linux-raid.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bill Davidsen <davidsen@tmr.com>
To: Matt Garman <matthew.garman@gmail.com>
Cc: linux-raid <linux-raid@vger.kernel.org>
Subject: Re: OT: sas controllers/expanders and "desktop" drives?
Date: Sat, 20 Mar 2010 13:34:50 -0400	[thread overview]
Message-ID: <4BA5073A.1030802@tmr.com> (raw)
In-Reply-To: <20100320164748.GA31221@septictank.raw-sewage.fake>

Matt Garman wrote:
> On Sat, Mar 20, 2010 at 11:34:19AM -0400, Bill Davidsen wrote:
>   
>> Matt Garman wrote:
>>     
>>> For example, from the FAQ for that SuperMicro chassis[3]:
>>>
>>>    Question: We want to use the new Seagate 1.5TB ST31500341AS
>>>    drives in combination with a storage solution based on your
>>>    846TQ chassis, is this compatible?
>>>
>>>    Answer: This ST31500341AS hard drive is designed to be a
>>>    Desktop type hard drive, not an Enterprise hard drive which
>>>    is needed in this setup.  We don't recommend this
>>>    combination, so please use the Seagate 1TB ST31000340NS drive
>>>    instead. This drive is an Enterprise drive and validated for
>>>    our 846TQ chassis.
>>>
>>> In short, we want to build a server with 24 drives.  The desktop
>>> drives are significantly cheaper than the enterprise drives.  But
>>> getting 24 ports is cheapest with these SAS controllers +
>>> expanders.
>>>       
>>  
>> I have the feeling that what they meant is that the drives have
>> firmware intended for desktop use and will not work well in RAID
>> usage. This has been discussed here before, the desktop drives
>> have firmware which tries repeatedly to recover data on a failed
>> read, and which therefore may hang for 20-30 sec before reporting
>> an error or finally returning the data without an error
>> indication.
>>     
>
> That makes sense.  For what it's worth, I found out that the
> expander chip on the SuperMicro 846 case is the LSI SASx36[1].
> (Note that there are several versions of this case; the E1 and E2
> are the ones that have the expander.  The TQ version doesn't have
> the expander chip.)
>
>   
>> Until/unless Neil changes his idea that timeout should be in the
>> driver or the driver, these drives work poorly with RAID.
>>     
>
> For this particular application, my requirements are a little
> different---I actually just want a bunch of individual drives; I
> won't be using RAID.  (But there seems to be a lot of folks on this
> list with a wide range of hardware experience, and I was hoping to
> catch someone with experience with this particular configuration.)
>
> Anyway, we went ahead and ordered the hardware---LSI SAS3081E-R
> card, case with the LSI SASX36 expander, and both WD Green and
> Samsung EcoGreen drives (both desktop class).  I'll be able to
> follow up shortly and report how well this configuration works (or
> doesn't) for standalone usage.
>
> As for RAID usage (md or hardware): just to be clear, my
> understanding is that the biggest problem is that, even if the
> desktop drives work fine for non-RAID use, they'll still be
> "difficult" in an array.  That is, they take too long to report
> errors, and thus get kicked out of the array.
>
> >From my perspective, this isn't a true hardware incompatibility;
> it's a software (or firmware) "mismatch".  If this is correct, then
> the "real" answer to the FAQ I copied above is more nuanced.  The
> current answer should be qualified with "for use in a RAID", and
> then go on to suggest that the drives should be ok for standalone
> use.
>   

See the information provided by Randy Terbush on this, you may be able 
to just change the ERT if you have the right drives.

-- 
Bill Davidsen <davidsen@tmr.com>
  "We can't solve today's problems by using the same thinking we
   used in creating them." - Einstein


  reply	other threads:[~2010-03-20 17:34 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-03-17 14:36 OT: sas controllers/expanders and "desktop" drives? Matt Garman
2010-03-20 15:34 ` Bill Davidsen
2010-03-20 16:47   ` Matt Garman
2010-03-20 17:34     ` Bill Davidsen [this message]
2010-03-20 19:24     ` Richard Scobie
2010-03-21  0:42 ` Simon Matthews
2010-04-03  5:37 ` Thomas Fjellstrom

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4BA5073A.1030802@tmr.com \
    --to=davidsen@tmr.com \
    --cc=linux-raid@vger.kernel.org \
    --cc=matthew.garman@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).