Anyway, thats printed in the owners manual, and in some cases forms part of the warranty. I don't recall ever reading an owners manual for a hard drive that states to give it multi passes and subject it to constant read/writes for a week or more before actually using it for data. I can appreciate some people wanting to be all geeky about it and follow the running in practices that were used several years ago, and may still be used in mission critical environments today, thats fine for them. But for storing my copy of Fast and Furious 7 and the holiday photos (that will also be on blu ray, qnap and the cloud), I think I can live without it.
Hey, that's totally your choice. I haven't always tested my hard drives before using them until recently.
I bet you don't always follow *all* of the manufacturer's recommendations either. Ever buy a disk and put it in your computer within 8 hours of getting it home? You probably broke some of their recommendations. If you read all that fine print that comes with a hard drive, they tell you to allow time for the hard drive to reach equilibrium with room temperature to prevent thermal forces from potentially shocking the drive. So every time you move your system from location A to location B, and the two locations have different temps, you should be allowing time for the temperature of all of the metal to warm up/cool down. I've seen (and heard) plenty of people that pick up that Newegg box with a hard drive it it off their porch while snow is on the ground, and 15 minutes later have it running in their desktop. Major no-no.
Feel free to say "I told you so" if I launch a thread in the next few weeks saying my drives died, but the whole point of building redundancy and having backups, is to protect against a drive failure in the first place.
And what purpose would that serve? We're above that kind of behavior. There's also no guarantee that if you burn in a hard drive it won't fail later. As an electrician in the Navy for 8 years, you can test a breaker and verify it works right now, but it is no guarantee that the breaker won't fail to open the next time an overcurrent condition exists. And yes, I've seen breakers tested that later failed, even with 4 times their rated current flowing throw it for several minutes (yes, a fire ensued).
This burn in testing may weed out drives that are going to fail early, but it doesn't give any guarantee that the drive wont ever fail. So the system is built to withstand a drive failing a year from now or a couple hours from now, the process is just the same, lift the dead drive and replace with another one.
Nobody here will ever try to argue that it guarantees a drive won't ever fail. That's just a straw man argument because uncertainty is part of the reality when manufacturing products. Ever heard of the
bathtub curve? If not, read up on it. Notice the very high failure rates during the initial phase of use? THAT is all you are trying to weed out. Plenty of people here have built zpools and within a week or two had enough drives fail to lose their zpool. That kind of scenario is at a MUCH higher risk during the early life of new manufactured parts. Some have theorized that the chances of multiple drive failures at more than 1000x higher than drives that are beyond the typical "infant mortality" stage.
So you are simply hedging off that you've ensured you won't lose your zpool. Nothing more and nothing less.
When a drive fails, do you leave your array with that faulty drive in while you spend a month burning in the spare before using, or do you have a "pre burnt in" spare?
If you read up on burning in disks you'd read that you should always test your spares immediately before putting them on a shelf (or leaving them installed as hot-spares). This ensures that you aren't storing a spare drive that is going to fail shortly after you install it (infant mortality). If you are particularly unlucky, you'll have a disk that fails due to infant mortality AND it will be out of warranty, so you basically paid for a paper weight when you bought the drive
because you didn't test it.
Im not wanting to start arguments or debunk what people are doing when they burn in their drives, I just dont get the logic of building a system that is designed to cope with failure whenever it happens, and then spending a month testing for a failure.
It's not the fact that we're trying to avoid coping with failures. We're trying to avoid having to cope with multiple simultaneous failures resulting in data loss during that initial period where the hard drives can (and do) suffer from much higher than average disk failures.
Ultimately it is totally your choice as to how much time you want to spend burning in your hardware. Some find no value added, others find it very important and consider it part of intial setup of their server. It is your data that potentially at risk, for better or for worse.
When I bought 10x6TB WD Reds last year, just minutes after they went up for sale by Newegg, I got 10 drives. I had zero drives that failed immediately out of the box, and no drives have failed since either. I also pulled them all out of the box and left them to sit overnight before I powered them on for the first time. I also had them spinning for 2-3 weeks (burnin was performed during some of those weeks) because I was just too busy to start migrating the data (most people would argue that if you are so desperate to start moving data as soon as the disks arrived you deserve what you get because you didn't plan ahead better).
