I love living dangerously. The clues are in my previous entry from just over a year ago. However 'nearly' running out of fuel is hardly work related, so this entry should hit the nail on the head. I left the office with a laptop, a Dell Lattitude D600. This gave me VPN access into the office, although our IT temp keeps referring to it in full, Virtual Private Network connection, which is a mouthful during the course of several support calls every day. I left my desktop machine on (completely out of character, as I ensure my machine is shut down every evening a few minutes after 6pm like clockwork.) As an office based employee, I have no need for remote access. It is impossible for me to work from home, although if many big companies can move their entire call centre operation to the sub-continent why cannot a helpdesk person work from home? It is a strange dilemma, but I feel on a matter of principle the situation should be thus. In the office I work, when I am at home, it is my time. I completely switch off. Well this evening I digressed from this point purely to get some urgent work done.
Here I was, in my living room, watching Monday's edition of Hollyoaks on E4, while connected to my wireless network, logged into my desktop and effectively sitting at my desk in the office. I smiled at the beauty of everything working, but knew I was apprehensive about the final part of the process. My colleague had been stuck in an all day meeting, so had only been able to briefly give me basic instructions over MSN during his lunchbreak. I understood what needed to be done and had access to the scripts I would need to run. It was now or never.
While I was at University I considered SQL to be predominately a command line based programming language. What did you expect? I had to a SQL trainer on Oracle. You can imagine I was rather surprised by the number of tools and utilities provided by that little known company from Redmond, Washington. While I do not doubt that all developers require fully armed GUI, it was a shock that so little time spent was spent swimming in lines of code. The heart of the system is Enterprise Manager, a console I was introduced to within the first month of joining the company and instantly took me into the realm of 2nd Line Support. Although it is very difficult to manage your SQL Server 2000 instances blind, it is possible to go without. This is the program which could be considered a double edged sword in our fight against the Evil Sresu. With the knowledge and training it can help you fix a numerous amount of problems with a SQL instance, from replication to suspect databases. In the wrong hands, well it could be a disaster.
Here I was, on a Friday evening, with an objective to save myself a bucketload of grief on Monday morning. Or I could cause myself a whole lot of grief by carrying out such an operation outside of office hours. I took the plunge. This was a risk worth taking, for it was only on a QC instance. What is the worse that could happen? Having spent over an hour and a half downloading a backup database from a client server, I had the simple task of disabling replication, which is in my opinion far too simple. Tools --> Disable Publishing. The next job was easy, a simple restore over the existing database. Something I have become accustomed to, since learning the trade some eight months ago. This was the easy part, believe me. For a moment I thought should I stop or should I go ahead and restore replication onto the instance. What was the worse damage I could do? I had the scripts and knew what to expect? Afterall I had seen several developers restore replication in the office. This was different. I was on my own, surfing dangerous waters to coin a phrase.
Well it kind of worked. Replication restored but I could not run any of the snapshot agents. Why? Well I soon discovered the answer, I had not changed one line in the script for a server. So it was trying to write files to a network share that did not exist. Oh no. Although I was concerned about the live server instances, a quick check proved they were all running with no problems. I could resolve this, with some work. Yet I decided against it. The demo database could wait until Monday morning, I was quite positive that one of the developers would easily be able to fix the problem within a few minutes, rather than me wasting hours trying to find fix the issue on a trial and error period. Knowing all too well that an error would mean going right back to the beginning, dropping replication and starting again. I could have done that, but just felt in my hands, I could potentially do more harm than good. The lyrics to a Green Day song come to mind.