Voice of the DBA

Category: Editorial

  • Backups Over Time

    I’ve written about backups at various times, including responsibility, problems with systems, and home protection. This is a subject that is important to me, as I consider the responsibility for protecting data to be first and foremost. Security, performance and more all matter, but backup and restores are the most important items to pay attention to.
    As data professionals, whether looking at organizational systems or our personal data, we will find that our backup strategies change over time. We may also find that systems come and go, and it can be easy to forget about older systems. I know I’ve had to track down tapes and restore decommissioned systems years after they were reformatted or powered off.
    I ran across an interesting post from someone that went through and found their old backup media and moved it all to newer media, as well as cloud storage. While I’m not sure that I really want to go through old hard drives and keep old code or data, it’s an interesting exercise to think about.
    Do you worry about losing data from old backups? This probably applies more to home systems than organizations, but perhaps you have regulatory requirements to keep seven (or more) years of backups around. Maybe you want to be sure that your old code, projects, pictures, and more are saved elsewhere. Maybe you even want to ensure that newer formats of media are being used.
    What techniques might you use to accomplish this? I know I have a Windows Home Server that receives copies of pictures/music/video and a Crashplan account that backs up various machines in the house. That seems to work well, though I do consider taking those pictures/video and putting them on DVDs for long term storage out of the house. I’m hoping that .jpg and other formats don’t go out of popularity anytime soon.
    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 2.8MB) podcast or subscribe to the feed at iTunes and LibSyn.

  • Guess the RTM Build

    I’m going to take a page from Joey D’Antoni and start a contest. We’ve gotten quite a few releases of SQL Server 2016, including now 4 Release Candidates (RC). RC3 was released recently, and noted as being feature complete. I am assuming this might is the last RC, though certainly any big bugs might result in an RC4 (or 5).

    This week, for a bit of fun, I’m going to start a pool for the final RTM Release number. If you post in the comment thread for this editorial, and guess correctly, I’ll send you a prize. Likely this will be something from Amazon, though I haven’t picked anything yet. Suggestions are welcome, though I am thinking some new SSD. We all need more fast storage.

    I’ll use the Price is Right rules, so the closest person, without going over. Ties will result in the first person posting winning, based on thread order. If the final build is 13.00.9500.12, and you post 13.00.9500.13, you’re out of the running.

    The RCs have been almost 100 builds apart. However there’s no guarantee that the RTM build will be follow any pattern, though it seems like previous releases have ended on an easy to remember build number like 12.00.2000, 11.00.2100, 10.0.1600, etc.

    Take your guess, and if you’re right, I’ll contact you with a private message to redeem your prize. Contest ends when SQL Server 2016 RTMs.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 2.4MB) podcast or subscribe to the feed at iTunes and LibSyn.

  • Crack that Encrypted Data

    Ransomware appears to be gaining some traction as a new trend. For awhile in my career, it was virus programs designed to send to all your contacts. Then it was infections to use your computer as part of a bot net. Now it’s encrypting your files and demanding a payment to get the password.

    I’m starting to think that a)I need to ensure I have solid, better backups on all devices, and b) I should pay attention and be aware of decryption programs. I’d love to say that I could build a decryption program, like someone did, but as much as I’m interesting in encryption and study it, that’s a little out of my skillset wheelhouse.

    I’m actually starting to think that this might be a way that people in communities, like the SQL Server community, can help each other. We can be aware of potential ransomware threats, like the one that hit this hospital, and potentially share ways to recover from the incident, or even decrypt the drives. In fact, I suspect it might be worth keeping a system handy to practice decryption techniques, if you can determine the attack vector.

    I’m sure many organizations wouldn’t want to share details of attacks and infections, but this is exactly the type of information that we, as data professionals, should be sharing. It’s incredibly difficult to keep up with all the threats and attacks, not to mention the techniques to recover systems. I’d urge all of you to ask your employers if you can at least help others, even if you can’t disclose how or where you gained the knowledge. If nothing else, the information needs to be shared more publicly to allow us to better protect our systems and be effective custodians of data.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 2.2MB) podcast or subscribe to the feed at iTunes and LibSyn.

  • The Hidden Trigger

    I ran across a post recently where someone asked about using a trigger for a particular issue. There were quite a few responses, but one person brought up the fact that triggers have various downsides, all of which I agreed with. However I didn’t see one mentioned, which is a big one with me.

    Triggers are hidden.

    It’s not that triggers are hidden from DBAs by default, after all, there’s a triggers folder under each table in the Object Explorer, but how often do you open that folder? It’s just not obvious when a trigger exists on a table. The use of triggers is too uncommon, and it’s easy to forget they exist, even if you’re the person that added them to a database.

    This is really my one big complaint about triggers. I can’t easily determine when a trigger is firing, which sometimes leads to issues. I’ve troubleshot strange things in SQL Servers, only to realize hours later that a trigger is firing based on some action.

    Perhaps it’s me, but do many of you think about triggers first when something strange happens in SQL Server? I think I’ve gotten better over the years at looking for them, but I still get stymied and waste time trying to debug issues. I really wish that a “triggers” tab would appear in SSMS, maybe after the Execution Plan tab, whenever a trigger fired. It could have a copy of inserted and deleted, or even just the trigger code. That alone might clue me in quicker.

    I think triggers can be valuable in database systems, but because they seem to be so rarely used, I’ve grown to distrust and dislike them over the years.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 1.9MB) podcast or subscribe to the feed at iTunes and LibSyn.