How many hours / years before you replace a drive set?

[Jon Schilling]

Hey Shawn,

Run SMART status, (Areca, ATTO & Highpoint cards all have this feature) on each of the 16 drives. If you get 4 or more drives that report read/write errors, time to replace. The enterprise class drives you indicated generally have a MTBF of about 5 years, but a little “oil change/maintenance” every once in awhile certainly doesn’t hurt.

Jonathan Schilling
Vertical Sales Manager
Proavio
12221 Florence Ave.
Santa Fe Springs, CA 90670
Dir: 562-777-3498
Main: 562-777-3488 X106
Fax: 562-777-3499
Email: jon@proavio.com

[Shawn Larkin]

Thanks Jon,

I guess the question then is:

Do you upgrade the drives (after doing some data juggling) OR do you jump on some new technology?

We are using the R380, which is SAS / SATA 3.0. There are newer options now: Thunderbolt, SAS / SATA 6.0, etc.

What do big companies do: recycle and swap drives or just add new stuff?

Any best practices?

Since we don’t really need the speed bump on our network, I thought replacing drives made the most sense.

[Alex Gerulaitis]

[Shawn Larkin] “Since we don’t really need the speed bump on our network, I thought replacing drives made the most sense.”

Totally.

[Shawn Larkin] “What do big companies do: recycle and swap drives or just add new stuff?”

AFAIK both depending on circumstances; your existing tech is still quite fast, viable and compatible with current and future tech releases (SAS 6G, 12G), so there very little point in retiring it.

Further, don’t have to swap the whole drive set just because one drive is close to, or past SMART thresholds: swap just that drive. Drives can work 10+ years without showing much wear.

Alex Gerulaitis
Systems Integrator
DV411 – Los Angeles, CA

[Simon Blackledge]

Were still running R380’s. Great cards.

Built a new server and used a R680 though.

Both our chassis were upgraded to 16Bay 6gig SAS expanders.

The 8bay EnhanceTech running off a R380 in a suite after 4 years still running strong. Hitachi Ultrastars in all chassis.

In 16 bays I tend to run Raid6 with a hot spare.

Like what was suggested look for bad blocks and replace those if worried.

I tend to let things run their course and drive upgrade are normally for data size req ie going from 1TB to 2 or 4TB drives way before drives die.. normally need more storage every 24 months.

It’s easier to let them run their course now though as we can just buy another chassis and expand.

[Shawn Larkin]

This is all great information.

When I runs SMART, I see a ton of figures and columns.

What is the key columns and figures (and how do I interpret them)?

[Jon Schilling]

Shawn,

It should be the column pertaining to “errors” a high number would be bad, while a lower number would be interpreted as “good”.

Some cards will have a listing like this
10/120
meaning 10 errors out of 120 allowable errors.
120 being the threshold set by the manufacturer.

I uploaded a .pdf from ATTO for you here: https://f1.creativecow.net/4514/atto-smart-info

Check this out Shawn: https://www.macupdate.com/app/mac/24875/smart-utility/

Someone that has used that software indicated: “I get an answer I can trust in a few seconds, in contrast to Apple’s Disk Utility that generally warns when it’s to late.”

Jonathan Schilling
Vertical Sales Manager
Proavio Storage by Enhance Technology Inc.
12221 Florence Ave.
Santa Fe Springs, CA 90670
Dir: 562-777-3498
Main: 562-777-3488 X106
Fax: 562-777-3499
Email: jon@proavio.com

[Fred Jodry]

(The answer after the question).

How many hours / years before you replace a drive set? by, Shawn Larkin on Aug 8, 2012 at 2:23:46 pm
Is there a standard for when you swap out say 16 drives in a a rackmount?
We use an ATTO 380 RAID Card and a 2006 MacPro and 16 Hitachi Enterprise Class (Ultrastars) Drives that have been spinning for around 3 years / 26,280 hours now inside of a single enclosure (that sits in a rackmount). All 16 make up a single LUN / Volume.
Any words of wisdom? Thanks in advance.

Two ways: 1. Use the “SMART software” methods above to digitally find data errors or strained, about to crash that everyone has advised above.
2. Contact the hard drives manufacturer to obtain a “connection extender sled” to hook up a hard drive getting tested at the time pushed outboard where you can reach it. Then, the manufacturer can tell you where to solder small brass or copper wire loops onto the circuit card so that an oscilloscope probe can test signal envelopes for actual magnetic playback (after the preamplifier) off the disc, as well as a few critical servo waveforms. This method is the most direct answer. Also make a few pickoff loops for the backplane and the cables from the power supply. (Old hard drives can be recycled, or sometimes rescued, by sending them to Drive Solutions in California).

[Matt Geier]

Hi Shawn,

When it comes down to drives and storage, you’ll find the manufactures will vary in how dependable they are made to run.

Honestly I never like to suggest someone outright replace an investment in any hardware. There’s always going to be something newer and better anyway.

Does it make sense for the budget to add another storage array or swap out all the existing drives just to have all the old ones sitting around doing nothing? I think adding another array is the logical move.

Speaking in terms of what I see for reliability, Seagate drives tend to fail more then WD, and WD drives tend to fail more then Hitachi drives. Of course one never really knows due to all the factors that play a role in the process. My point is that if you have 200 drives vs a place with 20 drives, the 200 drives user will probably see more failures more quickly then the place with 20 drives.

What’s the problem your faced with right now? Are you worried about your drives failing in the near future or are you trying to position budgeting for upgrades to newer tech and bigger drives?

Just curious.

Matt Geier
(Video Networking Solutions Expert)
(Creative Design Workflow Consultant)
(Social Media Networks Consultant)
(Technical Video Industry Sales Consultant)