Do you always burn-in your hard drives?

[Deleted47050]

Just curious to see how many of you always burn-in their new hard drives. I didn't want to clutter the other drive burn-in thread with this question, let's keep that one for the technical instructions.

So, do you always burn-in your new drives?

 

[Spearfoot]

Yes!

 

[nojohnny101]

I have to admit, I do not but I know I should. The next rebuild I do (aka replace all my current 3TBs in the main box or all the current 2TB in the backup box with bigger drives) I will most certainly. It does take longer and delays the satisfaction of getting instantly new expanded storage, but the benefits of ruling out early failures and such I know is worth it.

The earth will not crash if you don't, but it certainly is wise to take every precaution especially if you have data you care about (which is of course backed up I assume). Rant over!

 

[Ericloewe]

Yes, definitely. Just yesterday, I had to replace a disk in my backup server. I had a spare on the shelf, ready to go, which I had burned in together with the rest of the disks when I bought them.

 

[HardChargin]

For systems hosting data that I care about I burn drives in as well.

Another thing to consider regarding burn-in is getting bad/weak drives ID'd ASAP so you can get them exchanged/replaced, preferably with new vs. refurb.

 

[Mr_N]

yep, even the spares so they're ready to go!

 

[Redcoat]

Now I've learned, yes. And tmux is my friend for badblocks.

 

[joeschmuck]

In all honesty, generally I run a single SMART Long test on them. If they pass then I'm good to go. I don't have a spare which is new, all my spares are old but I did test those with badblocks before putting them on a shelf. They are good as an emergency spare until a new drive arrives.

 

[Ericloewe]

In all honesty, generally I run a single SMART Long test on them. If they pass then I'm good to go. I don't have a spare which is new, all my spares are old but I did test those with badblocks before putting them on a shelf. They are good as an emergency spare until a new drive arrives.

*Points finger*
BLASPHEMER!

...

...

To be honest, that's what I typically do with client drives. I guess that's another reason to acquire a third server. I'd use my old WHS 2011, but it doesn't have IPMI.

 

[joeschmuck]

Got a laugh out of me.

 

[Dice]

I do.
Also in the case where 'used drives' are repurposed into the FreeNAS box. I know it is some additional wear and tear, but - remembering the entire point of doing a proper burn in procedure is designed to break/fail hardware before being committed to the system.

 

[Linkman]

Yes, I do now.
Even on my desktop HDDs, as well as MemTest on any new sticks of RAM. Never did those before starting to run FreeNAS, but now I have been bitten by the server bug, by the end of the month, every box I have will have burned in drives, MemTested RAM, and ECC RAM to boot.
(Have to update my sig then too.)

 

[Deleted47050]

Yes, I do now.
Even on my desktop HDDs, as well as MemTest on any new sticks of RAM. Never did those before starting to run FreeNAS, but now I have been bitten by the server bug, by the end of the month, every box I have will have burned in drives, MemTested RAM, and ECC RAM to boot.
(Have to update my sig then too.)

Server bug?

 

[LotLits]

I recently rebuilt my NAS and used the convenient burn-in script from qwertymodo. It took about 3 days to do my 4TB drives but it seemed to work well.

I once didn't do a burn-in and had a drive fail within a month or so. It might have not been caught in the burn-in either but better safe than sorry I suppose.

 

[joeschmuck]

I once didn't do a burn-in and had a drive fail within a month or so. It might have not been caught in the burn-in either but better safe than sorry I suppose.

That is true. While a burn-in test may not catch every early failure, we are all hoping that it catches the majority of them. I have no idea if there are any statistics which state if burning in actually catches many infant mortality issues. I would expect that the manufacturer would have conducted some surface testing before the drive was released from the factory.

 

[Linkman]

Server bug?

Been bitten by the server bug == all of my boxes are now servers. ECC RAM, and at least some sort of remote management.

FreeNAS box is a white box Supermicro MB + Xeon E3 CPU + ECC RAM with IPMI. My secondary box / secondary server is an HPE Proliant ML10 + Sky Lake i3 CPU + ECC RAM with Intel AMT, and my new desktop is a Dell T20 + Xeon E3 + ECC RAM + PCIe sound card, using processor graphics, with Intel AMT.

All the RAM was MemTested for a minimum of 24 hours, and all the drives were burned in with badblocks per the usual recommendations here.

And I still need to update my sig,

 

[SweetAndLow]

Yes and I also take about a month to burn in everything. This includes motherboard, CPU, memory, fans, PSU and HDDs. I run the server and monitor temps the entire time.

This thread also reminds me I don't have a spare drive on the shelf that is burned in. I'll go do that now.

Sent from my Nexus 5X using Tapatalk

 

[Linkman]

I'll add, given @SweetAndLow 's comment above, that I did monitor temps and occasionally the voltages while burning in the FreeNAS box, since I was pulling all those components together. On the other two, since they were OEM / turnkey systems other than drives, I just burned in the RAM and drives, and really didn't bother with temps, etc. since I was within the OEM's hardware recommendations on number of drives, etc. YMMV on that level of risk :)

 

[Deleted47050]

Yes and I also take about a month to burn in everything. This includes motherboard, CPU, memory, fans, PSU and HDDs. I run the server and monitor temps the entire time.

This thread also reminds me I don't have a spare drive on the shelf that is burned in. I'll go do that now.

Sent from my Nexus 5X using Tapatalk

That's some impressive dedication.

 

[ByteTamer]

I would expect that the manufacturer would have conducted some surface testing before the drive was released from the factory.

I ordered (4) 2T HGST 5400 rpm drives for my NAS build. HGST has a very good reputation for reliability. I ordered 2 one month and 2 the following month. I then used SpinRite to test each drive. I had 1 drive fail in the first batch so I returned it for replacement. I also had 1 drive in the second batch fail the burn in test. That was also returned for replacement so I am a little skeptical of the amount of testing the manufacturers do. However, it did take around 12 hours of testing before the failures appeared.

The 4 drives that survived the initial testing now have about 40,000 hours of error free operation in my NAS. I am definitely a fan of burn in testing.