This blog post is going to be relatively short and sweet as my expertise in the storage realm is limited. Yesterday I had the pleasure of learning some new and interesting things about that I thought was interesting and would share with you.
For聽 this month鈥檚 meme is hosted by Mike Walsh of (). For my post I don鈥檛 have a solution so much as a nugget of information to along I found interesting. I was speaking to a consultant yesterday about a few things and the topic of his experiences with Oracle DBAs-vs-SQL Server DBAs in terms of his experience with them in regards to storage (SAN) consultations. Clearly this perked my ears up and I asked him to explain. He went on to tell me that in his experience he鈥檚 seen that the Oracle DBAs he鈥檚 come across come across as rather paranoid and never believe anything he tells them despite showing them whitepapers direct from storage vendor on the matter. On this particular matter we were talking about how NetApp has a best-practice recommendation that seems rather contradictory and (rightfully so) the DBAs were skeptical and would continue asking the same question over and over again despite having it already answered鈥ver and over again. What鈥檚 that recommendation you ask? Well in NetApp world they have what are called Aggregates which are聽 nothing more multiple RAID groups. Here鈥檚 the excerpt from Wikipedia about it:
NetApp supports either , , or disk drives, which it groups into (Redundant Array of Inexpensive Disks or Redundant Array of Independent Disks) groups of up to 28 (26 data disks plus 2 parity disks). Multiple RAID groups form an “aggregate”; and within aggregates Data ONTAP operating system sets up “flexible volumes” to actually store data that users can access. An alternative is “Traditional volumes” where one or more RAID groups form a single static volume. Flexible volumes offer the advantage that many of them can be created on a single aggregate and resized at any time. Smaller volumes can then share all of the spindles available to the underlying aggregate. Traditional volumes and aggregates can only be expanded, never contracted. However, Traditional volumes can (theoretically) handle slightly higher I/O throughput than flexible volumes (with the same number of spindles), as they do not have to go through an additional viritualisation layer to talk to the underlying disk.
Ok, so what鈥檚 so different about that? Well that鈥檚 not the part that鈥檚 interesting. What鈥檚 interesting is explicitly states:
For Oracle databases it is recommended that you pool all your disks into a single large aggregate and use FlexVol volumes for your database datafiles and logfiles as described below. This provides the benefit of much simpler administration, particularly for growing and reducing volume sizes without affecting performance. For more details on exact layout recommendations, refer to [2].
Now think about that for a minute. As a SQL Server DBA you鈥檙e probably having a mental breakdown as I did when first slapped with this one as they鈥檙e essentially telling you throw all your eggs in the same basket, its better for you. Well this is where our conversation got interesting as he started breaking down for me exactly how Aggregates worked, how NetApp鈥檚 algorithms function, and WHY this best practice exists and isn鈥檛 as bad as it appears at first glance. Apparently because of the way NetApp鈥檚 Aggregates work the more you expand your Aggregate (read also: add more disks) you鈥檙e actually helping improve performance as you鈥檙e adding more spindles to it and will help performance along. At this point of this post you storage guys are probably ready to tear me a new one as I may or may not be explaining this correctly/accurate to which I re-state, 鈥淚鈥檓 not a storage guy, I鈥檓 a DBA learning something new and attempting to relay this information as best as I understood it.鈥�
Which brings me to the point of my post. As a DBA crazy things like a best practice recommendation that doesn鈥檛 make sense can and will come up in your career. Should you question them? Without a doubt! After all, it鈥檚 your bacon on the line after these guys are gone. The important part however is the learning. Ask questions, realize the differences between technologies and understand the how鈥檚 and whys. In this post I talked about NetApp鈥檚 solution but EMC works differently as well as has different terminology. It may not be your job to be a SAN admin but as a DBA I think its essential to understand all the technologies involved in your configuration and work with those responsible to come up with the best solution that works for you. There are plenty of resources out there to garner knowledge from, they鈥檙e just a quick 鈥�
Follow the hashtag on Twitter to check out everyone’s posts.