This is just a quick blog entry to draw attention to a freshly-released EMC white paper. I don’t aim to turn my blog into an announcement board for such things, but this one is worth it. When I posted my announcement that I’d left Oracle Server Technologies (Exadata development) to join EMC I should have also made the point of how much interest I have in virtualization.
Virtualization (Done Right) and Oracle Database 11g Direct NFS
My convictions regarding the importance of virtualization in modern data center architecture are quite strong. That’s a significant portion of the motivation behind why I left Oracle to join EMC and one of the reasons I really like this paper. But that’s not all. The paper also centers on Oracle Direct NFS technology. Regular readers know my strong backing of dNFS as well as my long history with the technology. New readers can visit the past on that matter by reading my many posts on the topic.
The following is a short list of items covered in the paper:
- FAST (Fully Automated Storage Tiering). I cannot say enough about the importance of taking care to cache both reads and writes! Imagine that! Not all caching schemes possess that critical attribute! If you try really hard you can probably think of a few dynamic caching solutions that are really good at caching clean data and offer no benefit for writes.
- Oracle Real Application Clusters OLTP performance with the Oracle Database 11g Direct NFS feature with and without FAST (Fully Automated Storage Tiering) on physical servers.
- Oracle Real Application Clusters OLTP performance with the Oracle Database 11g Direct NFS feature with and without FAST (Fully Automated Storage Tiering) on a VMware vShere virtual servers.
- Loads of configuration how-to’s
- Coverage of live migration from physical to virtualized Oracle Real Application Clusters. This is my personal favorite in this paper!
Here is a link to the paper: EMC Performance for Oracle – EMC VNX, Enterprise Flash Drives, FAST Cache, VMware vShere
I wanted folks to get this paper as soon as possible so they’d have something good to read during their flight to Oracle OpenWorld 2011. If you read it you might come up with some difficult questions to pose to the EMC folks in Booth 901 🙂 Go ahead, give it a go. Tell them I said, “Hi” because this will be the first OOW I’ve missed since 1997.
Kevin Closson's Blog: Platforms, Databases and Storage 
“… this will be the first OOW I’ve missed since 1997”
Sorry to hear that – a missed beer opportunity
Doug,
That is indeed a sad thing! You know me…I wouldn’t miss the show unless something very important got in the way…turns out I’ll be “hiking” at high altitude “looking for food.”
Wow! How timely was that paper? Thanks a zillion, Kevin!
Noons,
Do tell!
Currently undergoing the pain of configuring an Oracle DW into a VMAX. Being “spoon fed” totally irrelevant and inaccurate “white papers” by local marketeers disguised as “experts”. That paper is gonna help me a LOT! You guys really need to look into what the heck is going on in Australia, the level of technical support and consultancy we get here from EMC is unbelievably bad… Sorry, don’t want to sound like I’m bashing the company: I like EMC products and recommend them at every instance. This is just a local expertise problem.
Noons,
As always, thanks for stopping by… I’ll see if I can’t figure out to whom your comments should be forwarded…
BTW, any good fish lately?
Narh, haven’t been out much around here. But in East Timor during the last holiday? Go check my album in Facebook, Timor 2011. And hold your chin! 😉
“FAST Cache was enabled for the eight LUNs that contained the database data files.”
Why not to enable FAST Cache on two LUNs that contained the database redo log files?
Lliya,
I wasn’t on the project but I’ll volunteer why I wouldn’t employ FAST for redo: It doesn’t need it and cache technology like flash is best used for data that gets re-visited. Redo I/O is large sequential…that’s perfect for round spinning media.
Consider:
https://kevinclosson.wordpress.com/2007/07/21/manly-men-only-use-solid-state-disk-for-redo-logging-lgwr-io-is-simple-but-not-lgwr-processing/
Lliya taken from the paper!
“In line with EMC recommendations, FAST Cache was enabled for the Oracle data files only. Oracle archive files and redo log files have a predictable workload composed mainly of sequential writes. These can be efficiently handled by the array’s write cache and assigned HDDs. Enabling FAST Cache on these files is neither beneficial nor cost effective.”