Because Heath said he checks this every morning...

Tue May 15 17:29:27 CDT 2009

Good morning Heath Seals...

Amdahl's Law applies to people too.

Sun Jul 23 04:56:13 CDT 2006

I recently switched employers (four months ago), and am already switching again. This is in no small part due to how my current employer manages problems. Every single issue is a crisis, because every single issue causes mass confusion. Well, mass confusion is probably a little harsh, but let me give you an example.:

I get a call that we may have a problem with a server because someone is not seeing traffic across it. They had called the helpdesk, who then yelled at the shift manager, who then yelled at me. Yes, yelled. I currently work in a command center that would make the Cobra Commander or Dr. Strangelove weep. I'm not familiar with this particular system, and have no idea where to start looking (you'll see why if you keep reading) so I sametime the admins in New Jersey, and then the shift manager has me call them. Three phone calls to NJ and two sametime chats, and by trial and error (guessing which data store the information is in, looking, and repeating) I find the application owner for the application in question. I call him, and he tells me that it's not a problem. I relay this to the manager, who relays it to the helpdesk who tells the person who thinks there should be traffic. Now the application owner and the user disagree, so a conference call is started with me, and all people involved. The final solution: The is not and never was a problem. There's an hour of my life that was well spent.

The problem is communication. Let's have some fun with math: The direct communication pathways between a number of individuals (nodes) can be represented by:

 >   (i - 1)
a quick table of this for a small group is:

and so on. As you can see as you add people, the number of possible communication pathways a single node need deal with grows faster than the nodes themselves. And this model assumes that every node (person) talks directly to one another, and that the do not route messages through one-another, if messages can be routed through people "Tell Jim to tell Bob," then it grows exponentially. So one solution to this is to implement centralized nodes that can "broadcast" (for lack of a better term) to more than one person at a time. So instead of a phone call to each node, all nodes call into a conference call. Or instead of sending an instant message to each node, you enter Sametime chat rooms. Now while this seems like a good solution, Let's just list the possible "broadcast" mediums and their narrowcast alternatives (if any):

1) Conference call/Phone call
2) Sametime Meeting/Sametime side chat
3) Lotus Notes email - Multiple Recipients (tech pages) / Single Recipient
4) Lotus Databases (To many to count)
5) Lotus Universal Transmittal System (Yes ticketing systems convey information)
6) ServiceCenter 5 & 6
7) Shift Turnover (in Lotus Notes)
8) Web pages (one per group http://ssg http://ots http://ccg etc.)
9) Run Books/Word documents/Excel spreadsheets such as contingency test plans and server location
10) Meetings/Talking person to person

There are more, but If these aren't enough to make my point, then more won't help. Furthermore the majority of these communication channels are "dirty" which means that a signifigant percentage of bad information gets conveyed along with the good, and there is no mechanism to correct them after the fact. Websites and runbooks over 6 years old with no date on them anywhere, so the reader can't even tell if the information is stale or not. So when bad information is conveyed, then people start picking up the phone and calling to get clarification.
The result in a large organizaton is that the majority of the time is spent trying to figure out what needs to be done, or even what is going on. Even the most simple problems quickly escalate to incidents, as more and more people are contacted through one of these narrowcast/broadcast mediums until the one person who knows what is going on and how to resolve it is contacted, and fixes it. Does this sound familiar?
Well that's great, but what can be done to fix it? They've been doing things this way so long they're used to it.

1) Use a Wiki.
All documentation that needs to stay around more than a week (Web pages, Run Books, server layouts, etc) need to be in a wiki. Why? Because this information will need to be updated (really, it will) and when the reader finds an error, they need to be able to annotate the documentation as needed. Having the reader tell their manager to tell the author's manager to tell the author is insane. With a wiki, they can fix the error "on the fly", and the author can be notified (by email) that someone has edited his or her page, and review it without getting 2 other people involved. Wiki's also keep track of the complete history of a document and tell you how old the document is, and who made what changes on every revision. As a result, your documentation gets better over time, and you decrease the number of people awakened in the middle of the night to clarify bad documentation. It is also a one-stop searchable repository for your documentation.

2) Stop using email groups and set up a listserv.
Email is a great colaborative tool, but it fragments your conversations into dozens or more of small inboxes all over your network. With Lotus notes, you can't even sort by subject or do a global search the contents of the disparate folders, I can't search your email for something I might have removed. A listserv will pool all of the emails on a given problem, archive them, and they can be indexed for searching by everyone. This will immediately cut down on "clarification calls" and anything covered in the list that needs to be more permanent can be added to the wiki. Entire list threads can be added to the wiki. You will never have to solve the same problem twice.

3) Lose Sametime and set up an IRC server.
Sametime has to be one of the worst chat clients ever produced. Anytime you have to message someone to ask them to invite you into a chat, you should realize something is wrong. Every time you accidentally hit <esc> and it closes your window without prompting, you should realize something is wrong. But those annoyances aside, there is a reason IRC has been around as long as it has: It's a great utility. You can create chat's on the fly, so you can even have one chat per problem ticket, and log the chat (automatically) and save the contents of it to the wiki. Don't even get me started on the usage of bots . They are very useful. And users can use the irc client of thier choice to connect. Believe me when I tell you that user customization increases efficency, that's a whole rant in itself. But being able to log the chats and go back and see "how did we solve that problem the last time?" and then, yes, you guessed it, add it to the wiki.

4) Get a ticketing system with an open API
If you haven't figured it out yet, the theme of this rant is to unify your documentation. Having a ticketing system that makes you load a fat client to get information out of is rediculous, you should be able to link to open and closed tickets in a browser, then you can reference them in the wiki, and your tickets aren't just thrown out after they are implemented/resolved. They instead become a living part of your documentation system. An open standards web-based ticketing system is your best bet. Something like bugzilla or trac (which has a wiki built in) would work nicely.

5) Stop using the phone.
As for the phone calls, it would be nice if those would go away altogether. They can't be logged, searched later or used to update your documentation after the fact unless someone is transcribing them. Phone calls are a horrible problem management tool. But if effort is spent unifying the wiki/listserv/irc into a single colaborative tool, you can minimize your need for phone calls. Maybe a system for logging phone calls could be developed using a tool like Asterisk, which would also give you voip capabilities. The entire conversation could be replayed and even timestreched and frequency compensated like in MythTV, the technology exists, but isn't ready for prime-time like the wiki/listserv/irc are.

If you think of your organization as a large distributed parallel computer system, with each person as a processor, then the wiki becomes your "shared memory" and every "processor" can read and write to it. It will decrease the latency in propagating *accurate* information out to different nodes, thereby increasing accuracy and efficiency. Or is that too nerdy of a metaphor?

Suggested Reading:
"The Mythical Man Month" -- Frederick P. Brooks

Controlling Your Network.

Tue Apr 25 09:35:27 CDT 2006

    Your network is out of control, or more accurately, at the edge of out of control. That may seem a little harsh, but allow me to explain. First let's address the term "control", it has been defined as, "the power to direct or determine." In a system it would mean the ability to know (determine) what state your system is in, and the ability (power) to make (direct) that system to another state. In this context, "system" does not mean a single computer system (although it could) instead we want to think of all of the systems on the network as a single system.

    Let's look at a simple (and automated) control system, the thermostat in your house. You know what temperature you want it to be, lets say 73 degrees. Now you set the thermostat by adjusting a spring, upon which sits a mercury switch. The spring is actually a bimetallic material that expands and contracts when the heat energy in the room decreases or increases, respectively. This will cause the mercury to flow to one end of the glass tube or the other. At one end of the glass tube are the contacts for the heating system and at the other end there are contacts for the air conditioner. If the mercury has pooled up in the middle then no contacts will be closed. So if the heat energy in the room increases beyond the set temperature the air conditioner is enabled and if it becomes too cool the heater is enabled.

    Heaters and air conditioners are really "dumb" they are either on or off. If you didn't have a thermostat, you could maintain the temperature in your house by putting a thermometer on the wall, and have someone sit in front of it. You will need to give this person some instructions as to how to maintain the room temperature:

  • If it gets to 72 degrees then turn on the heater.
  • If the temperature gets over 72 degrees, turn off the heater.
  • If it gets to 74 degrees, turn on the air conditioner.
  • If the temperature gets under 74, turn off the air conditioner.

  • If the heating and cooling people were in different groups, then the heating instructions could be in one manual (or web site) and the cooling instructions in/on another.

        Now you can't keep one person there all day so you will need 5 people, for 1st, 2nd, 3rd shifts and weekend days and weekend nights. Actually, you will need at least twice this many people, (10) so the shifts can be covered in the event of vacation or sickness. Furthermore, if these devices do not operate with a simple on/off switch, you might have some additional people, who actually build, repair and operate the heating and cooling systems on call, and give these 10 people the phone numbers of the 4 repairmen (one for heating and one for cooling and their redundant counterparts) and have one of these thermostat watchers call one (or two) of the repair men to enable or disable a heating or cooling device. You will also need 2-3 managers to manage these 14 people bringing the total to 17, plus yourself. So it would take about 18 people to maintain the temperature in a single room.

        Luckily these same people could do this on several rooms, or an entire hotel, so it scales quite well in this case. But that doesn't change the fact that these 18 people could be replaced by bimetallic strips with mercury in glass tubes affixed to them. Mercury doesn't get tired and make mistakes like a human is notorious for doing. This system may work well for maintaining a single temperature for multiple rooms (provided there is a way to get the individual thermostat numbers to the thermostat watchers) but it becomes increasingly complex if there are rooms that need to be at different temperatures, if temperatures need to vary during the day or week, and in situations the heaters or air conditioners need to be taken offline for maintenance. The thermostat watchers have to be intuitively aware of all of these possibilities, and when to expect the systems to not respond or when to expect the temperatures in the rooms to drift out of tolerance.

        Now if you add even more complexity humidity, light levels, smoke detectors, phone lines and/or data ports in the room, then it quickly becomes untenable. You end up with the simplest incorrect configuration leading to several people being woken up at all hours, trying to determine if the problem is being caused by the room being to humid or the room being too hot. Now I believe this analogy has been thoroughly driven into the ground.

        The "no thermostats" method is how your company is currently controlling the system (the network.) My analogy seems quite absurd, because no sane person would do that for a thermostat, so let's expand the scope of my argument to services running on systems. The simplest service is telnet, but that one doesn't impact the business directly, so let's use the second simplest service: HTTP. Let's say a weblogic instance is down. Currently, if the web server stops listening on a port or stops serving up pages, Your company knows about it in one of two ways: the first way is if one of our monitoring systems (your thermometers) is configured to notice it. Be it TBSM, Tivoli Enterprise Console, HP OpenView, SiteScope, or Topaz or something else for which you paid a lot of money. You see a status page of some type change to red, or get an automated email through Exchange, or Lotus Notes. So in effect, your looking at least one thermometer, in larger organazations two or more. In most cases the service needs to be stopped and re-started. So you either take the action ourselves or some "thermometer watcher" calls the person or group responsible for the application and then ask them to stop and restart it. The second way you know about a problem is when a customer complains. This is the "bad" way. It comes in over the phone or as an email or Sametime message and it's too late. The business has been impacted. So how do you make a bimetallic strip and mercury bubble that can restart a service, or take other actions to keep the business from ever being impacted?

        In engineering, there is a concept of a "control system" It has a minimum of three parts: the sensor, the corrective actuator (or just "actuator" for short) and the system itself. The sensor detects an aspect of the system and determines if it is out of tolerance. If so, it sends a signal to the actuator that is proportional to how far out of tolerance the systems aspect is. The actuator then takes a proportional action, and the system's aspect in question is moved. The sensor, which is always operating, is constantly feeding the actuator, which is constantly changing the system's state, only stopping when the sensor detects that the difference between the actual system state and the defined system state is zero. (Actually, it doesn't stop. The feedback loop is still intact, but the proportional actuator action to a zero difference is zero, so the actuator does nothing.)

        Computers can do this quite easily, as a matter of fact, analog control systems are a thing of the past. Most control systems are now digital, and don't even require the computing power required to play a game of "Pong." So being that we have more than one computer at our disposal, there is no reason that these systems should not be completely self-correcting. We only need three things:

    1.To know what the state of the system should be (the system)
    2.To know how to see deviations from the defined state of the system (the sensor)
    3.To know how to restore the system to the defined state if it is incorrect ( the actuator)

    In the thermostat example: the room is the system, the bimetallic strip is the sensor, and the heater/air conditioner is the actuator.

        Based on what I have seen in my short time working for large corporations, the current monitoring strategy focuses only on the sensor. That is what all of the monitoring tools money can buy are. If you have not clearly defined the state of the system, then any number of sensor readings are meaningless. Now, for employees who have been here for a few years or more, who know (or have a feeling for) what the state of the system is, the sensor readings can be useful. For new personnel, they are effectively gibberish. If the system state is clearly defined, then anyone with any systems knowledge can log on to a system and see if the system is not in that state. Furthermore, if corrective actions for sub-states are defined, then the same new personnel could correct system problems without the benefit of instruction. Without this list of corrective sub-state actions, the only way to know how to correct a problem is to have experienced the problem and been instructed on how to correct it. The implications of this are disastrous, as the only way you can have completely trained new staff members is to have every single thing that can go wrong, to do so, at least once while they are on shift. To completely train your 10 staff members, everything in your enterprise has to break 10 times, and twice on every shift. So you are left with having untrained personnel (if you're lucky, and nothing ever breaks) or you will have a very competent staff, but a horrible reputation for service uptime. It's simply not workable.

    Know the state of the system.     For every sub-system on the network (a sub-system would be a host machine) you need to know what should be running, what shouldn't be running, what services are listening and to what file systems the processes are bound. This would be the minimum system state that needs to be defined. But before this is useful, you need to know how many hosts you have, their operating systems, where they are located and so on? The best method, in my experience (and opinion), to store this data is in an LDAP directory.

        Why LDAP? LDAP (the Lightweight Directory Access Protocol) is a protocol for reading and writing x500 based directory services. It can be or is by default used by most operating systems. Novell's eDirectory is one such implementation as is Microsoft's Active Directory Service. Finding support for accessing an LDAP data store is trivial, no matter what operating system one uses. LDAP is the de facto replacement for NIS and NIS+ on enterprise networks everywhere. It is also completely extensible, and new schemas may be created just as easily as new tables in a database. Directories also perform significantly better than databases in read heavy environments (where reads are at least an order of magnitude more frequent than writes.)

        How does this help us solve problems? First, let me digress into what makes up an application. In its purest form, an application is a collection of files, some of which get loaded into memory and executed as binary code and some are read into memory as data. The application is just a collection of files on a system or set of systems. In the latter case, these files are spread across multiple machines, and often operating systems, that use some type of networking protocol to hand data off to one-another. Most of these files are confined to a few file systems on each host involved. Understanding host file system, and networking protocol relationships is key to rapidly solving problems.

        What I propose is setting up a map, using a LDAP database, to map application file system, hosts, and network protocols to the three letter acronym group that is responsible for maintaining and resolving issues on the correlating application or set of applications. Most LDAP implementations provides a host schema that include, host, IP address, description and manager fields, and these can be extended to hold any number of attributes and lists, such as file systems, and owners of said file systems. We can extend these to hold any information that may be useful in solving a problem, such as ports in LISTEN states, open files and more. We can effectively define the state of properly operating machine inside of the directory along with the group to contact when this state is out of tolerance. An extremely simple host record might look like:

        dn: cn=file01,ou=Hosts,dc=yourdomain,dc=com
        objectClass: top
        objectClass: ipHost
        objectClass: device
        objectClass: customhost
        descCpu: 1
        descRam: 524210176
        operatingSystem: Solaris 8
        uniqueId: a8c0e73d
        hardwarePlatform: sun4u
        filesystem: /opt/iplanet WEB
        filesystem: /opt/sybase/SYBSOMETHING DBA
        Service: TCP:8080 WEB

    Now if a problem comes in involving file01 such as a failed maestro/cron/ at job or connection refused, we can use an LDAP search utility to know that the problem is owned by the DBA group or WEB group depending on which host or file system is in question. Furthermore there are several utilities that can access and update the information in an LDAP directory, so tools can be written (quickly) to index and search the information.

    One can also create custom record entries for sensors and actuators, and even combine them.

        dn: cn=weblogic, ou=Sensor,dc=yourdomain,dc=com
        objectClass: sensor
        objectClass: actuator
        host: file01
        descSensor:  Weblogic Threads
        cmdSensor: /bin/ps -ef | /bin/grep "wls" | wc -l
        valueSensor: 3
        descActuator: Bounce Weblogic
        cmdActuator: rsh ${HOST} /etc/init.d/weblogic restart
        tryActuator: 5
        delayTry: 60
        ownerGroup: WEB

    This way, there is a clear definition of what should be running where, and how to correct it, how many times to try, how long to wait before attempting another restart, and what group/user to contact if it fails all attempts. This is only an example. How much information we tie to a sensor and/or actuator is completely arbitrary. And this is really going above and beyond the scope of identity management. This would be the desired endgame to completely streamline problem resolution, if not completely automate it.

        To prototype a system like this, we would need a single virtual machine in the ESX server farm running Red Hat Enterprise Linux 3 ES or RHEL 4 ES, with Open LDAP installed, and time to add the host and file system records and develop a search and index web page running on Apache. Both Open LDAP and Apache ship with Red Hat Linux so no additional cost other than the 8-12 GB install of RHEL and the cost of a RHEL license. Overall this is an insignificant cost for what we would get in return, the rapid resolution of system issues handled by new support personnel. Alternatively, we could implement this on Solaris using the Sun One Directory Server and iPlanet or apache. We could have a prototype up and running in about a month, with at least partial data inserted.

        If you can define the system state, you will have taken a significant step towards rapid or automatic problem resolution. If you cannot define the system state, or remove the mentality that it is "unknowable" then having all of the sensors in the world will do nothing for you. And until the actuators are closely tied to the sensors, the only reliable actuator you will have is to "get someone on the phone." and the "Charlie Foxtrot" that follows.


    Mon Jul 11 11:49:43 CDT 2005

    Some of you might not have heard, I wrecked the pod.
    So Here are the Frequently Asked Questions a.k.a. the PodWreck FAQ:

    Are you OK?
        Yes, I was a little shaken up, but completely uninjured.

    How is your cat, Hecuba?
        Furious. She wasn't happy about the trip in the first place. This didn't help.

    When did this happen?
        Friday July 8, 2005 at 11:05 CDT

    What Happened?
        I was supposed to go to a wedding in Germany. I was going to be gone 10 days, so I would need someone to watch Hecuba, and the pod, so I decided to dirve the pod to Danville and park it at my parent's place and have them drive me to Atlanta. This would allow me to fly with my friends to Germany (we had got on the same flights.)
        As I was driving down I-65 S. in Tennessee, and At mile marker 27 just before the Lynnsville/Cornersville exit, there is place where the road cuts between two mountains. As a result, the wind whips up through it pretty fast. The wind hit the side of the pod, and it started to sway left. I tried to compensate, but then it started to sway hard right I wanted to get in front of the trailer, so I needed to move to the right to get the trailer behind me. But there was a tractor-trailer in that lane. (My front bumper was lined up with its rear bumper.) To avoid hitting it, I had to slow down. Now, I know you *never* slow down in a skid or a sway, but it was the only way to avoid hitting the other vehicle, so I found myself in a catch-22 situation and opted for "maybe jacknifing" instead of "definitely hitting a semi."
       Needless to say, "maybe jacknifing" became a full slide and the trailer left the road. Later, we saw three trenches parallel to the road where the triple-axle was in full-slide in the dirt. It was in this slide when we assume all three tires broke-seal, the rims dug into the dirt and the pod flipped. Luckily the tongue ripped off of the ball, or the truck would have rolled with it. The safety-chains held and pulled the truck behind the spinning pod, and the whole mass came to a stop 180 degrees later.

    Does this happen a lot?
    According to one of the Tennessee Highway Patrolmen that stopped, this has happened 5 times in the past year it this same place. He called it a "vaccuum for trailers." While we were sorting through the wreckage, a man who had wrecked his trailer in the exact same spot *last week* drove by and told us as much. I don't remember seeing a "trailer wind warning" sign, but if there isn't one, their should be, and it should have those flashy-lights on it. My brother is a truck driver (long-haul) and when we told him approximately where I wrecked, he said, he knew the spot, because he has had trouble with wind on his 18 wheeler there before.

    What about your mobile computer lab?
        Yeah, you remember when I had that? That was cool. That was also three days ago.

    What about everything you own?
        Apparently the guys I work with are the kind of guys who will go to the wall for a friend. Or at least drive an hour to stand in the 95 degree Tennessee sun and drag junk onto one pile, and clothes and canned food that didn't explode onto another. My undying thanks to the guys who came out and got heat exhaustion/heat cramps with me, so we could sift through the wreckage.

    How many times should I say "I'm realy sorry" or "That really sucks, dude." before you'll understand how bad I feel for you?
        Once. Repeating it just makes me think about how bad this really is, and I need to stay focused to stay sane.

    Now What? Aren't you homeless? What about your job?
        Several of the guys I work with offered to put me up until I get back on my feet. I will be staying with one of them so I can go to work when my "German Vacation" is over. I'm not planning more than a few steps ahead right now. Step 1 is doing laundry, as everything I own was covered in cat litter. Step 2 is seeing which hard-drives survived, and getting the data off of them. After that, it's back to work, and look for a place to live, I'm not sure about a new pod or not. Pod-Life was pretty cool, that is, right up until the podsplosion.

    Was the pod insured?
        Yeah, but when the Allstate assessor went to the lot where it was towed, they said it wasn't there. I can only guess they were looking for a airstream trailer and no a scrap pile of aluminum. So given that premise they would miss it.

    Spring in the pod.

    Mon Apr 4 22:37:17 CDT 2005

    It didn't really occur to me when I got the pod that there might be hidden advantages to effectively camping all year. Well, there are disadvantages as well. Winter pretty much sucks, especially when you lose your injector to you primary propane heater, followed by your secondary. Luckily there is a non-vented tertiary heater that can be used for short periods of time without fear of asphyxiation. It was enough to heat up the pod so the ceramic heaters (electrc) could maintain the temperature. But now spring is here, which means I can open the pod windows (hatches) and enjoy the weather. Of course the majority of my neighbors are an vacation or retired. So It's pretty much a festive atmpsphere year-round. But now that spring is here, It's even moreso. So I work a late day today, and I get home and sit down at the console and and prepare to make sure the snapshots have run properly. Before I know it's 9 PM and I realize I haven't eaten. I get up to mke some soup (I don't really cook much. It's a waste of good system time) and one of my neighbors brings me a tenderloin, and peas. They were having a cookout. This wasn't the first time this has happened either. It was pretty awesome. I mean how often do perfect strangers come by and bring you a steak? I mean c'mon, free steak.

    What's new?

    Sat Jan 22 11:47:09 CST 2005

    Well it has been over six months since I last updated the site, so I figured it's time. By now I'm sure no one is even looking at the site anymore, having long quit caring. That makes now the perfect time to update it. I have been contracting as a Unix Systems Admin for CAT Financial in Nashville, TN since July, and they have been keeping me pretty busy. Le Charget is having bodywork done and so that means spending more money on it, as well as it being techincally "inoperable" as with no paint anywhere, no vinyl on the roof, and the bumpers and trim missing it isn't much to drive. Pod Life has become pretty routine. I am now referred to as "The Pod Guy" around the office. So I registered the domain and one of my co-workers made me a logo. Apparently I look a great deal like Peter Griffin from Fox's The Family Guy Eh? what can you do? I took some more pod photos (at the bottom) mainly of the new MythTV/Powefile jukebox system.

    When it rains, it pours. (Cell phone number change)

    Fri Jun 18 13:10:39 CDT 2004

    So I'm getting out of my car last night and just as I'm slamming my car door, my phone slips. It manages to fall between the car door and the door frame, just in time for the door to close. I hear a *crunch*. Well I get bad service where the pod is located, so I figure why not try to switch from TDMA to GSM, It's not like the cell phone companies are expanding the TDMA networks anyway. So I can either buy a new phone, or process it as a relocation and get a discount. Being that I am out of work, discounts are good. So $150 later I have a spiffy new flip-phone and a shiny new number. (256) 566-5866. I will try and contact as many people as I can, but I have to start with *EVERY DAMN JOB SITE ON THE WEB* But I'll rant about them a little bit later. Keep in mind that I probably don't have your number If I don't call you once a week, as all of that information is in my dead phone.

    There are Muscle Cars in hell.

    Wed Jun 16 19:49:58 CDT 2004

    Well, in my own little personal hell, that is. I called the mechanic to see what the status is on Le Charget . I put in in the shop on Feb 25th, so I am happy to hear that it's ready to go. I'm not so happy to hear that one of the mechanics wrecked it into something. So I get a ride (3 hours) down to Montgomery. Now three hours doesn't seem like a long time, but when all you can think about is how bad your car has been wrecked, it's an eternity. So I get there and the damage is pretty damn bad and I am not a happy camper. I need to know how much the damage is so I drive to get a damage appraisal. Here's an interesting fact: Most body shops will not even talk to you if you have a vintage car. Who knew? The third body shop I tried directed me to a shop that specializes in classic muscle-cars. I run out of gas on my way to this shop. Now it's important to note that the car had *at least* a half tank of gas when I brought it in to the shop. I was in the mechanized infantry for 4 years and am permanently mentally scarred to the point of being paranoid when it comes to running out of "class 3". As a result, I never have less than a half-tank of gas. I also feel it is important to note that there is a $21 GAS itemization on my bill. This means they went through *at least* 25 gallons of gasoline while my car was in their possession. That could be another rant. I "top the car off", and the paranoia subsides. I make my way to the only place in town where I can get a quote. $1,759.50 + $325 for the drivers-side floor carpet the mechanics tore. The chrome on the bumper has a groove in it so it needs re-chromed or replaced. The pictures tell the rest of the story. Now on the way back to the mechanic's shop, I notice my amp meter is pegged, that means it's pushing more than 40 AMPs, not volts, AMPs. Now I can run everything in my pod on 30 AMPs if that is any indication of the severity of the problem. And the driver's side window will not go down. This is important too, because I had brought the car to the shop, four months ago, to have the near-melted wiring harness replaced, and to get the windows working. So It took them four months and a little over $2000 to not fix my car, and wreck it, *twice*. But I'm not bitter. So I have to leave the car there, because it's if it ran 40+ amps for the entire 3 hour trip home, I would die in a firey explosion. That's not high on my TODO list. But if this keeps up, I might raise it a few pegs. So Larry and I will be going down next week with the trailer, so we can bring the charger back in it's natural state, on the bed of a trailer in tow.

    Pod Life

    Fri Apr 16 22:45:10 CDT 2004

    I took some photos of the pod, It still needs some work but I'm pretty much living in it now. I plan to lock down the computers, so they will not move during transit, but other than that It's ready to move. So should I find work in anywhere I can be moved and set up in about 48 hours. The computers are operational, but I'm currently leeching broadband from the parents. It's Starband and as a UNIX guy, who spends most of his time in a remote shell, I must say the latency is unbearable. There is up to three seconds of lag between keystrokes. So I type it blind and backspace when there is an error. It's tedious. If is wasn't for openwebmail, I'd have to stop using email, pine is my primary client and it's unusable in a remote shell.

    And I'm back....

    Wed Apr 14 14:06:47 CDT 2004

    Well, I've got the public content back up. The bookshelf will be down until I get work or find someone willing to host it. My UML linode has limited space and such. I'll see what I can do, to get it back up, though looking for work is my priority right now. I'm going to try to get some pod pictures up this weekend, if anyone cares. Mostly they will be provided to answer the "How can you live in a pod?" questions and the like.

    Temporary Layoffs....Good Times...

    Tue Mar 16 07:31:01 CDT 2004

    Due to finiancial situation at my previous place of employment, I am no longer employed. It will take me a litte bit to get the content off of the old site, what with moving out of Montgomery and all. The server previously serving this site is powered down and in storage. I spend most of my day looking for work, and the weekends rennovating the pod. If anyone knows someone in need of a systems administrator who pretty much eats, breathes and sleeps for this computer stuff, pass them along my resume Have pod, will travel. Thanks.

    Thanks Bill.

    Tue Jul 1 07:31:01 CDT 2003

    I was rousted out of bed at 5:30 this morning by the chirping of computer-lab crickets, every UPS in the house was beeping. I have a flashlight that I keep right next to my bed in the event of a power outage. I realized at this point that I had moved my flashlight so that I could do see into a computer in the living room, grrr. So I look for my lighter, a zippo I carry because I'm comforted by the fact that I can make fire, I guess. It's out of fuel, grarrrr. So the 7 candles I have at hand are now useless. Now I normally keep my lighter fuel on the counter near the kitchen sink, but I had moved it so that I might cut some plastic with string. Don't ask. Luckily I know I have another flashlight in Le' Charget, so I head out to the garage, locate it, and it won't come on, *ugh*, so I toss the batteries and stumble around the kitchen until I find some fresh ones. No joy, The bulb is bad. *rage building* So I have to grope around in the dark in my workspace using my cell phone backlight for illumination, if you can call it that, until I find my lighter fluid. I manage to locate it and promptly refill the zippo blind. Now it has plenty of fuel. It still won't light. The wick is scorched. Now I'm livid, and the UPSes are beeping like crazy. I manage to shut down most of them with ctrl-alt-del, wait for floppy light and power off. I locate my leatherman and manage to pull the wick out, and it lights. I now have candlepower. And as soon as I light the first candle, I see my flashlight. I would go back to sleep, but I'm now furious. So I drive to work, only to realize I left my driver side rear window down last night, so I get to splash in the puddle that is my floorboard all the way to the office. At least it's not Monday.

    The GPL and The Common Cold

    Wed May 21 09:53:53 CDT 2003

    Following my current involvement in a flame war on the mailing list I decided to type up the following for the similie impaired If I offended people on the list, then they probably needed to be offended. But religious devotion should not be a replacement for logic... Even if it is for a cause as good as linux advocacy.

    PREMISE A: Any two objects that share a similar behavior or property can be said to be alike.
    PREMISE B: (object one) The GPL applies itself to any code to which it it linked.
               (object two) A virus applies itself to any tissue to which it is conjoined.
               The two lines above describe *similar* behavior between two objects. 
               (not exact, explicit behavior, but similar behavior)
               The GPL and a virus share a similar behavior.
    Conclusion: The two objects, The GPL and a virus can be said to be alike.
    Therfore: The statement "The GPL is like a virus." is correct.

    Neither consent, desire, nor the properties of the LGPL, Apache, or any other Non-GPLed license nor pro-opensource ideals have any effect on the premises or logic above. Note, that usage of a similie does not mean that the GPL is a virus, just that is shares a property or behavior with one. If one of the premises above is incorrect, by all means, let me know. Hell, I'll even post and addendum right here. But don't email me if you cannot refute a premise above. Telling me how great the GPL is doesn't effect the logic above. I'm a big fan of the GPL. I use Linux on 7 of my 8 computers including my primary workstation. But I am also capable of emotional detachment on this topic, which allows me to analyse the facts. Sad to say this is not true for some of the people on the list.


    Tue May 27 10:44:09 CDT 2003

    Derek and I have discussed this more since the flamewar. The conclusion we've come to is that while it is true that the GPL can be said to behave like a virus, It is not the most apt analogy, and the negative stigma to the word 'virus' makes the statement inflamitory. it is just as apt (if not more so, depending on your point of view) to say, "The GPL is like Hershey's syrup swirling about in the delicious creamy milk that is your code." But both statements have a bias.

    Site Migration

    Tue May 6 17:33:48 CDT 2003

    Well I'm finally moving over to the new system. Basically a bunch of home grown mod_perl scripts as slashcode has become so bloated as to be useless. It's too bad, I really liked the look and feel of slashdot, but I felt I needed to get away from it. So expect everything to be broken for the next couple of weeks as I get organized. I don't have a lot of free time to migrate the functions, so I'll be doing it when I can. While the migration is going on, the old (and undoubtably broken) site will be available at: *note: this server will be going up and down*

    Google Search


    ASCII Face
    Le Charget
    My Resume
    Scrap Book
    SysAdmin Rules
    Web Mail
    Terror Alert
    Terror Alert Level
    Job Search
    Job Bank
    Net Temps
    RedHat Network
    Driver Guide
    Google News
    Daily Kitten!
    The Register
    Local Weather
    Solar Activity
    Urban Legends
    Yahoo Mail
    Rotten Tomatoes
    Opry Mills 20
    Ebay Search
    [ Find it ]
    Dancing Bologna
    Because most every home page has some.
    Because most every home page has some.

    Debian Apache