Sounds like I’m bragging, right? There is a free book involved. Or maybe you don’t care because it’s all NoSQLs and stuff and you’re a SQL Server DBA. And that’s where we differ.
When I first heard about NoSQL databases, I had the same reaction that a lot of people are having right now: disbelief and mockery. I remember making fun of MySQL when I first ran into it. It was such an odd database: it didn’t have foreign keys, joins didn’t work, it sometimes ate all of your data, and writes put locks on tables. It was no comparison for SQL Server, Oracle, or PostgreSQL. When I hit publish, this blog post is going to get dumped into a MySQL database on my server. I’ll probably hit up slashdot or read some blogs. That’s going to hit MySQL as well. These days, I try to keep a passing familiarity with MySQL because you never know when you’re going to need to use such a cheap, powerful, tool.
I see a filthy commie on the other side of this intertubes
This is what I think about NoSQL databases; especially Hadoop.
So, I’m going to Hadoop World and I’m going to learn as much as I can about Hadoop and the technology that surrounds it. The fun part is that I’m going to take some notes. And I’m going to share those notes.
That’s all well and good. As Senator Joe McCarthy used to say, “sharing is caring and caring is for commies.”
So, instead of just going and learning about what is interesting to me, I want to know what kind of questions DBAs have about this new-fangled Hadoop thing. I want to find the answers to the questions that people have about Hadoop, its place in the enterprise, and how it may or may not change the DBA’s job.
So, sound off in the comments. If you’ve got a question, type in the box and hit send. I’ll give you props when I post my write up of Hadoop World and I’ll do my best to find the answer.
I owe this entire idea to Andy Warren (blog | twitter). He mentioned that he was curious about Hadoop and that sparked this blog post. Here are his questions:
- How do you install it?
- What login security model(s) does it have
- When does it make sense to use Hadoop vs SQL Server?
- What tools are provided to manage it?
Comments
I guess my biggest question, and not trying to be dum here, but is there really a reason to mess with Hadoop outside of a need for massive scaling or analyzing (via MapReduce) mega-data? I’m as curious about shiny new things as the next guy, but just don’t want to indulge in shiny things for the sake of the shininess. So far the area noSQL ‘shines’ in seems to be scaling and dealing with huge datasets (Terabytes at least).
Have fun, and yes I’m jealous!
Here’s mine: what’s the ‘sitch’ as far as consistency of writes with Hadoop? Is it a system basically to deliver massive amounts of mostly read-only data? Or does it deal with concurrent writes and transactions, locking etc. If so, how?
Sounds interesting. Wish I could attend. If I had the opportunity to attend a Hadoop overview session at a user group, code camp, or SQL Saturday, I would (hint, hint).
“Is Hadoop webscale?”
Seriously, though. I can’t really think of any questions right now… Maybe “When is eventual consistency alright? I can think of web session management, where else should we be considering it?” ? That’s it for serious questions for now. Looking forward to the follow up post regardless.
Trackbacks
One Trackback