This is Richard Siemers's Typepad Profile.
Join Typepad and start following Richard Siemers's activity
Join Now!
Already a member? Sign In
Richard Siemers
Dallas/Ft Worth
Storage Administrator
Recent Activity
From what I have seen of the T800s, disk shelves are only connected to a single node pair, meaning each disk ia accessible by only 2 of the 4 controllers. Your disk shelves end up being evenly divided between your node pairs just like in your "2 controller split between 2 array" picture. So in essence, a 4 Node 3Par is a federation of two dual-controller systems in one rack, is it not?
1 reply
Richard Siemers is now following the storage anarchist
May 8, 2010
@Nigel @John F. I agree, zero reclamation by itself should have a "your mileage may vary" disclaimer. Migrations done via file system level copies will be thin by default, but migrations done with disk imaging or block by block mirroring may have the most to gain from reclamation, assuming the source blocks were zeros that got allocated on copy. I like John's point about zeroing out non-zero free space... secure erase tools come to mind as a cheap/free was to overwrite the "free space" of a lun with zeros prior to doing a zero reclamation. I also ponder the actual contents of those 700 gig sql MDF files that are half empty... is the white space truly all zeros? Netapp has an interesting solution to that by de-duping the white space.
1 reply
The discussion on page size seemed to leave out some related structural elements like pool size increments for growth, the 3Par pool (CPG) grows at 16 to 32 GB per "extent", while the luns inside the pool grow at much smaller increments (i heard 16k). So HDS beat 3Par to the punch on "zero reclamation"?? What we really want is "filesystem free space" detection/reclamation. Netapp can already do this with a host based agent.
1 reply
Services revenue? 3Par sells professional services? I just installed a T800 with 140 TB to replace 6 soon-to-be EOSL clariions and 2 DS4500s. I never saw any room for professional services. nor was any even pitched during the pre or post sales engagements. The system went from crates to production ready for hosts in 5 hours. I've been an EMC DMX/Clariion admin for 9+ years, and have worked on several hardware replacement projects. The amount of discussion required for BIN file creation can be 5 hours by itself... actual bin file creation is another task all together.
1 reply
I'm not saying thats my definition of Enterprise Storage... I'm an open systems guy. I'm just trying to look at the question and Tony's answer with an open mind and provided my interpretation. If you ask a career mainframe guy the same question, "What does Enterprise Storage mean?"... pretty sure they will require Ficon and and ESCON... This question and answer are subject to perception and interpretation.
Toggle Commented Jun 3, 2009 on Enterprise Storage? at Storagebod's Blog
^^^ As if to imply "Enterprise Storage" is Ficon and Escon.
Toggle Commented Jun 3, 2009 on Enterprise Storage? at Storagebod's Blog
Perhaps "External Storage Virtualization that is Enterprise Storage" means that it's the only virtualization product that uses external (3rd party external to the USP) storage and presents it up to a mainframe with the performance and reliability of typical mainframe Enterprise Storage.
Toggle Commented Jun 3, 2009 on Enterprise Storage? at Storagebod's Blog
@Virtual Provisioning Common storage pools: Really? I can get this on my DMX2000 today? I guess I'm over due for a talk with my SE/CE. Re: the clariion and thin provisioning, I was turned off by the imposed limitations of how many spindles could be allocated to thin provisioning. It seemed purposely gimped to discourage use imho.
1 reply
The link referencing the EMC blog is broken. Your article hits home as we are running 6 clariions in production, and presently evaluating a single 3par T800 to replace them. I have a folder full of excel spreadsheets that confirm your points re: the clariion metalun architecture. However, I feel you have over simplified the steps to manage the InServ, especially over time as space gets used up in the system (growth warnings, allocation alerts, capacity limits, all that mojo that protects an over subscribed storage system from the things we're paranoid about). Clariion thin provisioning doesn't use the old school metalun/raid group architecture for its thin disk pools if I recall correctly. It would be useful to see a comparison of 3par's Thin Provisioning architecture versus the Clariion's and Netapp's implementation of thin provisioning... if your looking for content ideas. Please, no more washers on a peg board lol.
1 reply