Testing the Postgres JDBC driver for compatibility with the Java API…
This was a fun little project… install Postgres, an open source commercial grade RDBMS that I’d never worked with before (although I have extensive Sybase SQL Server DBA experience, and a modest level of exposure to MySQL from a few informal personal projects), get it configured and running, then install the J2SE SDK and the J2EE SDK, get them configured and running, and then get the Sun API Test Suite up and running, which first involves getting the JDBC driver for Postgres properly configured and installed. Not a terribly complex project, actually, since it appears that software packaging and auto-configuration has come a long way since I first logged onto a HP 835 (PA-RISC machine running HPUX) as root back in 1994. Until, of course, you realize that less than half-a-dozen people have actually done what you’re doing before, and that various and sundry minor but crucial steps in the installation and configuration instructions are skipped over in the documentation. Not to mention the whole process of fumbling around with technologies and software that you’ve never dealt with before, combined with the usual aggravating “doh!” type errors that consume significant chunks of time before you figure them out.
Anyway, this has been my life as a sysadmin - go do “this”, and by the way, can you have it done yesterday… “this” being something complex and almost completely new to my experience, involving much trial and error experimentation while I grope around for an understanding of what is going on.
Started the first day I logged in as root on the HP 835 I mentioned above… just me, this humming little beige cube, and a bookcase full of manuals. And of course, I immediately ran into a brick wall… the PPP software that came with the OS version (HPUX 8.x something) didn’t work (we had a 19.2k serial connection hard-wired from our ISP [who was in the same building] into our 120 square foot office; we intended to use to hook up to the Internet); after some investigation, I determined that the PPP implementation supplied with that version of HPUX was flat out broken - this, according to HP themselves, and that the only solution was a wholesale upgrade to HPUX 9.04. So, that was my first major project as a Unix Sysadmin - a complete OS install from scratch. Fun. I never looked back. My personal philosophy: someone actually intended this software to work, and to be installed by other people, god bless them, and so I just have to put myself in their head and figure out what they thought was so totally obvious that it didn’t bear mentioning… at worst case, there is always the option of calling technical support (although usually you’ll know more than they do) or digging around on the Internet to see what brick walls others have run into. :)
I love being a Unix SysAdmin, I couldn’t think of a job more fun that mucking around all day inside the innards of some half-documented piece of open source software trying to track down and document some obscure bug, or sitting on the phone for half an hour trying to convince some yahoo in first-level support that yes I do know what I’m doing and could you please escalate this so that I can talk to someone with a clue? :) Seriously, there is no bigger ego trip than knowing that 99.999% of the people out (if not even fewer) could sit down in front of a computer and do what you can. It is *all* about problem solving, troubleshooting, talent, and an intuitive understanding of how this stuff works, so much that half of it is unconscious, and sometimes you have no clue about what led you to try something, other than a hunch based on some half-remembered analogous problem three years ago. :)
So: Give me a job please. Sr. Unix SysAdmin for hire.