Self <[EMAIL PROTECTED]> http://www.netcom.com/~kmself
Evangelist, Opensales, Inc. http://www.opensales.org
What part of "Gestalt" don't you understand? Debian GNU/Linux rocks!
http://gestalt-system.sourceforge.net/ K5: http://www.kuro5hin.org
GPG fingerprint: F932 8B25 5FDD 2528 D595 DC61 3847 889F 55F2 B9B0
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: No space left on device
Date: Wed, 5 Jul 2000 01:35:58 -0700
On or about Fri, 30 Jun 2000 21:18:19 +0100, [EMAIL PROTECTED] scrivened:
> [EMAIL PROTECTED] did eloquently scribble:
>> On or about Fri, 30 Jun 2000 01:53:18 GMT, Anthony Tekatch <[EMAIL PROTECTED]>
>scrivened:
>>> Thank you for your help.
>>> I used fsck.ext2 on /dev/hda5 and there were many problems that it
>>> reported including "deleted inode xxxxxx has zero dtime".
>> You should generally be able to fix these with:
>> $ fdisk -y
> fsck -y you mean?
Yeah, yeah. Fortunately I've got the dwimnwis shell here (early beta)
-- that's do what I mean, not what I say.
--
Karsten M. Self <[EMAIL PROTECTED]> http://www.netcom.com/~kmself
Evangelist, Opensales, Inc. http://www.opensales.org
What part of "Gestalt" don't you understand? Debian GNU/Linux rocks!
http://gestalt-system.sourceforge.net/ K5: http://www.kuro5hin.org
GPG fingerprint: F932 8B25 5FDD 2528 D595 DC61 3847 889F 55F2 B9B0
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: umounting a drive on shutdown
Date: Tue, 4 Jul 2000 22:41:46 -0700
On or about Mon, 03 Jul 2000 14:42:32 +0000, Andrew Williams
<[email protected]> scrivened:
> Hello - can anyone help with this? It's getting on my fscking nerves!
> I've bought a new hard disk and mounted the /usr directory on it. I
> stuck the mount command in one of the rc.d files so that it's there when
> I log in. However, when I try and umount it, even if it's the first
> command I type after logging in it says the device is in use. I don't
> like rebooting with it still mounted - does anyone have any ideas??
In addition to the other suggestion to add the filesystem to /etc/fstab and
remove the mount command from the rc.d scripts, it's generally possible to
mount a drive read-only even while it's currently active:
$ mount -o remount,ro /usr
If the drive complains about being in use, you can track active files with:
$ fuser -mv /usr
man mount and fuser for more info.
--
Karsten M. Self <[EMAIL PROTECTED]> http://www.netcom.com/~kmself
Evangelist, Opensales, Inc. http://www.opensales.org
What part of "Gestalt" don't you understand? Debian GNU/Linux rocks!
http://gestalt-system.sourceforge.net/ K5: http://www.kuro5hin.org
GPG fingerprint: F932 8B25 5FDD 2528 D595 DC61 3847 889F 55F2 B9B0
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Debian 2.1 upgrade failed !!!
Date: Wed, 5 Jul 2000 01:34:26 -0700
On or about 3 Jul 2000 19:46:50 GMT, Barry Samuels <[EMAIL PROTECTED]> scrivened:
> Having been running Debian 2,1 for some time now I decided to
> take the plunge and upgrade to Frozen. I now have a ystem that
> won't run.
> Here is what I did, via HTTP, after making the relevant changes
> to my sources.list:
<snip>
> 8. 2. apt-get -f -u dist-upgrade completes apparently without
> problems.
> 9. Re-boot. Notice that kernel 2.0.36 is still loading, not
> 2.2.12 as it should be. Get as far as checking modules then
> error messages, about undefined symbols, fly past at a rate of
> knots. Machine eventually hangs.
Can you boot single? At boot prompt (press, IIRC, <shift> to get a LILO
prompt), select your desired kernel and type something like:
vmlinuz single <other options>
...where "<other options>" are optional other boot options you might wish
to specify. See the BootPrompt-HOWTO for more info.
> So it appears that, in its wisdom, apt-get did not upgrade the
> kernel. Strange! It appeared that, from the output of
> dist-upgrade, all libraries and other parts of the system were
> upgraded.
It's also possible that the kernel was upgraded but that lilo wasn't
re-run. What kernel images do you have under /boot? However, in general,
I don't believe that kernels are included in Debian upgrades, though I'm
not sure of the truth or reasons here.
> Having checked the list of packages, which are still on the
> drive, there is no mention of kernel image. The only package
> with the word 'kernel' in it is kernel-package. There is no
> mention of kernel image in the output from apt-get -f -u
> --simulate dist-upgrade.
> I though that I might have been able to cure the problem, as I
> have Debian 2.1 on my laptop, and I have the kernel source for
> 2.2.13, by creating a kernel image, together with a modules
> directory, and copying them to the offending Debian partition
> using OS/2.
Rather than this, see if you can't get make-kpkg. This debianizes the
kernel build/install process. Pretty slick, mostly useful.
> I have tried Tomsrtbt floppy as that has both lilo and libc on
> it. So after mounting my Debian partition on /mnt and entering
> 'lilo -r /mnt ' lilo does run but displays the following:
> 'first boot sector doesn't have a valid LILO signature'.
Confirm that you're doing what you think you're doing. lilo by default
tries to use the *current* boot sector (whatever TRB booted from,
presumably floppy) as its boot sector. I usually chroot to my standard
root partition, then run lilo.
You might also post fdisk output from the relevant disk.
> I wrote all 3 Debian partitions to a 4GB SCSI tape using 'dd'
> before I started so I assume that I could do the reverse and
> write the images back again to restore my original system.
Probably -- if you restore your physical geometry ***exactly***. dd is
*not* a general-purpose archival tool, though it's good for things like
disk images. You're much better off with a tar archive, cpio, afio, or
dump.
> I do not, however, understand what went wrong and even if I
> restored to my original system I don't see why it wouldn't
> happen again. I would certainly prefer not to have to go
> through that 8 hour download again.
Actually, since dd *is* a disk image, it very likely *would* happen again
-- you're replicating every bit on the system.
--
Karsten M. Self <[EMAIL PROTECTED]> http://www.netcom.com/~kmself
Evangelist, Opensales, Inc. http://www.opensales.org
What part of "Gestalt" don't you understand? Debian GNU/Linux rocks!
http://gestalt-system.sourceforge.net/ K5: http://www.kuro5hin.org
GPG fingerprint: F932 8B25 5FDD 2528 D595 DC61 3847 889F 55F2 B9B0
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: repeat over several lines
Date: Tue, 4 Jul 2000 21:53:33 -0700
On or about Mon, 03 Jul 2000 12:07:28 +0000, Ian Mortimer <[EMAIL PROTECTED]> scrivened:
> Hi all,
> how do I repeat a command over several lines in vi ?
> eg. I want to insert a <li> tag at the beginning of 110 lines without
> resorting to .j.j.j.j. etc.
Using a numbered range:
(in command mode)
:.,+109g/^/s//<li>/
...note that to repeat a command over 110 lines, you specify start of
range, then 109 additional lines.
If you prefer marking both the start and end of the range:
(in command mode)
ma # marks line as "a"
(move to end of range)
mb # marks line as "b"
:'a,'bg/^/s//<li>/
If you have some pattern at beginning of line you want to replace with a
<li> tag:
:'a,'bg/^pattern/s//<li>/
To prepend "<li>" to the pattern:
:'a,'bg/^\(pattern\)/s//<li>\1/
...vi is cool. But you'll be twisted by it. <g>
--
Karsten M. Self <[EMAIL PROTECTED]> http://www.netcom.com/~kmself
Evangelist, Opensales, Inc. http://www.opensales.org
What part of "Gestalt" don't you understand? Debian GNU/Linux rocks!
http://gestalt-system.sourceforge.net/ K5: http://www.kuro5hin.org
GPG fingerprint: F932 8B25 5FDD 2528 D595 DC61 3847 889F 55F2 B9B0
------------------------------
From: Brad Tarver <[EMAIL PROTECTED]>
Subject: Re: How do I uncompress .tar.bz2?
Date: Thu, 6 Jul 2000 23:14:57 -0500
if you are using redhat (which uses a newer version of GNU tar), use:
tar -xfI filename.tar.bz2
if you are using slackware, use:
tar -xfy filename.tar.bz2
On Fri, 9 Jun 2000, Mail and News lists wrote:
| If you are not using too old a distribution it should be on your box. try
| bunzip2 file.bz2
|
| [EMAIL PROTECTED] wrote:
| > does one uncompress .tar.bz2 exactly? I'm used to .tar.gz or some other
|
|
------------------------------
Crossposted-To: alt.os.linux
From: Brad Tarver <[EMAIL PROTECTED]>
Subject: Re: Startup Fortunes
Date: Thu, 6 Jul 2000 23:23:37 -0500
at the bottom of /etc/profile:
echo
/usr/games/fortune
echo
or put /usr/games in your path in the PATH line at the beginning of
/etc/profile, then put:
echo
fortune
echo
at the end of /etc/profile
On Sat, 10 Jun 2000, Jeff Craig wrote:
| Okay, this is really kind of not important, but I am pretty sure I saw
| how to set up linux to run the fortune program everytime a user logs on,
| supplying them with a random quip, piece of knowledge, whatever. I
| can't remember how to do it, and I was hoping someone else would be able
| to tell me. I haven't been able to find it, and I've tried.
|
| TIA
|
| --
| Jeff
| --------------------
| "The Earth is the cradle of the mind,
| but man cannot stay in the cradle forever."
| -- Konstantin Tsiolkovsky
|
|
|
|
------------------------------
From: Stuart Rauh <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,comp.os.linux.redhat
Subject: NTP trouble...
Date: Fri, 07 Jul 2000 04:30:05 GMT
Using Redhat 6.2 and Xntp-5.93-14 to keep the server clock on time.
Xntpd doesn't complain but doesn't update the clock either. When trying
to update manually with "ntpdate clock.psu.edu" it responds with "No
server suitable for synchronization found".
I can ping clock.psu.edu from the Linux box. An NTP client on a Win98
machine connected through IP masq on the Linux box updates fine. I just
can't get the Linux box to sync up!
Any suggestions???
--
Stuart Rauh
[EMAIL PROTECTED]
------------------------------
Crossposted-To: comp.unix.aix,comp.unix.solaris,comp.sys.hp.hpux,comp.sys.dec
From: "Andrew N. McGuire " <[EMAIL PROTECTED]>
Subject: Re: (JOB) Seeking Unix Engineers for Migration/Porting
Date: Thu, 6 Jul 2000 23:30:19 -0500
On Thu, 6 Jul 2000, Kevin Galloway quoth:
[] Unix Engineers
[] ============
[] We have a need for high grade Unix people with 5+ years C/C++ programming
[] experience under AIX, Solaris, HP-UX, Tru64 or Linux.
First, 'UNIX' ne 'Linux'. Second, this, nor any of the other groups to
which you posted are meant for jop postings. So, in essence, go bother
people who want to be bothered with job postings. You may also want to
consider brushing up on your Netiquette as well.
Regards,
anm
--
/*------------------------------------------------------------------------.
| Andrew N. McGuire |
| [EMAIL PROTECTED] |
| perl -le'print map?"(.*)"?&&($_=$1)&&s](\w+)]\u$1]g&&$_=>`perldoc -qj`' |
`------------------------------------------------------------------------*/
------------------------------
From: root <[EMAIL PROTECTED]>
Subject: why make clean before making kernel?
Date: 06 Jul 2000 21:47:52 -0700
To build a new kernel I do something like this:
make xconfig && make dep && make clean && make && make zImage && \
make modules && make modules_install
As far as I know this is standard procedure.
But why the make clean? Isn't that missing the whole point of
makefiles?
-chris
------------------------------
From: [EMAIL PROTECTED] (Robert Nichols)
Subject: Re: what changed my /tmp's write permissions?
Date: Fri, 7 Jul 2000 02:48:22 GMT
Note: E-mailed *and* posted.
In article <[EMAIL PROTECTED]>,
Neil Zanella <[EMAIL PROTECTED]> wrote:
[SNIP]
:[nzanella@tulip nzanella]$ ls -ld /tmp
:drwxrwxr-x 17 nzanella nzanella 12288 Jul 6 14:39 /tmp
:[nzanella@tulip nzanella]$
[SNIP]
:How is it possible that an accident occurred to change not
:only the permissions but also the ownership of the directory?
:I do not use any particular backup/restore tools other than
:tar and gzip. Could it be that some buggy script with the setuid
:bit changed the permissions and ownerships?
When run with root permissions, 'tar' will set both ownership and
permissions on files and directories. It might seem odd to find your
own user name listed as the owner, but since your numeric UID is
probably 101, this just means that you extracted a tarball created by
someone with UID=101 on some other machine. Or, you might have
extracted one of your own tar archives, but I'm presuming you'd remember
having done that.
Linux does not support setuid on shell scripts.
--
Bob Nichols [EMAIL PROTECTED]
PGP public key 1024/9A9C7955
Key fingerprint = 2F E5 82 F8 5D 06 A2 59 20 65 44 68 87 EC A7 D7
------------------------------
From: "Brian" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.setup
Subject: Re: The Big Dogs and the Tech Shitzus.
Date: Fri, 07 Jul 2000 05:23:22 GMT
Hey C.J.:
C.J. wrote in message <3964d807$0$8310$[EMAIL PROTECTED]>...
>In article <6TW85.81871$[EMAIL PROTECTED]>, "Brian"
<[EMAIL PROTECTED]> wrote:
>>Seems you be happy with Redhat - works for me!
>I've been pretty satisfied in general, but I'm considering
>looking around a little at other distros.. maybe revisit
>Slackware.
I have been using Slackware for over 6 years - guess that makes me a loyal
user. I have tried Redhat (OK), Debian (Excellent) and a number of the
derivatives (Redhat - Caldera, Debian - Corel) and prefer Corel for the
excellent desktop and GUI apps.
>I getting a little irritated with the
>"Microsoft-ish" practice of deciding what settings I want
>when I upgrade versions of RPMs or of RedHat.
Been there. Done that. Got the T-Shirt.
>My last update
>was a patched Apache which murdered my web server by changing
>the httpd.conf in a way that broke it. Unlike Windows,
>however, it did leave a backup copy of the config
>which I was able to compare to find what it had changed.
I prefer upgrading by hand - I live for tarballs! In the end it is very
quick compared to even a couple minutes of trouble-shooting!
>Hehe.. The first PC I actually owned was a Timex Sinclair
>ZX81. It was a single-unit with a membrane keyboard that
>used your TV for a video display and a tape recorder for
>storage. It had a wopping 2k RAM but I eventually got the
>upgrade to 16k. I later upgraded to a Comodore 64 and
>wondered how programs could realistically take up more
>than 64k of RAM.
I still have a couple Sinclair/Timexs around - they are actually becoming
collectable if you have the manuals, box and memory modules. I skipped the
C64 (my kids got one) and went for the classic IBM PC (64K memory on board).
This puppy has builtin BASIC interpreter, cassette tape mass storage
(really - for BASIC), came with a 320K full-height floppy disc ($300 extra)
later upgraded to 360K.
>>There are many utilities, services and facilities under Linux that have no
>>equal on any Windows OS - they just require something of a learning
process.
>Agreed. There are already things that *I* can do with Linux that simply
are
>not available with Windows or are too dificult to implement. For the
non-tech
>users I deal with (mostly family), short of multimedia support the
>windows-or-linux issue is a toss-up. Unfortunately, the few things they
can't
>do on Linux (like make free long distance phone calls) are important enough
>that they won't switch yet.
Agreed. sigh...
>Speaking of free/cheap long-distance calling, when is even one of these
>services going to support Linux? Several mention "some day" supporting
Mac,
>or that Mac support is in the works, but none cover Linux except how to
enable
>Windows clients to work through a IP-MASQ server.
?
Doesn't Dialpad.com run on the browser in a java virtual machine? Will it
not run on Netscape?
At any rate, I always have a couple nodes that dual-boot into W98 for just
such occasions.
Best regards,
Brian
------------------------------
** 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
******************************
