[funsec] Former Seagate engineer: Company destroyed evidence

Dan Kaminsky dan at doxpara.com
Mon Jan 4 03:36:29 CST 2010


The funny thing is I remember seeing Seagate drives with Lifetime  
warranty and thinking, Wow!  They must have really gotten their  
reliability shit together!

Nope. Just aping what a company with a good product would do, leaving  
the cleanup to the next guy.  Short term thinking like this is a real,  
REAL problem.



On Jan 3, 2010, at 9:19 PM, "Tomas L. Byrnes" <tomb at byrneit.net> wrote:

> The evil is the deny, divert, annoy.... in response to well documented
> failures.
>
> Who cares how cheap a drive is if it trashes a month of data?
>
> In my experience, over the last 12 months;
>
> 2 WD failures, 3 year old drives that have experienced god-awful heavy
> duty (the primary ThreatSTOP servers) on power cycles. Raid 10 and  
> Raid
> 5 on 3Ware, no data loss.
>
> 13 Seagate failures (more AS than ES, but 2 ES bricked as well, at the
> same time, which is what caused the data loss and corrupted the 1TB  
> RAID
> 5), totally random, all drives less than a year old. Single spindle,
> Raid 0, Raid 10, Raid 5EE; USB/Sata; Sata, Sataraid; Nvidia, Intel,
> Adaptec. Nearly a Terabyte of corrupted data, 50GB (approx) data loss.
>
> 0 Hitachi (have several in multiple machines).
>
> The Seagates all had to go through multiple RMA cycles until I finally
> got 7200.12s, which seem to be stable.
>
> SMART provides no warning, because the problem isn't with the drive or
> the drive controller.
>
> The only indication of impending failure is ever increasing "Aborted
> commands" on the SATA interface, because the problem is Seagate has a
> lousy (don't know if it's cheap hardware or buggy software, and don't
> care which) SATA interface, which has nothing to do with AS or ES, and
> everything to do with bad management.
>
> Seagate: Yet another example of how the race to the bottom strategy
> ruined yet another once-great American Business (GM, Citigroup, GE, US
> Steel, etc.).
>
> Can we PLEASE put pre-teens in charge of US companies instead of  
> Harvard
> MBAs?
>
> http://articles.moneycentral.msn.com/learn-how-to-invest/are-you-a-smart
> er-investor-than-a-5th-grader.aspx
>
>
>
>> -----Original Message-----
>> From: funsec-bounces at linuxbox.org [mailto:funsec- 
>> bounces at linuxbox.org]
>> On Behalf Of Peter Evans
>> Sent: Sunday, January 03, 2010 3:33 PM
>> To: funsec at linuxbox.org
>> Subject: Re: [funsec] Former Seagate engineer: Company destroyed
>> evidence
>>
>> On Sun, Jan 03, 2010 at 02:32:46PM -0800, Tomas L. Byrnes wrote:
>>> I've had a 50% failure rate of 7200.11s in Desktops (ICH10 and
> Nvidia
>>> MCP55 RAID 10) and low use servers (Adaptec RAID 5EE), leading to
>>> serious data loss.
>>
>>    They're all evil, just some are more evil than others.
>>    Hitachi, for example, has a high suicide // drop dead rate.
>>
>>    Strangely enough, I had a spate of weird arse failures with
>> drives
>>    in raid 1 on ICH10 boards. In one case, the cause was actually
>>    reproducible running just the disk test and the root issue was
>> memory.
>>    (burnin test pro from passmark, they fixed it for me too![1])
>>
>>    We have about 500 80/160g drives deployed all over fukuoka,
>>    over the past 1y9m about 5 have failed. (I have them at the
> ofis,
>>    I can check what models if people care.)
>>
>>    The drives are a mix of WD (early 80g boxes going out) and
>>    hitachi/seagate. With probably half being hitachi.
>>
>>    seagate 500g are my current drive of choice for ofis boxes,
>> because
>>    they cost about the same as 10 boxes of cereal, or so.
>>
>>
>>    P
>>
>>
>>
>> [1] release 1017 is entirely my fault. excellent support guys there!
>>
>>
>> _______________________________________________
>> Fun and Misc security discussion for OT posts.
>> https://linuxbox.org/cgi-bin/mailman/listinfo/funsec
>> Note: funsec is a public and open mailing list.
>
> _______________________________________________
> Fun and Misc security discussion for OT posts.
> https://linuxbox.org/cgi-bin/mailman/listinfo/funsec
> Note: funsec is a public and open mailing list.


More information about the funsec mailing list