aru

The reason for having the backup on the second drive is in case the first one crashes, or vice-versa. I just has this happen to me and it is not fun. This setup would allow you to mirror important data on two drives in case one goes down. You lose your install, but at least all your critical files are nice and safe on the second drive. I think you misunderstood me :D I meant that, IMHO, the 'correct' place to put a backup partition is under the /var tree, in fact I have my backup partition in a different disk, but mounted at /var/backup/ while /var itself is another parition ;) Filesystem Hierarchy Standard

The files I talked about were part of the lilo-21.7.5-1mdk rpm package. Sorry, but as I'm still using Mdk 8.1 I have those files here. I didn't thougth that the lilo package had changed that much. Maybe you might want to search for that version of lilo rpm packgage in an 8.1 mirror or search for the files them selves at the home site of lilo... just a thought Also, removing your current rpm package and installing the last tarball version of lilo should work since I'm booting in textmode here on my slackware using the lastest version of lilo.

Divide the /home in two and create an /usr/local partition on the free half. That should will empty a lot your /usr partition :)

No, it is a program that allows automatic dialogue with interactive programs. Very useful, for example, to automate a web site (i.e. at lycos) through ftp Don't tell me that you prefer such mounstruosity than this cute icon --> :mrgreen: a nut :P

I insist in that I have no idea of rsync, but I trust you and I believe that the commands are right, so this should work: /usr/bin/rsync -avz --exclude *.{SID,example,tdb} -e ssh /etc/samba/ MedusaII:/etc/samba/ about the password, since rsync seems to use ssh and ssh can be configured to run w/o password, you might want to look for that sorry for the repost, but my connection sucks right now

another idea could be the use of 'expect' to write the password automatically ---ramfree!!! that icon scares me a lot!!!! please don't use it about the syntax, don't use quotes at all (notice that I don't have any idea about rsync but bash doesn't expand the "*" whithin quotes in that context)

Your layout is perfect (except that I'd put www and backup into /var -- e.g. /var/www and /var/backup) But 2-3 gigs to / ???? That's a big waste of space, isn't it? If the opt dir is not used, then 400Mb for / is far enough. If opt is used for example for KDE, then the best thing to do is a new /opt partition. The less disk access in / the less problems you'll have in the future. My advice is: / --> the minimum space needed 300-500mb (safer) /usr --> A big partition (here go all the installed programs) /usr/local --> more or less like /usr (useful because of updates) /home --> the size will be up to you /var --> here goes all the dynamic system stuff, such as backups, http, ftp, logs, ... the size will depend on the usage of the system /tmp --> for security reasons is a good idea to have a dedicated partition (the size doesn't need to be big since you can clean it periodically) /opt --> (maybe useful, for KDE I think) ofcourse there are special cases, e.g. a dedicated ftp partition, but that goes under /var tree.

Putting that at the very beginning you are offending all the GNU part of "GNU/Linux" operating system :P I have to go now, but I promise to read the full article and put my feed back If you have some spare time, visit the GNU-Linux FAQ to have a better understanding of what GNU/Linux is: the kernel (named linux), the operating system (named GNU/Linux), and the Distributions (or ways of packaging programs --among many other things-- for the end user) :) http://www.gnu.org/gnu/gnu-linux-faq.html

I don't know if you already know this, but as it has been my last bash-discovery, here it goes: ---> at least this works for bash: In my Mdk box: [arusabal@mdkhost ~]$ echo $MACHTYPE i586-mandrake-linux-gnu and in my Slackware box: [arusabal@slackhost ~]$ echo $MACHTYPE i386-slackware-linux-gnu ... now I have to find out where does bash get that info...

Putting that at the very beginning you are offending all the GNU part of "GNU/Linux" operating system :P I have to go now, but I promise to read the full article and put my feed back

those little examples should work not in every linux, but in every modern UNIX (well, here I'm maybe exaggerating a bit), because all the commands used are plain standard unix tools (maybe except the "-T" flag used with 'df' which is a GNU option). Also AFAIK /etc/fstab and /etc/mtab exists on every Linux So don't worry :wink:

Just a note: The /etc/fstab file may have commented lines, which might (improbable but possible) induce errors. So to be sure, you'll need to add a step to pre-process that file in order to remove the commented lines (ie using sed). Or, better, use /etc/mtab which doesn't include commented lines, and it must show at any time the root partition (but which is not true for other partitions such as /usr /home... because mtab only shows mounted partitions).

well, that are good news as it can be achieved even just using 'df': mount_point="/" filesystem="ext2" if df -T $mount_point | grep -q $filesystem; then echo "partition at $mount_point is ${filesystem}: WARNING" fi If you want to parse /etc/fstab (I insist in that you can also parse mtab), this is a cool way: #!/bin/bash mount_point="/" # name of the partition that we are asking for. filesystem="ext2" # name of the fs wanted. root_fs=($(awk '$2 == "'$mount_point'" {print $1" "$3}' /etc/fstab)) # now ${root_fs[@]} is an array with the device name of the root partition and the filesystem that it uses (as in "/dev/hda1 ext2"). if [ "${root_fs[1]}" == "$filesystem" ]; then # checks that the root partition is ext2 echo "The partition ${root_fs[0]}, mounted at ${mount_point} is ${filesystem}: WARNING" fi I admit that there are simplier ways of doing the second script, but since you are learning, I thought that maybe you're interested in some advanced shell tips, such as conditional parsing with awk or arrays in bash ;) PS: the 'WARNING" in the output is just the 'dramatic note'. I know that you are interested in the "/" being "ext2" so the warning here is out of place :P

OK if you are talking about 'primary' partitions versus 'logical', AFAIK, there can be 4 primary partitions (lest say hda1 to hda4), and one of them can be 'extended' in order to hold 'logical' partitions, named starting from number 5 (hda5 ...). So If I'm not wrong you can just parse the fstab file and look for the numbers of the partitions (/dev/hda# --first field of fstab--) and then check if they are ext2 (--third field--). Notice that the new nomenclature of the devfs is no more "/dev/hda1", now the partitions can also be named as "/dev/ide/host0/bus0/target0/lun0/part1", so you should know that in order to parse the fstab file. (Also you might want to take a look to fdisk(8 )). But if you are talking about 'primary' as the 'root' partition, then you'll need to parse the /etc/fstab file looking for '/' to match the second field. Sorry, but as I don't understand why are wrong filesystems other than ext2, and I don't know what are you trying to do, I can't give you more accurate help. :P If you need help on parsing the fstab file, please ask me, I will enjoy helping you.

would you please be a little bit more clear :) That is a quite confusing way of doing something that is really simple. You just want to check if both strings "/home" and "/usr" are in /etc/fstab, so instead of counting the number of times that those expresions appear on that file, check if they are or not, w/o any output. (grep -q) Notice that most of file processing or text editing commands can handle files their-selves without the need of cat. The command that you want is: partition_list="/etc/fstab" if grep -q "/home" $partition_list && grep -q "/usr" $partition_list; then your code fi that 'if' statement evalues the return status of the two grep commands, if the return status of BOTH commands is '0' meaning that grep found the string (no error, true), then it runs 'your code'. Note that instead of using fstab (static) you can use /etc/mtab (dynamic), but that would depend on what do you want to achieve. hope that helps

if you want the distribution name (or release name) on a mandrake then: [arusabal@localhost ~]$ cat /etc/mandrake-release Mandrake Linux release 8.1 (Vitamin) para XXXX (i586) or simply do 'cat /etc/issue' In both cases you'll get the name of the mandrake release, which is pretty useless. What is important is what paul said ('/proc/version' or 'uname') because there you'll get the kernel version, the compiler used... which is what really matters. I don't think there is a standard file where to put the name of a distro (even there is no need to do it at all). That just depends on the people who made the distro (IMHO). Looking at the man page of each command/program/daemon, you can check which parameter is used to return the version of the program (usually -v or -V). other useful hints are: rpm -qa | grep package_name Will let you know the rpm version of the package (which is normaly related to the version of the software it provides). To get a deeper detailed info, do: rpm -qi package_name

I supose that you meant 'BACKGROUND' ;)

I don't know either because I only use gkrellm. But about sleeping a bit before launching the apps, you might want to check this faq for the reason of using that old trick: comp.windows.x: Getting more performance out of X. FAQ; just a little quote:

:lol: he is a very good guy, even he never got banned from my part when I was a mod in the old board (except one or two PMs). He is a bit impulsive, but that's all :) bvc, I knew you'll appreciate my post :D

from a resources point of view is better to use 'exec' than calling fluxbox directly, because with exec you'll substitute the current shell (the session you started in .xinitrc - #!/bin/bash -) with fluxbox (or whatever wm). Otherwise you'll have a bash session running in the background waiting the end signal from fluxbox to exit itself. so it's a bit more correct to launch the wm with 'exec'. Do you remember the club-nihil's discussion about fluxbox and .xinitrc? If I don't remember bad there we discussed why was wrong the 'exec fluxbox' in DOlson's fluxbox's init scripts and why was a good idea to use it in a .xinitrc file :)

you might want to check that the default route is set through your internet interface and links to your ISP: Connection establishes, holds, but you can't get anywhere (MUO docs)

run it on the slit, as gkrellm -w No idea, I have no experience with it Read the whole thread at club-nihil (and save it if matters since it will soon disapear):starting things (like grkellm) with fluxbox.

:shock: sorry, I had to do this : [arusabal@localhost ~]$ man XF86Config ....... Option "BlankTime" "time" sets the inactivity timeout for the blanking phase of the screensaver. time is in minutes. This is equivalent to the Xserver's `-s' flag, and the value can be changed at run-time with xset(1). Default: 10 minutes. Option "StandbyTime" "time" sets the inactivity timeout for the "standby" phase of DPMS mode. time is in minutes, and the value can be changed at run-time with xset(1). Default: 20 minutes. This is only suitable for VESA DPMS compatible monitors, and may not be supported by all video drivers. It is only enabled for screens that have the "DPMS" option set (see the MONITOR section below). Option "SuspendTime" "time" sets the inactivity timeout for the "suspend" phase of DPMS mode. time is in minutes, and the value can be changed at run-time with xset(1). Default: 30 minutes. This is only suitable for VESA DPMS compatible monitors, and may not be supported by all video drivers. It is only enabled for screens that have the "DPMS" option set (see the MONITOR section below). Option "OffTime" "time" sets the inactivity timeout for the "off" phase of DPMS mode. time is in minutes, and the value can be changed at run-time with xset(1). Default: 40 minutes. This is only suitable for VESA DPMS com patible monitors, and may not be supported by all video drivers. It is only enabled for screens that have the "DPMS" option set (see the MONITOR section below). ....... No offense intended, but the options are there. Anyhow, I'm glad to see that you find a way to do it with xset :D

... for example, you can look to your XF86Config file and checkout the options: "BlankTime", "StandbyTime", "SuspendTime", and "OffTime". Take a look to the XF86Config(5x) man page to see which are their default values. Notice that those options might not appear in the XF86Config file, but X will behave following their defaults. good news are that you can change those settings "on the fly" using xset(1)

how is the dialmode set? [root@localhost ~]# isdnctrl dialmode ippp0 Dial mode for ippp0: manual Check out that it is not in "auto" mode (in "auto" whatever outgoing package can restart the connection) If that was the problem you'll need to change it back to "manual" through your isdn configurator, the config file "/etc/isdn/profile/link/myisp", or the particular script that initalizes your ippp interface (sorry, but I can't be more accurate since I initialize my interface through a self-made dedicated script). If matters, the command that sets the manual dialmode is: isdnctrl dialmode ippp0 manual