I very rarely blog about an application, however sometimes there is an app or two that make my life and job a lot easier to manager. I use a lot of mysql databases, and managing them is something I like to, lets say…not do.
I simply don’t like navigating through databases and tables from the mysql command line editor, and I really hate the PHP gui tool that it comes with…its just so cumbersome. I’ve been using navicat of late, but I hate the idea that I have to pay for it. I mean lets face it, why the heck would you want to pay for a tool that manages an open-source application! Its irony at its greatest. So in my early morning googling, I stumbled across a diamond among the db management tools.
Read more…
Uncategorized
As the cobwebs fill up on my blog just in time for spring, I thought it may best to clean it up and put some fresh stuff…So reporting is a kind of a big deal at these big corporations. Seems like there is always someone higher in the food chain who wants a spreadsheet to tell them what you’re actually doing, and if you’re hitting your targets. As annoying and cumbersome as it may be, it’s very important that you produce and provide statistics that are accurate, and up to date, so I’ve got a quick solution that will make any boss somewhat happy…
Read more…
Uncategorized
Often, when I begin a new gig, I have a few little tools I install to make my job that little bit more efficient.
One of these would be a wamp server and the ruby installer, along with the usual gems I often utilise. The wamp server is great to store data for scripts I’m writing, and a little ruby here and there never hurt anybody!
It occurred to me that I’d never actually blogged how to get ruby and mysql connectivity. As with anything ruby related, its easy but it does require a few key steps..
Read more…
Uncategorized
A strange problem occurred when I was running one of my selenium test suites. The test would run fine for my first testcase (that is, the first “method”). If you know a little about how selenium works, you’ll know that the teardown method is called after each method.
The teardown basically grabs any errors that occurs and kills the browser session, thus opening a new session in your next method. The issue that was occurring was this annoying JavaScript error being pulled from Internet Explorer, saying it cannot run scripts on the page. It halted my script and I had no way to get to the following methods…so after a little digging around I found a solution…
Read more…
Uncategorized
So I’ve been dabling with Selenium again…I love open source, but the lack of documentation is very frustrating. Hitting F1 in Loadrunner was taken for granted, so if I find something useful, as a fellow open-sourcer, I’m going to share it. Hopefully this can be of some use to you all.
Enjoy!
Read more…
Uncategorized
Quite often I get personally contacted by people requiring some assistance or just wanting to ask me for my advice. It’s a nice ego boost, but also a very important aspect of the industry that should be maintained. My beliefs are we should do what we can to assist others, and with the use of the internet, we have a great medium to share and distribute information, with such things like twitter, Google talk and, well my blog, are all making it possible for a total stranger to speak to another and get the help he/she needs.
Recently, a now friend of mine, Roberto Brusa struck up a conversation about changing the system time in vugen. He was making a system call, which is fine, its a great way to do it, but he did mention that he did not want the annoying cmd pop up and probed me for a possible alternative. Now I’ve never done this before but after bouncing ideas off each other I suggested that in theory, it should be possible, through the usage of the win32api, to manipulate the tme…We found that this statment was true, and proved it with the following…
Read more…
Uncategorized
From time to time I’m asked to do an odd job here and there, so I have a whole bunch of little snippets of scripts that I have created for people, to make their jobs faster, easier and more efficient. This last one, I was pretty pleased with, as it is very dynamic in what it can do.
Basically, the requirements were very vague–
“We need to find a range of values in a block of text.
The block of text will sometimes be 2 lines, or can be up to 20 lines.
The files we are searching through are in all different formats.
We would like to do multiple searches in one action.
We need to search for multiple values in a search, and only pass if all values are found within the line count we specify.
We would like our search data to be in a separate file we can build and run the search against this.
We would like to have total control over what file to search through.
We need to know the line number it was found on, when it is found.
If possible, can we run this on a windows box, and on a mainframe server to which we can’t install apps”
So with all these in mind, I was off to write a script. I wanted to use ruby, but knew it wasn’t native to mainframe’s, so I opted for Perl instead. My Perl skills are not what I would like them to be, so feel free to update/criticise/chop my code 
Read more…
Uncategorized
perl
Quite often, I don’t get a read only ssh account to a server, so I can capture server metrics during runtime in loadrunner. This usually means I need to request the data to be sent to me after the test, which the sys admin will very easily do. The only problem here, is that a vmstat log is ugly, and difficult to read in raw text not to mention, had to analyse.
Test managers are generally very busy, and want a report with pretty pictures, giving them a view on what was happening during the run. Loadrunner analysis is a great tool for displaying such data, and has an import ability, for such problems, but first, we must convert our vmstat/sar/iostat file into a format that is easily graph-able.
Read more…
Uncategorized
Performance and load testing has many input requirements. The most important and difficult to obtain, is clean and valid input data. I find that it’s often harder to request for data, than it is to just make my own. Unless the data needs to be aged, or processed somewhat, then you are stuck at the back of the queue. So think not of vugen/loadrunner as only a performance testing tool, but a gateway to process information on the application at a rapid rate. This is why I favour vugen/loadrunner as a data creation tool.
Once I set up my scripts they can often be reused throughout the lifespan of the project creating adhoc data when I need, quickly and rapidly.
Storing the data is something I’ve experimented with over the years, and have found that a mysql database is superior to store, validate, and call your from, all with a nice little wamp server, but more about that another time. If you want to just store your data in a CSV, text file, or any other format to your liking, you can do easily. I warn you, it’s clunky but it does work.
Read more…
Uncategorized
When saving paramaters through the web_reg_save_param, the ord=all option is very useful in finding multiple instances of your desired item.
Generally you can work out which ordinal you want to use, and call your parameter later using {paramaterValue_15} for the 16th ordinal (starts at 0). But what if you want to use every single ordinal found?.
Why? Well, why not! — you could have a situation where every ordinal found is a value which is the trail of a url, and you want to navigate to each subsequent url that this page returned. Since the ORD is an array, we should be able to call it easily via paramaterValue[#], right?. Wrong…this simply doesn’t work in vugen, however with a little bit of funky C coding, the answer is quite simple.
Read more…
Uncategorized
