How to find + stop SQL injection attacks

There’s a lot of stuff out there about SQL injection attacks, but there’s not much that will help you figure out how to stop these attacks from occurring.

First, let’s talk about what a SQL Injection Attack really is. Some people think it’s a virus of sorts, that is “inside your site.” Not the case. These are bot attacks by other virus infected computers. They simply use a brute force approach of scanning URLs that take POST/GET inputs and attempt to send their own data to them.

So, how do you track these down and stop them? For web sites powered by Microsoft’s IIS, here are our suggestions:

1. Look at your IIS logs
   Try searching for the word “DECLARE” or “EXECUTE.” If you’ve been hit by an attack, these will more than likely show up in your IIS logs — at least for any attack that was attempted using “GET” posts. If you do find any instances of “DECLARE” or “EXECUTE” these are the pages to start with.
2. Use centralized database connection handling
   Simple, make a centralized file (e.g. connection.asp if you are using ASP) that handles all of your DB access. This way, it’s easier to make sure that you are SQL encoding your pages. You can easily search queries for “DECLARE” and “EXECUTE” and stop the attacks dead in their tracks.
3. Implement a site wide solution
   If you are running your own server, we highly recommend ISAPI_Rewrite from HeliconTech (http://www.helicontech.com/isapi_rewrite). This is an ISAPI filter that allows you to do a variety of things, including scan URL data. This will stop 99% of attacks without changing ANY code on your site!\
4. Never use “sa” as your database user, create a user for the database you’re working from and then remove privileges to read the master dbo. This prevents the attacker from “sniffing” your database structure, however, these attacks have evolved so that sanitized stored procedure based attacks happen even with these types of security in place. See http://www.ngssoftware.com/papers/more_advanced_sql_injection.pdf for more information on this and other ideas in preventing SQL Injection attacks.

If you’ve ever been scratching your head wondering where the leak in your programming is (or have taken over a project from someone else) then the best way to determine through which page the attack happened is by checking out the IIS logs.

There are many scripts available to clean up, but the best tip is to backup hourly, and to follow best practices (some of which I highlighted here). Good luck

How to setup a Windows 2003 Web Server

Every time we setup a W2K3 box as a Web Server we do the following

Stage 1
1.1 Copy i386 to c:\
1.2 Configure Windows Permissions on all partitions
- Leave
Administrators (Full)
System (Full)
Remove Everything Else
1.3 Rename Administrator
1.4 Rename Machine
1.5 Reboot
1.6 Install Windows Updates (Not the .NET Frameworks)
1.7 Add IP’s to TCP/IP
1.8 Add DNS suffix to TCP/IP (Only if required)
1.9 Disable Shutdown Event Tracker
1.10 Reboot

Stage 2
2.1 Install Support Tools
2.2 Install Resource Kit Tools
2.3 Install KB908521 (Not needed if w2k3 SP2)
2.4 Install SNMP
2.5 Install SNMP Informant
2.6 Configure Windows Time
2.7 Reboot

Stage 3
3.1 Install R2 (Only if you want it)
3.2 Reboot

Stage 4
4.1 Setup Automatic Updates
4.2 Setup MS DNS
4.3 Set TCP/IP to Local DNS
4.4 MSTDC Fix

Stage 5
5.1 Install IIS
5.2 Allow direct Metabase Edit
5.3 Set IIS IP’s to be listened on
5.4 Restart IIS
5.5 Set Default Documents
5.6 Set W3SVC Logs Params
5.7 Install FastCGI
5.8 Install PHP 5.1 & 5.2 (FastCGI Mode)
5.9 Install Zend Optimizer (32bit edition)
5.10 Install Perl
5.11 Install Python
5.12 Configure IIS SMTP
5.13 Disable IIS SMTP Socket Pooling
- http://www.isaserver.org/tutorials/i…etpooling.html
5.14 Reboot

Step 6
6.1 Install .NET Framework 2.0
6.2 Install .NET Framework 2.0 SP1
6.3 Install .NET AJAX 1.0
6.4 Install .NET Framework 3.5 (Make sure you use the installer without SP1)
6.5 Install .NET Framework 3.5 SP1 (Only if using Helm 4.1)
6.6 Reboot

Step 7
7.1 Install ISAPI_ReWrite
7.2 Install JMail.NET
7.3 Install W3JMail
7.4 Install Hotlink Blocker
7.5 Install IIS Password
7.6 Install Winrar
7.7 Install Persists ASP Email
7.8 Install Persists ASP JPeg
7.9 Install Persists ASP Upload
7.10 Reboot

Naturally if the Box isn’t going to be a DNS Server you wouldn’t install MSDNS, instructions apply to both 32bit and 64bit w2k3. I’m not going to explain each step in detail the info is just so you can see what order we do things. It has been very well tested and we haven’t had any issues

At each step we test everything we install so we make sure PHP is working correctly before proceeding onto the next step. If you need to install MySQL or MSSQL I’d do it at Step 8. Once everything is setup you would install Helm as a remote or control.

Calling a Web Service from classic ASP

1
2
3
4
5
6
7
8

Function sendToWebService(WebService,Data)
    Dim SoapRequest
    Set SoapRequest = Server.CreateObject("Msxml2.XMLHTTP")
    SoapRequest.Open "POST",WebService , False
    SoapRequest.setRequestHeader "Content-Type", "application/x-www-form-urlencoded"
    SoapRequest.Send Data
    sendToWebService = SoapRequest.responseText Set SoapRequest = Nothing
End Function

Data would look like:

1

Data = "firstparameter=yourdata&secondparameter=yourdata" and the WebService would look like: WebService = "http://" & trim(Request.ServerVariables("SERVER_NAME")) & "service/service.asmx/webmethodfunction"

Replace webmethodfunction with your function name and bob is your uncle.

The difference between a Sub and a Function

During the last several years in the programming field I have seen Junior’s (even more seasoned guys) make the simple mistake of using Function’s for everything. Some even used Sub’s… I suspect it’s because they didn’t know any better.

True, nowadays with the advent of .NET and PHP languages that instill (or force for that matter) you to use the correct syntax for each moment, I feel as hobbyists who won’t necessarily go to school to learn this, might not be clued up.

Simply stated. a Function returns a value, a Sub does not.

An example of this in Classic ASP:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

Dim myVariable : myVariable = "AcidRaZor"
 
If checkIfTrue(myVariable) Then
Response.Write "This is indeed true"
End If
 
Function checkIfTrue(var)
Dim ret : ret = false
If var = "AcidRaZor" Then
return = true
End If
 
checkIfTrue = ret
 
End Function

An example of a Sub would be:

1
2
3
4
5
6
7

Dim myVariable : myVariable = "AcidRaZor"
 
saveThisName myVariable
 
Sub saveThisName(var)
Session("nameSaved") = var
End Sub

This spills over (and hopefully answers some of your questions) into the new versions of .NET and beyond. Returning variables from Function’s or just processing data with Sub’s. Soon the student will become the teacher

Date Defaults – What every programmer should know

A few years back when I started programming my first task was to code an intranet system able to track meetings and certain tasks employees did throughout the day based on a calendar. This was before I knew what power SQL Server 2000 (and now 2005/2008) had and how easy it was working within those systems.

MS Access (which changed a lot since 1999) had one flaw. It always kept the date formats in mm/dd/yyyy (America is the “world” hahahahaha) so this made my job increasingly difficult trying to handle the dates on a South African based time (dd/mm/yyyy)

I struggled hard to understand the shameless act to which MS Access butchered my date inputs but I couldn’t help but think that this CAN’T happen to all programmers!!

After a lot of heart-ache and a lot of sleepless nights I finally found the silver bullet to the werewolf that was MS Access (and other servers worldwide running on different timezone settings)

dd-MMM-yyyy!!! Translated into: 01-Jan-1999 for example, you can insert ANY date with 100% surety that your sleep WILL guaranteed. This universal date format will work in ANY (yes you read it! ANY) database known to man and even though your raw data may look like 09/02/1999 in the database, it will always insert the correct date in the correct format in the database irrespective of what timezone the server is set on!

Classic ASP exampes of reading the date out is:

1

Day(date()) & "-" & Left(MonthName(Month(Date())),3) & "-" & Year(Date())

SQL Server example would use DateName instead of MonthName (which can be abbreviated without the Left() function I had to use in ASP) and of course, date() = getdate() in SQL Server

I would always recommend reading it out of the database (whenever possible) with stored procedures in the correct format to begin with. But it’s up to you to decide how to use this gem of knowledge I just thrown at you. And if you’re reading this thinking “But I already knew this!” then think again. Unfortunately, in 2009, there are still quite a few programmers unable to grasp simple concepts like timezones and date handling.

Remember kids. Making mistakes when you’re still new in the game is one thing. Brandishing a CV (Resume) of 3 years+ experience and still making these mistakes is just plain wrong… you chose a field where you have to evolve and look for answers constantly, don’t stagnate… I’ll haunt your dreams if you do