Re: [expert] Mdk71 - su question !
"José Ignacio Alvear Zapata" [EMAIL PROTECTED] writes: Hello Ernie, i login as superuser! (root) on console 1 !! Then i want to switch to the user frank. su frank and i get the /dev/tty1 Permission denied. but the user id has changed... Why i get this message?? (with Mdk 7.1). Hi... If you are at root's $HOME, when you su to another user you are breaking permission rules (the other user should not have x, r or w permissions there). That's what the message is telling you (I think). Try "su - frank" instead of just "su frank" "man su" for details Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] Defrag counterpart in MDK
Civileme [EMAIL PROTECTED] writes: On Mon, 12 Jun 2000, you wrote: Dear all: Could any body know there is any utility like Defrag under Windows for optimization of Hard disk in Mandrake 6.0 or RPM in higher version? Thanks in advance, Wei Quan Tian You can find a defragmenter at www.freshmeat.net It works much better on the old ext filesystem. The algorithm for setting up files in the ext2 filesystem already keeps fragmentation very low. More importantly since the original post was concerned with imporving disk performance may I suggest that you check out the manual page for hdparm which is used to tweak hard drive performance paramaters, be warned though if you have dodgy hardware it may cause data corruption. It works great for me ;) My settings for your information: /sbin/hdparm -d 1 -c 1 -A 1 -m 16 -k 1 /dev/hda Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] Ducking the DUL
Glyn Millington [EMAIL PROTECTED] writes: My situation is as described below. When sending to Freeserve and one or two others I get a message that refers me to mail-abuse.org and specifically to the DUL list. Either my IP address is being used by someone evil or Freeserve just will not accept anything but mail through an ISP smtp server. It' the latter, well sort of. Freeserve won't accept email from people who are on dial-up connections, for the reasons why see http://www.mail-abuse.org/dul/ I wondered if there was a way of dealing with this - registering somewhere, whatever. Maybe, presumably the DUL list doesn't contain *every* ISP's dial-up IP's you could try going to a different ISP and try from there (You'd then have a different IP which may not be blocked by DUL) Just a thought Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] 7.1 Beta-3 VS. The official 7.1
"TK Kim" [EMAIL PROTECTED] writes: I've been happly using 7.1 beta-3 for a couple of weeks while waithing for the official release of 7.1. So as soon as it was made avail, I downloard it and installed. And to my surprise, it actually seemed not as good as beta-3. There were seg-fault upon closing the control centre in GNOME, and the system in general seems to be actually slow. Don't know why, but even with enough RAM, I would be actually using the swap with settings of the system being identical as the beta-3. Is there anyone who tried both and find the similar result as I did? For now, I am gonna use beta-3 instead until I found out more about it. I'd appreciate your inputs. As far as I know BETA-3 contains exactly the same packages as the final release, maybe you got a corrupted download? Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] DISCUSION Upgrade (fwd)
Denis HAVLIK [EMAIL PROTECTED] writes: What do you expect from upgrade? It'd be nice to "upgrade in a window" like with Helix. Ie run a nice little app which checks your systems tells you which apps can be upgraded safely, lets you pick which ones to upgrade and goes and does it. It should also tell you other things that it thinks need upgrading but might damage your system/can't be upgraded while running (kernel etc :) ) and prompts you to reboot and do a "normal" upgrade if you want to upgrade those packages. Probably quite tricky, but you did ask ... :) Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] glib gtk+
Ace Frehley [EMAIL PROTECTED] writes: Hi, I am trying to install the PAN usenet news reader and one of its requirements is that I update glib and gtk+ to 1.28. First I get glib to install correctly but flail when installing gtk+. When upgrading gtk+ and performing the first step by running the ./configure script it'll fail stating that it has found an older version of the glib. It wants me to correct a path in my LD_LIBRARY_PATH variable and or edit my /etc/ld.so.conf to reflect my updated glib... Not sure what I should add or subtract to correct this problem... and need info to correct this... What has happened is that the old glib is still installed in the /usr hierarchy, while your newly compiled glib resides in /usr/local. /usr is checked by default on your system, while /usr/local is probably not. There are many ways to fix this. Since you're probably upgrading from 1.2.x anyway I would suggest installing the new glib/gtk+ over the old ones (Others will no doubt argue this point ...) If you want to upgrade your system wide glib and gtk then simply add "--prefix=/usr" to you "configure" line and re-install, e.g. $ cd glib-1.2.8 $ ./configure --prefix=/usr $ make make install $ cd ../gtk+-1.2.8 $ ./configure --prefix=/usr $ make make install In the past on single user (Read "desktop") systems I've simply placed the following in my .bashrc: export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH export PATH=/usr/local/bin:$PATH This has the advantage that only you will see the new installation, other users will continue to use the old version under /usr Or alternatively check out helix gnome which will upgrade your packages for you, and provide you with the latest and greatest of many applications http://www.helixcode.com Regards Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] helix-gnome
leith [EMAIL PROTECTED] writes: I have a problem installing helix-gnome. The instructions are straight forward. After logging on as root, I ran the script and after download was complete I got these messages. Is there something wrong??? Installer extracted into /tmp/installer.2299. Running Gdk-WARNING **: local not supported by C library Xlib: connection to ":0.0" refused by server Xlib: Client is not authorized to connect to Server Gtk-WARNING **: cannot open display: :0.0 ==Cleaning up temporary files...done. The root user on your machine does not have permission to start X-based applications on your screen. To fix this you need to do two things: 1) (As the user you started X with) type "xhost localhost" 2) As the root user before you run the go-gnome script type "export DISPLAY=:0" Number 1 says to your X server that any user is allowed to display programs as long as they are on the localhost (Your machine), if you care about security you should probably do "xhost -localhost" after the install to turn this off again Number 2 says that any X-based applications you run in you current shell they should contact display number 0 on the local machine to appear on. (By the way this will only work if you run bash as your shell, other shells set variables in different ways ...) This should solve your problem Regards Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] ld.so problems
[EMAIL PROTECTED] writes: On 31 May, Charles Curley wrote: -Device BootStart EndBlocks Id System - /dev/hda1 1 277 2094088+ 6 FAT16 - /dev/hda2 278 556 21092405 Extended - /dev/hda3 * 557 563 52920 83 Linux - /dev/hda4 564 1268 5329800 83 Linux - /dev/hda5 278 556 2109208+ 7 HPFS/NTFS - Other, your partition table is royally clobbered. I don't get it? Why? Kinda figgered that... hda1-4 should be either primary or extended partitions. They are aren't they? 1,3,4 are primary partitions, 2 is extended. Logical partitions should be hda5 and up. Yep hda5 is a logical partition of hda2 ...hence why hda5: it occupies the same space on the drive as hda2. Hmmm... I don't understand. Only the 2nd partition is supposed to be extended. That's an NT (or FAT?) partition. It was created (or at least modified) with Partition Magic. Correctly as far as I can see Erm, you're still stuck with the fact you have no swap partition though :( Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] ld.so problems
Charles Curley [EMAIL PROTECTED] writes: hda2 is an extended partition, both according to the file system ID and what it does: it holds other, logical, partitions within it. hda5 is labeled 7, NTFS or HPFS, but it really is an extended partition. I don't know how it got created or labled, but it should never have been created, and, having been created an extended partition, it should have been IDed as such. As far as I can see, and from what I understand it's a logical partition, and it is perfectly legal!. Logical partitions are IDed as what they are (In this case NTFS), exactly the same way as primary partitions are. The very fact that it is "hda5" says that it is logical, it doesn't need any further IDing! Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] ld.so problems
John Aldrich [EMAIL PROTECTED] writes: On Wed, 31 May 2000, you wrote: [EMAIL PROTECTED] writes: On 31 May, Charles Curley wrote: -Device BootStart EndBlocks Id System - /dev/hda1 1 277 2094088+ 6 FAT16 - /dev/hda2 278 556 21092405 Extended - /dev/hda3 * 557 563 52920 83 Linux - /dev/hda4 564 1268 5329800 83 Linux - /dev/hda5 278 556 2109208+ 7 HPFS/NTFS hda1-4 should be either primary or extended partitions. They are aren't they? 1,3,4 are primary partitions, 2 is extended. AFAIK, you can NOT have a primary partition within an extended partition. You can only have LOGICAL partitions within an extended partition. I never said you could! We have this: +--+ | hda1 | | Primary | +--+ | hda2 | | Extended | | +--+ | | |hda5 | | | |Logical | | | +--+ | +--+ | hda3 | | Primary | +--+ | hda4 | | Primary | +--+ -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] volume not in /etc/fstab - mounts anyway??
Gavin Clark [EMAIL PROTECTED] writes: hi, I just set up a couple of new scsi drives with diskdrake. I set up with one big partition each for testing. I can use them - read/write files, etc. but something is wierd: How do you know you're writing on them? Do you intend mounting them some place? Perhaps, for example, /mnt/scsi1 If you intend doing this then you must have a directory /mnt/scsi1 You can write to this directory because it exists on your current disks. Until you mount the scsi drives on that directory you are *NOT* writing to them. they mount on startup but there is no entry for them in /etc/fstab I doubt it ... It doesn't work like that ... df doesn't see them - it does list my hda paritions. They're not mounted then. Try writing a large file (Say 2 megs) where you think you're writing to the disks and watch 2 megs disappear off your hda partitions ... also the module that drives the card (aic7xxx.o) is not loaded. In which case they're not mounted (Unless of course your kernel has support for them compiled in, but I doubt that ...) #umount /test1 comes back with: umount: /test1 : not mounted Then they're not mounted ... how can I be using these volumes if they're not really mounted?!? You're not, see above. I realise that this is not the best explanation but it should make *some* sense ... Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] Where's security in Mandrake???
Mario Galan [EMAIL PROTECTED] writes: But yesterday I was having a look at www.securityfocus.com and I was very surprised when I saw that there are several known security holes in our distro that aren't listed in the updates page. You're unlikely to get a detailed response unless you mention what these problems where ... the less time someone has to spend working out what your problem is, the more time they have to working out how to fix it. (That's a point I'm making not necessarily to Mario, but to mailing list people in general (Also see Pj's recent mail)) - What should I do to have a secure system?. Well, I know I could be reading bugtraq list and security docs all day but I have no time for such things. There is much more to security than updating all packages. I'd suggest getting one of the many books on security and reading that as a start point, for example: - Only run necessary services + Disable those you don't need (ipchains, inetd.conf, hosts.deny) + Even better don't even install them! - Limit access to services you do need to those that need it - Read package documentation (Especially things as what user a service runs as!) - Is Mandrake a secure distribution? No distribution is *secure* as such, it's a question of whether it is *securable* and that is where Linux (Not any particular distro) wins over other OS's (No flamewar intended ...) Lee -- | Lee WillisFixed cost Internet access available @ | Application Developerhttp://www.plus.net | PlusNet Technologies Ltd +--"PlusNet - The smarter way to Internet"--
Re: Re(2): [expert] Where's security in Mandrake??? - on services
Kirk McElhearn [EMAIL PROTECTED] writes: On 19/05/00, at 10:23, Lee Willis [EMAIL PROTECTED] said: - Only run necessary services + Disable those you don't need (ipchains, inetd.conf, hosts.deny) How do I know which ones I need? There are many services that are installed by default whose cryptic names mean nothing to me. Ask :) What is your machine doing? Is it a web server, and email server, file server, print server, desktop machine? Lee -- | Lee Willis Fixed cost Internet access available @ | Application Developer http://www.plus.net | PlusNet Technologies Ltd +---"PlusNet - The smarter way to Internet"
Re: [expert] ipchains logging
Vincent Danen wrote: What file, specifically, does ipchains log to when I use -l? It says it writes to syslog but when I look in /var/log, I don't see anything that it writes to and I'm trying to debug a firewall script... any info is appreciated. Typically it is /var/log/messages unless you have specified otherwise in /etc/syslog.conf Hope that helps Lee
Re: [expert] apache ... httpd.conf
hans wrote: I have attached my httpd.conf. Maybe, I have not seen a possible failer. But I cannot find out, what's the reason, why a connection to "localhost" is NOT possible. [httpd.conf] # Port: The port to which the standalone server listens. For # ports 1023, you will need httpd to be run as root initially. # Port 8080 Web browsers by default send requests to port 80, but your apache server is listening on port 8080. To check this is the problem try entering: http://127.0.0.1:8080/ into your web browser and you should get a response. To fix it simply change the line that says "Port 8080" to "Port 80" (Without the quotes of course ...) and restart apache with "/etc/rc.d/init.d/httpd restart" or "/usr/local/apache/bin/apachectl restart" depending on how you installed (I didn't really follow the rest of the thread, sorry :( ) Lee -- Lee Willis [EMAIL PROTECTED]
Re: [expert] BASH Shell games
Charles Curley wrote: Why is this failing? I must admit I didn't even look at your example beyond thinking that it was overkill ;) As far as I can see what you are looking for is cp -u sourcefile destination file which does exactly what you want with only one commnd! "man cp" for details. Hope this helps Lee -- Lee Willis [EMAIL PROTECTED] Application Developer PlusNet Technologies Ltd http://www.plus.net.uk