So this is a thing: I was nominated as an outstanding volunteer for PASS for April 2018. I really appreciate that someone thought I was doing something outstanding. I am bad at both accepting compliments and self promotion. I really believe in the PASS as a positive extension of the SQL community. I know that I wouldn’t be where I am today without the generous and helpful people who want to help myself and others do and be their best meeting the data needs of our respective employers/clients/organizations. It’s only natural that you turn around and help those around you that want to do better for themselves as well. I’m excited for the growth of our users group and the success of our SQL Saturday: I could not have done it alone.
My road to tools has been both short and long. Short because I started out (just a few years ago) and didn’t think I needed much more than a connection to the database and my own brain. Long because I also had not yet had the misfortune of discovering how much pain good tooling can eliminate. It started one crisp winter day as bright eyed junior DBA when I was informed I would be supporting an internal deployment that evening. The first that had occurred since I started about 5 months prior. Our internal databases are relatively small and mostly simple so none of this was particularly frightening.
“They’ll give you a deployment script to run on the database. Make sure you take a backup so you can roll back if you need to, but it shouldn’t take more than a few minutes of your evening. Good luck,” they told me. With an update like that what could go wrong? Pretty much most of the things.
The developer provided me with a script. I dutifully took the database backup, ran it and waited for him to confirm the update had applied successfully.
“Somethings wrong. Hang on a sec. I forgot to include some changes in the script.” Within a few minutes I had a second script to run in addition to the first, which I quickly run against the database. A pause.
“No, it’s still not working as expected. Um, let me check some things…” they sent me. This went on for 30 minutes, an hour. I couldn’t make the call to roll it back and cancel the upgrade and the developer was researching what he might have missed in his change script. I asked often if he had anyway to to tell me a total list of objects he expected to have changed for the deployment to work (“Who needs a staging environment?” past me complained). A “simple” deployment that I had expected to maybe spend 30 minutes (including logging in and waiting around for the start time) dragged on to close to 2 hours. Eventually it was completed and we went our separate, frustrated ways for the evening. The next morning I started researching tools for diffing databases and found Redgate’s SQL Compare. Within a week I had approval to purchase their toolbelt and the next time there was a deployment we scheduled in time for them to review the differences between their test environment and production to come up with a deployment script. It took a little longer but I was able to finally convince the development team they needed their own tooling to manage their database code but after about 18 months we had them using Ready Roll to manage database migrations and now a few years later we are finally getting to the point where we can use a build agent like Octopus Deploy so none of us DBAs have to get up at all.
Besides SQL Compare the other tool that I really notice when I don’t have it (absence makes my heart so much more sad) is SQL Prompt. Having to work on a machine without it now makes me feel like an idiot now that I’m used to have easy access to a library of code snippets (and automatic code formatting: tabs forever!).
It’s not essential (yet) but recently I’ve been trying to use PowerShell more when I can and quickly discovered dbatools which does a lot of the heavy lifting and provides immediate access to a lot of management tasks. My favorite database migration: we used to manually copy backups from one server to another and restore them or detach a database and copy the mdf/ldf to the new server and re-attach which was a pain: I can never remember the exact syntax, I end up having to RDP to the server to move files: ugh. But with a single line of powershell I can take a backup (or backups) to a network location, restore that(those) backup(s) without leaving my own desktop and I’ve barely scratched the surface of what I can do with it: highly recommended!
I’ve been taking notes hoping to share some lessons learned from our deployment of Red Gate’s SQL Clone but they kept it so deliciously simply and the documentation and worked examples for PowerShell are difficultto improve upon (also we aren’t doing anything too terrible fancy yet either). My only gotcha was learning that if you don’t use the redgate extension (.SQB) for backups that SQL Clone won’t recognize that it was compressed and encrypted with SQL Backup (we’ve wrapped an existing backup procedure around the SQL Backup proc that preserved our old naming conventions when we switched to the new backup tool). The work around was simple: replace the backup file extension prior to creating the image from the backup. If you think that you could benefit from cloning your databases to replicate (and mask) production data I encourage you to check it out. In the meantime I’ll see if I can find a use case or work around that needs some more documentation as we roll it out for our test group.