Re: Upgraded Potato, now I can't telnet/ssh in
Benjamin Suto <[EMAIL PROTECTED]> writes: > I've been keeping up to date with the Potato release, and I've recently > upgraded to the newest distribution sometime yesterday, March 12. > > Now, when I try to telnet/ssh in, I get an immediate disconnection, and > the following message in my system logs. > > Mar 13 23:00:12 edgy inetd[27942]: getpwnam: root: No such user I think you need to restart inetd and sshd. I had this same problem, so I used it as an opportunity to upgrade to 2.2.3 and reboot, and then everything worked again. -- Jeff Lessem.
Re: Horrid question: ssh or ssl-telnet for Windoze(95)/Doz
George Bonser <[EMAIL PROTECTED]> writes: > There is an ssh client for Windows. Look around, you will find it. I might > start my search at rootshell.com since they seem to have a lot of links to > the manufacturer of ssh. Remember, do not use ssh-2.x, use 1.x. I use (well, I set up for other people...) TeraTerm Pro with SSH patches. The urls are http://hp.vector.co.jp/authors/VA002416/teraterm.html http://www.zip.com.au/~roca/ttssh.html It is very easy to install and setup. The only problems I run into are that occasionally the teraterm.ini sets itself read-only, and to automatically use ssh it is necessary to edit teraterm.ini and set the use ssh line (near the bottom of the file) to 1. -- Jeff Lessem.
Re: Corel® WordPerfect® 8 for Linux® ishere! (fwd)
John Stevenson <[EMAIL PROTECTED]> writes: > Actually this does not work, as you have to agree to the licence > first. Go here to start with: > > http://www.download.com/PC/Result/TitleDetail/0,4,0-42505-g,1000.html?st.dl.fd.hl.g42505 Even easier, just go to ftp://ftp.cdrom.com/pub/simtelnet/cnet/win95/business/GUILG and get GUILG00.GZ, which is a non-compressed tar file, not a gzipped file. I tar -x'ed it into a directory, ran Runme, and it seems to work, mostly. I have not yet been able to register it, but hopefully Corel's traffic will lighten in the next 90 days. -- Jeff Lessem.
Installing identical Debian setups on multiple machines
The first week in March I will be running the computers at a workshop with 25+ PC's all running Linux. All machines will be configured absolutely identical (except, of course, for a few files in /etc, /var, etc.). I will install Debian on one machine and get it setup exactly the way I want. I then need to clone this machine to the other 25 or so computers. I am soliciting opinions on the best way to do this, any advice? Last year when I ran the same workshop I used the Slackware distribution, and installed slackware on each machine to the point where I NFS mounted the Slackware package files. At that time I broke out of the normal install and untared a copy of the clone system from the NFS mount. This worked pretty well because of the way Slackware is installed---boot with two disks, and it is then network ready. Debian however, seems to not be network ready until the tar image is copied off the four base disks. Is there any way to avoid this? Ideally, I would like to trick the Debian install to untar a clone of the complete install from an NFS server, instead of the standard base system from the floppy set. If this is not possible, I guess the next best thing is to go through the base install and then untar the clone instead of running dselect after the first reboot. I really want to avoid any hardware based solutions (put two disks in a machine and tar/untar from one to the other), because I would have to do it all myself. Any solution I can break down into a simple script and and give it to some PFY's with the instructions, ``do this on them there computer,'' is acceptable. Do you folks foresee any other problems I should be concerned about? Thanks, Jeff Lessem. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
mkinitrd in Debian
I noticed that mkinitrd (for building initial ram disks) doesn't seem to be in a Debian package. Have I just overlooked it and it is really hiding in some deb file, or is it really not around? I can install it from scratch easily enough, but I was wondering if there was some reason that this fairly basic Linux tool wasn't present in the thousands of packages that make up a Debian distribution. -- Jeff Lessem.
Re: 2.4.0-test kernels?
In your message of: Tue, 03 Oct 2000 15:44:31 CDT, you write: >Can anyone provide some info on how to get working with a 2.4 kernel? About a month ago I started running 2.4-testX on an otherwise potato system. I had to upgrade modutils, nfs-common, and nfs-kernel-server to make things work. If you aren't running kernel nfs, then you probably will only need modutils. Fortunately I still had the .deb's on the machine, so you can download them at http://ibgwww.colorado.edu/~lessem/2.4/ They are dated 26 Aug 2000, so they aren't the latest, but they did dpkg -i right onto an otherwise potato system. -- Jeff Lessem.
/usr/share, why bother?
I am very much in favor of the strongly adhered to filesystem layout that Debian uses. Having all of the configuration files in /etc, means that when I got a new laptop the only part of the OS I saved from the old to the new was /etc. Stuff that changes is in /var, that is great, right where it should be. Architecture specific stuff not on / goes in /usr. This leaves all of the non architecture specific stuff (text, html, perl scripts, emacs lisp, etc.) to go under /usr/share. I use /usr/local/share to have common TeX, etc. on completely different architectures. Makes things great for system management. I only need to install a new bibstyle in one place and my Linux and Tru64 boxes all find it. This is how it should be. My problem though, is that Debian only seems to go halfway. Just as I cross mount /usr/local/share between all my computers, I want to also cross mount /usr/share between all my Debian boxes. In order for this to make any sense though, there needs to be some sort of dpkg setting to say, "don't install /usr/share on this machine, it is a client." I know, this can lead to problems where a package is installed on a client, but not on the master server. That is my problem as the administrator though, and shouldn't be anybody else's concern. Am I just dense, and such a setting is fully documented and I just need to rtfm? I know, disk space is real cheap, so wasting a few hundred meg per machine isn't a big deal, and there isn't really anything in /usr/share that I, as the administrator, need to be changing, but it is just the principle of the thing. I guess what I mean is, why call it share if it isn't meant to be shared?
Re: Mirrors <<< Re: Debian Weekly News - October 25th, 2000
In your message of: Wed, 25 Oct 2000 18:20:27 EDT, you write: >According to [1], the mirrors will be afected, so to create them it is said the >some "exclude" tweaking must be done. Does anyone have the info -or the script- >for what exactly is meant >Mirroring Potato, getting ready for 2.2_r1 I use the "mirror" perl script to maintain a local mirror of i386 potato and woody. So that I do not download files for all of the architectures I am not interested in, my mirror configuration file for the Debian archives has lines like exclude_patt+|binary-sparc|disks-sparc in it. In the future (if I understand things correctly) there will no longer be binary-$arch directories; all of the binaries will be in the same directories. So that I don't download the .deb files for the architectures I do not want I will have to add some lines like exclude_patt+|sparc.deb|sparc.changes
Re: [OT] Mozilla and JunkBuster
I had to stop using junkbuster a while ago because Mozilla really didn't like using it. Something about http 1.0 and 1.1, but for a long time Mozilla just plain didn't work with proxies. I ended up doing some work arounds to keep the ads from getting too annoying. Obviously, use the Mozilla cookie management stuff. I think this is one of the best features of Mozilla. Block ads from dedicated ad servers using the Mozilla image loading management. I right click on ads, select copy image location, paste that into a scratch buffer and if it is something like ads.annoying.com then I right click on the image and select block images. The previous step only blocks some ads because many load from www.someplacegood.com/ads or something that Mozilla's simple blocking model can't handle. To deal with these I turned off animated gifs in Mozilla. There might be an easier way to do this, but this method works. Open libnsgif.so with a hex editor (emacs hexl-find-file) and search for NETSCAPE2.0 and change a letter or two of that string. Make sure you keep the same number of characters though. The gif library looks for this string in gifs, and if it finds it then the gif animation is looped forever, but if the string is not found the animation is played once and stopped. Since doing these things I really haven't missed junkbuster.
Large file support in Woody
Now that woody is based on glibc2.2 I thought I would play with large file support. Perhaps this was a mistake, as I have managed to create a 3GB file, but I can't remove it. Any access to the file with the fileutils package (ls, rm, etc.) returns with "Value too large for defined data type". I thought maybe fileutils had to just be rebuilt against the new glibc, but after an apt-get source --build fileutils, I still get the same error. I am sure I am missing something quite obvious, so any pointers in the right direction would be greatly appreciated. -- Thanks, Jeff Lessem.
Re: Large file support in Woody
In your message of: Fri, 17 Nov 2000 22:39:01 +0100, you write: >there's a hardcoded 2GB limit in libc. You've to patch libc >and recompile yourself. My boss made that (for writing something >like 70 GB files), but he's not here currently, so I cannot ask him. I thought this limit was removed in for glibc 2.2. I did a dd if=/dev/zero of=bigfile bs=1024k count=3072 and it worked fine. Previously this blew up at 2GB. I am sure I am mistaken about some of this, which is causing my problem.
Re: Large file support in Woody
In your message of: Sat, 18 Nov 2000 08:45:41 +0800, you write: >You have to patch the 2.2.17 kernel or you have to use 2.4 kernel. Yes, I am using 2.4.0-test10. -- Jeff Lessem.
Re: bug: "neighbor table overflow" message
[EMAIL PROTECTED] (Josef Dalcolmo) writes: > Hello, > > I am getting a "neighbor table overflow message" during installation of > woody from 2001-11-09 and also afterwards > Does anyone know where this message comes from? The times I have seen this message it was because the network loopback interfact (lo) wasn't configured. Try running "ifconfig lo up" and see if that fixes the problem. -- Jeff Lessem.
PPP stops passing data after 100kb
ppp/options) Nov 6 11:46:35 linux pppd[5157]: noipx^I^I# (from /etc/ppp/options) Nov 6 11:46:35 linux pppd[5157]: pppd 2.4.1 started by a_ppp, uid 0 Any advice would be greatly appreciated. I apologize if you see this message multiple times, but I am sending it to several different lists, as I am totally at a loss on how to proceed. Hopefully the solution is some brain fart on my case, such as echo 1 > /proc/sys/net/ipv4/ppp_should_work -- Jeff Lessem.
Re: About a laptop
In your message of: Mon, 24 Sep 2001 22:06:19 EDT, you write: >On Mon, Sep 24, 2001 at 09:38:31AM -0700, Lucho Debianero wrote: >| Im planning to buy a dell inspiron new 8100, anybody >| knows if there is any problem with woody specially >| with the internal-pci modem and ethernet? > >Do you have the model number? That usually helps when researching >hardware compatibility. The internal modem will probably work if it is based on the Lucent/Agere chipset. Look at http://www.heby.de/ltmodem for the drivers. The builtin ethernet will almost definately work, regardless of what type it is. I understand it is difficult to find out exactly which modem/ethernet card Dell is using at any particular time. Once you have the laptop, lspci should tell you what hardware it contains. >As I was looking up info on the suspend-to-disk feature I found some I got suspend to disk working by setting up a partition with lphdisk from http://www.procyon.com/~pda/lphdisk/. Then pressing fn-a will suspend to disk. I found this to be a nearly worthless feature though. Writing out a 384MB+ memory image takes a several minutes, and reading it back in takes much longer than booting from scratch. Even with the normal suspend, I have left my laptop asleep for 24 hours and still only used less than 1/3 of one battery.
Re: Galeon spinner/toolbar broken?
In your message of: 22 May 2002 17:37:17 EDT, you write: >Hello. I seem to have broken galeon's tool bar/spinner. I believe it was >due to the gdk_imlib packages installed when I tried out kde3 (see >#debian-kde). Yes, the same thing happened to me. It took me a while to even notice it, so I just kept using galeon without the spinner. I did as you did, and removed kde3, and reinstalled imlib and other things that had come with kde3, but none of that worked. The next day mozilla 1+rc2-3 came into sid, and after installing that my spinner came back. I am not entirely sure what exactly fixed the problem, but you might trying reinstalling or upgrading mozilla and its dependencies. What really bothers me though, is that I can't get anti-aliased fonts to work in galeon on my laptop, where they actually are useful. On my laptop, anti-aliased fonts work fine with mozilla, and on my desktop anti-aliased fonts work fine with galeon (and mozilla). I followed the directions for setting anti-aliasing in galeon and mozilla, but it doesn't work. I copied over the configuration stuff from my desktop to my laptop, but they still don't work. I know they are still pretty much an unsupported bonus feature, but it does disturb me that it works on one machine but not the other, when they are both pretty similar Debian Sid machines. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
New apache in sid (was) Re: Apache Exploit Released - where is an update for Woody?
I just ran apt-get install apache/unstable apache-common/unstable on an otherwise testing system and apache 1.3.24-3 was replaced with 1.3.26-1. So far php4 and mod_ssl from testing still seem to be working. Server Version: Apache/1.3.26 (Unix) Debian GNU/Linux mod_ssl/2.8.7 OpenSSL/0.9.6c PHP/4.1.2 Server Built: Jun 19 2002 12:00:21 Of course in order for the /unstable bit to work you have to have the appropriate deb lines in your /etc/apt/sources.list. deb http://my.favorite.mirror/foo/bar/ unstable main non-free contrib Adding APT::Default-Release "testing"; to /etc/apt/apt.conf will keep the system at the testing level, even though it knows about unstable packages. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Can one "fake" a Debian package?
Pollywog <[EMAIL PROTECTED]> writes: > > I have QT 1.42 installed from the source tarball. Unfortunately, my KDE > > installation was in the form of Debian packages, and now I get > > "kde-whatever depends on qt-142" errors from dselect and apt-get. How can > > I convince my system that QT is installed? Is there a file I can edit? > > > > - thanks, Bill > > I had this problem too and got around it by installing the qt debs AND the > source as well. > > I don't know if that is a correct way to do things, but it worked for me. I just created a local package that provides (in the dpkg sense) whatever packages I do not want to install from deb files, for whatever reason. I created a directory tree and file: local/DEBIAN/control where control is: Package: local Version: 1.00 Architecture: all Essential: no Provides: tex, latex, texinfo, dvips, dvipsk, xdvi, metafont Maintainer: "Jeff Lessem" <[EMAIL PROTECTED]> Description: Provides for packages that I prefer to maintain myself This is just my example. You probably want to look through /var/lib/dpkg/available to see what packages qt Provides:. Then from ../local, run "dpkg-deb -b local" and local.deb will be created. Then use "dpkg -i local.deb" to install your packages. It contains no files, it simply tells dpkg that whatever packages are listed on Provides: are installed. -- Jeff Lessem.
