[Tfug] Crappy USB LVM/Software RAID10 performance?

Jordan Aberle jordan.aberle at gmail.com
Sat May 30 15:21:22 MST 2009


I have several clients running either Raid5 or Raid1, Unless you have a real
hardware raid controller it is almost completely unreliable.  I have ran
into these situations more than once with either software raid or hardware
raid (which by the way is still software raid if you are using the built in
controller on the board).

Situation one:
Raid 1 (mirror) (seen this software and hardware level)
One hard drive will fail but the raid is still in state (health).
(Software raid and or / bios level raid.)

Situation two:
Raid 1 (mirror) (software raid, have not seen this issue on hardware raid
yet)
It is replicating but not replicating correctly (data corruption on one hard
drive)

Situation three:
Raid 5 (extremely slow but more reliable than mirror)

My point: Invest in a real hardware raid controller card, do not trust built
in raid at the motherboard level or software raid, yes it can work but I
have seen too many failures to trust it.  I have seen several different
brands of built in raid controllers fail miserably.  Luckily all of the
listed scenarios were fine due to backups, but it could have been a data
loss scenario.

Make sure you have a backup solution of some type, raid by definition is
very reliable I just haven't seen this to be true in the field.

I'm a little bitter when talking about raid excuse me..
-Jordan

**(Software raid: referring to either a raid at the kernel level or a
windows server raid)

On Fri, May 29, 2009 at 10:10 PM, Bowie J. Poag <bpoag at comcast.net> wrote:

>
> Nope. I stopped doing things for free when I got a job. :)
>
>
>
> Jeffry Johnston wrote:
>
>> I trust that you fixed those Wikipedia errors.
>> Jeff
>>
>> On Fri, May 29, 2009 at 4:59 PM, Bowie J. Poag <bpoag at comcast.net> wrote:
>>
>>
>>> Damn straight. Should I ever enter into a disagreement while motoring, I
>>> threaten to reach for my glove compartment. Where I keep my gloves.
>>>
>>>
>>> In Zack's defense, the whole subject of RAID levels is absolutely
>>> peppered
>>> with bad information and bad definitions out there written by people who
>>> don't work directly with the technology, only know the subject
>>> third-handedly, or otherwise don't know what the hell they're talking
>>> about
>>> in general. I can point you to a half-dozen pages that say RAID5 is the
>>> safest RAID level to store your data at---a statement that is
>>> categorically
>>> and empirically false. That doesn't stop it from making the rounds. I'm
>>> looking at the Wikipedia page, and even it gets it wrong in some places.
>>>
>>> (Example: Single disk failures in RAID5 arrays do NOT affect storage
>>> capacity, no more than my car disappears when I get a flat tire. In
>>> RAID5,
>>> The contents of the missing/dead drive are reproduced logically via
>>> parity
>>> reconstruction. The capacity of the array does not change.)
>>>
>>> Cheers,
>>> Bowie
>>>
>>>
>>>
>>
>> _______________________________________________
>> Tucson Free Unix Group - tfug at tfug.org
>> Subscription Options:
>> http://www.tfug.org/mailman/listinfo/tfug_tfug.org
>>
>>
>>
>
>
> _______________________________________________
> Tucson Free Unix Group - tfug at tfug.org
> Subscription Options:
> http://www.tfug.org/mailman/listinfo/tfug_tfug.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://tfug.org/pipermail/tfug_tfug.org/attachments/20090530/e13557fc/attachment-0002.html>


More information about the tfug mailing list