Hard Drive failures

[Bozo]

Hard drive paper from Google.

http://216.239.37.132/papers/disk_failures.pdf

Bozo :joker:

 

[LunarMist]

I saw that linked somewhere, too. Interestingly there was no heat factor failure correlation up to 50 °C.

 

[Bozo]

I saw that linked somewhere, too. Interestingly there was no heat factor failure correlation up to 50 °C.

\. ??

Bozo :joker:

 

[LunarMist]

???

 

[timwhit]

http://slashdot.org/

 

[Stereodude]

Too bad they didn't break down the information by brand.

 

[Pradeep]

Going with what's at work (obviously nothing like the quantity that Google has).

We have a bunch of IBM storage arrays for PACS (X-rays and ultrasound).

TotalStorage DS4500:

210 146GB 10K SCSI drives in 15 EXP700 chassis for online storage (used heavily). Might get one failure every 2-3 weeks, this is just an amber light, so often it's a predictive failure and never actually bites the dust before being replaced.

224 250GB 7200rpm SATA drives in 16 EXP100 chassis, for near line storage (used on occasion, always spun up). Never seen a failure in 12 months.

IME the 10K 72/146GB drives tend to die reasonably often, moreso in 1U servers.

The SSA drives in our IBM Shark also seems quite reliable, might get one drive replaced every 3-4 months. The Shark is our SAN with PPRC enabled so data is duplicated in real time to our backup datacenter. The PACS units synchronise with their duplicates during off-peak hours.

 

[LunarMist]

\. ??

Bozo :joker:

It should be backslashdot then.

 

[LunarMist]

Too bad they didn't break down the information by brand.

I doubt that info could be made public without approval from the suppliers. Usually there are some contractual obligations to that effect. Sometimes it is in the Quality Agreement.

 

[Stereodude]

I doubt that info could be made public without approval from the suppliers. Usually there are some contractual obligations to that effect. Sometimes it is in the Quality Agreement.

Dare to dream...