UPDATE  – Sorry for the RSS spam, I have been playing with twitter (and frankly am failing to see its draw) and setup twitter tools to integrate my blog with my twitter microblog. I apparently left the create daily twitter summaries on. My appologies for the RSS spam.

–ColinSimilar Posts:

• Fibre Channel over Ethernet is taking off
• Cisco Nexus 5020 and 5010 FCOE video ordering guide
• Nexus 5020 – Consolidated 10 Gig Ethernet and 4 Gig Fibre Channel
• Cisco Unified Computing System Quoting and Configuration with Netformx
• Is your network ready for Cloud Computing with Virtual Infrastructure 4?
• Cisco Nexus 7000 DataCenter switch released – Welcome to DataCenter 3.0

--Colin McNamara

Twitter Updates for 2008-06-05

GoDaddy, one of the nations largest registrars and hosting providers is distributing bloated and possibly insecure code that will cause the average user to more then double their hosting costs. In the pages below you will learn exactly what GoDaddy is doing to your server, how their support staff will try to upsell you, and the steps you need to take to ensure proper operation of your Virtual Dedicated Server.

Background

One of my new years resolutions this year was to consolidate hosting accounts into one virtual server (hosted). I had my domains, and my old hosting with Godaddy already so it was a no brainer to try out one of GoDaddys VDS (Virtual Dedicated Servers).

I went with their 29.99 a month package, with Centos5, unlimited domains, 10Gb disk, and 256 MB of memory. This should be perfectly fine for hosting a couple MySql driven sites, and a couple gallery instances. Let me emphasise this is only handling 4 active domains, two of which only have static HTML.

Provisioning was a breeze, from order to shell account only took 4 hours. I was provided with shell access, pre-configured yum repositories, and this web control panel – simple control panel, or TurboPanel (seems to have two different names). I was able to pop into GoDaddy’s control panel interface with a direct link from their server manager console, and was setting up domains in no time. (Let me throw this caveat out though, don’t buy this product for your mom’s hosting…. the documentation is horrible, and by horrible I mean non-existant).

So I get my server all set up a couple weeks ago, transfer all my files from 2 Cups Solutions and set up my new www.colinmcnamara.com site. Things go just fine, I changed over to wordpress as a CMS and am totally thrilled. My applications and email are working perfectly. Plus, I have a shell account at GoDaddy which is a very handy thing to have as a network engineer. Things are going so well, that I give my buddy Rick a Christmas present and get ricksdavis.com and el-cinco.net for him, and host it on my GoDaddy VDS.I go ahead and purchase the domain through godaddy’s domain manager. This is obviously GoDaddy’s core competency, and goes flawlessly as usual. Next step, I go into the Turbo Panel web interface. Let me give you a little background on TurboPanel. This is the “free” equivalent “to plesk or cpanel. It is actually ok for automating your domain provisioning, though if you are a normal skill level user, I would recommend paying the extra 9.99 a month for Cpanel.

Server Error

I open up my TurboPanel interface and go to provision ricksdavis.com into the domain manager. It comes back with the least descriptive error I have ever received.

Server Error
We are sorry, the system has encountered an error while processing your request.

Home

If you continue to receive this error, please contact your system administrator.

Your URL: /domain/edit.do

Error details:

CommandFailedException: Unable to get min/max uids
at c.g.t.f.systems.user.LinuxUserSubsystem.loadUids:825
at c.g.t.f.systems.user.LinuxUserSubsystem.getMinUid:780
at c.g.t.f.systems.user.LinuxUserSubsystem.loadUserInfo:670
at c.g.t.f.systems.user.LinuxUserSubsystem.getUserInfo:646
at c.g.t.w.actions.domain.ActionDomainEdit.process:84
at c.g.t.w.actions.AbstractSpringAction.execute:118
…
at c.g.t.w.filters.AuthorizedResourceFilter.doFilter:38
…
at c.g.t.w.filters.RequestPopulationFilter.doFilter:117
…

This is the most descriptive error ever right? it tells you what is wrong, has a link to the support system, and gives you actionable information…. I would say a resounding NO. This is a classic example of why friends don’t let friends program in Java. When I got this error last night, I was scratching my head. As an engineer, the first thing I will look at is the last change to the system. Coincidentally I had installed awstats two nights before, and looking at my change logs, I saw that I had upgraded my perl version. So, with no fast response to the support email from GoDaddy I chose to put in a server re provision request (fully automated) and restore from my backups. That process took about an hour, but afterwards I was back online with no errors. Eureka! I found it (I thought). I provisioned Rick’s domain, wordpress, gallery2 etc and then went to bed.

GoDaddy Support Response

Fast forward to this morning, and I finally recieve an email reply from godaddy support. The email is pasted below-

(I have replaced the agents name with John Doe. Tech support is a hard job and I see no reason to highlight him specifically)

Before re-provisioning the server, I went ahead and tried the old three finger salute (reboot) the error still existed. So even if this email would have came to me on time, it would not have helped.

But that is besides the point. lets dig into the solution

The agent suggested the following fix -

Restart tomcat – which I was NOT using for any of my web applications (not an ejb guy)

/etc/init.d/tomcat55 restart

Restart TurboPanel (or simple control panel now – they haven’t updated their init scripts)

/etc/init.d/turbopanel restart

GoDaddy tried to upsell me, instead of fixing their code

He gives the standard, run less stuff on your server speech (remember, I only have 4 domains on this server)

Now here is the kicker – To remedy this issue long-term, you could either setup a server with 512mb RAM, or upgrade to a Dedicated server.

What the heck is with that? I should not need an upgrade with only 4 domains on a server. Especially when in the setup the default Cpanel implies support of 30 domains. Is this Tech Support or a Sales Call?

But, since my server was working fine I don’t pay much attention to the email and move on with my life.

Fast forward an hour, and I am show Rick how to access all the features of his new site, and I figure that I need to change an email account on his domain to forward to his old account. Fine, this should take two minutes. I log onto the TurboPanel interface to put the email forward in.. and there it, a big useless error screen. What the heck is with that?

So this time I actually read the email, and try the fix. Results = nothing. GoDaddy offers shell access so I log in, run top and filter for %memory used. Low and behold, there is only 8142 bytes of memory free, and a Java process owned by root is using 300Mb of virtual memory, and 132Mb of real memory, and Tomcat is using 115Mb of virtual memory and 86Mb of real memory.

Let me translate this into English – Godaddy’s control panel application was using 218 Megabytes of the 256 Megabytes of memory I had purchased. That left me with 34 Megabytes of memory . Let me clarify this, I had paid for a virtual server with 256Mb of ram, up to 1000Mb bursted (which I think is their code for swap). I am only running 4 domains on this server, and two pop3 email servers. This should not be a problem.So what is the cause of the problem? I can sum it up, crappy Java programming. Someone decided to write this program in Java (probably easier to outsource) instead of optimising it to run on lean systems. Their code effectively takes up all the available memory. And on top of that, they are are running a webserver process as root… yes as root. It is like asking for your server to get hacked.

Now, that I am done ranting, let me highlight how to fix this problem.

If you are on Windows use the following procedure to get shell access to your GoDaddy VDS -

• you will want to download a ssh client called putty – Download Here
• Copy this file to your desktop, double click putty.exe , and you should see something like this -

• In the host name field I have www.yourdomain.com . replace yourdomain with your domain name.
• Click on the open button on the bottom right, and a shell should pop up, along with a warning that looks like this (you can click OK to the warning)

• Now skip past the linux section

If you are running Linux or Unix start here -

• Open up a command line terminal
• ssh using your godaddy simple control panel username example -

ssh your-godaddy-control-panel-username@www.yourdomain.com

• If this is your first time connecting to this server, you will be prompted to accept an unkown ssh key into known_hosts, choose yes to accept

Both Windows, Linux and Unix Continue Here –

• Enter your the password you use to access your GoDaddy Simple Control Panel Interface
• You will be presented with what looks like a DOS window, this is called a secure shell terminal. Type in the username you use to access your Godaddy Simple Control Panel and then hit enter

• Now type in the password that you use to access your Simple Control Panel Interface and hit enter

• Congratulations, if you see the window below you are now shelled into your virtual dedicated server.

Change to the Root user

• Now that you are shelled into a Linux device you need to escalate your privileges to get administrator level access. In the Unix world this user is known as Root. You can change to this user, and get full system privileges by using the following command.

su – root

• You will be presented with a password prompt, enter in the same password that you have used to log into your Simple Control Panel Interface
• You are now root, be careful with what command you enter under this user, as you can do some damage if you are careless

Clean out your servers memory

• Most recent Redhat direvatives (including Centos) utilize a tool called yum to add and remove packages. This is also true with your linux servers at GoDaddy.
• you need to install a tool called memhog, it is part of a package called numactl that is normally used to assign specific process’s to specific cpu’s in a multi-core system. We will be using it today to fix GoDaddy’s memory hogging application

yum install numactl

• choose yes to all the prompts, and numactl will automatically be downloaded and installed on your server

Stop GoDaddy Simple Control Panel, and Tomcat, and tell them not to start automatically when your server restarts.

• Tomcat is a special type of webserver for Java based applications. Godaddy uses it to run their control panel interface. 99.99999 percent of users will not need to use Tomcat. If you do need to use Tomcat then you are a technical user and will know what to do.
• In Linux, server applications are called daemons. The are executed by init scripts. We will use these scripts to turn off these server applications
• Turn off the TurboPanel daemon ( this is the process that runs your simple control panel web interface

/etc/init.d/turbopanel stop

• Next we need to turn off Tomcat

/etc/init.d/tomcat55 stop

• Now that we have these services turned off, we need to make sure that they don’t come back when we reboot the server. We can do this by using the chkconfig command.
• Stop the Simple Control Panel Interface from starting automatically by executing the following command

chkconfig turbopanel off

• Stop the Tomcat server from starting automatically by executing the following command

chkconfig tomcat55 off

Clean the mess GoDaddy made of your servers memory

• A couple steps back we installed numactl. The executable we wanted out of this package is memhog. Issue the following command to take your memory back. This command will overwrite 200 megabytes of your memory, allowing the rest of your applications to get access to that memory.

memhog 200m

• This will Clear out the memory that GoDaddy’s application took over, and allow the rest of your daemons to run fine.

Great, my server is running better now. But I want to use my Simple Control Panel Interface. How do I do that?

• Easy, all you need to do is temporarily start the turbopanel daemon. When you are done making changes, you can turn it off again

/etc/init.d/tomcat55 start

/etc/init.d/turbopanel start

• When you are done, don’t forget to turn it off

/etc/init.d/turbopanel stop

/etc/init.d/tomcat55 stop

*** UPDATE *** *** UPDATE ***

I managed to stay with Godaddy and use their Virtual Dedicated Servers for 12 months. However I have moved to a new hosting provider (rimuhosting) in the spring of 2009. My last straw with GoDaddy occurred when I noticed that even with my memory utilization under control my applications were performing poorly. I shelled into verify and found that the virtual disk I/O was severly limited.

Performance issues –

Simple command to determine directory size on GoDaddy VPS

[root@ip-72-167-15-128 home]# time du -hs
1.3G    .
real    0m49.013s
user    0m0.011s
sys    0m0.062s

Simple command to determine directory size on Rimuhosting VPS
[root@colinmcnamara home]# time du -hs
1.3G    .
real    0m0.343s
user    0m0.050s
sys    0m0.290s

Translated into english, it took 49 seconds to execute a simple command on my GoDaddy server, and it too 0.34 seconds (under 1 second) to execute the same command on my virtual server at Rimuhosting.

On top of that, I was paying $32 dollars a month to GoDaddy for a server with 256 megs of ram as well as a FTP backup account. When I created my short list of hosts, I noticed that I was not getting the best deal by staying with GoDaddy.

Pricing at Rimuhosting

• Virtual Dedicated Server 190 Megabytes RAM – $19.95
• Virtual Dedicated Server 400 Megabytes RAM – $29.95
• Virtual Dedicated Server 900 Megabytes RAM – $39.95
• Virtual Dedicated Server 1150+ Megabytes RAM – $49.95

One thing you may notice, is that for the same price as the base GoDaddy VDS ($29.95) you get 400 Megabytes of RAM instead of 256. (and they won’t put on a memory sucking turbopanel app) On top of that, things that I had to pay extra for at GoDaddy such as backup space, backup mail relays, and DNS service came bundled for free.

I made the choice to move to Rimuhosting, and the service has been superior to GoDaddy in every single way. If you get sick of GoDaddy like I did, you might want to check Rimuhosting out, I think you’ll like them.Similar Posts:

• What should I do this thanksgiving break?
• Zone based IOS firewalls
• Routers can email you when they go down
• New features in VMware 3.1
• Altor Virtual Network Security Analyzer (VNSA) integrated with Cisco’s Nexus 1000v for VMware
• Identity aware networking using Cisco TrustSec

--Colin McNamara

Why GoDaddy Linux Virtual Dedicated Hosting Sucks & How to Fix It

On the fire maps of the San Diego Witch fire, we are pretty much under the big red spot (directly under lake hodges, to the west of the 15). We moved to 4s ranch in June , on the intersection of rancho bernardo and ralphs ranch road. (pretty much right under the big red spot). We left on on monday at around 10 am with fire on 3 sides of our community.We got out with a couple changes of clothes, important documents, and the kids. Almost everything else is replaceable.

the fire directly upwind of us, coming up the ridge -

Getting out of 4s ranch was crazy though, traffic was jammed up through camino del sur, and the police were not helping. We ended up sneaking out through a backroad and coming out through Del Mar. Our final destination (and temporary home base) is the Ramada plaza hotel in Anaheim.

The winds were blowing 60 mph +

While it is rough having to flee from our home, we are much more fortunate then those people stuck at Qualcom, or Del Mar, etc. While they have been sleeping on the cold cement for two days my family has enjoyed two queen size beds + a fold out couch. Kylie is a little shell shocked, I think fleeing from her home was a bit much, while Chris is just being a normal little two year old terror.

Fireman going after the fire where it wrapped around our neighborhood -

We have also been very fortunate to have many supportive friends. I thank everyone of them for all the support and help they have given. It makes things much easier when you know that you aren’t alone in this struggle. We have a couple standing offers of homes to stay in, a home cooked meal (from the best mexican food cook I have ever met) waiting for us tonight, clothes / diapers for Chris, pretty much everything we could have wanted.

The view as we were leaving (sorry about the blur, we were in the truck)

During this whole ordeal I there have of course been challenges, but I am amazed how everyone is pulling together. In the hour before the police forced the evacuation neighbors were helping each other with information, people were grabbing hopping in trucks and running to help with the fire lines. While it was a post apocalyptic scene with the choking smoke, ash everyone, and blood red sun, people were helping each other. It is to see the best of people when they are under the worst conditions.

The blood red Sun

The two big questions every day are – Is my home still there ? and If it is there, when can I go back? As of right now the evacuation order still stands for my neighborhood. It is directly downwind (about 1 kilometer) from the homes in Rancho Bernardo they have been showing on the news. When we left, fire was coming up the Ridge that separates our neighborhoods. On signonsandiego.com ’s firemap it looks like we may have been spared, but we are still in the middle and downwind of the major fire. The only thing that will signal a change an all clear is for the santa anna winds to subside.

neighbor checking taking pictures of the fire prior to the evacuation order

One thing that is important in any emergency is to go through a lessons learned.

The things that worked well were -

1. Having a 4×4, it allowed us to choose creative escape routes, and avoid the possibility of getting stuck in traffic with the fire behind us.

2. Having decent non-perishable snacks (granola, raisins, etc)

3. Having lots of water bottles laying around – dehydration is a big issue with the dry air.

4. having travel kits ready – because I travel a bit, I normally have my carry-on bags prepped.

5. Bringing a toy, and a little video player for Chris – this saved lots of heartache for everyone

Things that I could have done better -

1. Inspected my wife, daughters, and sons bags before we left – they didn’t pack enough clothes, my daughter forgot things like underwear, socks, etc.

2. Had the key to the breachlock on my shotgun handy – i couldn’t find the key to unlock my weapon. If this would have devolved to Katrina standards then this would have been essential. In retrospect I would have done better with a semi-auto handgun. It is easier to transport and conceal while traveling.

3. I should have had better radios. I am a ham operator, have my license. But I have always used peoples base stations tethered through the Internet (basically a big ip based microphone / speaker connected to their transceiver. Ham operators are a godsend in situations like this, and having a mobile unit in my truck would have added valuable information in a time when we had none.

Colin McNamara
CCIE #18233

Copyright ©2008 | Colin McNamara | CCIE 18233 | All Rights Reserved”

Colin McNamara
CCIE #18233

http://www.2cups.com

“The difficult we do immediately, the impossible just takes a little longer.”Similar Posts:

• San Diego WildFIre Update
• Colin has moved to the San Francisco Bay Area
• How to succede in 2007 – By Tim O’Reilly
• How close did the fire come?
• My Twitter tweets this week
• The emergence of MDS features in Cisco’s datacenter networking equipment

--Colin McNamara

Witch fire in San Diego – Notes from an evacuee

Cisco has finally included zone based firewalling in the IOS firewall feature set. The configuration guide can be found here -

Zone Based Firewall Design and Configuration Guide

The things that really got me interested are -

1. It is VRF aware (works well with network virtualization strategies)
2. No more CBAC’s
3. Policing built into firewalling classes
4. Content inspection including HTTP,P2P, and Instant Messenger

I think the biggest plus for this release is that IOS firewalls are finally following the general trend of zone based firewalling. By moving this way, configuration errors resulting in lax controls are likely to be minimized.

Excerpts from the documentation -

Cisco IOS Software Release 12.4(6)T introduced a new configuration model for the Cisco IOS Firewall feature set. This new configuration model offers intuitive policies for multiple-interface routers, increased granularity of firewall policy application, and a default deny-all policy that prohibits traffic between firewall zones until an explicit policy is applied to allow desirable traffic.

Nearly all firewall features implemented prior to Cisco IOS Software Release 12.4(6)T are supported in the new zone-based policy inspection interface; supported features are as follows:

•Stateful packet inspection

•Application inspection

–HTTP

–Post Office Protocol (POP3), Internet Mail Access Protocol (IMAP), Simple Mail Transfer Protocol/Enhanced Simple Mail Transfer Protocol (SMTP/ESMTP)

–Sun RPC

•VRF-aware Cisco IOS Firewall

•URL filtering

•Denial-of-service (DoS) mitigation

Zone-based policy firewall generally improves Cisco IOS performance for most firewall inspection activities.

The only Cisco IOS Firewall features that are not supported in zone-based policy firewall in Cisco IOS Software Release 12.4(6)T are as follows:

•Authentication proxy

•Stateful firewall failover

•Unified firewall MIB

Zone-based policy firewall completely changes the way you configure a Cisco IOS Firewall.

The first major change to the firewall configuration is the introduction of zone-based configuration. Cisco IOS Firewall is the first Cisco IOS Software threat defense feature to implement a zone configuration model. Other features might adopt the zone model over time. The classical Cisco IOS Firewall stateful inspection/context-based access control (CBAC) interface-based configuration model employing the ip inspect command set will be maintained for a period of time, but few, if any, new features will be configurable with the classical command-line interface (CLI). Zone-policy firewall does not use the stateful inspection/CBAC commands. The two configuration models can be used concurrently on routers but not combined on interfaces; an interface cannot be configured as a security zone member as well as being configured for ip inspect simultaneously.

Zones establish the security borders of your network. A zone defines a boundary where traffic is subjected to policy restrictions as it crosses to another region of your network. Zone-Policy Firewall’s default policy between zones is to deny all. If no policy is explicitly configured, all traffic moving between zones is blocked. This is a significant departure from stateful inspection’s model, in which traffic was implicitly allowed unless it was explicitly blocked with an access control list (ACL).

The second major change is the introduction of a new configuration policy language known as CPL. Users familiar with the Cisco IOS Software Modular quality-of-service (QoS) CLI (MQC) might recognize the format being similar to QoS’s use of class maps to specify which traffic will be affected by the action applied in a policy map.

Colin McNamara
Copyright ©2008 | Colin McNamara | CCIE 18233 | All Rights Reserved”Similar Posts:

• Cisco is using Linux virtualization and 40 core CPU’s for its next generation routers
• Cool new features in 12.4(15)T
• Routers can email you when they go down
• Identity aware networking using Cisco TrustSec
• Measuring and mitigating risk involved with sharing virtual infrastructure between DMZ and Internal environments
• Cisco NX-OS 4.0 | Next Generation Internet Operating System

--Colin McNamara

Zone based IOS firewalls

My background is all over the place. It goes from running a small ISP when I was 18 to managing international CallManager clusters, to enterprise data center design and migrations. I swear that I have tech ADD. I see something new and I am like.. oooooh shiny, lets do a sniff and see how it works.

I actually attempted the Route Switch lab twice in 2002 (still had token ring and dlsw still) but got activated for the war before I could get my number. I can tell you.. that sucked horribly.

So, why did I choose to get my Storage CCIE first instead of finishing up my Route Switch first? Well, over the past couple years I have done a significant amount of IP storage (NFS, ISCSI, ATA over Ethernet, etc) including building my own IP storage heads based on linux. I had also done a little bit of fiber channel work. I felt that I had a significant advantage compared to most network engineers in the storage networking space. And, I think more importantly, the topics that I had to study were new and fresh. Whereas when I cracked open my Route Switch books.. I honestly wasn’t to excited about it at the time ( I think I was holding a grudge from my first attempts in 2002).

The partner e-learning central portion of Cisco has an Excellent lab access to labgear.net. It gives you 2 hour segments of time with 2 mds’s (both with IPS blades), 1 2 port jbod, and 2 2 port servers. They also have introductory tutorials for most major technology segments.

I did all of those, along with the every E-learning class that was on PEC. I think the major ones were the Design, Support, and CASSI classes.I took that, sat for the design and support specialist exams, and took my written last December.

After doing all that work, I was honestly 70% there. I took a little break for january and febuary (There was no open lab dates) and then started hitting the labs I think at the end of febuary.

For the majority of my practice I used the labs available through PEC. I also was able to weasel my way into 8 hour sessions every sunday from our channel SE. (I owe him plenty of drinks at networkers this year) For those 8 hour sessions I had labs created to summarize the major technology areas and to be as evil as possible to myself.

I had one attempt which didn’t work out as well as I would have liked. In RTP the lab starts at 7:15. This means if you are from the west coast like I am you will be getting up at 3:00 am in the morning for your lab. That royally kicked my but. I diverted from my attack plan and ended up
running out of time. It was a classic example of letting the lab run you, instead of you running the lab.

Luckily a data opened up just over a month later. I spent that time just working on my speed (speed is the secret sauce). I migrated to using Fabric Manager (gui interface) instead of command line. After 2 practice runs my time had dropped 25%. I also flew out to RTP 3 days before, and made sure to relax and get on east coast time.

For my final prep I got access to iementor’s lab. Their lab is excellent. I cannot reinforce enough how much it contributed to my confidence in that final week. Roman was really cool about working with an existing candidate to get me squeezed in. They are really cool guys, and they have the only workbook on the market right now.

So, I used their lab 2 days before mine, and then just chilled the day before. The day of the lab I had 70 points by lunch, and had completed configuration of the lab by 12:30. I spent the rest of the day reviewing my configurations (found 2 errors), fixing one bug, and generally harassing the proctor to make sure I didn’t misinterpret anything.

So.. if I had it all to do over again what would I do. I would still do the PEC stuff, and continue to primarily live in the Doc CD. Though I would probably pony up the cash and buy the iementor book and more lab time with them.

Copyright ©2008 | Colin McNamara | CCIE 18233 | All Rights Reserved”

Colin McNamara
CCIE #18233
“The difficult we do immediately, the impossible just takes a little longer.”Similar Posts:

• What does it take to pass the CCIE exam?
• Fibre Channel over Ethernet is taking off
• And it begins again – On the road to my CCIE in Storage
• Cisco is using Linux virtualization and 40 core CPU’s for its next generation routers
• Simplifying your Data Center with Cisco’s Nexus 2000 Fabric Extender (FEX)
• My CCIE Storage Shopping List

--Colin McNamara

Why was Storage Networking my first CCIE? And What did I do to prepare?

2. Update the Real Time System config generator

Existing Features-
Generates Content Switching configs from application data
Has a web interface for development access
Notifies help desk
New Features -
Create firewall definition files
Create nagios probe config
Integrate XML submission probe generation
Thoughts
Nice to have, but 90% of the functionality is already there

3. Update Nagios system config generator (nagiosuploader)

Existing Features-
Grabs system info when run locally
Posts configs to monitoring server for processing
New features
Host auditing to determine services monitored
Clean up error condition handling
Web interface for config generation
Thoughts
Having a web interface for users to fill out would lower my workload

4. Automate web application error reporting

Existing features -
None, well.. manually executed text processing on log files. I wouldn’t call that an application
New features -
Automation of customer / error code matrix grep statements
Output in XML / http format web page and email
Text only output in commented form <– text –> for command line mail clients
Thoughts -
Proves the point that certain people can be replaced by very small shell scripts (Unix humor)
Should be a big with with Client services
Could drive the application “digital dashboard”
Should be a fun to practice using AWK for formatting data in a standard output
Serious danger of scope creep
May be better implemented in a MySQL back-end

5. Do some personal growth items

Finish the Vmware image for the 7206 emulator and post to Vmware Technology Network for general use (delete IOS files before uploading)
Read some more of Long’s storage book
Write a sample chapter for Stay at Home Photography
Get GTK pod to run under my cmcnamar account, not root (I think the IPOD’s file system doesn’t like changing ownership or file attributes)
Redo the 2cups homepage
Clean the Garage……..
Copyright ©2008 | Colin McNamara | CCIE 18233 | All Rights Reserved”

Colin McNamara
CCIE #18233

http://www.2cups.com

“The difficult we do immediately, the impossible just takes a little longer.”Similar Posts:

• User experience testing – enhanced
• Google Translate API and speech recognition
• Usability features in Cisco’s Nexus 7000
• Why GoDaddy Linux Virtual Dedicated Hosting Sucks & How to Fix It
• RSS feeds – an intranet aggregation solution?
• Routers can email you when they go down

--Colin McNamara

What should I do this thanksgiving break?

This is great information to have, however when you have a multi tiered application it becomes harder to get a true status of your application stack. Luckily most new applications nowadays incorporate a HTTP interfaces.
However, this interface responding to a http get request generally does not indicate full application functionality. To get that level of information out you usually need to present some authentication to the web app, execute a click through, etc.

Now this has always presented a problem for me. I can check the availability of each level of the application stack. What I can’t do is verify that the application is truly working properly. This puts a major blind spot in my monitoring strategies. Luckily I just came across a very cool application. This sweet application is webinject – http://www.webinject.org/
It can be configured to graph application performance to mrtg, or output to nagios network monitor. Below is some information about this project.

What is WebInject?
WebInject is a free tool for automated testing of web applications and services. It can be used to test individual system components that have HTTP interfaces (JSP, ASP, CGI, PHP, Servlets, HTML Forms, etc), and can be used as a test harness to create a suite of [HTTP level] automated functional, acceptance, and regression tests. A test harness, also referred to as a test driver or a test framework, allows you to run many test cases and collect/report your results. WebInject offers real-time results display and may also be used for monitoring system response times.
WebInject can be used as a complete test framework that is controlled by the WebInject User Interface (GUI). Optionally, it can be used as a standalone test runner (text/console application) which can be integrated and called from other test frameworks or applications.

Programming Language and Platforms
WebInject uses an XML API (interface) for defining and loading test cases. You can use WebInject without ever seeing it’s internal implementation (no scripting or programming necessary to use it).
WebInject is written in Perl and can run on any platform that a Perl interpreter can be installed on (MS Windows, GNU/Linux, BSD, Solaris, MAC OS, and many more). Currently, binary executables of WebInject are only available for MS Windows. If you would like to run on other platforms, you must have a Perl interpreter and run it from the Perl source code.

Test Cases
Test cases are written in XML files, using XML elements and attributes, and passed to the WebInject engine for execution against the application/service under test. This abstracts the internals of WebInject’s implementation away from the non-technical tester, while using an open architecture [written in Perl] for those that require more customization or modifications.

Results/Reporting
Result reports are generated in HTML (for viewing) and XML (for tranformation by external programs). These detailed results include pass/fail status, errors, response times, etc. Results are also displayed in a window on the User Interface if you are running the WebInject GUI, and are sent to the STDOUT channel if you are running the WebInject Engine as a standalone (console) application.

Service-Level Monitoring
HTTP response times can be collected and monitored in real-time during test execution. Timer statistics are calculated and displayed in a monitor window during runtime. When used along with gnuplot (a plotting utility), a response time graph is generated and updated in real-time as the test runs. This is used to verify responses from the web application or web service under test are within an acceptable range (to meet your SLA or quality of service criteria). This also enables WebInject to be run as a performance probe for application/service monitoring.
WebInject can also be integrated as a plugin for external monitoring systems. In this case, it is used in console mode as an intelligent test agent that returns status and response times to your external program.
For real-time monitoring of your web applications or web services, WebInject is able to run in a mode that makes it compatible with Nagios. Nagios is an open source host, service, and network monitoring program.
For graphical trending of web service-levels over a long period of time, WebInject is able to run in a mode that makes it compatible with MRTG. MRTG (Multi Router Traffic Grapher) is an open source tool for collecting, storing, and graphing time-series data.

–Colin
Copyright ©2008 | Colin McNamara | CCIE 18233 | All Rights Reserved”

Colin McNamara
CCIE #18233

http://www.2cups.com

“The difficult we do immediately, the impossible just takes a little longer.”Similar Posts:

• What should I do this thanksgiving break?
• Application Extension API notes – Cisco Live 2008
• Why GoDaddy Linux Virtual Dedicated Hosting Sucks & How to Fix It
• Is your network ready for Cloud Computing with Virtual Infrastructure 4?
• New Buzzword Enterprise Service Bus
• RSS feeds – an intranet aggregation solution?

--Colin McNamara

User experience testing – enhanced