My first blog entry after I started my website was Lessons Learned in Web Hosting Provider Selection . Looks like I am still learning lessons. You would hope (mistake - 1) that your web hosting provider would remind you when it's time to renew your domain and website registration. Well, I was reminded by my provider BUT on the day when my domain and my account were due to expire - A big 1 year anniversary surprise! I made the payment right away and assumed everything was taken care of (mistake 2). Just an hour ago, I realized my website was down, I saw a weird page when I went to the website. It was some search page with a link "why am I seeing this?"
When I had started my website, my provider provided ONLY web support, which would require me to open a trouble ticket using their website and/or participate in the discussion forum. I had (and I still don't have) 24x7 phone support (sort of mistake - 3).
The problem was that my domain had expired because even though the payment was received and approved, they had some "difficulty" renewing the domain. Fortunately, my provider now provides 24x7 chat support and surprisingly, some one was available to help. He worked on my problem right away and made the website available. Why wasn't I sent payment reminder email early enough? I don't know. Hopefully, the rep I spoke to will "look into it."
The lesson learned? It's hard to remember to pay your website bill since it's a yearly thing. But, do remember it. Make it an outlook reminder, make it a yahoo reminder, setup a yearly bill payment (if you bank offers it), register a credit card with your provider but don't rely on your provider to remind you. If your website goes down, you may loose business. Worse yet if your domain expires and becomes available, some one might take it.
Here is an amendment to my first blog entry . Before you sign up, ask your provider how do they renew the account? Automatically or do you have to do this manually? What happens if you fail to make the payment on time? Do you loose your domain and your content right away or do you have grace period?
Russ Olsen has posted 10 Unix things every Java developer should know on http://www.javalobby.org/articles/10things-unix/, +1 to that. I am going to post 5 more (didn't feel like adding unnecessary ones to make the magic number 10). Here are my picks in no particular order:
Know file permissions: Unix assigns read, write and execute permissions to owner, group and others. If you are going to be running scripts, installing software or simply copying files from another resource, you will need to understand and most likely change file permissions. Good luck learning chmod 755, chmod a+x, chown -R and more
Know how to use ssh (secure shell) and the tools that use ssh to allow login to remote machines. There are times when Unix servers don't allow telnet and/or ftp but allow ssh. Sometimes, you also need to download some files and use scp (secure copy or copy over ssh) to move it to the server. Checkout Putty and Cygwin
Know basic file operations: You should at least know how to create, delete, copy, move and rename files. It's pretty much a crime to change configuration files without saving a backup first. Learn basic vi editor commands. I like this page for vi help: http://cac.uvi.edu/miscfaq/vi-cheat.html
Know various archive utilities: Most of the stuff you download for Unix will be in tar and/or gz format. It is a good idea to learn to work with those utilities. Just knowing how to extract an archive or create one isn't enough. For example, how would you extract an archive in a different directory?
Learn to use the log files: If I were listing the items in any order, this would have been number one! Learn to tail a log file so that you can keep an eye on it while it's changing, learn to list log files by timestamp so that you are sure to look at the last one, learn which log file is updated when. For example, on my WebSphere Portal project, I have to look at trace.log, SystemOut.log, SystemErr.log and wps_$timestamp.log files in different situations. If the server doesn't start properly, I have to look at SystemOut.log and startServer.log. Occasionally, you will have to send log files to support group to get their help fix a problem. You may also need to update properties files that control how log files are updated and what is logged
Alright, here is a bonus one: Learn to type fast (this one needs no explanation)
Agile methodology is the best choice for my project. I am a project manager/senior architect/architect on a project and I have to finish the project in a short time so I will use agile methodology.
Agile in this case doesn't mean pair programming. I can't afford to have two developers working on the same piece of functionality. Agile in this case doesn't mean frequent testing. My developers can't afford to waste time writing test cases. Agile in this case doesn't mean automated testing. We neither have time nor skills to plan, design and create an automated testing framework. Agile in this case doesn't mean frequent interaction with the client/customer/end user/stakeholder. Instead, we have put barriers so that customers don't bother us in our work. After all, we need to accomplish a lot in a short time!
Agile on this project means no documentation will be done. We have a tight budget, we can simply ask questions when we have to. Did we have barriers in communicating with the client? Oh no. It's okay to ask questions. Just make sure they don't interrupt your work. We can't afford to waste time in coaching them. Agile on this project means no upfront architecture and planning will be done. Design? That will certainly not be done. You guys are all experienced developers. You should not have to be told what to do. Coding standards, documentation standards, naming standards? Forget all that. We are going to be measured on whether we can deliver the product not how well the code is documented. Agile on this project means iterations are not important even though we have them. We will simply not define any milestones. We will not try to release incomplete yet working piece of software at the end of the iteration. Iterations are simply dates on the project plan.
The project isn't meeting dates. We are not moving any faster with agile. There are still some unknowns and we haven't delivered a thing. The developers are confused and requirements are changing. Why are we using agile again? Shhhshshsh… we are running behind because you are not working hard enough. Stop talking and keep coding.
One of the most commonly used version control system is CVS and one of the most commonly used IDE is eclipse (Rational Application Developer - RAD, in my case). Surprisingly, Google isn't much of a help when you do searches like "rad cvs," "rad setting up cvs," "ibm rational application developer setting up cvs" etc. so I thought I would document the process I follow:
In most cases cvs repository resides on a server. To use cvs from within Rational Application Developer (or eclipse IDE), you first need to setup the repository and then share your project(s) using that repository. Let's review the process step by step:
Yesterday I ranted about the adamant db2 uncatalog. Today my ranting
continues with 'ibmdirctl' - a command used to stop/start IBM Tivoli
Directory Server. This is more of a gotcha than a rant though. I was
testing my server script to automatically shutdown the directory server
when the machine shuts down and bring the directory server back up when
the machine is booted.
