Voice of the DBA

Tag: Database Weekly

  • Lots of data in RAM

    I remember buying my first computer after college. At the time I was working part time and in graduate studies part time. My aging high school era Apple II wasn’t capable of handling the load, not to mention its 300baud modem wasn’t a lot of fun. I could likely have gotten a faster modem with a serial connection, but most of the modern Internet providers wanted Windows or a Macintosh OS.

    I decided to just spend the money and get a computer. Being a bit of a geek, I wanted to customize things a bit and get the most bang for my buck. At that time, I paged through Computer Shopper, looking for deals on parts and trying to find the cheapest 486 CPU I could. I made various choices, one of which was the RAM. I remember thinking hard at that time about whether I could get by with 2 1MB modules or did I need to go ahead and max out the motherboard and get 4. These days I think about GB, not MB, and 2 or 4 isn’t a good number.

    One of the first servers I built at one job was with 8MB chips, getting 32MB into a machine, which felt like a lot. Certainly more than I’d ever had in a computer before. I also remember helping write a PO for a Netware server that cost US$250,000, one with 256MB of RAM. It was the size of a washing machine and filled with 32MB HDDs, making it by far the largest machine I’d seen to date.

    That’s a nostalgic view back to 1992, when I was working on my CNE certification. I always think about that large, very expensive machine when I see some leap of hardware. This past week, it was a DD4 Gen-Z memory module that has 256GB of RAM in a stick. That’s not only a larger scalar value than the RAM sticks I started working with, but it’s an order of magnitude more memory. Imagine 4 of these sticks giving you a terabyte of RAM in a desktop. Who among us wouldn’t want this for our desktop SQL Server development?

    This level of tech isn’t ready for most of us, but it’s not that far off. 32GB laptops aren’t common, but they are easy to find. You can even get 64GB of RAM now if you want to. How long before we see 128GB+ desktop and laptop machines? I have no idea, as we seem to be in a bit of a hardware lull. Most developers run 8GB or 16GB, and it seems like this has been the standard for quite a few years.

    I don’t know if we’ll start to see some vendors pushing to add more RAM. Certainly the cloud has changed things with lots of processing and storage occurring off the local system. At this point I’d expect that three years after buying my last laptop, I’d be able to get 32GB in the same form factor for the price I paid a few years ago, but that’s not the case. The machine supports 32GB, but it’s pricey, the same price it was 3 years ago.

    I wonder how many of you are still power hungry, wanting more powerful machines. Or has the level of performance you get from your i7 on Windows or OSX been good enough the last few years? I think this machine works great for me, whether running VMs or containers. 16GB does the job well, and I’d look to get the same level of hardware when I replace this laptop.

    Steve Jones

  • Changing Hardware

    I’ve been an Intel user for most of my life. While I have owned Mac machines, and I did once have an AMD desktop for the kids, I have pretty much stuck with Intel as my CPU of choice. Certainly I’ve had other processors in mobile devices, but for work, I have almost always purchased servers, desktops, or laptops with Intel CPUs.

    I remember the time in the 80s and 90s when there were lawsuits, advertisements, and innuendo put out by the companies, as well as fans of either side. I knew people that were sure Intel was the only way to go, fervently claiming that I’d have all sorts of issues trying to run my software on anything other than an Intel CPU. Just as insistent where others that assured me AMD worked fine and Intel was an overpriced waste of money.

    Over the years, I’ve seen Glenn Berry note that SQL Server runs better on Intel, because of the superior single-threaded performance on Intel machines. He’s changed his mind recently, and talks about why in a post this week. Glenn does advocate moving to new hardware, which often can give you better performance over older hardware at a lower consumption. If you are stuck on an older 2008/2012/2014 system, perhaps a hardware refresh is worth considering.

    Would any of you consider moving to AMD CPUs for your SQL Server instances? Is a hardware refresh something you might consider instead of upgrading software versions? If you got a 30-40% performance increase from a modern processor over a 2012/2013 era one, would that be worth the cost of changing hardware?

    We know that changing hardware isn’t as simple or easy as replacing one part with another. Likely you’d have a fair amount of labor to change out a bit of server hardware, even if you didn’t have to change any of your storage. With the growth and advantages of the cloud, is this something to even consider these days?

    I don’t know, though I tend to think that for most organizations, it’s not worth the time and hassle to update hardware without updating software. Perhaps you feel differently, and if so, tell us why.

    Some of you might be like me, not even really thinking about what hardware choices are out there. You let someone else worry about the physical machines and you just ask for a certain level of RAM, a number of cores, and a certain amount of storage performance. That’s how I’ve built SQL Servers in the past, which is really what I do now in the cloud.

    Steve Jones

  • More Linux than Windows

    Not for me. I definitely have more Windows running than Linux. Well, at least if you count minutes. If we’re counting instances of kernels, there might be times I have more running with Linux kernels in containers. However, on most days I have 1 or 2 Linux systems running and 3-4 Windows ones.
    That’s not the case in Azure, where the majority of instances are on Linux, not Windows. Microsoft announced this recently, noting that the use of Linux is growing for them. If you follow their work in the development area, this isn’t surprising. They are doing more and more OSS integration with their products or with other work. We have VS Code and Azure Data Studio, as well as .NET Core, running on Linux.
    We’ve had PowerShell on the Windows platform for over a decade. In that time, it has improved and grown, and become a default way for many people to manage their systems at scale, or certainly, in a consistent way. I’ve been surprised with how pervasive PowerShell has become. That’s why I’ve been surprised in the last year at //build/ and other conferences when I see Microsoft employees promoting bash as the new thing.
    I have nothing against bash, though I was more a korn shell person in University. I do like working in it at times, and I find myself doing that more and more. What I don’t like is a push away from PowerShell, especially as I’ve just started to get comfortable with -eq and -gt. I hope that as Microsoft moves towards more Linux versions of tools that they continue to keep PoSh as a first class citizen, on all platforms.
    Linux lost the desktop war. I’m pretty sure that’s the case, as I haven’t seen a Linux laptop or desktop in years. I see lots of MacOS, but more Windows, especially with the amazing hardware advances manufacturers have made in the last 4 years with the 2-in-1 format. However, I start to wonder if Linux will win the server war, or at least, become a much larger percentage of the back ends in the next ten years. I think it might, handling .NET and SQL Server loads in many cases.
    Steve Jones

  • Microsoft, Think DevOps First

    This post from Melissa Coates is a good example of not thinking through your product architecture early on. The short version of the post is that Power Bi Desktop is a better place, perhaps really the only place, to author your Power BI Reports. Melissa covers a few potential issues while using the Power BI online editor, but the big item is that you cannot easily prevent conflicts and track versions there.

    Power BI started as the online service, with the Power BI Desktop tool seeming like a bit of an afterthought. At least from the marketing perspective as all of the early demos and media from Microsoft was about the online report service. Since then it’s evolved to become the primary way we will build reports in SQL Server moving forward, which I think is a good move.

    While I can understand the developers at Microsoft not really thinking that the service would on-premises and maybe not considering the need to provide some text format for the reports, I can’t understand why they didn’t learn from the Integration Services team and realize that binary versions of programmable items don’t make sense. We need a format that can be easily versioned, and maybe more importantly, stored in a VCS and diff’d in a way humans can understand.

    When building a format for storing code, please consider the need to work in a team and version the changes made. This means any format should not be a) binary and b) difficult to decode. Separate visual elements from logical elements and ensure a text version of this can be examined by developers. You can use XML, JSON, YAML, or any other text based format, but choose something that makes sense. Even if you add your own extension, ensure that standard tools that work with code can use this.

    I do know the PBIX format is a ZIP file, but zip files don’t easily integrate with a VCS. We could use hooks to extract/rebuild out files on commit/checkout, but that’s cumbersome and silly. I’d rather that the PBIX were a folder with the files inside. Users, including my Mom, can zip a folder and email that if needed. To me, that would have been a better structure from the beginning.

    Microsoft is supposed to be a company providing platforms that we build upon and use in our work. The decisions for the Power BI service seem to be poorly thought through with that in mind. I’d urge them to create a baseline set of rules for future products that consider DevOps, teams, and the need to track code.

    Steve Jones