best budget RAID controller for low IOPS VM environment

Posted on 2014-09-01
Last Modified: 2014-09-01

I am trying to determine what RAID controller will get me what I want without throwing money at the issue.  My situation is this, I have several servers that need a RAID controller to be VMWare compatible.  I can make use of these systems as replication or mass storage devices.  Very few active VMs will be used. expected stats are:

-IOPS will rarely exceed 500, never exceed 800.
-RAID5 desired, but RAID 10 is ok
-not mission critical servers
-array 1 will have 8 x 300GB 15k SAS (server 1) - this array may hit 800 IOPS under unusual conditions
-array 2 will have 4 x 500GB 7.2k desktop drives (server 2) - this array will only get server replications that max at 50MB/s
-array 3 will have 8 x 500GB 7.2 desktop drives (server 3) - this array will only get server replications that max at 50MB/s

What I have been experiencing with Adaptec 5805 RAID controllers in RAID5 is that the arrays degrade under high usage (i think).  What i can tell you is that between two servers, each with 8 15k SAS drives, the arrays have degraded no fewer than 20 times (but I do have 8 active VMs on them).  I'm a bit gun shy of even considering Adaptec as they don't even acknowledge their product could have this problem.  

I have been told LSI is the top dog with RAID controllers these days and their reps gave me a model number of 9271-8i or 9266-8i.  Can anyone confirm that this is about the minimum i would need for the 8 x 15k SAS drives?

If any controllers can handle the desktop drives well with VMWare 5.5 that would also be valuable info.

I know this is lengthy, so I'll say in advanced, thank you for taking the time to read and respond to this.
Question by:AdvNetSol
    LVL 46

    Assisted Solution

    What is your budget, and is your backplane SAS-2 (dual-port 6Gbit instead of 3Gbit?)

    Most of the money that separates a budget from a mid-range is in intellectual property and how well they behave when the inevitable happens -- unreadable blocks and failed drive.   The $100 controllers would much rather just destroy your data and they don't have multiple levels of failsafes that keep it safe.   Spend more, and you get battery backed up NVRAM, so a bad block or failed drive doesn't result in partial loss of data.  Spend more and the controller will use idle time to look for and repair bad blocks.  Spend more and you can have a hot spare kick in when the drive is sick, rather than dies (so you have a fallback). Spend more and you have NVRAM that saves metadata for disks on both controller and all the disks so it doesn't lose the configuration after a power surge.

    Spend more and you can do RAID6 without dying performance wise.  Spend more and data goes on the backplane at 6Gbits vs 3Gbits, because you are mixing/matching SAS & SATA.

    Spend the most you can afford on a RAID controller that is in the HCL, not for the performance, but for the safety and availability.

    P.S. you MUST MUST MUST use true enterprise class SATA drives.  No cheap WD consumer drives in the $100 range. The firmware isn't compatible with things like the MegaRAID class.  It puts you at extreme risk of data loss if doing anything other than RAID1.

    Author Comment

    I don't need a battery although i do understand the advantages.
    Don't need a hot spare
    RAID6 is a little overkill, RAID5 or 10 will be fine.  I'm not worried about having complete data loss on these drives, but i do want to avoid having to rebuild them once or twice a month.
    speed (6gb vs 3gb) isn't a huge concern, if it isn't much more i would go that route ($50 or $100) so the card has more use in the future.
    Not sure what HCL is

    I would like to prevent problems after a power surge.
    finding bad blocks in idle time is great, but not critical.

    My budget will be what it needs to be.  i am showing the 9271-8i cards for $500ish which doesn't scare me at all.  If I have to spend $1000 each so be it, i'll spend what i have to spend to avoid rebuilding a RAID array when it can be prevented.  I would consider the maximum to be 1 rebuild ever 6 months.

    regarding desktop SATA drives, i do know the way they report back to a controller is different than enterprise drives in that the timing and logic on how to determine a block is bad is different between the two.  I have learned the hard way as to why you don't use desktop drives for anything that needs to be reliable.

    again, thank you for spending the time to write back.  It sounds like you have a deep understanding of the technologies so I'm looking forward to your response.
    LVL 46

    Accepted Solution

    HCL is the Hardware Compatibility List at VMWARE. You need to choose a controller they have drivers for.  Warning - the cheap crap controllers won't be certified anyway, so the problem may be moot.

    The 9271-8i is overkill.  You can pick up older LSI SAS/SATA controllers with RAID10 support for $50 on ebay (3Gbit limit, which will not make you happy).  They are light on the features, no RAID5 but do the job.  Any LSI 38xx family controller with the -IR type firmware will work.  (The same cards have -IT class firmware which is all JBOD) Just a matter of getting the connectors right and the bus interface.  (Make sure VMWARE supports it, of course)

    The Host Bus adapters don't have hardware RAID5 support, but RAID10 and RAID1 have zero overhead to speak of so you don't need it.  

    Look here

    $99  buy it now, 6Gbit SAS/SATA support, and the IR [integrated RAID] firmware.
    Can't go wrong with this.

    Write Comment

    Please enter a first name

    Please enter a last name

    We will never share this with anyone.

    Featured Post

    How your wiki can always stay up-to-date

    Quip doubles as a “living” wiki and a project management tool that evolves with your organization. As you finish projects in Quip, the work remains, easily accessible to all team members, new and old.
    - Increase transparency
    - Onboard new hires faster
    - Access from mobile/offline

    Usually shares are where we want them for our users and we tend to take them for granted. There are times, however, when those shares may disappear causing difficulty for your users. One of the first things to try is searching for files that shou…
    Hyper-convergence systems have taken the IT world by storm and have quickly started to change our point of view of how the data center should and could be architected. In this article, I’ll explain the benefits of employing a hyper-converged system …
    This video discusses moving either the default database or any database to a new volume.
    Get a first impression of how PRTG looks and learn how it works.   This video is a short introduction to PRTG, as an initial overview or as a quick start for new PRTG users.

    759 members asked questions and received personalized solutions in the past 7 days.

    Join the community of 500,000 technology professionals and ask your questions.

    Join & Ask a Question

    Need Help in Real-Time?

    Connect with top rated Experts

    12 Experts available now in Live!

    Get 1:1 Help Now