10 Reasons To Quit Your Job Right Now!

May 28, 2012

via TechCrunch by James Altucher on 5/26/12


The game is over. That game where they get to hire you for 40 years, pay you far less than you create, and then give you a gold watch, and then you get bored, you get depressed, and you die alone.

It wasn’t that fun of a game anyway.

When I had a corporate job I would wake up depressed. I couldn’t move out of bed. The sun would be coming in. A cat on the fire escape staring at me through the window. Even it was more excited to be alive than me. And, by the way, I had the best job in the world. I interviewed prostitutes for a living at three in the morning.

But they were going to kill me in my cubicle.

In 2009 I asked about 10 Fortune 500 CEOs, “did you just use this crisis as an excuse to fire all the people you were afraid to fire before.” Only one said “of course” instantly. The others had to drink more. But then it was admitted: you’re all dead weight and there’s no loyalty.

We’ve entered the “Choose Yourself” era. The era without middlemen. Without The Other telling you your bonus, your salary, your movie can be made, your book published, your company funded, your life validated. The era where you have to always be planning your escape. Where you create your platforms on twitter, facebook, quora, pinterest,  blogging, vlogging, itunes, and wherever else and every day you Create and you Innovate and you Sell for yourself. You Eat what you Kill. And your rewards are commensurate with how sharp your teeth are.

Most people need to begin planning their exit strategy RIGHT NOW:

So here’s the 10 reasons you need to quit your job right now. And below that I have the methods for doing it.

1) Safety. In ancient history you would start as the shoeshine boy, move to the mailroom, impress someone with your go-get-it attitude, become an assistant account executive, move up the ranks, move horizontally to another company, get promoted again, move vertically, horizontally, zig zag across corporate America and eventually retire with your IRA savings. The myth was over in 2008. It never really existed but now we know it’s a myth. You were addicted to the stability. The white picket fence. Getting away from home for ten hours a day. I understand. But it was an addiction. And the fix is gone. Your job was never safe. And it’s less safe now than it was yesterday. A billion people in China need a job and they are gunning for your cubicle.

2) Home. Everyone thinks they need a safe job so they can save up to buy a home and also qualify for a mortgage. Mortgage lenders at the banks like people who are like them – other people locked in cubicle prison.  They want to see an income statement. A tax return. A credit check. A stability check. A note from your therapist. Everything that proves you are a reliable human just like them. Well now you don’t need to worry about that. Here’s why you should never own a home in the first place. Save yourself the stress.

3) College. Everyone thinks they need to save up to send their kids to college. Depending on how many kids you have and where you want them to go to college it could cost millions. Well now you don’t need to send your kids to college. So you don’t need to stress about that money anymore.

4) Your boss. Most people don’t like their boss. Its like any relationship. Most of the time you get into a relationship for the wrong reasons. You were too young. You didn’t know what you wanted. You really loved the other girl but she rejected you. Eventually you’re unhappy. And if you don’t get out, you become miserable and scarred for life. That’s why 1 in 2 marriages end in divorce. That’s why you need to quit your job.

5) Your coworkers. Look around. Are these the people you were meant to spend the rest of your life with. You will spend more time with them then you will spend with your children.

6) Fear. We have such a high unemployment rate, people are afraid if they leave the job they are miserable at, they won’t be able to get a job. This is true if you just walk into your boss’s office and pee on his desk and get fired.  But its not true if you prepare well. note that we’ve just had 26 consecutive months in a row of private sector job growth. Much of that is people working at one or two-person companies (i.e. “startups”).  More on that in a bit.

7) The Work. Most people don’t like the work they do. They spend 4 years going to college, another few years in graduate school, and then they think they have to use that law degree, business degree, architecture degree and then guess what? They hate it. They made a bad decision when they were 18. They chose “LAW”. Or “ECON”. But they don’t want to admit it. They feel guilty. They are in debt. A trillion dollars in debt backed by the US government. No problem. Read on.

8) Bad things happen. You start to get depressed and you don’t know why. You start to feel like your life didn’t add up to what it should’ve. SOMETHING WENT WRONG. You start to physically ache. You get nervous about bonuses, promotions, who gave credit to who? You play politics (an ugly game), you fantasize about selling diet pills (Tim Ferris did it!), adults yell at you for irrational reasons, you have sex with another girl at work. Now work is like one big sexually transmitted disease.  And it gets worse and worse. You don’t want to look back at your life and say, “man, those were the worst 45 years of my life.” That wouldn’t feel good.

9) The economy is about to boom.  I don’t care if you believe this or not. Stop reading the newspaper so much. The newspapers are trying to scare you. Bernanke just printed up a trillion dollars and airlifted it onto the US economy. Who is going to scoop that up. You in your cubicle? Think again. And just what is a “Greece”? Is it that tiny country with the economy the size of Rhode Island that other countries have been supporting since Augustus paid all their bills in 20 BC? Just what are they?

10) Your job has clamped your creativity. You do the same thing every day. You want to be jolted, refreshed, rejuvenated.

Note: I will grant some people love their jobs. This is not for them but the 90% who don’t.

But, you say: you still need to support yourself, you still need to support your family, you can’t just walk into your boss’s office and quit.

Good point. You need to prepare. Its like training for the Olympics if you feel now is the time to move on from your job. You need to be physically ready, emotionally (don’t quit your job and get divorced on the same day for instance), mentally (get your idea muscle in shape) and spirituall all ready.

The posts that will help you quit your job. To quit, at least follow the ideas in the first post:

–          How to be the Luckiest Man Alive in 4 Easy Steps

–          What to do if you were Fired Today

–          The 100 Rules for Being an Entrepreneur

In the above link, it’s not about starting a business. It’s about finding what your frontier is, how to explore it, how to test the waters and move beyond it. I’m not saying I can do this. I’ve hit my boundary so many times and bounced off that I have six broken noses to show for it.

Some notes on this post:

Note #0: Why is this on techcrunch? Because all people want to know that they have a choice. That they can eat what they kill. That in the “Choose Yourself” era it’s ok to make the leap into the unknown, in the abyss, do your startup, save the world, deliver value, invent, create, make money, and have fun. You don’t have to do what is expected of you.

Note #1: I get a lot of criticisms from anonymous people in the message boards. Claudia begs me, “Don’t look at the message boards unless you talk to me first.” Because she knows I’m an addict. I tell her ‘ok’ but I know I’m going to look. Because that’s what addicts do. I’m not selling anything. If you want any of my books and can’t afford then write me and I’ll send for free. I’m not pushing any agenda. I have nothing to gain by you quitting your job.

Note #2: Sometimes people criticize the “list” format in these posts. “10 reasons” for this. “10 reasons” for that. About 20% of my posts are lists. Not so much. Read “Happiness this Second”  for a recent non-list post. And Charlton Heston clearly didn’t mind lists when he came down from Mt. Sinai with “The 10 Commandments”, the very first blog post. 3500 years later and still getting clicks.

(the first blogger)

I don’t mind when people critique me when they’ve lost, quit, or have been fired from as many jobs as I have. Or lost a home. Tried to raise two kids with almost nothing. Been as desperately unhappy as sometimes I’ve been. This doesn’t qualify me for anything, of course. Maybe it disqualifies me. Who cares?  A lot of people have had much worse than me. And I’ve been very blessed as well. I’ve been able to come back.

Sometimes  you can build back up. And sometimes you just think, “How the hell did this happen to me again”.  My goal in these posts is to help people maybe think for a split second they can reduce some stress in their lives, they don’t have to go through what I went through, they can throw themselves into experience and still come back alive, and at the end of the day, they can use some of these ideas to live a better and more fulfilling life. I’ve had that experience and I like to write about it.

Later tonight I’m going to give my two daughters, ten and thirteen years old, two choices and ONLY two choices. Either they watch “Star Wars” with me or they watch “Schindler’s List”. And if they don’t like either choice then maybe I’ll just sit by the TV with some ice cream and watch all by myself.



rich.horwood@gmail.com sent you a link to content of interest

April 30, 2012
rich.horwood@gmail.com sent you a link to the following content:

Book Excerpt: Bruce Perry’s Fitness For Geeks

The sender also included this note:


It’s Almost as If Facebook Apps Should Be LESS Obnoxious!

April 26, 2012

Interesting read on cloud computing

November 18, 2011

Top 10 DTrace scripts for Mac OS X

October 11, 2011

via Brendan's blog by Brendan Gregg on 10/10/11

Since version 10.5 “Leopard”, Mac OS X has had DTrace, a tool used for performance analysis and troubleshooting. It provides data for Apple’s Instruments tool, as well as a collection of command line tools that are implemented as DTrace scripts. I’m familiar with the latter as I wrote the originals for the DTraceToolkit, which Apple then customized and enhanced for Mac OS X where they are shipped by default (great!). I use them regularly to answer this question:

why is my MacBook slow?

I work in an office where everyone has MacBook Pros, and “why is my MacBook slow?” is a common question. Applications can become slow or unresponsive while waiting for CPU work, memory requests or disk I/O to complete.

For people who try to ignore the slowdown, the question can become:

why is my MacBook fan so loud?

Standard performance analysis tools like Activity Monitor and top(1) (and any third-party tools based on the same foundation) can’t tell you some key information about activity on your system, such as how much CPU consumption is caused by short-lived processes, or which processes are causing disk I/O. DTrace, however, can see (just about) everything.

In this post, I’ll cover the top ten Mac OS X DTrace scripts that I use for figuring out why laptops are slow or why applications are misbehaving. Most of these scripts are already installed, a few are from the new DTrace book.

How to use DTrace

If you’ve never run a DTrace script before or even used the command line, here’s a basic walkthrough:

Open Terminal

Run the “Terminal” application. You can find it here in Finder:

You can also type “terminal” in Spotlight (the magnifying glass in the top right corner of your Mac’s screen), which should find it. I usually drag it to my Dock so it’s easy to find later:

Customize Terminal

When you first run Terminal, it’ll probably look like this:

I find the default font small and hard to read. The size of the window in terms of characters is also small (80 columns by 24 rows), presumably to pay homage to original Unix terminals of that size. (Why Unix? Mac OS X is Unix under the hood: the Darwin kernel).

You can adjust the font to your liking in Terminal->Preferences->Text. I use Monaco 13pt, with “Antialias Text” on. Under the “Window” tab is the default size, 80×24, which you can also increase later by clicking and dragging the bottom right corner of the terminal. For example, here’s my screen as I write this blog post (in a terminal-based text editor).

Running top

While DTrace can see everything, there are some things already covered by easy-to-use (and easy-to-type) tools, like top(1). Now that you have Terminal running, type “top -o cpu” and hit enter, which will refresh the screen showing top running processes. Type “q” to exit the top program. The output will look something like this:

Brendan-2:~ brendan$ top -o cpu [...] PID COMMAND %CPU TIME #TH #WQ #PORT #MREGS RPRVT RSHRD RSIZE 67254- firefox-bin 98.8 04:53:45 62/1 1 743 9935+ 374M+ 78M 637M+ 86550 top 15.3 00:28.60 1/1 0 24 33 1748K 264K 2324K 1068 Terminal 13.2 17:38:19 5 1 358 1326 37M 67M 61M+ 61501- Google Chrom 7.9 13:40.89 6 1 93- 556- 45M- 73M 77M- [...]

I’ve truncated the header block to just include the columns. Look at the “%CPU” column to see which processes are making the CPUs busy, and the “RSIZE” column to see who is consuming main memory. The busiest (CPU) process will be at the top, since we sorted on cpu (“-o cpu”). The top was “firefox-bin” (Mozilla Firefox) at 98.8% CPU, which is in terms of a single processor (this has two). If firefox stayed that high you could look for the responsible tab and close it down, or restart Firefox.

After top, I turn to DTrace.

Running DTrace

DTrace requires admin privileges, so to use it you’ll usually need to type in a password to authenticate, provided you have administrator access (if you aren’t sure you do, click here to see how to check).

You can run DTrace by prefixing your DTrace commands with “sudo”, which will prompt for the password the first time around (but not for some time after that). When a DTrace command is running, you usually type Ctrl-C to end it. Here’s an example:

Brendan-2:~ brendan$ sudo iosnoop Password: password UID PID D BLOCK SIZE COMM PATHNAME 503 67261 W 384070496 73728 TweetDeck ??/Cookies/Cookies.plist_tmp_67261_0.dat 503 67261 W 384070640 4096 TweetDeck ??/Local Store/td_26_brendangregg.db-journal 503 67261 W 384070640 4096 TweetDeck ??/Local Store/td_26_brendangregg.db-journal 503 67261 W 308056800 4096 TweetDeck ??/Local Store/td_26_brendangregg.db 503 67261 W 308056856 4096 TweetDeck ??/Local Store/td_26_brendangregg.db ^C brendan@macbook:~>

If sudo says “permission denied”, check your admin status.

and here’s what happens if you forgot the “sudo”:

Brendan-2:~ brendan$ iosnoop dtrace: failed to initialize dtrace: DTrace requires additional privileges


The scripts follow. A note on style: if the script ends with “.d”, it’s a basic DTrace script. If it doesn’t, then it’s a shell-wrapped script that provides command line options. Use “-h” to list them.

And if you’d like to learn DTrace, I’d recommend starting by reading the basic ones.

1. iosnoop

This “traces” disk I/O execution live. Each time a disk I/O completes, a line of output is printed to summarize it, including process name and filename details:

Brendan-2:~ brendan$ sudo iosnoop UID PID D BLOCK SIZE COMM PATHNAME 503 54079 R 286522800 4096 Google Chrome ??/Cache/data_2 503 54079 R 286522808 4096 Google Chrome ??/Cache/data_2 503 34852 W 385001216 53248 Adium ??/Default/.dat8824.01b 503 65002 W 308056800 4096 TweetDeck ??/Local Store/td_26_brendangregg.db 503 65002 W 308056864 4096 TweetDeck ??/Local Store/td_26_brendangregg.db 503 65002 W 385001320 4096 TweetDeck ??/Local Store/td_26_brendangregg.db-journal 503 65002 W 385001320 4096 TweetDeck ??/Local Store/td_26_brendangregg.db-journal 503 65002 W 385001320 4096 TweetDeck ??/Local Store/td_26_brendangregg.db-journal 503 54079 W 385001384 12288 Google Chrome ??/Default/Cookies-journal 503 54079 W 385001384 4096 Google Chrome ??/Default/Cookies-journal 503 54079 W 134993856 4096 Google Chrome ??/Default/Cookies 503 54079 W 134994056 4096 Google Chrome ??/Default/Cookies 503 54079 W 134994176 4096 Google Chrome ??/Default/Cookies 503 54079 W 134994224 8192 Google Chrome ??/Default/Cookies [...]

This lets you instantly find out which applications are using the disk, and what files they are reading or writing to. Disk I/O is typically slow (for non-SSD disks), so an application calling frequent disk I/O (a dozen per second or more) may run slowly as it waits for the disk I/O to complete.

The output columns show: UID = user ID, PID = process ID (unique identifier for the process), D = direction (R = read, W = write), BLOCK = location on disk, SIZE = I/O size in bytes, COMM = process name, PATHNAME = trailing portion of file pathname.

In that output I caught Google Chrome reading from a cache file (“data_2″), and writing to cookie files (“Cookies-journal” and “Cookies”). TweetDeck also wrote to database files (“td_26_brendangregg.db” and “td_26_brendangregg.db-journal”). The “??” is where the path information ends for iosnoop (if you are a darwin programmer and want to take a swing at improving that, see the fi_pathname translator in /usr/lib/dtrace/io.d).

The “-h” option lists options (don’t need “sudo” for this):

Brendan-2:~ brendan$ iosnoop -h USAGE: iosnoop [-a|-A|-DeghiNostv] [-d device] [-f filename] [-m mount_point] [-n name] [-p PID] iosnoop # default output -a # print all data (mostly) -A # dump all data, space delimited -D # print time delta, us (elapsed) -e # print device name -g # print command arguments -i # print device instance -N # print major and minor numbers -o # print disk delta time, us -s # print start time, us -t # print completion time, us -v # print completion time, string -d device # instance name to snoop -f filename # snoop this file only -m mount_point # this FS only -n name # this process name only -p PID # this PID only eg, iosnoop -v # human readable timestamps iosnoop -N # print major and minor numbers iosnoop -m / # snoop events on filesystem / only

For tricky performance issues I often use “-stoD” to get start and end timestamps for each I/O in microseconds, and a couple of different types of I/O time calculations.

In hindsight, I should have called it diskiosnoop, since “io” could refer to different locations in the kernel I/O stack.

2. hfsslower.d

This script answers an iosnoop FAQ: why dosen’t iosnoop see my application disk I/O?

The reason is that applications rarely request disk I/O directly, rather, they access a file system that does disk I/O on their behalf. To increase performance, the file system will usually try to cache as much file data as possible in main memory (DRAM). The application may (by some notion) think that it’s doing disk I/O, but it’s actually reading from very fast DRAM, thanks to the file system. Writes can also buffer in DRAM and write to disk later, which also speeds up application performance.

The hfsslower.d script measures I/O before it is processed by the HFS+ file system (Apple’s current default file system). iosnoop measures I/O after the file system, and only if it reaches disk:

A couple of points:

  • hfsslower.d will see a lot more I/O than iosnoop, as it includes file system cache hits.
  • hfsslower.d better reflects application performance, as it measures the same latency that the application directly suffered.

The hfsslower.d script is from the DTrace book, and can be found here. To run it, you’ll need to create a text file containing the script (or pull it from the DTrace book tarball), and make the file executable from Terminal by running “chmod 755 hfsslower.d”.

This script takes an argument which is the minimum number of milliseconds to show I/O for. Here’s tracing 1ms HFS+ I/O and slower:

Brendan-2:~ brendan$ sudo ./dtbook_scripts/Chap5/hfsslower.d 1 TIME PROCESS D KB ms FILE 2011 Sep 27 19:00:15 Google Chrome R 0 15 data_5 2011 Sep 27 19:00:15 Google Chrome R 0 10 data_0 2011 Sep 27 19:00:15 Google Chrome R 0 10 data_0 2011 Sep 27 19:00:15 Google Chrome R 0 10 data_5 2011 Sep 27 19:00:21 TweetDeck W 0 1 td_26_brendangregg.db-journal 2011 Sep 27 19:00:25 Adium R 0 18 Smile.png 2011 Sep 27 19:01:08 firefox-bin W 0 17 _CACHE_001_ 2011 Sep 27 19:01:36 firefox-bin W 0 9 _CACHE_001_ 2011 Sep 27 19:01:36 firefox-bin W 2 14 _CACHE_002_ 2011 Sep 27 19:01:36 firefox-bin W 0 7 _CACHE_001_ 2011 Sep 27 19:01:37 firefox-bin W 0 1 _CACHE_001_ 2011 Sep 27 19:01:40 firefox-bin W 0 6 _CACHE_001_ 2011 Sep 27 19:01:46 firefox-bin W 0 14 _CACHE_001_ 2011 Sep 27 19:01:49 firefox-bin R 15 1 _CACHE_003_ 2011 Sep 27 19:01:49 firefox-bin W 12 7 _CACHE_003_ 2011 Sep 27 19:01:49 firefox-bin W 0 4 _CACHE_001_ 2011 Sep 27 19:01:51 firefox-bin W 0 9 _CACHE_001_ 2011 Sep 27 19:01:52 firefox-bin R 21 27 Times.dfont/..namedfork/rsrc 2011 Sep 27 19:01:52 fontd R 0 16 annex_aux 2011 Sep 27 19:01:52 firefox-bin W 0 5 _CACHE_001_ [...]

Columns are: TIME = time of I/O completion, PROCESS = application name, D = direction (R = read, W = write), KB = I/O size in Kbytes, ms = I/O latency in milliseconds, FILE = filename.

If you use the argument “0″, it will trace everything. If I’m chasing down slow I/O, I’ll often use an argument of “10″ for I/O slower than 10 milliseconds.

At this point you may think: if you just care about slow I/O, then just use iosnoop. That works to a point, but there can be slow I/O caused by something other than disks (file system lock contention, for example). The other advantage of the hfsslower.d script is that the measured latency matches the application pain suffered, whereas at the disk level you can only assume a correlation.

3. execsnoop

This traces the execution of new processes. This is great at identifying short-lived processes that may be caused by misbehaving applications and can slow down your system. These short-lived processes are usually too quick to be picked up by standard monitoring tools like the Activity Monitor or top(1).

To demonstrate this tool, here’s what happens when you type “man ls”:

Brendan-2:~ brendan$ sudo execsnoop -v STRTIME UID PID PPID ARGS 2011 Sep 28 20:19:18 0 67234 66312 man 2011 Sep 28 20:19:18 0 67234 66312 man 2011 Sep 28 20:19:18 0 67235 67234 sh 2011 Sep 28 20:19:18 0 67235 67234 gzip 2011 Sep 28 20:19:19 0 67236 67234 sh 2011 Sep 28 20:19:18 0 67234 66312 man 2011 Sep 28 20:19:19 0 67236 67234 gzip 2011 Sep 28 20:19:19 0 67234 66312 man 2011 Sep 28 20:19:19 0 67244 67243 less 2011 Sep 28 20:19:19 0 67242 67239 sh 2011 Sep 28 20:19:19 0 67242 67239 gzip 2011 Sep 28 20:19:19 0 67240 67238 tbl 2011 Sep 28 20:19:19 0 67241 67238 groff 2011 Sep 28 20:19:19 0 67245 67241 troff 2011 Sep 28 20:19:19 0 67246 67241 grotty

This prints a line for each new process that is executed, in a rolling output similar to the previous tools. The fields are: STRTIME = (string) timestamp, UID = user ID, PID = process ID, PPID = parent process ID, ARGS = process name (should be process + arguments, but that doesn’t yet work on Mac OS X; if you want to debug, see pr_psargs in /usr/lib/dtrace/proc.d).

Here’s what happens when I turn AirPort (wifi) off, then on:

Brendan-2:~ brendan$ sudo execsnoop -v STRTIME UID PID PPID ARGS 2011 Sep 28 20:28:00 0 67204 1 airportd 2011 Sep 28 20:28:01 503 67205 140 fontworker 2011 Sep 28 20:28:01 0 67206 1 mDNSResponderHel 2011 Sep 28 20:28:06 0 49 1 autofsd 2011 Sep 28 20:28:10 0 67208 37 ManagedClient 2011 Sep 28 20:28:31 503 67209 140 fontworker 2011 Sep 28 20:28:43 0 67210 1 airportd 2011 Sep 28 20:28:49 0 67211 1 mDNSResponderHel 2011 Sep 28 20:28:49 503 67212 140 fontworker 2011 Sep 28 20:28:51 0 67214 1 kerberosautoconf 2011 Sep 28 20:28:51 0 67215 1 kerberosautoconf 2011 Sep 28 20:28:56 0 49 1 autofsd 2011 Sep 28 20:29:01 0 67219 37 ManagedClient 2011 Sep 28 20:29:04 0 67220 1 ocspd 2011 Sep 28 20:29:16 503 67221 54079 Google Chrome He 2011 Sep 28 20:29:19 503 67222 140 fontworker

Near the end I opened up a new tab in Mozilla Firefox and another in Google Chrome. You can see a fundamental difference: Chrome creates a new process for that tab, Firefox doesn’t.

execsnoop has various options, use -h to list them all. I used -v above, to print the time.

4. opensnoop

This traces file opens and prints various details, including the time and error code when using “-ve”. I usually use it to look for failed opens, which can be a sign of misconfigured applications. Discovering their config files and resource files can also be useful.

Brendan-2:~ brendan$ sudo opensnoop -ve STRTIME UID PID COMM FD ERR PATH 2011 Sep 30 01:44:15 0 11 DirectoryServic -1 2 /var/db/dslocal/nodes/Default/hosts 2011 Sep 30 01:44:15 0 11 DirectoryServic 20 0 /etc/hosts 2011 Sep 30 01:44:15 0 85198 dtrace 5 0 /etc/localtime 2011 Sep 30 01:44:15 503 146 SystemUIServer 19 0 /Users/brendan/Library/Preferences/com.apple.menuextra.clock.plist.47C2X9o 2011 Sep 30 01:44:15 503 146 SystemUIServer 19 0 /System/Library/CoreServices/Menu Extras/TimeMachine.menu/Contents/Resources/TMRotatingArrow.pdf 2011 Sep 30 01:44:15 503 146 SystemUIServer 19 0 /System/Library/CoreServices/Menu Extras/Clock.menu/Contents/Resources/Clock.pdf 2011 Sep 30 01:44:15 503 146 SystemUIServer 19 0 /System/Library/CoreServices/Menu Extras/TimeMachine.menu/Contents/Resources/TMRotatingArrow.pdf 2011 Sep 30 01:44:17 503 67261 TweetDeck 17 0 /Users/brendan/Library/Cookies/Cookies.plist_tmp_67261_0.dat 2011 Sep 30 01:44:17 503 67261 TweetDeck 17 0 /Users/brendan/Library/Cookies/Cookies.plist 2011 Sep 30 01:44:18 503 54079 Google Chrome 72 0 /var/folders/bU/bU0WYlnwE6KKnMBrvI5QUU+++TQ/-Tmp-/.com.google.Chrome.sQdN3D 2011 Sep 30 01:44:19 0 36 mds 9 0 . ^C

While tracing, I changed the clock in the top bar from digital:

to analogue:

The files opened to performed this can be seen by the “SystemUIServer” process, beginning with a Preferences file and then some PDFs. Huh? PDFs? What are those PDFs?

That’s bizarre. I didn’t know that those tiny icons were implemented as PDFs!

5. dtruss

The previous two tools, opensnoop and execsnoop, operate by tracing specific system calls. A system call (or “syscall”) is what an application performs to request the operating system kernel to perform privileged work, including process creation, file operations, and other I/O (eg, disk or network I/O). Syscalls are a great target for analysis with DTrace, since examining them often provides a pretty good picture of what an application is trying to do. They can also provide useful details including byte counts, file and process names, error codes, and latency.

The dtruss tool traces all types of system calls, which is very useful for general debugging, especially since Mac OS X doesn’t come with a standard syscall tracer (like Linux’s “strace” or Solaris’s “truss”). One advantage of dtruss over those other tools is that dtruss can trace multiple processes at the same time, matching on the process name “-n”. For example, tracing Firefox via its process name “firefox-bin”:

Brendan-2:~ brendan$ sudo dtruss -n firefox-bin         PID/THRD SYSCALL(args)                  = return 67254/0x720286: write_nocancel(0x7, "8", 0x1)                 = 1 0 67254/0x720286: lseek(0x24, 0x75500, 0x0)                 = 480512 0 67254/0x720286: read_nocancel(0x24, "", 0x200)                 = 512 0 67254/0x720286: write_nocancel(0x7, "8", 0x1)                 = 1 0 67254/0x720286: lseek(0x24, 0x73B00, 0x0)                 = 473856 0 67254/0x720286: read_nocancel(0x24, "", 0x200)                 = 512 0 67254/0x746a81: select_nocancel(0x3D, 0xB04909B8, 0xB0490938)                 = 1 0 67254/0x746a81: read_nocancel(0x6, "8", 0x400)                 = 1 0 67254/0x746a81: recvfrom_nocancel(0x36, 0xB0490C3F, 0x1)                 = -1 Err#35 67254/0x746a81: select_nocancel(0x3D, 0xB0490998, 0xB0490918)                 = 1 0 67254/0x746a81: sendto_nocancel(0x36, 0x20A51008, 0x292)                 = 658 0 67254/0x720286: write_nocancel(0x7, "8", 0x1)                 = 1 0 67254/0x720286: lseek(0x24, 0x74000, 0x0)                 = 475136 0 [...]

The output will be many pages, as applications commonly make frequent system calls.

dtruss can also launch and trace a program. Here the humble “ls -l” command was traced:

Brendan-2:~ brendan$ sudo dtruss ls -l hfsslower.d -rwxr-xr-x 1 brendan staff 1152 Jan 28 2011 hfsslower.d SYSCALL(args)                  = return ioctl(0x3, 0x80086804, 0x7FFF5FBFD710)                 = 0 0 close(0x3)                 = 0 0 stat64("/usr/lib/libstdc++.6.dylib", 0x7FFF5FBFCB20, 0x7FFF5FBFD160)                 = 0 0 mmap(0x0, 0xD000, 0x3, 0x1002, 0x1000000, 0x7FFF00000001)                 = 0x20000 0 [...100 lines truncated...] open_nocancel("/etc/sysinfo.conf", 0x0, 0x1B6)                 = -1 Err#2 lstat64_extended(0x100100768, 0x7FFF5FBFE3F0, 0x100101580)                 = 0 0 listxattr(0x100100768, 0x0, 0x0)                 = 0 0 fstat64(0x1, 0x7FFF5FBFD960, 0x7FFF5FBFDA2C)                 = 0 0 ioctl(0x1, 0x4004667A, 0x7FFF5FBFD9AC)                 = 0 0 access("/etc/localtime", 0x4, 0x0)                 = 0 0 open_nocancel("/etc/localtime", 0x0, 0x0)                 = 4 0 fstat64(0x4, 0x7FFF5FBFB8B0, 0x0)                 = 0 0 read_nocancel(0x4, "TZif", 0x2A64)                 = 1017 0 close_nocancel(0x4)                 = 0 0 write_nocancel(0x1, "-rwxr-xr-x 1 brendan staff 1152 Jan 28 2011 hfsslower.d\n", 0x3C)                 = 60 0 fchdir(0x3, 0x100100000, 0xFC080)                 = 0 0 fchdir(0x3, 0x0, 0x100800000)                 = 0 0 close_nocancel(0x3)                 = 0 0

100 lines were trimmed to keep that example short.

dtruss supports various options, including the printing of system call timing for use when analyzing performance.

Brendan-2:~ brendan$ dtruss -h USAGE: dtruss [-acdefholLs] [-t syscall] { -p PID | -n name | command } -p PID # examine this PID -n name # examine this process name -t syscall # examine this syscall only -a # print all details -c # print syscall counts -d # print relative times (us) -e # print elapsed times (us) -f # follow children -l # force printing pid/lwpid -o # print on cpu times -s # print stack backtraces -L # don't print pid/lwpid -b bufsize # dynamic variable buf size eg, dtruss df -h # run and examine "df -h" dtruss -p 1871 # examine PID 1871 dtruss -n tar # examine all processes called "tar" dtruss -f test.sh # run test.sh and follow children

I’ve used the “-e” option many times to figure out what’s slowing down an application. It shows the elapsed time for the system call in microseconds:

Brendan-2:~ brendan$ dtruss -e ls -l hfsslower.d  -rwxr-xr-x 1 brendan staff 1152 Jan 28 2011 hfsslower.d ELAPSD SYSCALL(args)                  = return 245 mmap(0x10000B000, 0x2000, 0x5, 0x12, 0x3, 0x7FFF00000001)                 = 0xB000 0 11 mmap(0x10000D000, 0x1000, 0x3, 0x12, 0x3, 0x7FFF00000001)                 = 0xD000 0 10 mmap(0x10000E000, 0x1F10, 0x1, 0x12, 0x3, 0x7FFF00000001)                 = 0xE000 0 31 open("/dev/dtracehelper", 0x2, 0x7FFF5FC45370)                 = 3 0 40 __sysctl(0x7FFF5FBFD5B0, 0x2, 0x7FFF5FBFD5DC)                 = 0 0

I then look down the “ELAPSD” column for the largest times.

A dtruss FAQ is: why is the output in the wrong order?

As with other DTrace “tracing” style tools (that print lines of output as they occur) the output can be shuffled slightly due to CPU buffering on multi-CPU systems. The “-d” option for relative times (since program start) is useful for cases where the output order is important. You can then post-sort by that column so that the dtruss output is in the correct order. (For example, by redirecting the output to a file, then using the sort(1) command on that column.)

Ryan Dahl (creator of node.js) had been using dtruss so frequently recently that he made some enhancements and posted a newer version on github here.

6. soconnect_mac.d

This script is from the DTrace book, and traces outbound TCP connections along with details:

Brendan-2:~ brendan$ sudo ./dtbook/Chap6/soconnect_mac.d PID PROCESS FAM ADDRESS PORT LAT(us) RESULT 45343 firefox-bin 2 443 735 In progress 65002 TweetDeck 2 80 94 In progress 65002 TweetDeck 2 80 89 In progress 54079 Google Chrome 2 80 76 In progress 45343 firefox-bin 2 80 752 In progress 45343 firefox-bin 2 80 88 In progress 45343 firefox-bin 2 80 85 In progress 45343 firefox-bin 2 80 39 In progress 45343 firefox-bin 2 80 22 In progress 45343 firefox-bin 2 80 20 In progress 27 ntpd 2 123 24 Success [...]

It’s a quick way to find out what applications are connecting to whom on the Internet. The script is here.

Columns are: PID = process ID, PROCESS = process name, FAM = protocol family (2 = IPv4), ADDRESS = IP address, PORT = TCP destination port, LAT(us) latency of the connect() system call in microseconds, RESULT = return of the connect().

Many connections these days are “non-blocking”, so the latency appears low and has the result “In progress”. The actual connection completes sometime later (the timing can be traced using some more DTrace, when desired).

There is a companion tool, soaccept_mac.d, which shows inbound TCP connections. If you ran both regularly, you may catch something performing networking that shouldn’t be (e.g., spyware, virus, …).

7. errinfo

This tool provides a summary of which system calls were failing, showing the process name, error code, and short description of the error:

Brendan-2:~ brendan$ sudo errinfo -c Tracing... Hit Ctrl-C to end. ^C EXEC SYSCALL ERR COUNT DESC TweetDeck lstat64 2 1 No such file or directory TweetDeck select 9 1 Bad file descriptor TweetDeck stat64 2 1 No such file or directory cupsd unlink 2 1 No such file or directory firefox-bin connect_nocancel 36 1 Operation now in progress launchd mkdir 17 1 File exists launchd open_nocancel 2 1 No such file or directory [...] Google Chrome 0 35 Google Chrome 0 35 mdworker getattrlist 2 24 No such file or directory mdworker mkdir 13 24 Permission denied TweetDeck access 2 33 No such file or directory TweetDeck read 35 35 Resource temporarily unavailable mDNSResponder recvmsg 35 38 Resource temporarily unavailable mdworker stat64 2 47 No such file or directory gpg-agent read_nocancel 35 60 Resource temporarily unavailable thnuclnt ioctl 25 60 Inappropriate ioctl for device ntpd __pthread_canceled 22 69 Invalid argument ntpd sigreturn -2 69 ntpd sigsuspend 4 69 Interrupted system call Google Chrome 0 0 activitymonitor proc_info 3 1495 No such process

This can be another quick way to track down failing or misconfigured applications. Note that (usually) most system call errors are not a problem: it can be normal for system calls to fail, and the calling application handles that condition correctly. errinfo is particularly useful when the application didn’t handle the failure correctly, and the system calls are silently failing.

8. bitesize.d

This is a simple DTrace script that characterizes the disk I/O workload, showing a distribution of the size of the I/O in bytes along with the application name:

Brendan-2:~ brendan$ sudo bitesize.d Tracing... Hit Ctrl-C to end. ^C PID CMD 31502 vim value ------------- Distribution ------------- count 2048 | 0 4096 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ 4 8192 | 0 16384 |@@@@@@@@ 1 32768 | 0 54079 Google Chrome value ------------- Distribution ------------- count 2048 | 0 4096 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ 18 8192 |@@@@@@@ 4 16384 | 0 86143 bsdtar value ------------- Distribution ------------- count 2048 | 0 4096 |@@@ 19 8192 |@@@@ 23 16384 |@@ 12 32768 |@@@@@@@@@@@@@@@@@@@@@@@ 151 65536 |@@@@ 25 131072 |@@ 10 262144 |@@ 10 524288 |@ 8 1048576 | 3 2097152 | 0

The “value” column shows the minimum size of the I/O in bytes, and the “count” column shows the number of I/O in that range: between the minimum size and the next size shown.

To follow an example: the “bsdtar” command shown above performed 151 disk I/Os that were between 32768 bytes and 65535 bytes in size (32K to 64K), which was where most of its I/O fell. The text plot in the middle provides a visualization for the distribution of the events, which shows a spike that corresponds to this 32K to 64K range. Google Chrome did a couple of dozen I/O between 4 and 16 Kbytes.

Disk I/O is an expensive operation, so it’s usually preferable for it to be larger in size to improve the throughput to the device.

9. iotop

This presents the same data as iosnoop, but in a summarized way similar to top(1). It’s handy when disk I/O is so frequent that iosnoop is too verbose, and you want a high level summary of which process is rattling the disks. Here I’ve used “-CP” to not clear the screen and provide a rolling output (-C), and to show disk busy percentages (-P):

Brendan-2:~ brendan$ sudo iotop -CP 1 Tracing... Please wait. 2011 Sep 27 19:33:11, load: 0.74, disk_r: 0 KB, disk_w: 57344 KB UID PID PPID CMD DEVICE MAJ MIN D %I/O 0 66661 66312 cp ?? 14 2 W 98 2011 Sep 27 19:33:12, load: 0.74, disk_r: 4 KB, disk_w: 55752 KB UID PID PPID CMD DEVICE MAJ MIN D %I/O 0 66661 66312 cp ?? 14 2 W 93 2011 Sep 27 19:33:13, load: 0.68, disk_r: 60 KB, disk_w: 54464 KB UID PID PPID CMD DEVICE MAJ MIN D %I/O 0 66661 66312 cp ?? 14 2 R 11 0 66661 66312 cp ?? 14 2 W 88 2011 Sep 27 19:33:14, load: 0.68, disk_r: 2560 KB, disk_w: 32928 KB UID PID PPID CMD DEVICE MAJ MIN D %I/O 503 65002 140 TweetDeck ?? 14 2 W 0 0 1 0 launchd ?? 14 2 W 1 0 66661 66312 cp ?? 14 2 R 11 0 66661 66312 cp ?? 14 2 W 71

The “??” is for where some details aren’t yet available to DTrace, for the same reason as was described for the iosnoop script (if you are a programmer, you can see what needs to be fixed in the /usr/lib/dtrace/io.d file).

If you get “dynvardrops”, sorry: back when I wrote this (6 years ago), I never saw a disk workload heavy enough to cause these. A remedy can be to increase the dynvarsize tunable by adding the following line:

#pragma D option dynvarsize=16m

near this one:

#pragma D option quiet

in /usr/bin/iotop. You may want to make a backup of iotop first (iotop.orig) or copy it somewhere else to run it.

10. maclife.d

This script is from the DTrace book, and traces the creation and deletion of files:

Brendan-2:~ brendan$ sudo ./dtbook/Chap5/maclife.d TIME(ms) UID PID PROCESS CALL DIR/FILE 3754594958 503 54079 Google Chrom VNOP_CREATE Chrome/.com.google.Chrome.K7I9jy 3754597703 503 54079 Google Chrom VNOP_CREATE -Tmp-/.com.google.Chrome.8MIKKL 3754597703 503 54079 Google Chrom VNOP_REMOVE -Tmp-/.com.google.Chrome.8MIKKL 3754597703 503 54079 Google Chrom VNOP_CREATE -Tmp-/.com.google.Chrome.TLlOje 3754597703 503 54079 Google Chrom VNOP_REMOVE -Tmp-/.com.google.Chrome.TLlOje 3754598365 503 54079 Google Chrom VNOP_CREATE -Tmp-/.com.google.Chrome.yIwTdE 3754598365 503 54079 Google Chrom VNOP_REMOVE -Tmp-/.com.google.Chrome.yIwTdE 3754603801 503 65002 TweetDeck VNOP_CREATE Cookies/Cookies.plist_tmp_65002_0.dat 3754605028 503 65002 TweetDeck VNOP_REMOVE Local Store/td_26_brendangregg.db-journal 3754605026 503 65002 TweetDeck VNOP_CREATE Local Store/td_26_brendangregg.db-journal 3754607674 503 65002 TweetDeck VNOP_CREATE Local Store/td_26_brendangregg.db-journal 3754607676 503 65002 TweetDeck VNOP_REMOVE Local Store/td_26_brendangregg.db-journal 3754609536 503 34852 Adium VNOP_CREATE Default/.dat8824.9fa 3754711583 503 17726 thnuclnt VNOP_CREATE thnuclnt-17717/thnuclnt.conf-fta17726 3754711587 503 17726 thnuclnt VNOP_REMOVE -Tmp-/4e8220118e918 3754711617 503 17726 thnuclnt VNOP_REMOVE -Tmp-/4e8220118ead8 3754711620 503 17726 thnuclnt VNOP_REMOVE -Tmp-/4e8220119621c 3754711623 503 17726 thnuclnt VNOP_REMOVE -Tmp-/4e82201196d7e 3754711624 503 17726 thnuclnt VNOP_REMOVE thnuclnt-17717/thnuclnt.conf-fta17726 3754612740 503 31502 vim VNOP_CREATE macosx/4913 3754612740 503 31502 vim VNOP_REMOVE macosx/4913 3754612741 503 31502 vim VNOP_CREATE macosx/top10dtrace.html 3754612741 503 31502 vim VNOP_REMOVE macosx/top10dtrace.html~ [...]

Interesting! While tracing I saved the file I was editing in vim, which is seen in the last four lines. This tells me that vim is creating and removing temporary files as part of the save process.

Also note that it looks like TweetDeck created the file twice before removing it (two VNOP_CREATEs followed by a VNOP_REMOVE). This isn’t the correct order, which can be seen by examining the TIME(ms) column. A side-effect of DTrace’s negligible performance impact design is that output can be slightly shuffled due to the way it collects data from per-CPU buffers. I often include a TIME column like that one, not just for the usefulness of knowing time, but also as a means to post sort the output.

Other Scripts

There are other tools ready-to-go in Mac OS X. If you run the “man -k dtrace” command in a Terminal (or “apropos dtrace”), you’ll see the following output:

Brendan-2:~ brendan$ man -k dtrace [...] bitesize.d(1m) - analyse disk I/O size by process. Uses DTrace cpuwalk.d(1m) - Measure which CPUs a process runs on. Uses DTrace creatbyproc.d(1m) - snoop creat()s by process name. Uses DTrace dappprof(1m) - profile user and lib function usage. Uses DTrace dapptrace(1m) - trace user and library function usage. Uses DTrace diskhits(1m) - disk access by file offset. Uses DTrace dispqlen.d(1m) - dispatcher queue length by CPU. Uses DTrace dtrace(1) - generic front-end to the DTrace facility dtruss(1m) - process syscall details. Uses DTrace errinfo(1m) - print errno for syscall fails. Uses DTrace execsnoop(1m) - snoop new process execution. Uses DTrace fddist(1m) - file descriptor usage distributions. Uses DTrace filebyproc.d(1m) - snoop opens by process name. Uses DTrace hotspot.d(1m) - print disk event by location. Uses DTrace httpdstat.d(1m) - realtime httpd statistics. Uses DTrace iofile.d(1m) - I/O wait time by file and process. Uses DTrace iofileb.d(1m) - I/O bytes by file and process. Uses DTrace iopattern(1m) - print disk I/O pattern. Uses DTrace iopending(1m) - plot number of pending disk events. Uses DTrace iosnoop(1m) - snoop I/O events as they occur. Uses DTrace iotop(1m) - display top disk I/O events by process. Uses DTrace kill.d(1m) - snoop process signals as they occur. Uses DTrace lastwords(1m) - print syscalls before exit. Uses DTrace loads.d(1m) - print load averages. Uses DTrace newproc.d(1m) - snoop new processes. Uses DTrace opensnoop(1m) - snoop file opens as they occur. Uses DTrace pathopens.d(1m) - full pathnames opened ok count. Uses DTrace pidpersec.d(1m) - print new PIDs per sec. Uses DTrace plockstat(1) - front-end to DTrace to print statistics about POSIX mutexes and read/write locks priclass.d(1m) - priority distribution by scheduling class. Uses DTrace pridist.d(1m) - process priority distribution. Uses DTrace procsystime(1m) - analyse system call times. Uses DTrace runocc.d(1m) - run queue occupancy by CPU. Uses DTrace rwbypid.d(1m) - read/write calls by PID. Uses DTrace rwbytype.d(1m) - read/write bytes by vnode type. Uses DTrace rwsnoop(1m) - snoop read/write events. Uses DTrace sampleproc(1m) - sample processes on the CPUs. Uses DTrace seeksize.d(1m) - print disk event seek report. Uses DTrace setuids.d(1m) - snoop setuid calls as they occur. Uses DTrace sigdist.d(1m) - signal distribution by process. Uses DTrace syscallbypid.d(1m) - syscalls by process ID. Uses DTrace syscallbyproc.d(1m) - syscalls by process name. Uses DTrace syscallbysysc.d(1m) - syscalls by syscall. Uses DTrace topsyscall(1m) - top syscalls by syscall name. Uses DTrace topsysproc(1m) - top syscalls by process name. Uses DTrace weblatency.d(1m) - website latency statistics. Uses DTrace

Plus more in the DTrace book.

And More…

DTrace isn’t just about running scripts; you can write your own custom scripts, run one-liners, and use higher-level tools that use DTrace behind the scenes, like Apple’s Instruments and Joyent’s Cloud Analytics. And it’s for more than just your MacBook: if you are using servers that have DTrace available, you can use it to diagnose their performance and issues too, including tracing kernel and application code. I regularly use it to see how a MySQL database interacts with a kernel file system, for example.

For more reading about DTrace, you can see my posts tagged dtrace, other blogs on dtrace.org, the original 410 page DTrace guide, 1100 page DTrace book, and my original DTrace page (which includes the DTraceToolkit). It’s a little old now (and I think prustat needs updating), but there’s also Matty’s Top Ten DTrace Scripts, which includes some of those above; and Greg Miller’s Exploring Leopard with DTrace, which includes Objective-C tracing.

Reporter had a stroke on live TV – Facebook scam

February 21, 2011

via Naked Security – Sophos by Naked Security – Sophos on 2/15/11

Serene BransonSerene Branson, a CBS Los Angeles newsreader, became an unwilling YouTube star overnight after speculation spread that she had suffered a stroke while presenting from the Grammy Awards.

The footage of Serene Branson stumbling over her words quickly became viral, as users on Facebook and Twitter passed the link on to each other.

Although reports indicate that Miss Branson was not hospitalised and is “feeling fine”, interest in the video snippet continues to bubble away – and now scammers are exploiting the news story.

If you see a message like the following posted from one of your Facebook friend’s accounts don’t click on the link.

Omg this reporter had a stroke on live tv check it out [LINK]

If you do make the mistake of clicking on the link – perhaps out of morbid curiousity to watch Serene Brandon struggling in her piece to camera – you will be presented with a screen like the following claiming that what you are about to do is use a “verified app”.

Of course, the Facebook app is in reality a third-party rogue application, designed to make money for the scammers who instigated the scheme.

The scammers’ plan is to exploit interest in the Serene Brandon video, by tricking users into approving an application that will be able to access profiles and post messages onto the walls of Facebook accounts.

Clicking “Approve” is a bad idea, but many people fall for social engineering tricks like this all too easily.

What you probably don’t realise is that behind-the-scenes your own Facebook page has published the link to your online friends and family, encouraging them to also click on the link.

In this way the link spreads virally, increasing the opportunities for the scammers to make money.

And how do they make money? By presenting you with the all-too-familiar survey scam before you can watch the video footage. If you complete the survey, the scammers earn a small amount of commission – and you’ve helped them generate even more by sharing the link virally via your Facebook page.

If you made the mistake of approving the rogue application you should remove it immediately, and remove the offending messages from your Facebook profile before your friends are also roped into the scam.

If you have been hit by scams like this on Facebook, and are struggling to clean-up your profile, here’s a YouTube video I made which describes what steps you need to take:

(Enjoy this video? You can check out more on the SophosLabs YouTube channel and subscribe if you like)

Make sure that you keep informed about the latest scams spreading fast across Facebook and other internet attacks. Join the Sophos page on Facebook, where over 60,000 people regularly share information on threats and discuss the latest security news.

You could also do a lot worse than check out our best practices for better privacy and security on Facebook guide.

Dr. Foster’s 30 Laws of Decision Making

October 24, 2010

#1: Focus on the most important thing (Are you giving full weight to the most important issue?)
#2: Don't decide until you're ready (Why not give yourself more time?)
#3: Look for all the good things that can happen (What's the best outcome you can expect?)
#4: Choose it or lose it (What big decision are you not making now?)
#5: Base your decision on self-acceptance (Does your decision fit the real you?)
#6: Look ahead (How will your decision play out over time?)
#7: Turn big decisions into a series of little decisions (What small step can you take toward a decision?)
#8: You always havebetter options (What are better options beyond those you've checked out?)
#9: Get what you need to feel safe (What are you doing to take care of your safety needs?)
#10: Do what you really want (Are you sure you've let your heart's desire hold sway?)
#11: If it ain't simple, it ain't gonna work (Are you making things more complicated than they need to be?)
#12: Have a hopeful heart and a cautious head (Have you balanced hoping for the best with protecting yourself from the worst?)
#13: Tune in to what you're saying to yourself (Have you remembered to really listen to yuorself?)
#14: Never let a lower priority outweigh a higher priority (Are your priorities all out of whack?)
#15: No matter what, don't get trapped (Are you protecting yourself from getting stuck?) [For more examples of important traps, see 

#FFP05: Harry Browne's Freedom Principles.]
#16: Know your Achilles' heel (What are the top ten bad habits in making decisions?)

(1) Biting off more than you can chew
(2) Detail mania
(3) Fear
(4) Losing touch with you
(5) The green-eyed monster [envy]
(6) Keeping on keeping on [persistence with what can't work]
(7) Acting without thinking
(8) Dithering
(9) Taking the path of least resistance
(10) Not believing that things can be better than they are.
[Personally, I would add the following to the above list:
(11) Poor diet/unhealthy lifestyle
(12) Allowing your Self 1 to interfere with your decisions
(13) Living your life as a "passive passenger" — obedience or "following the crowd" — let others decide for you — automatic acceptance or rejection of information
(14) Irresponsibility — blaming others or factors outside yourself
(15) Dishonesty
(16) Coercion — initiating force, threat of force, or fraud in order to impose your will upon others — see Why You Must Recognize and Understand Coercion
(17) "Thinking" with your sexual organs rather than your brain — see #TL05AA: The Breeding Motivation: What You Can Do About It
(18) Unrealistic expectations
(19) Not considering the hidden bad side of people — see #TL15A: The Good and the Bad
(20) Unwillingness to make drastic changes
(21) Ignoring or neglecting absolute essentials
(22) Denial (of personal disadvantages)
(23) Procrastination and failure to follow through (lack of persistence)
(24) Failure to be a life-long learner and keep abreast of new developments
(25) Failure to distinguish between what you can and can't control
(26) Overconfidence.]

#17: Always take your own best advice (Would you tell your best friend to do this?)
#18: Appreciate the newness of each situation (Are you in danger of applying old learning to s new decision?)
#19: Make yourself proud (Which decision will give you self-respect?)
#20: Pay attention to the big, fat, obvious issues (What are you overlooking?)
#21: Never forget why you made your decision (Are you judging yourself with different criteria from what you first used?)
#22: Know what's real (What realities will affect your decision?)
#23: Get what you need to make your decision a success (Are you equipped to carry out this decision?)
#24: Find and follow an expert (Are you making sure you're relying on smart advisers?)
#25: Keep an open mind (Are you caught in attitudes that keep you from seeing what's best for you?)
#26: Take care of the basics (Will your decision address your basic needs?)
#27: Some things you "know" are wrong (Have you checked your facts?)
#28: You don't have to run from risk (Do you know what the risks are and how to protect yourself?)
#29: Following through makes decisions great (Will you do a great job carrying out your decisions?)
#30: Make decisions to make things wonderful (Will you get something that's wonderful?)