Voice of the DBA

Author: way0utwest

  • Migrating Schema Between Platforms

    PostgreSQL is a popular relational database, one that was designed to be an object relational system, based on standards, and allowing for extensibility. I’ve always thought this was a great system, exceeding SQL Server in some ways, though falling short in others. It’s popular, and certainly used enough so that Azure and Amazon Aurora include PostgreSQL compatability products.

    I ran across an interesting post, where AWS has a Schema Conversion Tool that will help you migrate from Oracle to Aurora running PostgreSQL. I think this makes some sense as there are some capabilities in PostgreSQL that make moving Oracle code easy. AWS also has a Migration Service to move SQL Server to MySQL. I’m not sure how smoothly that goes, but SQL isn’t a terribly complex language and perhaps the conversion isn’t too bad, though I’m sure this isn’t painless.

    I wonder how many people will choose to move their database systems to a new platform. Whether you’re on a commercial database, such as SQL Server or Oracle, or on a free one (MySQL, PostgreSQL), changing the underlying platform is disruptive for many applications. I’m sure there are some software packages that could change a driver and connection string and their code might work, but I bet that’s relatively few systems.

    For most of us, the cost of change is high. New skills need to be learned and new code written. Staff may leave, requiring an organization to hire new employees, which is always a bit of a gamble. The entire process is also a disruption that prevents other work from being completed in a timely manner. Often a large ROI is needed to justify the change, though that’s not always the case as sometimes management decides to change for other reasons, regardless of cost, ignoring any ROI. I have even seen some developers psuh to make the platform change, working many extra hours because of some desire to port to a new platform for dubious reasons.

    I do see plenty of companies looking to move away from Oracle systems, with a very large cost for support and maintenance. While SQL Server isn’t always cheap, it is much less expensive and Microsoft has programs to help port Oracle to SQL Server. Companies may also choose to move to MySQL or PostgreSQL from Oracle, accepting the disruption and costs with a long view towards reducing their yearly outlay for database services.

    I have migrated platforms before, and it’s been an ordeal each time. I haven’t ported code to a new platform, and not sure I would. Perhaps if we were abandoning a codebase and rewriting the application I’d consider it, but in most cases, the cost of staff and time to rework code makes this an expensive proposition. If you have different thoughts or experiences, let us know today.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 4.2MB) podcast or subscribe to the feed at iTunes and Libsyn.

  • DevOps – Downloading a File from the Internet with PoSh

    One of the things we need to do as data professionals is move data files around. Often we’ll get a local path, but in looking for public data sets, I wanted to get a file from the Internet. In this case, a rather large file.

    I could have put the URL in a browser, but the file was slow to load, and while waiting ten minutes or so for the download to complete and doing a “Save As” is quick, it isn’t easily repeatable. Plus, since I needed to get a few files, and might need to do it again, I thought a PoSh download would be better.

    A quick search turned up a few options. I could use System.Net.WebClient, or I could use Invoke-WebRequest. I decided to use the latter because it could use credentials or even parse the file prior to doing some save.

    I just had a simple item, so I used:

    Invoke-WebRequest -Uri “http://labrosa.ee.columbia.edu/millionsong/sites/default/files/AdditionalFiles/unique_tracks.txt” -OutFile “e:\Downloads\unique_tracks.txt”

    I could easily have wrapped this in a function to simulate a copy command, and I may do that at some point. For now, I can drop this in a file, copy/paste, and change a few filenames.

    If this were part of a regular process, such as getting files from a remote web server, I could easily automate this in a task on some server. For now, this is a quick, easy way to get a file from the Internet without a browser.

  • DevOps Can Help

    Amazon had a load balancer failure in 2012. The analysis of the event shows that there were missing data in the devices that caused issues. The restore of data from these devices is complex, way more complex with less mature tools than most database platforms. The result was a nearly 10 hour period of time when some customers were experiencing issues.

    In 2016, Gliffy had three days of downtime from a database error. In this case, an admin was updating a replicated system, but failed to sever a link with the primary node. Forgetting this step caused a data removal on the node, which replicated to the secondary nodes. They discovered the restore and replay of logs would take many days due to the size. They hadn’t practiced a DR situation in some time, and were not prepared for the delays.

    Digital Ocean received alerts earlier in 2017 that some services were not functioning. They traced this down to the primary database being deleted. The issue was a process used the wrong credentials for automated testing, and I’m guessing that part of the testing was removing and rebuilding a database. Five hours across the middle of the night resulted in the main database being restored, and a couple more hours to get replicas caught up.

    In the first two cases, there were issues with the deployment of changes to systems, as well as inadequate backup and restore processes. In both of these cases, I would argue that a good DevOps process would have automated the way the code was deployed, including ensuring that steps weren’t forgotten or predeployment backups captured the state of configuration. DevOps includes the “Ops” changes and should ensure that all state information is captured and stored in a VCS. If this had been done, it’s possible that these companies wouldn’t have had these issues.

    In the last case, certainly whoever sets up a system is responsible for using the correct credentials. While it’s easy to say that a developer or tester shouldn’t know the production credentials, but it’s entirely possible that the person that configured the process would have the credentials. I don’t know what to do here, as the first test of this might cause the issue. Maybe a second set of eyes is important for security changes in automated systems? That certainly could be part of your DevOps process. What I’d like here is two factor authentication for all security setup, including for SQL Server.

    DevOps isn’t a prescriptive set of things that someone does. Whenever I talk with people about DevOps and they give reasons why a particular step I’ve demonstrated won’t work for them, I tell them to stop doing that step. After all, the way you implement DevOps doesn’t have to match what I did. We each need to do what works for our environment, and ensure we have some consistency and repeatability in our process. Hopefully preventing downtime from simple mistakes.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 4.2MB) podcast or subscribe to the feed at iTunes and Libsyn.

  • The Need for DRE

    “Today’s database professionals must be engineers, not administrators”

    That’s a quote from an interview at O’Reilly on Database Reliability Engineering. I don’t want to quibble about whether those of us working in technology are engineers in name or not, but in function, I do think the world is changing to one where we need to do more than be caretakers of systems. We need to be actively involved in ensuring we have very reliable, well-designed systems that use the best practices and patterns of highly available and reliable database systems.

    There is advice to learn and work closer with developers. I like that, and I do think the idea of learning to script and automate tasks is important, perhaps more important all the time, for data professionals. Some of the advice to look towards the quick moving, limited lifetime of many software components today doesn’t make sense. A database, at least some parts of a database, don’t fit within that model. There are valid reasons why a datastore must provide some stability and persistence for an application and won’t fit within a model of multiple systems. Or at least not practically within a model.

    However, there is good advice in the piece to learn more about different types of database services and patterns, and consider other ways of implementing datastores other than a RDBMS. I wouldn’t necessarily abandon a RDBMS for some NoSQL store just because developers think it’s easier, but I would consider whether I actually need to very tight coupling between different parts of the database and complete consistency. I think there are a fair number of domains where a CQRS pattern or some distributed store would work well. The move to microservices might be an enabler for your business if you consider the advantages for OLTP type transactions.

    The flip side of all the power that many stores provide is that many also make reporting and aggregating information more difficult. I would venture that in many cases, a data warehouse (perhaps even a RDBMS-based or columnar store) is necessary, along with the ETL process necessary to keep it up to date. These aren’t simple processes, and take resources to build. For most of us, because we don’t work at extreme scales, I’m not sure it’s worth leaving a relational platform, but I do think that we can learn to evolve and enhance our relational databases faster with DevOps ideas and techniques.

    Steve Jones