As arrays become more and more dense and as energy costs trend upward, it becomes more imperative to take advantage of such features of Disk Spin Down Technology. As part of FLARE 29 on the EMC Clariion CX4 arrays, the ability to track disk idle time is not only an important feature but a necessity in multi-purpose arrays.

With the introduction of FAST, the dexterity to automatically tier to SATA as part of an archive or storage tiering initiative quickly becomes a win-win. Not only from the perspective of freeing up expensive FC or Flash disk, but also from the viewpoint of spinning down the disks supporting that rarely used archived data. This ultimately leads to reduced energy and cooling costs and summers in the Hamptons if you get my drift..

Now, here is an interesting table addressing DDSD and potential power savings. Of course take these numbers with a grain of salt, as these numbers are driven by application type and drive access. Note: this only takes into account power consumption from a drive perspective not any other componentry within the Clariion itself. Now the amount of power saved is directly harmonious to how much the disks spin in this idle mode. Only the electronics are active, the actual drive motor is not.

-Modes of Operation-

Well there really is only one with DSDT but further explanation is needed. So…the mode I speak of is Standby mode. The requirements for entering and exiting this mode, as well as limitations are quite simple and noted below. The drives can enter this mode when both Storage Processors report in that none of the drives in a RG have been accessed within a 30 minute interval, fixed value here..cannot change. But remember this, and this is important so listen up. Here is where LUN design and RG design are paramount. Senor FLARE only tracks I/O activity at the RG level NOT at the LUN level. Therefore if you have two LUNs within an RG, one is as active as a drunk bunny one is not, then chances are those drives will most likely never enter standby mode and you will be chumming around with egg on your face.  Anyway, how about those requirements for candidacy…

• Drives that are independent or absent from any RAID group are you guessed it, in like flynn
• Drives that are members of a  RAID group but have no LUNs bound on them, indeed.
• Drives that are configured as hot spares after being idle for 30 minutes..you betcha
• Only a “GO” on the CX4, sorry..
• Eligibility of the vault drives for spin down..ummm..bite your tongue
• Any layered app usage a top a RG is a recipe for non-supportedness..
• And, only 1TB SATA II drives under FLARE 29’s tutelage are spin-down qualified <—this will change soon..

-What You Need to Know-

This technology can be managed one of two ways. One, at the storage system level and two, at the RAID group level. What this implies is if enabled at the storage system level, all drives not bound within a RAID group immediately benefit from this feature (assuming they meet the above limits). If enabled at the RAID group level, well do the math, a little more effort pursues but greater efficiencies can be realized.

A number of error checking and sniffing mechanisms run periodically to ensure drive reliability and data integrity as this type of drive access is a bit against the grain. Furthermore, the barrage of testing that disk vendors and EMC themselves do, to me is complete enough to put your trust in it. Now I am glazing over a lot of the details so please feel free to read the hot and heavy here. So in closing, drawing attention to capabilities that you already own as a Clariion/Celerra customer helps you gain greater efficiencies without the added expense, that again is a win-win…

Related articles by Zemanta