[Noisebridge-discuss] Linux disk stress test & burn-in

[Corey McGuire]

++bonnie++

Bonnie should drain the memory caches on both the system and the drive which
will eliminate the possibility of a soft error that might checksum something
in ram as opposed to the disks.

OR!  You could soft mirror the drives in Linux using mdtools, fill them with
noise, and then compare them.  You need a tribunal for this or better, but
you can have a hundred disk mirror and linux wont care.

Linux soft mirroring should have no added load to your system WITH THE
EXCEPTION of the PCI and hard drive busses.

On Thu, Jul 7, 2011 at 8:50 PM, Ronald Cotoni <setient at gmail.com> wrote:

> +1 to bonnie++.  I was just going to mention that utility :)
>
>
> On Thu, Jul 7, 2011 at 6:48 PM, Seth David Schoen <schoen at loyalty.org>wrote:
>
>> travis+ml-noisebridge at subspacefield.org writes:
>>
>> > Hey all,
>> >
>> > So I have some HDDs I want to pop in a file server, but I want to
>> > stress-test them first to get them over the "bathtub curve".
>> >
>> > I assume I need to write to them over and over, and use SMART to
>> > monitor for problems.
>> >
>> > smartmon + dd = okay, seems decent
>> >
>> > 1) Does anyone know a good tutorial on SMART
>> >
>> > But it'd be better to check for errors... write various patterns to do
>> > this and then check the reads - right?
>> >
>> > 2) Does anyone know a Linux/BSD utility for this?
>> >
>> > Googled around to no avail.
>>
>> I think I want to suggest
>>
>> http://www.coker.com.au/bonnie++/
>>
>> which is meant as a speed tester, but should work the hard drive
>> pretty well.
>>
>> Someone on Launchpad also mentioned to me that you can use
>>
>> udisks --show-info
>>
>> instead of smartmontools to get SMART data from the command line.
>>
>> It seems to me that trying to check for errors is very unlikely to
>> find anything, because hard drives have extensive internal soft
>> error correction.  The probability of an error that gets detected
>> by the drive and reported to SMART must be _much_ higher than the
>> probability of an error where bad data silently reaches the
>> application.  Of course that kind of failure is possible, but it's
>> hard to imagine that you could elicit it in practice within a
>> few dozen hours of testing of a few hard drives!
>>
>> I would think on a modern drive the overwhelming majority of drive
>> errors do get reported through SMART, although I don't know how
>> to quantify "overwhelming majority".
>>
>> --
>> Seth David Schoen <schoen at loyalty.org>      |  No haiku patents
>>     http://www.loyalty.org/~schoen/        |  means I've no incentive to
>>  FD9A6AA28193A9F03D4BF4ADC11B36DC9C7DD150  |        -- Don Marti
>> _______________________________________________
>> Noisebridge-discuss mailing list
>> Noisebridge-discuss at lists.noisebridge.net
>> https://www.noisebridge.net/mailman/listinfo/noisebridge-discuss
>>
>
>
>
> --
> Ronald Cotoni
> Systems Engineer
>
>
> _______________________________________________
> Noisebridge-discuss mailing list
> Noisebridge-discuss at lists.noisebridge.net
> https://www.noisebridge.net/mailman/listinfo/noisebridge-discuss
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.noisebridge.net/pipermail/noisebridge-discuss/attachments/20110707/6b71b40e/attachment-0003.html>