Being a good Linux systems administrator

Being a good Linux systems administrator

 

Throughout out my journey as a Linux Systems administrator, I have met many interesting sysadmins. I believe after 13 years working as a Sysadmin and another 7 as IT in general, I know what ‘being a good sysadmin’ means and entails.

 

There are few words that come to mind.

·         Responsibility

·         Safety

·         Backup

·         Security

·         Hard working

·         Diligent

·         Self-motivated

 

A good sysadmin takes the system as its own responsibility and takes care of it as if it was their own.

When making a change, always think about the consequences or negative impact a simple change might cause.

Whenever modifying configuration files; always make a back copy before making changes.

Backs up the system and keep a fresh backup up their sleeves.

Always take security seriously and investigates logs, keep the system up-to-date with patches. There is no half-job, does not make their lives easier, the system integrity comes first.

Work, research, investigate, re-visit solutions, double check consultant work. Create procedures, planning is vital. Having a roll back procedure is crucial.

Never consider a job done; there are always things which can be improved.

A good sysadmin needs to spot things before hand, need to be watching and think ahead.

 

With so many requirements, a good sys admin must be self-motivated, to keep fresh, learning new technologies, turning things around quickly and provide good result.

 

I have met many sysadmins, very intelligent in certain areas, but lacking on enthusiasm or just lazy. Some don’t think about the consequences and just run commands without thinking twice.

 

I think it is also important to have creativity to solve problems, keep an attention to the detail, and have focus on the task.

 

I know experience comes with some mistakes, but a good sysadmin needs to be able to see when to avoid mistakes. There are mistakes and mistakes.

 

I think if you hit <ENTER> without double checking the command line and testing the result you are up for trouble.

 

I think a good rule is; test it before you do it on live.

Things are much easier now a day, we have more resources available for testing, and we don’t necessary need to invest loads of money to create a test environment.

With Virtualbox, ESXi and many more, we can easily and quickly test any OS, and create a small network.

There are many simulators out there for almost anything.

Take for instance Cisco one can use www.gns3.net and have loads of virtual cisco kit. There is almost a virtual appliance for any major player on the market. Netapp provides a virtual appliance for its Filer. So their homework is part of the job.

 

There are many different ways of doing the same thing, with that in mind I think it is hard to judge which one is right and which one is wrong.

 

 

By Renato Oliveira