vinum vs. DPT smartcacheIV raid

[Greg 'groggy' Lehey]

[Format recovered--see http://www.lemis.com/email/email-format.html]

> X-Mailer: SquirrelMail/1.4.3a

This seems to have difficulty wrapping quotes.

On Wednesday, 16 February 2005 at 10:52:24 -0500, Ean Kingston wrote:
>
>> On Wed, Feb 16, 2005 at 09:58:17AM -0500, Ean Kingston wrote:
>>>
>>>> I have a box with DPT PM2044 SmartCacheIV UW-SCSI PCI cards which
>>>> can do RAID-5 in hardware, but I'd have to use the DOS volume
>>>> manager to set up the array. I have heard reports that vinum
>>>> woudl be faster than using the native card. Is this true?
>>>
>>> Doubtful, though I have heard that there are some rare special
>>> circumstances where software raid can be faster.

Recall that there are no real hardware RAID controllers on the
market.  The difference is whether you have a special processor on the
controller card or not.  To determine which is faster, you need to
compare the hardware on the card and the hardware in the system.
>> The reason I asked is because
>> http://www.shub-internet.org/brad/FreeBSD/vinum.html
>
> I did not know that. Interesting read.

>> suggests vinum can be marginally better than the hardware raid on
>> the smartraid range of cards (which have an even faster processor
>> onboard than the smartcache range). The CPU platform is more or
>> less comparable.  Then again it is with old Fbsd, so I don't know
>> how accurate that is.

I'd guess that the version of FreeBSD is no particular relevance.

> You may have noticed that there were comments about not trusting
> vinum's RAID5 support in that article.

You'll also note that these claims are in no way substantiated.  It's
word of mouth:

> However, I still don't trust RAID-5 under vinum (it has had a long
> and colorful history of surprisingly negative interactions with
> software that it should not, such as "softupdates"),

There is in fact no substantiation whatsoever for this claim.
problems have been reported that suspected either Vinum or soft
updates, and which were frequently related to neither.  We have no
reason to believe that there was ever the kind of problem he's talking
about here.

> and I have not yet had a chance to test vinum under failure mode
> conditions (where at least one disk of a RAID set has failed).

I have.  It works.  There appears to be a bug in reintegrating disks
with a live file system, so it should be unmounted first.

(end of quotation).

> If you are using FreeBSD 5.3, the default is now gvinum (sort of
> second generation of vinum). The gvinum tools don't give you the
> ability to create RAID5 virutal disks so if that is what you want,
> you may not want to go with vinum or gvinum.

I'm not sure what you mean here.  I haven't tried RAID-5 on gvinum,
but it's the first time I've heard that it's not supported.

> Another thing to consider is if you use software RAID and your
> application gets CPU bound, you are going to take a double
> performance hit (both disk and cpu).

One or the other.  It's a tradeoff.

[54 lines extraneous text in original removed]

Greg
--
When replying to this message, please copy the original recipients.
If you don't, I may ignore the reply or reply to the original recipients.
For more information, see http://www.lemis.com/questions.html
See complete headers for address and phone numbers.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-questions/attachments/20050217/b1e34061/attachment.bin