James Lloyd

Blog of James Lloyd pro geek, and IT Evangelist

 

 
Home » projects » how-to-be-a-lazy-administrator
 

How to be a Lazy Administrator!

  • Category: Techy Stuff
  • March 18, 2008
  • James Lloyd
  • No comments yet
  • 257 views

  • 1 Star2 Stars3 Stars4 Stars5 Stars (1 votes, average: 5 out of 5)
    Loading ... Loading ...

    Introduction

    Lazy administration is the art of getting computers to do what their supposed to, all the boring stuff. While your computers are chewing away at all the mundane day to day tasks, you can concentrate on the interesting projects, that will ultimately make you greatly appreciated by you employer and make sure all you have left to do is the really interesting work. Its a self reinforcing goal, the closer you get to your goal the easier it is to achieve.

    To achieve this dream of a sys admin Utopia, you will need some resources. The first and most important of these resources is likely to be your or your direct reports time. If you day to day job keeps you on the back foot all the time (and if so where you finding the time to read this?) you need to create room so that you can start to implement your improvements.

    Where to Start?

    To be able to to move forward you naturally need to work out where you are currently. Being very close to your environment you will probably find you already have an idea of what needs to be done first. Don’t just dive in and try a mend the problem, take a step back and try to gather some facts. However if you find the current issue is that the exchange server is on Fire or currently being rained on by a faulty Air Con unit, a bit of common sense may tell you to deal with that right away.

    Communication

    Communication, not of the SMTP kind but that of the verbal type. Go and physically talk to your users and find out what their computer problems are. If you find their number one issue is they really really don’t like the corporate Backdrop you forced with a group policy, then your on to a winner. What is more likely is that you’ll come across comments like “The network is really slow” or “Application X is always down”. Don’t immediately dismiss these comments even though they have come from users some of which may have rung you earlier that morning to reset a password they changed less than 5 minutes previous. These users will sounds off to anyone who will listen and if their impression of your environment is bad, that news no matter how it was determined, will spread.

    To keep your users on side you only need to keep them informed. Sending out information about planned outages, and making sure everyone knows what is going on during an unplanned one. These types of communication should contain details on when, why and for how long as long as the communications are tuned to the expected audience, i.e. explaining that all the desktops will need to reboot at X a clock due to a patch being installed because the people that wrote the software in the first place are all idiots, isn’t going to do you any favours. Information should be high level and ideally focus on the benefit of the change your proposing to do.

    If you have an intranet, get a page on it showing the state of your environment with regard to the things that will matter to user. For example a traffic light system showing your logical systems. So your third party spam filter sendmail, and Exchange will be “Email”, the proxy and the Internet pipe will be “Internet/Porn”. Again here a traffic light system could potentially save you calls from the user, if page is regularly updated users will start to go to it when they have a perceived problem and not just pick up the phone.

    A good example would be Showing Email as amber and some text to explain that outgoing emails are currently leaving slowly due to the backlog of emails in the queue. This isn’t an outage but definitely a problem your users will detect and could perceive as an outage while they compound the issue further by sending thousands of test messages to their Gmail account & their friends.

    System Monitoring

    Monitoring your environment is the place to start so you can understand fully what exactly is going on. When asked “Whats the busiest time on your network” a classic response would be “about 9:00am, when everyone is logging in”. Then after looking at the network monitoring tools noting that the network is maxed out for a few hours after midnight because of the network backups. Know and understand your environment !

    Availability

    The first type of monitoring you should put in place is availability this is fairly simple and can be used to help to spot outages as they happen, so you can inform your users and then set about sort out the problem. Typical tests this type of monitoring will perform are:

    • Ping a host to check its alive
    • Telnet to a port to check its open
    • Check a service is running

    You should note that this type of monitoring while it will speed up your time to recovery will not help you find out why the problem happened in the first place.

    Capacity

    Capacity monitoring can be used to be more proactive with the prediction of failures and help with the fault diagnosis of an outage after the event. Typically capacity monitoring would used for items such as, but not limited to hard drives, CPU utilisation and network traffic. Building up trend data over a period of time will help you get familiar with how your environment is performing on a day to day basis. Once your aware of what is the “norm” for your environment, it will become very easy to spot anomalies that could indicate an impending failure. Again this sort of information can be used to inform your users that your aware that X isn’t performing as well as it might and your currently looking into the situation.

    Asset

    Implementing Your Monitoring System

    There is a tendency for people to get carried away with their swanky new monitoring system once its installed and try to start monitoring everything within their environment. This can be disastrous particularly if things are in a bad way. You’ll find you’ll get bombarded by alerts and ultimately start to ignore them. The best way to implement such a system is to pick a couple of the most critical systems, and monitor those. Once their on the system, fix all the issues that are picked up, and then move on to adding more. This approach will make sure that any alerts that you receive can be managed effectively.

    Get an enormous display to show your current status. Airing your dirty laundry can be quite helpful,


    Leave a Reply

     
    What i'm doing!
    • Ordered new laptop, Hoorah, like Darrens laptop except er, its mine 6 days ago
    • Dammit my wordpress theme is broke 3 weeks ago
    • wordpress turbo mode like super persuit mode but with out all the extra wings 2008-07-15
    • More updates...
  • Recent Posts

  • Categories

  • Archives




  •  
    View blog authority

    This site is powered by wordpress
    Wordpress Theme JLBP2 by James Lloyd, based upon the css framework Blueprint