I recently moved my website to a cheap VPS I had the fun journey of trying to get it work without grinding to a halt everytime I clicked on a page, which you ‘should’ have noticed is fixed.

Firstly I needed to be realistic this blog wouldn’t have more than two simultanious connections at once, with that known I could then edit the amount for apache server threads to an un-godly small number in /etc/apache2/apache2.conf

StartServers 2
MinSpareServers 2
MaxSpareServers 2
MaxClients 20
MaxRequestsPerChild 0

With that noted a quick look at ‘top’ shows MySQL eating lots of Virtual Memory a quick swap of the configuration file: /usr/share/doc/mysql-server-5.1/examples/my-small.cnf to /etc/mysql/my.cnf sorts out that problem. (Don’t forget to backup your originals)

And finally modifying the WordPress max memory limit as noted in this post
suhosin[10882]: ALERT – script tried to increase memory_limit to 268435456 bytes

This article has an updated version -> Here

Today at work a client he needed to access a cisco router but no-one new the password but we had the config files saved which included the MD5 hashes of the enable and enable secret passwords. So I set about cracking the hashes, which was done successfully btw, following on from that I then remembered a John the Ripper MPI patch which I had seen some time ago I ended getting it going at home and the results are good!

Directions for installation are as follows (on Fedora 11)

yum install mpich2 mpich2-devel
wget http://www.bindshell.net/tools/johntheripper/john-1.7.2-bp17-mpi8.tar.gz
tar -zxvf john-1.7.2-bp17-mpi8.tar.gz
cd john-1.7.2-bp17/src
make linux-x86-64 (for 64bit version duh!)
cd ../run
touch ~/.mpd.conf && echo "MPD_SECRETWORD=secret" > ~/.mpd.conf && chmod 600 ~/.mpd.conf
mpd &
mpiexec.py -n 4 -path ./ -wdir ./ ./john --test

Initial Benchmarking looks good. Below is 1 core vs 4 cores

Traditional DES [128/128 BS SSE2-16]
1591K c/s real, 1594K c/s virtual / 6131K c/s real, 6380K c/s virtual
BSDI DES (x725) [128/128 BS SSE2-16]
Many salts: 53222 c/s real, 53329 c/s virtual / Many salts: 207665 c/s real, 215407 c/s virtual
FreeBSD MD5 [32/64 X2]
Raw: 9718 c/s real, 9816 c/s virtual / Raw: 37720 c/s real, 39581 c/s virtual
mysql [mysql]
Raw: 2571K c/s real, 2571K c/s virtual / Raw: 9397K c/s real, 10254K c/s virtual

About a 100% increase per core! Life is good!

Recently I came across a new file format called RSDF, these appear to be txt files which have a bunch of links in them, so called link containers. I wanted to access the URLS in these files, but I didnt really want to entrust my computer to just “any” application. So after a bit of searching I came across this crafty german website from there I got the python script to decrypt the RSDF files, this Python script requires:

Python (doh!!)
Python-crypto
Probably something else also…..

from their its as simple as: drsdf.py rsdfcontainer.rsdf
and it outputs to your screen! So without further ado here it is:

    #!/usr/bin/env python
    # drsdf.py

    import binascii
    import base64
    from Crypto.Cipher import AES
    import sys

    # 8C 35 19 2D 96 4D C3 18 2C 6F 84 F3 25 22 39 EB 4A 32 0D 25

    file = sys.argv[1]

    file = file.replace(".ccf", ".rsdf")
    f = open(file, "r")
    data = f.read()
    f.close()

    f = open(file, "w")
    f.write(data.split("x00")[0])
    f.close()

    infile = sys.argv[1]
    Key = binascii.unhexlify('8C35192D964DC3182C6F84F3252239EB4A320D2500000000')

    IV = binascii.unhexlify('FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF')
    IV_Cipher = AES.new(Key,AES.MODE_ECB)
    IV = IV_Cipher.encrypt(IV)

    obj = AES.new(Key,AES.MODE_CFB,IV)

    rsdf = open(infile,'r')

    data = rsdf.read()
    data = binascii.unhexlify(''.join(data.split()))
    data = data.splitlines()

    for link in data:
    link = base64.b64decode(link)
    link = obj.decrypt(link)
    print link.replace('CCF: ','')

    rsdf.close()

Happy Downloading!

Had a client whose print queue would jam up and they couldn’t restart the printer spooler until you manually cleared the spooled documents. So I stole this piece of code and put it in a batch file. Works wonders!

@echo off
net stop "print spooler"
del /q "%SystemRoot%system32spoolPRINTERS*.*"
net start "print spooler"

Update: The reason the spooler was dying was because when he printed a PDF document he would close Acrobat before the document finished and in turned killed the Print Spooler. Because we charge by the hour, he didn’t want it to be investigated any further, which is fine by me!