Best of Technical Support
Our experts answer your technical questions.
Failure to Connect
My PPP connection is failing while using a modem to dial up to my ISP. My log indicates (after chat):
pppd:Serial Connection established pppd:Using interface ppp0 pppd: Connect: ppp0 /dev/ttyS2 pppd: LCP: timeout sending Config-Requests pppd: Modem Hangup pppd: Connection terminated pppd: Exit
My options file with and without these lines given separately had the same effect:
lock passive lcp-echo-failure 10 lcp-echo-interval 400
Chat (tried different values):
'' 'AT&F' (also used ATZ) 'OK' 'ATDTxxxxxx' 'CONNECT' '' 'login:' 'xxxxxx' 'password:' 'xxxxxx' 'TIMEOUT' '60'
pppd options (also tried 1500 for MTU and MRU:
hardflowcontrol=yes defroute=yes MRU=556
Do I need to include something more in my chat to make the modem wait longer ? Also, according to my ISP administrator, my machine is “not replying to LCP request”. My modem is a BestData 56K internal modem (#56SF), controller-based. I am using Red Hat 6.0 Intel. —Vijay Nunna, email@example.com
There isn’t anything else you should include in your chat script; however, some servers may be slow to start up the LCP communication, and thus your pppd daemon will timeout before then. To raise the number of LCP configure-request packets before pppd determines it is not responding, you can add the lcp-max-configure option to your /etc/ppp/options file, followed by the number of LCPs to send. For example, lcp-max-configure 30 would raise the number from the default of 10 to 30. You might also want to remove the—Andy Bradford, firstname.lastname@example.org
You need to add debug on your pppd command line and look at the /var/log/messages file. You also want to replace chat with chat -v. Those two combined will give you lots of debug output and hopefully a better idea of what is failing and where. —Marc Merlin, email@example.com
I am trying to install a package that includes a Ghostscript driver for my printer. I’m getting dependency issues. The files required (as printed by rpm) are:
libc.so.6(GLIBC2.0) libc.so.6(GLIBC2.1) libm.so.6(GLIBC2.0) libm.so.6(GLIBC2.1)
I’ve upgraded the libc libraries to glibc 2.1, but this seemed to have no effect. Looking at the dependency list, it also seems to need libc.so.6 and libm.so.6 without the GLIBC attached, and it has no problem finding these. I’ve looked at FTP sites for files with these names and cannot find them. Can anyone tell me what these are and if I need to find them, or whether I can tell rpm to ignore them? —Doug Morgan, firstname.lastname@example.org
Often, a dependency on libc6 or libm6 means only that the binaries were compiled to use these versions of the libraries—they can often be recompiled against an older libc/libm without any problems. To do this, grab the SRPM (it should have the extension .src.rpm), do
and install the resulting binary .rpm. Use this approach with caution, as there are applications that exhibit bugs under the older libc/libm (or won’t compile). Still, in my experience, it generally works just fine. —Scott Maxwell, maxwell@ScottMaxwell.org
Although you can install glibc 2.1 on a glibc 2.0 system, it may be better for you to upgrade your system to SuSE 6.3 instead. However, if you do have all the libraries installed, but for some reason the RPM database doesn’t agree, you can force the package to install with rpm –nodeps package.i386.rpm.
Another option is to get the .src.rpm version of your package and rebuild it for your system: rpm –rebuild package.src.rpm. The resulting package should be in /usr/src/redhat/RPMS/i386/. —Marc Merlin, email@example.com
I am thinking of building a dual-processor computer and wondering if Linux supports this. I have heard Linux is the best OS, and I have heard Windows NT is the only OS that supports multi-processing. —William Cason, firstname.lastname@example.org
Linux does indeed support SMP. In fact, I use a fast dual-CPU Linux box at work. —Scott Maxwell, maxwell@ScottMaxwell.org
Linux may or may not be “the best”, depending on what you’re trying to do and what you’re looking for. It does support multi-CPUs on recent kernels (2.2.x), and some distributions like Red Hat support SMP at install time. Some require you to compile your own kernel. —Marc Merlin, email@example.com
I run Linux Red Hat 6.0 and Netscape 4.7, and I am struggling to install Real Player. I successfully installed (using RPM) Real Player 5.0 (rvplayer) and verified using GNOME; RPM shows the program is installed correctly. In the Netscape options, I set up MIME for file RA and RAM to use rvplayer. When I click on an Internet site with radio, nothing happens. There is a message on a Linux newsgroup, saying there seems to be a problem with rvplayer 5.0 and the 2.2. —Yossi HaYored, firstname.lastname@example.org
That is correct; you need the beta version of the G2 player. You can get it from proforma.real.com/mario/player/player.html. —Marc Merlin, email@example.com
I had trouble getting rvplayer 5.0 to work on a 2.2.5 kernel myself, but rvplayer 6.0 worked just fine for me. —Scott Maxwell, maxwell@ScottMaxwell.org
Colorful Monitors Ensue
I have been trying to install Linux and have had one successful installation using a monitor other than my personal one. When I took it home to my monitor, it gave me a blank screen, and the monitor’s on light goes from green to maroon, indicating a problem after it entered its graphical interface. When I try to reinstall Linux Mandrake or Caldera 2.2, it gives me a blank screen when it does its graphical interface or when I try to probe for a video card. I have a Compaq V50 monitor and Creative Labs 4MB video card. When I was able to get a successful installation, I was using a Micron monitor. Could there be a compatibility issue, and will there be a way to get Linux installed at all? This monitor configuration works with Windows just fine. —Andy Kissner, firstname.lastname@example.org
What’s apparently happening is that X is trying to drive your monitor too hard. For this reason, graphical installs are not necessarily going to work. At least with Red Hat, you can ask for a text-mode install when you boot the first floppy. Some distributions, like Debian, will install only in text mode, which is fine for you, and others expect to see a VESA-2-compliant video card and a monitor that can accept suitable refresh rates. Your best bet, once a distribution is installed, is to edit XF86Config (usually in /etc/X11 or /etc) and reduce the value of those two parameters:
HorizSync 30-50 VertRefresh 50-70
The values given here may work, but you should really put in the ones specified in your monitor’s manual. —Marc Merlin, email@example.com
I actually have several questions.
1. How do I access my Zip drive? I suppose I could try mounting and unmounting everything in /dev, but that seems particularly ungraceful.
2. How do I log out? I can switch to superuser and shut down, but letting my kids and wife be superuser defeats at least one reason for preferring Linux over Windows.
3. I tried to set my default shell to tcsh instead of bash. It doesn’t seem to have taken. In particular, my .cshrc file (or .tcshrc) is not read at login. —Tim Allison, firstname.lastname@example.org
1. The answer is very different, depending on whether your Zip is parallel, IDE or SCSI. There are two HOWTOs which you should read: ZIP-Drive and ZIP-Install.
2. In a shell, type logout. In X, you have to find the logout option. It depends on which window manager you are using and whether you use XDM (runlevel 5). When in doubt, CTRL–ALT–BACKSPACE will kill your X session, but that’s the sledgehammer approach.
3. To change your shell, you can use chsh or simply edit /etc/passwd and change the last field for root. Make sure that whatever shell you specify exists. To find out which shell you are currently running, try echo $SHELL. If you are indeed running tcsh, try doing an echo in ~/.tcshrc to see if it is run or not, and consult the tcsh man page which explains which files are being run, depending on whether your shell is a login shell or not. —Marc Merlin, email@example.com