Tactical Tech Tips

Urban Disaster Preparedness

2010-12-21T20:32:00.000-08:00

Below are some simple stockpiling suggestions to help city dwellers survive a disaster.

Why should I stockpile food, water, and other supplies?

Because bad things can happen. Natural disasters such as earthquakes, volcanoes, hurricanes, floods, fires, or severe storms could force you to survive on your own for up to two or three weeks. Flu pandemic, terrorist attack by nuclear / biological /chemical /radiation, or economic collapse can breakdown society for several weeks or months. Grocery stores may be empty. The power grid, water/gas lines, transportation, fuel, and highways may all be broken or shut down. If you wait until you need it, it's a sure bet others have beat you to it.

Won’t the government provide for me in an emergency?

Nope, FEMA clearly states it's your responsibility to provide food, water, and other supplies for at least the first three days of a disaster. In extreme cases, you could be on your own for up to several weeks or even months.

Isn’t stockpiling costly or wasteful?
Nope. It's good insurance for surviving a disaster, and inexpensive if done right.

Tier 1: How do I cover basic 2-3 day stockpile needs?
Food shouldn't be a concern because most people typically have more than 2-3 days worth of food on hand, and could go that long without eating if necessary. The biggest need is having sufficient water on-hand. Use tap water in Walmart six-gallon water containers ($9), or use buckets with lids. To purify, either boil the water or add 8 drops of bleach per gallon.

Tier 2: How do I cover 6-8 week stockpile needs for more severe disasters?

For food, the simple approach is to buy bulk items having a long shelf-life. You could buy and store your own rice, beans, pasta, and other dry or canned goods; but Costco makes it easy. They have an Emergency food supply ($90) with a 20-year shelf-life that provides 275 meals (feed a family of four 2 meals a day for one month). They also have emergency rice ($50) and pinto bean ($60) buckets.

Costco offers a solid emergency supply kit ($140) that should cover most basic needs. It includes 90 servings of food, a water filtration system, hand crank flashlight/radio/cell phone charger, and typical survival supplies (matches, cord, space blankets, water pouches, compass, multi-tool, etc.). This can be your "go-bag."

Tier 3: What other stuff might be worth stockpiling?
* Fire: lighter; matches
* Fuel: extra propane tanks; kerosene; gasoline; wood
* Light: flashlights (hand crank?); candles; lantern; wicks
* Cooking: vegetable oil; cook stove; cast iron pans
* Flavoring: salt; sugar; honey
* Protein: tuna/chicken cans; boxes of oatmeal; dry milk
* Carb's: dry pasta w/cans of tomato sauce; crackers; dry cereal
* Toiletries: buy large Costco quantities for everyday use, always have spares
* First-Aid¹: band-aids, vitamins, iodine (f/radiation), rubbing alcohol, hydrogen peroxide
* Drugs: aspirin, analgesics, antibiotics, antifungals, antihistamines, antivirals, and steroids
* Water: buckets w/lid; bleach; filter (Britta?); identify local source
* Power: batteries; rechargeable batteries; solar recharger
* Weapons: (for defense) shotgun, pistol, rifle, ammo, bat, machete, sword, nun-chuks...
* Shelter: tarp; stakes; rope; twine; sleeping bag; blankets
* Seeds: if extreme TSHTF then start your own vegetable garden
* Help/Barter: buy more of the above than you need to help neighbors and for trading

Other Information and Planning:
* Family Communication Plan²: who to call out of state, etc.
* Utility Shut-Off³: where and how to shut-off the water, power, and gas
* Escape Plan²: when, where, and how to bug out to if the city becomes too dangerous
* Safety Skills: how to provide CPR, dress a wound, properly prepare water, etc.
* Literature: keep books and paper copies of disaster-specific survival tips (no internet?)

How do I store it?
In a cool, dark, place with low humidity and an ideal temperature of 60°F. I like buckets and lids (~$4) from Home Depot or Lowes because they stack well, seal well, and are portable. Ziploc bags inside the buckets keep contents air tight for a longer life. Use 3" mailing labels and a black felt pen to label buckets and ziploc bags.

Notes:
¹ First Aid: Ready.gov or 72hours.org
² FEMA Emergency Plans
³ Utilities: 72hours.org, FEMA

Release of MSSQL.SchemaDiff

2010-08-04T22:30:00.000-07:00

Last 2.5 months' spare time has been spent building MSSQL.DataMask. Just released v1.0.0.5 after several weeks of beta feedback. Very fun project. Use it multiple times a day at my real job when doing database updates and ensuring dev, test, staging, and production db's all in synch. Use it to ensure full setup script database matches udpate script database.

Download from: http://www.wintestgear.com/products/MSSQLSchemaDiff/MSSQLSchemaDiff.html

Description: MSSQL Schema Diff is a free database comparison tool used to quickly and easily locate structural differences between two SQL Server databases. Results are clearly displayed with MSSQL objects in a familiar SQL Management Studio tree layout and property comparisons in a matching data grid. Intuitive icons and color codes make it easy to spot differences. Filter buttons make it easy to narrow down to just what is different or missing from one or the other database. Usage is simple; you select two MSSQL databases, click Compare, then browse the differences

Use MSSQL.SchemaDiff to:

• Troubleshoot and find errors caused by database differences
• Ensure consistency between dev, test, staging, and production database environments
• Learn what changed between database versions

How Do I Uninstall a .NET Windows Service If the Executable Was Moved or Deleted?

2010-05-20T16:33:00.000-07:00

Was working on a project for another team at work. Decided to re-arrange folders and re-pull from source control. But after moving the folder, ran into problems. Wanted to uninstall the service and reinstall.

After some searching, happened across the right keywords in a Google search. Turns out David Klein has several answers at his blog: http://ddkonline.blogspot.com/2008/07/cannot-uninstall-net-windows-service.html.

Saved here for future reference. Thanks David.

How Do I Add My Company Logo Icon to Browser URL's

2010-04-21T16:31:00.000-07:00

Create a 16 x 16 icon. Name it "favicon.ico". Place it in your root folder.

That's all.

Note that the icon is only downloaded the first time a user visits your web page. Thus, if you cannot test view the icon initially after placing it...then clear your browser cache and try again...it will work.

Note that you can also embed a 32 x 32 icon in the same file as the 16 x 16 for clearer decktop shortcuts and favorites.

Beta Release of MSSQL.DataMask

2010-04-20T22:27:00.000-07:00

I've been buried for 3 months spending all spare time creating MSSQL.DataMask. Just released it in beta for feedback. Will go through several weeks of testing and bug fixing.

Download from: http://www.wintestgear.com/products/MSSQLDataMask/MSSQLDataMask.html

Description: If you need rich test data to develop, test, or outsource your project, you've probably tried either generating from scratch, or cloning from production data. Auto-generating the test data is difficult and error prone for all but the simplest databases due to complex variances, frequencies, and data interdependencies. Business and legal obligations such as HIPAA require that production data clones be thoroughly sanitized of protected personal or health information.

MSSQL.DataMask is a simple, free tool that will quickly sanitize a clone of your production database into a safe, secure test database. Once built, the process is easily repeatable to refresh your test data from production. You can either load and re-run a set of data masks from the application, or generate a fully documented tSQL script to modify, run, or schedule as your needs dictate.

Scrub: Overwrite all rows of a column with the "same value"
• Scrub.Null: Overwrite with null
• Scrub.EmptyString: Overwrite with empty string ("")
• Scrub.FixedString: Overwrite with a text you enter
• Scrub.Zero: Overwrite with zero (0)
• Scrub.FixedNumber: Overwrite with a number you enter
• Scrub.FixedDate: Overwrite with a date/time you enter

Substitute: Overwrite all rows of a column with a "unique value"
• Substitute.LoremGibberish: Overwrite with a random lorem ipsum gibberish text (retains original size)
• Substitute.GUID: Overwrite with a random GUID (16-character text only)
• Substitute.RandomNumber: Overwrite with a random number between min/max points you specify
• Substitute.RowNumber: Overwrite with a row number (and text you specify if n/varchar)
• Substitute.FromList: Overwrite with a random selection from value list you specify

Transform: Overwrite all rows of a column with an "obsfucated original value"
• Transform.Scramble: Overwrite with scrambled original value (encrypt, then truncate at initial length)
• Transform.Hash: Overwrite with hashed original value (MD5, always 16-bytes)
• Transform.Prefix/Suffix: Prefix and suffix original value (with text you specify, can leave blank)
• Transform.Find/Replace: Find & Replace text within original value (you enter find/replace text)
• Transform.AsteriskMask: Overwrite first (n) characters with asterisks (like credit card statements)
• Transform.ShuffleRecords: Randomly shuffle all rows within column (my favorite for most test data)
• Transform.NumericVariance: Randomly de/increment original numeric value by percent you select
• Transform.DateVariance: Randomly de/increment original date value by number days you select
• Transform.DateTimeVariance: Randomly de/increment original date & time values by your selections
• Transform.Truncate: Cutoff original value at character length you specify

Test Case Manager Lite

2010-01-30T14:50:00.000-08:00

Do you develop and test software...if so you may often run into these questions...

1. Are we ready to ship?
2. How much longer will it take to test?
3. If we release now, what is our exposure (lack of test coverage)?
4. Are we systematically regression testing existing functionality, or adhoc testing differently each time?
5. Are we capturing every bug in a test case to prevent recurrence?
6. How is our quality trending from test cycle to test cycle?
7. Do we have proven written test cases on which to start automating?

Just released TCMLite as a free, simple, highly configurable test case management tool enabling you to better answer these questions. Use TCMLite to design, write, and execute your test cases. Your entire test team can concurrently access the same test case repository (XLS file) using Excel's "shared workbooks" feature (Help section 1.4).

Best of the Best Software Testing Articles

2010-01-16T21:09:00.000-08:00

Here are some of the best testing articles I've come across over the years...

How Do You Know When You Are Done Testing? article
James Bach's Test Automation Snake Oil article
IBM 28 best Practices for software testing
Brian Marick's Classic Testing Mistakes article
Brian Marick's Failure Improvement...how to refine a bug article
James Bach's What is Exploratory Testing? article
Wikipedia for a great overview of Software Testing

How do I quickly get record counts for all MSSQL tables?

2010-01-12T22:11:00.000-08:00

Frequently when testing, you will need a quick way to get record counts for all tables in a given database. For example, you can take before and after snapshots of all tables' record counts to quickly ensure the proper work was done on the right tables.

The following two-line SQL statement will automatically generate a much larger, custom-built SQL script that can be copied and executed to fetch counts for all tables in the target database.

Follow these basic steps to use it:

STEP 1: Open SQL Server Management Studio

STEP 2: Open a "New Query" pointing to target database of your choice

STEP 3: Copy and run the generic "SQL Script Generator" statement below:
   SELECT 'SELECT ''' + TABLE_NAME + '''
     AS TableName, COUNT(*) AS
     RecordCount FROM ' + TABLE_NAME
   FROM INFORMATION_SCHEMA.TABLES

STEP 4: Copy "Auto-Generated-SQL-Script" (AGSS) from below (yellow execution results)

STEP 5: Paste the AGSS from Figure 2 (yellow highlight) into a new Query window

STEP 6: Execute the AGSS to see table counts for the entire database as shown below (green highlight)

How do I quickly get MSSQL table sizes?

2010-01-11T21:54:00.000-08:00

Frequently when testing, you will need a quick way to get the size (in kilobytes) for all tables in a given database. For example, you can take before and after snapshots of all tables' sizes to quickly ensure the proper work was done on the right tables.

The following two-line SQL statement will automatically generate a much larger, custom-built SQL script that can be copied and executed to fetch sizes for all tables in the target database.

Follow these basic steps to use it:

STEP 1: Open SQL Server Management Studio

STEP 2: Open a "New Query" pointing to target database of your choice

STEP 3: Copy and run the generic "SQL Script Generator" statement below:
    SELECT 'EXEC sp_spaceused '
    + TABLE_NAME + ';'
    FROM INFORMATION_SCHEMA.TABLES

STEP 4: Copy "Auto-Generated-SQL-Script" (AGSS) from text above (yellow execution results)

STEP 5: Paste the AGSS from Figure 2 (yellow highlight) into a new Query window

STEP 6: Execute the AGSS to see table counts for the entire database as shown in Figure 3 (green highlight)

How do I flush the MSSQL cache for performance testing?

2010-01-10T21:29:00.000-08:00

When performance tuning, you often follow a cycle of changing the tSQL, running it in query analyzer, jotting down the new execution time...and repeating the cycle again and again to reduce the time.

Because MSSQL caches results, your performance timings can be inaccurate. For example, the first run does the work and takes say 35 seconds. All subsequent runs might take 1 second or less because they are not doing any work. The subsequent runs are simply fetching the previously cached results.

This script resolves the problem by flushing the cache. Each subsequent run after a cache flush forces MSSQL to do real work thereby ensuring accurate performance test measurements.

/* Clear MSSQL before running performance tests */

-- 1) Clear buffer cache;
-- write dirty pages to disk & clean buffers.
CHECKPOINT

-- 2) Remove clean buffers from buffer pool
DBCC DROPCLEANBUFFERS

-- 3) Remove elements from stored proc cache
DBCC FREEPROCCACHE

Free Tool for Comparing MSSQL Databases

2010-01-09T20:27:00.000-08:00

There is frequently a need to compare two databases when developing or testing a data-driven application.

You'll want answers to questions:

Did a table change?
Did some fields get added, deleted, renamed, or have a type change?
Did views or triggers or indexes get modified or added?
Did source code inside a stored procedure change?
Is the testing schema different from the development schema?
Have there been any changes in the development schema since last week?

Having written a freeware diff tool for Microsoft Access back in 1999, I was considering writing a diff tool for MSSQL. The first step in my project was market research to determine whether there is a need for a freeware MSSQL Diff tool.

My project ended here in a quick blog because it turns out there are some excellent free MSSQL diffing tools on the market already. After reviewing several, I found DBComparer to be the best and added it to my Freeware Tool Server list. DBComparer does a comprehensive job of comparing schemas, including 22 object types ranging from tables to views to stored procedures. It scripts out the differences in a separate window, and has an easily navigable tree to walk through all objects' differences.

Although DBComparer does a fine job of comparing schemas, it does not compare data between two databases. If you need more functionality such as this, then consider purchasing products such as Microsoft Visual Studio Team Edition (built-in schema compare), RedGate SQL Compare, RedGate Data Compare, ApexSQL Diff , or Apex Data Diff.

Oops, I've Been Breaking Many .NET Coding Standards!

2010-01-02T20:08:00.000-08:00

Back in the classic VB3-6/VBA days, I heavily used hungarian notation as it was the naming convention standard. Assumed I should continue following the same standard when naming my objects in .NET. Found out this is wrong and that I've been violating several .NET / C# best practices.

I happened across Steve Orr's website which pointed me to several useful MSDN articles.

Some great takeaways regarding what I've been doing wrong include:
1) Naming Conventions, General:

a) Do not use underscores, hyphens, etc.
b) Use CLR type names, not language-specific names...that means use Int16, not C#'s short
c) Do not prefix with "C" for class or "O" for object like used to in classic VB

2) Naming Conventions, Upper Case:

a) Do not use Hungarian notation
b) All objects are pascal-case, except parameters which are camel-case
c) Two letter acronyms are always uppercase unless they are at the beginning of a parameter name, then they match and all stay lowercase
d) Three+ letter acronyms are treated like words (first letter only upper case)

3) Namespaces

a) Avoid too many or too deep namespaces
b) Separate advanced namespaces from common namespace by suffixing with "xxx.Advanced"

4) Static Classes:*2

a) Use static classes sparingly, they are not a miscellaneous bucket

5) Property Design

a) Return original value if error encountered

6) Parameter Design:

a) Always place out parameters to the right after byval and ref parms...even if inconsistent overloading order

Pascal-Case:
First letter of every word is uppercase (e.g.: SampleObjectName)

Camel-Case:
Same as pascal case, except first letter always lower case (e.g.: sampleObjectName)

Wait! Don't throw out that old computer that won't boot

2010-01-01T21:38:00.000-08:00

The fix can be as simple as a $4 CMOS (watch) battery.

My evening backups failed two nights ago. A ping didn't respond. When I hooked up a monitor, the Windows logo was frozen onscreen.

Okay, no problem, BSOD's or freezes can happen on old hardware running windows. Reboot should do the trick.

I rebooted and the system powered down. It powered back up. The powersupply fan came on... Good. The harddisk and keyboard lights flickered...Good.

But no...now a worse symptom. At this point, reboots either a) froze with a black screen 2 seconds into the bootup, or b) popped up the BIOS settings screen with an error about the CPU being recently changed.

Okay, should not be a problem...will just boot to linux OS on CD via Acronis backup. That'll bypass all Windows and harddrive problems.

But again no...would not boot to CD.

Sigh, time to spend an hour ripping out all the components. Thinned down to the bare minimum: keyboard, boot harddisk, and video card. Removed the extra RAM, the sound card, the extra USB card, floppy disk, even the CDROM.

But again no...would not boot.

Time to call my brother, the electrical engineer.

"Hey Q, I'm about to throw my old server away. I think the motherboard is bad. I stripped everything down to the bare minimum and the system still won't boot to linux on CD. Before I toss it, is there anything else I should try?"

My brother of course had an easy answer which turned out to be the right answer...

"Yes, it's probably a dead or dying CMOS battery. I've seen systems crash just like this when the voltage is below blah, blah, blah. You can confirm with a voltmeter by measuring blah, blah, blah. If that is the problem, then you can pick up a new CR2032 watch battery (that is all a CMOS battery is). They are just $4 at Fred Meyer, Kroger, or a similar store."

Sure enough, I put in the new battery and my system was healthy and running again.

P.S. - When the CMOS battery dies, your system may lose the BIOS settings for CPU type, extra HD info, etc. If enough crtical BIOS settings "change", it can trigger Microsoft Windows to go through a reactivation process. It happened to me...but that is another story.

So, don't toss out that old computer that you think has a fried motherboard...it could just be a $4 CMOS battery.

Nothing Like a Good Post Mortem

2009-12-30T06:08:00.000-08:00

Over the years, I have attended good post mortems...and not-so-good post mortems. Good post mortems tended to always have an agenda emailed out first, the feedback collated, good meeting guidelines, and a meeting that stays on topic.

I've incrementally improved the following links over the years. Hope you find these templates useful:
1. Post Mortem Template
2. Post Mortem Meeting Agenda

Why Is Our Site So Slow For Just One Customer?

2009-12-27T19:31:00.000-08:00

A customer reported page loads were exceeding 60-seconds; but they should have been 3-seconds or less.

Our site monitoring alerts didn't report a problem. Other customers online at the time didn't report a problem. Running our performance test battery didn't show a problem.

It was time to watch her reproduce the problem. We setup an online meeting and temporarily made her host. We observed her actual performance -- measuring it at 21 times slower than expected! Further testing revealed that she didn't have a problem from her home computer...

But why? Why just the corporate computers? Was it a faulty hardware? Was it machine configuration? Was it network configuration?

Next, we asked her to browse to http://www.speedtest.net/ and run the test.

We compared her test results against our baseline. Her latency (ping) was averaging 226ms...but the norm is around 35ms. Bingo! Latency was the culprit.

We soon discovered that this customer routed all corporate traffic through Atlanta. So, she was physically located 250 miles from us on the West Coast, but her traffic was routed 6,000 miles (round trip) across the U.S.

More hops and more switchgear to go through yielded larger latencies. A slow central proxy server in Atlanta also compounded the problem.

_______________________________________________
Content Acceleration:
Latency is often just an exacerbating symptom, not the root issue. Excess server requests are more typically the cause of poor performance (poor page design, or large session state management, etc.). If you can’t re-code, then research “content acceleration” or “web acceleration” as a viable work-around for poor latency. Content acceleration reduces the number of hops by caching content around the world and using faster, more direct links (less switchgear = less latency). Content acceleration also compresses the content where possible._______________________________________________
What is Speedtest.net?
Speedtest.net is the best speed test sites out there, performing over 1 million tests per day from hundreds of servers dispersed throughout the world. They perform three key measurements to determine the overall quality and performance of your Internet connection: download speed, upload speed, and latency.

Latency or Ping is the time it takes in milliseconds for a small piece of data to be sent from your computer to the Internet and back – the more switch gear or extra indirect hops to pass thru, the slower the connection (think flow velocity inside the pipe).
• Expect 20-30ms for geographically close servers.
• Expect 80ms for cross-country traffic (5,000km / 3,075mi).
• Greater than 100ms is bad.

Download Speed is the rate at which data is sent down from the Internet to your computer (think diameter of the pipe feeding in).
• Dialup: Expect 0.4-2Mbps
• DSL: Expect 1-2Mbps for standard, and 3-7Mbps for premium
• Cable: Expect 1-6Mbps for standard, and 8-20Mbps for premium

Upload Speed is the speed at which data is sent up from your computer to the Internet (think diameter of the pipe feeding out). Upload speed is typically many times slower than download speed (except dialup where they are equal).

_____________________________________

How Do I Use Speedtest.net?

(A) Browse to http://www.speedtest.net/.

(B) When you first enter the website, it checks your IP address to determine your approximate location on earth.

(C) Clicking the “Begin Test” button will run the test.

(D) Run results show the download rate, upload rate, and ping (latency).

(E) Download speed estimates show how long a typical MP3 file or video clip takes to download on your network.

For a more detailed analysis of latency and related delay metrics, visit http://www.pingtest.net/.

The site measures latency, jitter, and packet loss…and even includes a grade for your line. Here is a screenshot:

_______________________________________________

Other Notes:

Speedtest.net Requirements:
• Web browser
• Javascript
• Flash 8 installed

Avoid Inaccurate Results:
While you are running your speed test, be sure...
• No other computers are online stealing bandwidth
• Your test computer is not busy doing anything else
• You are direct connected (not over wireless)

Benchmarks:
You can compare your performance against the World Results section to browse top bandwidth speed by country, or narrow down statistics to very specific locations and even specific ISP’s.

Other Great Speed Test Sites:
• Audit My PC:
   http://www.auditmypc.com/internet-speed-test.asp
• CNET Speed Test:
   http://reviews.cnet.com/internet-speed-test/
• ComCast Speed Test:
   http://speedtest.comcast.net/