Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 20 Jun 1998 19:16:57 -0400 (EDT)
From:      Simon Shapiro <shimon@simon-shapiro.org>
To:        alex@nac.net
Cc:        freebsd-SCSI@FreeBSD.ORG, freebsd-questions@FreeBSD.ORG, Chris Parry <laotzu@juniper.net>
Subject:   RE: DPT support binaries - How to Setup
Message-ID:  <XFMail.980620191657.shimon@simon-shapiro.org>
In-Reply-To: <Pine.BSF.3.96.980619203317.24846M-100000@iago.nac.net>

next in thread | previous in thread | raw e-mail | index | archive | help

On 20-Jun-98 alex@nac.net wrote:
> 
> 
> Just some notes...
> 
> 
>> The DPT controller creates and manages RAID array in a manner totally
>> transparent to the (ANY) operating system.  Say, you have 45 disk
>> drives,
> 
> ANY operating system that has a drive for the SCSI card. Like, remember
> the FreeBSD didn't have a driver for it until recently.

Recently as in last 18 months :-)
The point is that if you install Doze, NiceTry, Unixware, SCSO, whatever,
the RAID array is the very same, compatible, interoprable, and just simply
there, regardless of the O/S.

>>    1   RAID-0 array to be used for swap, /tmp, /var/tmp, /usr/obj, etc.
>>        RAID-0 is very fast, but if any disk in the array fails, the
>>        whole
>>        array will lose its data.  For the indicated use, this is
>>        acceptable
>>        to us (Remember, this is just an example).
>>        We do not need an awful lot of space, but we need the speed of at
>>        least 15MB/Sec, so we will use 6 drives here.
> 
> Is this bright? Couldn't this panic the system in an event of a failure?

No.  It should not.  I use most my on-line systems this way for many years.
In any case, a fatal disk failure will crash your Unix system.  There
really is no real increased exposure here.  Besides, this was an example.
In reality one must choose the engineering solution desired based on
acceptable risks and compromise between benefits/costs/risks.  My point is
that the DPT technology (and other, comparable technologies) give you the
choice, Speed, performance, safety, capacity, etc.

>>    1   Huge RAID-0 array to contain news articles.  Again, we do not
>>    care if
>>        we loose the article.  This array needs to be big and as fast as
>>        possible.  We will use 33 disk drives here.
> 
> Also, I think this is a bad idea; at least make three 11-drive arrays
> (RAID-0 is ok, but this way if you lose *1* disk, you don't lose the
> ENTIRE spool, only a third -- and, you lose no storage space). With news
> servers like breeze, this is very easy.

You have missed my point :-)  This is an absurd example, set out to
demonstrate that it is practical and operational, and that the O/S does not
see disks, but RAID arrays, and that 84 disks appear as five, with the O/S
having no clue what happened.

Besides, for performance this arrangement is mostly silly anyway.


>>    1   Huge RAID-5 array to contain our E-mail.  We need reliability and
>>        capacity, so we will use 33 drives here.  In reality, RAID-5
>>        arrays
>>        are not so effective at this size, but this is just an exadurated
>>        example.
> 
> You are putting news and mail on the same machine?

Sure, why not?  I also put a FreeBSD mirror and a build environment to cut
CDs from. Remember, this is a SILLY ands ABSURD example.

>> What does the DPT do in case of disk failure?  
> 
> Usually kernel panic on bootup, but thats irrelevant.

Absolutely not true.  The DPT card has nothing to do with the panics you
quote.  Go read the code, and trace the panics.  I have.  There is only
one, identifyable combination of failures, which involves the USER doing
something against the clear instructions, that can cause the kernel to
panic in case of a RAID failure.  In this context, we might as well all
pack and go home:  You can always do something destruvtive and then
complain.

When in the army, we did not feel sorry for those who shot themselves in
the foot.  We tended medical services, but a purply heart you do not get
for doing something wrong, or careless.

>> Typically, the user who buys a $3,000.00 disk controller, attaching it
>> to
> 
> The most expensive 3334/UDW is like $1800.

True.  But with 64MB of ECC RAM, differential option, etc, it can retail
for more than $3,000.  Besides, I think you see my point.

Simon


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?XFMail.980620191657.shimon>