Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 26 Mar 2020 09:14:32 +0000
From:      Arthur Chance <freebsd@qeng-ho.org>
To:        David Christensen <dpchrist@holgerdanske.com>, freebsd-questions@freebsd.org
Subject:   Re: drive selection for disk arrays
Message-ID:  <24b684fe-35c4-5959-b02a-7c3ac1df3477@qeng-ho.org>
In-Reply-To: <713db821-8f69-b41a-75b7-a412a0824c43@holgerdanske.com>
References:  <20200325081814.GK35528@mithril.foucry.net> <713db821-8f69-b41a-75b7-a412a0824c43@holgerdanske.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On 25/03/2020 22:19, David Christensen wrote:
> Jacques Foucry wrote:
> 
>> ALWAYS (when it's possible) buy and use disks from different brand
>> (mix seagate, WD, etc..) in order to avoid same series and same MTBF.
>>
>> I know this to late in this case, but keep this in mind.
>>
>> I know this will not help in this case, please excuse my
>> intervention if it's inappropriate.
> 
> 
> To date, most of my arrays have been composed of similar drives.  But, I
> run a SOHO LAN and have limited experience.  I have been wondering about
> using dissimilar drives to prevent simultaneous common-mode failures.
> 
> 
> Backblaze publishes statistics for individual drives:
> 
> https://www.backblaze.com/blog/hard-drive-stats-q2-2019/
> 
> 
> I would be curious to read any data or reports comparing arrays of
> similar drives vs. arrays of dissimilar drives.
> 
> 
> Have anyone seen a failure involving multiple similar drives all failing
> in the same mode at the same time?

Not at exactly the same time, but the first time I ever built a zraid
array, shortly after zfs was first released, I used a batch of four
identical disks. About 5-6 years later they all failed over a period of
about 2-3 months. There was long enough between failures to buy a
replacement drive on next day delivery and recover, but this was on a
home file server, not a heavily loaded system. Nowadays I buy from a
couple of manufacturers based on Backblaze's latest statistics.

-- 
Violets are red
And roses are blue
When metamaterials
Alter their hue.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?24b684fe-35c4-5959-b02a-7c3ac1df3477>