cardassianscot

Thanks

I run a small network for a College. We have two networks with linux and windows machines on both, connected through a internet gateway running linux. Anyway my file server ran out of space on its main partition causing various problems. I traced the error to some very large log files (2 GB) in my /var/log/samba directory. I deleted the files and everything works fine. However, I am now wondering if one of the files was due to someone outside my firewall protected network trying to access my samba shares. Reason: one of the large files was for a machine called cc. Which is fine, I have a machine called cc on my network. However, I was trying to see what machines were working and I tried to ping this machine. However, cc is a windows machine so I shouldn't be able to ping it (I had forgotten this since I also use windows machine where it is possible to ping other machines on your localnetwork by name if they have file sharing enabled). However it returned an IP address outside my network. I get the same IP address from all the linux machines on my network but I get no IP address corresponding to cc when I try to ping it from my home computer (different ISP), I am loging in remotely from home so I cannot verify what I get from a windows machine on my network. I was able to trace the IP address to Seatlle. So my question, have I been hacked by this IP address or is there another less sinster explanation? Thanks for any help recieved.

Sorry my suggestions weren't much help. The only other thing I can suggest and this really is scrapping the bottom of the barrel, is changing the name of the host computer you are trying to connect from (the win98 SE ones), using all lower case letters only, less than 8 characters. Believe it or not I just couldn't get a win98 client to connect to my samba server although it could read shares from all the other windows machines on the network and see the shares from my Linux box. This was after a reinstall and I had changed the name of the machine to fit my new naming convention (calling all my staff computers after Babylon 5 characters), so I changed the name back to what it had been previously and it worked fine. Later I tried the same name on different computers and they didn't work with that name either. So my suggestion if changing the name of your computer. As to your other point. You can kill old shares by restarting smb. If you don't know how to do this from the command line but have SWAT installed point your web browser (running on the server) at http://localhost:901/

I have some suggestions you might like to try. The first one in the list has definitely caused me problems before with some versions of windows (I can't remember if it was win95 or win98 but I think it was win98), the others are just some suggestions that are along similar lines and may help or may not. Please also note that these apply to share names not file or directory names. 1. Some of your share names are too long. Try share names of 8 or less characters. 2. Try using all lowercase for your sharenames. I have had problems where one system is sending case independant data while the other is expecting case dependant data. 3. Get rid of hyphens and other non-alphanumeric characters from your sharenames. Hope these help, but like I said they might not.

Things to check (and do) if it doesn't work after you modify smb.conf (the first one is very obvious although not strictly neccessary, I think) 1. Restart samba 2. If you just use the basic setting (no user or group information) check that the user you are trying to mount the samba share as can actually access the files under linux. If not you won't be able to access them under windows or samba.

You need to add a new share. The easiest way is to use SWAT if you have it installed, point your browser at http://localhost:901/ (I think, somebody correct me if I got the port wrong I am writing this from winxp from memory). Or, you can manually add a section to /etc/smb.conf You need to add something like this for each of your directories. (This code adds a share called mp3 which gives read/write access to everyone with an account on the machine to the directory /data1. [mp3] path = /data1 writeable = Yes That's it however, if you want users to be able to edit files created by other uses then you will need to fiddle about with masks. Here is the above lines modified to include my masks. (This assumes that existings files are accessible by all users). [mp3] path = /data1 writeable = Yes create mask = 0777 force create mode = 0777 directory mask = 0777 force directory mode = 0777 inherit permissions = Yes Or if you want to restrict access to a particular group (which is very useful and what I use most often) then you want to use the following code, which restricts access to only the users in the group music. (The @ sign in the valid users line means group rather than user, I leave it as an exercise to work out how to restrict access to a single user). (This assumes that all existing files are read and writeable to the group music.) [mp3] path = /data1 valid users = @music force group = music writeable = Yes create mask = 0770 force create mode = 0770 directory mask = 0770 force directory mode = 0770 inherit permissions = Yes

Thanks Aru, I think I get what you were saying now. Sorry for not picking it up properly first time. It will need a bit more coding to get it working (I need to make sure that if a new print job is started before the first one has printed then both messages get through [this isn't a plea for help here, just outlining why this approach did not immediately strike me as easier]) but thank you for pointing me in the right direction. Or I may just change over to GDM, I haven't yet found the file or directory for running commands when logging out in KDM. I agree that it would be better to put my scripts into the print server and I have every intention of doing so, however I am waiting for CUPS 1.2 to do this since it has a facility for adding scripts to handle quotas. However, until then my users still deserve a working system and as I said my users just seem to ignore the email way. Once again thanks.

I do mean logining in through kdm and then loggin our through kdm. I know most processes are killed when you logout but for some reason this one isn't. (checked through flip to a text terminal and ps -A). This might have something to do with the type of script it is. It is essentially a network server which listens on a port and when it recieves data it forks a x-window to display a message. Having a killall script is a good idea but where do I put it or which script do I need to call it from to run on logout? Having a script that only runs when someone prints something might solve this one but complicates matters since 1) I have no idea how to start a script when someone prints something 2) the print server is not running on the local machine My script works by users printing to a file (the actual file is irrelevant but it means nothing prints) and then a script looks at the job files and finds out who printed, from what computer and how many pages. If the user has enough credits then it prints it to the real printer otherwise it blocks it. In anycase it sends a message to a program on the machine the file was printed from, which just pops up a window with the message in it. It's this program running on the local machine which I need to start and stop when a user logs in or out. Thanks

I have written an add on to the CUPS printing system in PERL which implements a quotas scheme. However, my old method of emailing success or failure and the reason was often ignored by users so I have designed a printer monitor system which pops up a windows with these messages. However, the system requires a program running in the back ground on the local machine which the print server connects to. To get this running properly I need to run this program as the local user when they log in and to stop the program when they log out, so that when another user logs in the program can be run as them. I am running Mandrake 9.0 with the GNOME desktop and the KDE login manager. I know I can launch a program when a user logs in through the session part of the GNOME configuration but this leaves the program running when the user logs out. How do I stop the program from running when the user logs out? Thanks

I have completely disabled supermount (well run supermount -i disable), checked /etc/fstab and then rebooted. I still get the same problem.

I also have a GF4 MX dual graphics card setup (second card Matrox Millenium PCI). It works fine in WINXP with 3D acceleration, however, I cannot get it to work properly in Linux. I can get it running without 3D with both graphics cards, however, as soon as I install NVidia's drivers I can only get the GF4 to work. If I keep the Matrox Millenium lines in my XF86Config-4 file then it refuses to go into X and if I change the GF4 stuff back to the pre-3d stuff then it also refueses to go into X. So in my experience, I either have to run without 3D or without my 2nd graphics card in Linux (and not in a way which you can switch easily between them). If anybody has got both of these running simulataeously please post how you did it.

Having had problems using supermount, I have switched to using autofs for floppy disk access although I still use supermount for CD access. When I install autofs it all works fine and continues to do so if I create a symbolic link to the floppy disk in most places. However, if I create a symbolic link from my automount floppy disk directory to /mnt/floppy, then the system hangs when I try to log in (it won't log in to GNOME or KDE, I can switch to a terminal Ctrl-Alt-F1 but trying to switch back to X gets me a blank screen from which I can't escape). I also can't stop or restart autofs, while it says it was successful it reports an error (sorry can't remember the exact error but I will recreate the problem and post if neccessary) and keeps a process running. This same error also causes the computer to fail to either halt or reboot (it hangs when it gets to stopping autofs). This problem only occurs with Mandrake 9.0 not 8.2. (This only happens after I access a floppy disk. But always happens after this, even if the computer is rebooted, or even turned off and then turned on again.) Obviously one way of stopping the problem is to link /misc/floppy to somewhere other than /mnt/floppy. However, KDE programs have floppy icons shortcuts in their open and save dialogues and I would like my users to be able to use them. Of course this could be causing the problem (even though the same problem happens in GNOME). So, does anyone know why this problem is occuring, or can I change the location of the floppy shortcut in KDE applications or disable this feature to avoid confusing my windows-loving users. Thanks

File > Properties > Statistics is the OpenOffice.org word count but it has limitations. It will not tell you how many words are in the selected text only in the whole document. If you want word count to work as it does in word, then you need to download the word count macro from OOExtras . Here's a direct link http://ooextras.sourceforge.net/modules.ph...op=getit&lid=28 You need to copy and paste the macro into the macros editor (tools > macros). I then assign it to a button on my toolbar (tools > customize > toolbars > customize). If you need more detailed instructions I can provide them.

Yes. That's what I'm saying.

Autofs is definitely the way to go. I had all sorts of problems with supermount when changing floppy disks. However, you will have to configure it manually. The changes I made were to edit /etc/auto.master to make the timeout 1 second for auto.misc. Then comment out the appropriate lines from /etc/auto.misc for my floppy disk and CD. I also had to add umask=0 to the options for the floppy disk so that all users could have write access to the disk (only in Mandrake 9 though, not 8.2). You also have to create /misc , which for some reason is not created by the install package. Then I create links to the standard mount points using ln -s /misc/floppy /mnt/floppy etc., altough this is mainly because I already have a lot of programs and links for lots of users setup to use /mnt/floppy . One word of warning, autofs only creates a mount point if you access it, ie., I use /misc/floppy to access my floppy drive but if I do ls /misc there is no entry for floppy, I have to either type ls /misc/floppy or cd there. This isn't much of a problem for command line use but might throw you if you usually access your files through konqueror or another graphical file browser. You either have to type in the directory in the address bar or use links (which cause a slight pause as you access a directory with a link as it checks wether you have a cd or floppy in your drives). Just a quick reminder autofs works unlike supermount. Hope this helps.

The easiest way is to use sftp. If you can already access your computer by ssh then there is no more configuration needed. sftp uses the same connection arguments as ssh. Once in you are given the usual ftp command line. get file copies file from the remote host to the local machine and put file copies file from the local machine to the remote machine