Voice of the DBA

Category: Editorial

  • Guidelines and Requirements

    I saw a post from Brent that Microsoft had changed the default memory guidance. At first glance I read this as they’d changed the default values, which would be interesting. However, this is a guideline, set to 75%. I also saw a few thoughts from Randolph West on LinkedIn, and quite a few comments. The comments were interesting in a few ways.

    It is easy to look at 75% and say that won’t work for this server that’s on my mind right now because I keep getting woken up. That might be true. However, the 75% number isn’t a hard requirement. It’s a guideline, a recommendation to ensure you have enough memory for the OS, but you’re trying to use most for SQL Server. Feel free to adjust it if you feel the need.

    There are certainly people who will also look at that number and then go to a DBA and say, “you’ve set this to 70% (or 85% or whatever) and that’s not what Microsoft says.” Which isn’t true. What the text says is this under the recommended column: “75% of available system memory not consumed by other processes, including other instances. For more detailed recommendations, see max server memory”

    If you go to the “max server memory” section, you see something else. It asks you to monitor before you set this, then do some calculations. Then it says: “This is a generic approximation, and your mileage might vary.”

    That’s a great statement. What they’ve written might not work for you. That’s true. Maybe you have little RAM and some other stuff on your server, so 75% might be way too high. Maybe you have 4TB of RAM, in which case, if you blindly set 75% you should be asked to work elsewhere. Anyone managing systems with 4TB of RAM should know how to monitor, measure, and then choose something different, which might be 85% of RAM.

    While there might be some requirements for managing database systems, there really are a lot of guidelines. You have to make decisions, which means you need some knowledge on which to make good decisions. If you don’t have that knowledge, or are unsure, ask others, ask the GenAI’s, conduct experiments, test things. That’s the job. Learn what you need to make things run better.

    Better being what your clients need, want, and desire.

    Steve Jones

    Listen to the podcast at Libsyn, Spotify, or iTunes.

    Note, podcasts are only available for a limited time online.

  • Concerns over AI Chat Privacy

    One of the major concerns for using GenAI tools is who is reading the data you submit as a prompt, and will this data be used in future training of the model? In other words, could someone using a future model access the data I put in a GenAI chat?

    It’s a valid concern, and not just because of the vendors. There is a lawsuit over the use of data by OpenAI, and a court has ordered all chats to be retained, including deleted ones. Since this is a lawsuit, there is always a chance that some of the data retained gets entered into a court document or even that it might be read aloud in court and captured in a transcript.

    This is a thorny data privacy issue that collides with the need for courts to maintain evidence for legal proceedings. I honestly don’t know what the answer here is, but I think this should be handled similarly to how code and other IP/proprietary items are handled as evidence. The challenge here is that there can be a lot of evidence, and I am not sure many legal organizations are really set up to handle and manage this much data.

    I’m also not sure who I think should pay for this. If two parties engage in a lawsuit, but the data is actually held in custody by a third party. Do we need to have some nominal charge rate for keeping data that some need to prove their case? Is there a need to force companies like OpenAI and others to hold data in chats for a certain period of time? Something like the 7 or 10 years that a lot of governments require for tax records?

    Our legal systems are outdated and ill-equipped to handle many ways in which the digital world differs from the analogue one. On one hand, I wish that lawmakers would work with advocates and technologists to update laws to make them more relevant, or perhaps, only applicable to digital data.

    On the other, I worry about adding new laws that create overhead, perhaps stifle innovation, or will be out of date as soon as they are passed.

    In 2025, we all should know that anything we send across the internet to another person or service could potentially be disclosed and abused. If that really bothers you, then beware of all free services and be cautious of the paid ones. Read the EULAs and decide if you can accept the risk.

    Because once you send it, your data is out of your control.

    Steve Jones

    Listen to the podcast at Libsyn, Spotify, or iTunes.

    Note, podcasts are only available for a limited time online.

  • Republish: How Much Can You Learn?

    It’s Labor Day in the US and I’m likely tackling some chores around the ranch during a stretch at home.

    I grabbed this editorial as many of us are labor. We work most of the week for a paycheck, trying to help our organizations and support our families. I re-read this editorial and thought about how I feel about How Much Can You Learn? I wrote this in 2016, nearly a decade ago.

    Andy and I still talk about this at times, not so much for us, but for people we work with. A surprisingly large number of people struggle to improve themselves. For a variety of reasons:

    • laziness
    • lack of motivation
    • lack of direction
    • busy with life
    • unconcerned about future employment
    • ??

    I don’t know why people wouldn’t be working to improve their skills. I need to rewrite this editorial for the modern, AI age.

  • Data Sovereignty in the Cloud

    I remember the court case years ago when the US government wanted to access data in Azure that was physically stored in Ireland. I wrote lightly about this and linked to the article back in 2020. This has typically been more of a concern for the EU (and other countries) than the US, but I’m sure there are organizations in the US that use the cloud and don’t want their data accessed by other countries’ governments.

    Recently, a Microsoft executive was asked about this in the French Senate. The Microsoft response was that they  (Microsoft) cannot guarantee data sovereignty for French customers. If the US government served a warrant under the Cloud Act, a US corporation would have to turn over the data.

    While the Internet is fantastic in many ways, especially the ability to communicate and collaborate with others all around the world, we still have physical countries and governments. The rules, regulations, and more vary across different countries, but apparently, those rules aren’t going to be enforced if the company doing business is based in the US.

    This might be good for the US, but not for customers in other countries. In my mind, it’s somewhat amazing that many other countries haven’t had organizations build clouds that work inside their borders. I’m especially surprised the EU hasn’t subsidized or assisted a company in growing as a cloud provider, though likely any organization that grew to a significant size was purchased by Amazon, Google, or Microsoft.

    This access apparently hasn’t happened to date, at least not in a way that is disclosed publicly. Proponents point out that this access would only be allowed with some level of evidence of a crime and a probable cause to access the data in question. That might be true, but we know from the past that a lot of government access to information is kept secret from public disclosure.

    The longer I work with data, the more I find that few organizations truly care about data privacy and protection. They give lip service to the idea, and despite employees sometimes wanting to treat data with care, profit (or other) motivations often override other considerations. I’ve mostly given up on worrying about strong data protections and accepted a significant amount of my data is likely being used by more organizations than I’d like.

    Steve Jones

    Listen to the podcast at Libsyn, Spotify, or iTunes.

    Note, podcasts are only available for a limited time online.