'Bugs' or Ommisions

  • Bruce Meyer
    Bruce Meyer

    I will place the items I find in this thread:

    First one is:
    I have the NST 1.20 installed on my laptop hard drive. I have used VNC to connect remotely, and then run firefox to access various items.

    In this setup, clicking on 'Windows utilities' results in a 404 not found. (\windows directory)

    I do have this one to document:
    I go down to 'measurement tools' and click on the 'Zander-Collection' and get a 404 not found.

    Side note:  I am used to RealVNC, but wanted to try TightVNC. it seems TightVNC doesn't allow copying and pasting between the session and the host. That's tough. I wanted to paste the exact message.

    --Bruce D. Meyer

    • Thanks for reporting these issues, it makes it a lot easier to clean things up when people take the time to report the problems they find.

      - The missing /windows utilities issue has been cleaned up in the CVS repository for the 1.2.1 release (just two days ago).

      - I will take out the Zander-Collection link as the URL no longer seems to exist. We need to find a good link checking tool which recursively checks all internal links (ones that resolve to a location somewhere under the starting point), and the presence of external links (but not recurse into them).

      - Not sure what to tell you about the cut/pasting from TightVNC. I see the same issue, I can't seem to cut/paste if I use vncviewer from TightVNC (even when I try using the F8 popup menu) where I can if I use the vncviewer from the realvnc package. We had a reason at one point for picking TightVNC over RealVNC (think it had something to do with color depth), but we may want to re-evaluate this.


    • Bruce Meyer
      Bruce Meyer

      I switched to RealVNC so i can copy and paste for better 'accuracy' (or was it laziness?)
      it immediately suppoorted the colors on my X window. I didn't have to switch to 'TRUE' colors to get rid of the unreadable fizziness that the default TightVNC game my X window. The next thing I did was run 'Bandidth' to measure what is going on while I download a large file, vpn on another machine etc. I was tail -f'ing messages and saw this Out of memory killed process item:
      (wish i could say when it happened in ration to the steps above)
      [root@probe server]# tail -f /var/log/messages
      Feb  8 19:28:18 probe kernel: Normal free:608kB min:312kB low:624kB high:936kB active:47268kB inactive:54276kB present:114624kB
      Feb  8 19:28:18 probe kernel: protections[]: 0 156 156
      Feb  8 19:28:18 probe kernel: HighMem free:0kB min:128kB low:256kB high:384kB active:0kB inactive:0kB present:0kB
      Feb  8 19:28:18 probe kernel: protections[]: 0 0 0
      Feb  8 19:28:18 probe kernel: DMA: 1*4kB 0*8kB 0*16kB 0*32kB 1*64kB 1*128kB 0*256kB 1*512kB 0*1024kB 0*2048kB 0*4096kB = 708kB
      Feb  8 19:28:18 probe kernel: Normal: 0*4kB 0*8kB 0*16kB 1*32kB 1*64kB 0*128kB 0*256kB 1*512kB 0*1024kB 0*2048kB 0*4096kB = 608kB
      Feb  8 19:28:18 probe kernel: HighMem: empty
      Feb  8 19:28:18 probe kernel: Swap cache: add 12374, delete 1509, find 23/24, race 0+0
      Feb  8 19:28:18 probe kernel: Out of Memory: Killed process 1414 (httpd).
      Feb  8 19:28:41 probe kernel: device eth0 entered promiscuous mode

      This is on an older i686 Compaq Laptop with a measley 128 MB of ram.

      • When you run out of memory, things fall apart. You are trying to do to much with only 128MB of memory (especially since I'm pretty sure the bandwidthd setup script will try to create a 64MB RAM disk by default).

        If this was a hard disk install, you can increase your virtual memory by creating a Linux swap partition, and using the auto_add_swap script when you want to make use of it (or, add the swap space to /etc/fstab if it will be permanently available).

        If you have a hard disk partition mounted, you can also make use of the -rdir option to many of the NST scripts to avoid the creation/use of a RAM disk altogether.

        • Bruce Meyer
          Bruce Meyer

          aHa! I had created a large swapfile with fdisk before installing NST, as I read something about that in the documentation. My fstab does say:
          [root@probe server]# cat /etc/fstab
          /dev/hda1       /               ext3            defaults                1 1
          none            /dev/pts        devpts          gid=5,mode=620          0 0
          none            /proc           proc            defaults                0 0
          none            /dev/shm        tmpfs           defaults                0 0
          /dev/cdrom      /mnt/cdrom      iso9660         noauto,owner,kudzu,ro   0 0
          /dev/fd0        /mnt/floppy     auto            noauto,owner,kudzu      0 0
          # Following swap partition found during install
          /dev/hda2       swap            swap            defaults                0 0
          # Uncomment if you create a new swap partition (change /dev/hda9 to partition)
          #/dev/hda9      swap    swap            defaults                0 0
          [root@probe server]#
          However, looking at the fdisk -l shows:
          Disk /dev/hda: 20.0 GB, 20003880960 bytes
          16 heads, 63 sectors/track, 38760 cylinders
          Units = cylinders of 1008 * 512 = 516096 bytes

             Device Boot      Start         End      Blocks   Id  System
          /dev/hda1               1       36000    18143968+  83  Linux
          /dev/hda2           36001       38760     1391040   82  Linux swap
          [root@probe server]#
          So it looks like it is being mounted. Even though I had that error, bandwidth did continue to run and reported results.  Hmmm. Ok, I am going to try and learn how to mount my DWL-650 using the drivers described to me earlier, and see if that resolves the issue with my pcmcia WiFi card. I'll report my experience once I am done.

          • fdisk -l just reports what partitions are available, not what is mounted or used as swap.

            Use "free" or "top" to see how much hard disk swap is being used. The output of "free" below shows a CD boot with no swap (three 0's on the swap line). In your situation, I'm guessing you'll see non-zeros.

            [pkb@localhost ~]$ ssh root@
            Last login: Tue Feb  8 20:31:24 2005 from

            = Linux Network Security Toolkit (NST v1.2.1) =

            [root@probe root]# free
                         total       used       free     shared    buffers     cached
            Mem:        515072     353236     161836          0     169268     125360
            -/+ buffers/cache:      58608     456464
            Swap:            0          0          0
            [root@probe root]#

            HOWEVER, I'm not positive on how RAM disks work. So, if you ran the "setup_bandwidthd" without specifying the "-rdir DIR" option, you may have consumed a lot of RAM. Try:

            - Creating a root level directory for the NST scripts to use (make sure that the directory is on a physical hard disk):

              mkdir /var/nst

            - Then use that directory as a argument to the -rdir option available on many of the NST scripts. ie:

              setup_bandwidthd -rdir /var/nst [OTHER OPTIONS]

            This may help in your low RAM situation.

            I still wouldn't try to do too much at once with only 128MB of memory, living with a lot of swapping is pretty painful.


            • Bruce Meyer
              Bruce Meyer

              Coo. Thanks. I forgot about top's note  about swap space, and free also. I once ran a full blow newsfeed of INN with a 486, small HD, and limited ram. For a year I thought that yellow light was like an extra power light. (HD light never flashed, it just stayed lit) Had to use a Berkely Watchdog to keep it rebooting when it would just freeze up. Who knew....

              anyway, I ran free and top, but I'll have to do it again when running 'stuff' and after having followed your RDIR instructions and see how it looks.
              [root@probe server]# free
                           total       used       free     shared    buffers     cached
              Mem:        126336     124580       1756          0      73780      17316
              -/+ buffers/cache:      33484      92852
              Swap:      1391032      48360    1342672
              [root@probe server]#
              Tasks:  44 total,   1 running,  42 sleeping,   0 stopped,   1 zombie
              Cpu(s):  0.8% us,  0.4% sy,  0.0% ni, 96.7% id,  2.0% wa,  0.0% hi,  0.0% si
              Mem:    126336k total,   124700k used,     1636k free,    73784k buffers
              Swap:  1391032k total,    48360k used,  1342672k free,    17340k cached

              ...back to work...

    • Bruce Meyer
      Bruce Meyer

      While looking for instruction as to where the mad-wifi drivers are and how to install them from the included files, I found this:


      This srcipt will try to install all network device loadable kernel modules

      • Got it - thanks.


    • Bruce Meyer
      Bruce Meyer

      Follwing the link to the Savannah Site for the Orinoco drivers brought me here:
      Driver Links     Orinoco Driver, Multiban Atheros Driver for WiFi (MADWIFI), NdisWrapper, Intel PRO/Wireless 2100 Driver, Intel PRO/Wireless 2200 Driver, The linux-wlan(tm) Company, 802.11g Prism GT

      When I went to he 'Orinoco Driver' link, I only found source code and CVS. But the file download here
      is empty. (clicked on FILES) I don't know if that is an oops, or if it is supposed to be empty.

      I thought I had found a second link (HOWTO's?) to a Orinoco driver which had everything, but I can't find it now.

      • The orinoco drivers are one of the "unusual" packages included into the NST distribution. In order to get the most up to date version (required for some wireless cards to go into monitor mode), we have to replace the kernel drivers with source code we check out from the CVS repository that hosts the orinoco project).

        Because of this, the download link is mis-leading (as there isn't a .tgz or .rpm file associated with the package). Its an artifact of the way we automate the build process.


    • Bruce Meyer
      Bruce Meyer

      Attempting to stop the pcmcia carmgr service I entered the following and got this interesting error message:

      [root@probe server]# /etc/rc.d/init.d/pcmcia stop
      Shutting down PCMCIA services:.
      error opening /etc/updfstab.conf: No such file or directory
      [root@probe server]#

      • I've entered a bug report for this (1120007) and will clean it up for the 1.2.1 release (looks like some sort of symbol to hardware mapping table that is part of the hwdata package).