Expected NAS HDD life 24/7 operation

OverDrive

OverDrive

OverDrive

Associate
Joined
28 Feb 2012
Posts
1,007
Location
Herts
I got my first NAS, a Synology DS215j back in Oct 2015 and installed 2x new 3TB WD Reds.

It's worked beautifully for 2.5 years..until this month!

On the 1st, the monthly disk health check reported a bad sector on Disk 1. I Performed an RMA on the drive and received a replacement yesterday. I installed it and had the NAS repair the RAID 1 configuration - all was well...so I thought.

Then last night, the system reported issues with Disk 2, lost connection to disk several times and then finally completely failing to recognise the disk. I've tried re-seating the drive and also hooking it up to my desktop - the drive spools up but is not detected on NAS or desktop.

This seems crazy - I've just setup a new RMA for Disk 2. I'm wondering if I've done something wrong here - I was very careful when removing and replacing Disk 1.

Anyway, the plan is now to replace Disk 2 and repair the RAID 1 volume and then I hope that will be it.

However, I was expecting these WD Red drives to be good for around 4-5 years of 24/7 use (warranty is 3 years), not fail after 2.5 years! My NAS is well ventilated and dust free and although on 24/7 is not used that heavily - occasional media streaming and file storage.

Also, annoyingly I can't wipe the HDD I am sending back as I did with the first as it is not being detected - will WD wipe the drive somehow?

What are other's experiences?
 
If you buy 2 drives from the same batch and treat them exactly the same as you would in RAID1 then they've experienced the same heat, and the same writes. Chances are they'll die at about the same time.

2.5 years doesn't seem long enough a life span for disks, 10 would be better.

If the disk is recognised if you put it in your PC you could try wiping it there.
 
In my NAS, the usual operating temp of Disk 1 is 27-28c and Disk 2 29-31c. I always assumed the notable difference in operating temp was down to the design of the NAS enclosure. However, I understand that 31c is still perfectly fine for a HDD.

Unfortunately the failed drive is not recognised by my PC either, despite spooling up. Maybe a problem with the circuit board on the HDD.

Anyway, it just seemed like too much of a coincidence that Disk 2 would fail on the same day I replaced Disk 1.. but Disk 2 is now truly dead. I really can't think it was anything I did though as I was very careful.

As you say, aside from the increased operating temp of Disk 2, it would have been subjected to the same read/rights as Disk 1 in RAID1. Also looking at the serial nos, the original pair were from the same batch - same dates too.

This has shaken my confidence in the NAS somewhat, which is a shame as it's been one piece of technology that I have been truly delighted with.
 
Does that NAS enclosure spin down the disks when idle, or do any other power saving magic?

Did a lot of research on this when building a DIY freenas build recently and plenty of stuff I read indicated that power saving and spinning disks down could shorten disk life
 
BongoHunter said:
Does that NAS enclosure spin down the disks when idle, or do any other power saving magic?

Did a lot of research on this when building a DIY freenas build recently and plenty of stuff I read indicated that power saving and spinning disks down could shorten disk life
Click to expand...

Depends on the drive and use-case. A standard desktop drive is usually rated for 8 hours a day x 5 days a week use, and up to 55TB read/write per year. Spinning down is actually good for these drives. NAS drives & most enterprise-level gear, on the other hand, you're absolutely right. They're designed to keep going 24x7 with higher workload ratings, and spinning them down excessively can cause unnecessary wear-and-tear on the drives.

It all comes back to choosing the right drive for the right job :)
 
The default setting in Synology DSM is to spool down the drives when not in use, but I found that they would spool up for one process or another several times a day, so I changed the setting to always on for this very reason - they are NAS drives after all. I changed this a few days after setting up the NAS back in 2015, so should not have been an issue with these disks.

Very good observation though.
 
Remember. RAID isn't a backup.
Click to expand...

Indeed, I have my NAS backed up to the 'cloud'. At the moment my NAS is running with just the one disk until the RMA replacement arrives, then I will rebuild the RAID1 volume and hopefully all will be back to normal.
 
Drives do not last a specific amount of time no matter what quality they are. It's all statistics. Most manufacturing faults work themselves out fairly quickly. Once a drive hits it's first birthday you are likely to see it last some time. But having said that, there is about a 10% chance it will fail every year. So a drive can happily fail after year tweo or it can just go on and on, it's all luck. I have to say that I disagree with the theory that same drives bought at the same time will last about the same. In the past I have found it to be completely random. And in fact I have never seen any evidence to suggest that high quality drives last any longer than lower quality ones. I guess if you buy thousands you might see a difference but for the average person buying a few it's all just random.
It depends on your storage and backup philosophy as to how you manage that failure rate. I also replace drives after about five to six years as well. I have to say that my experience defies the statistics since I have actually had very few failures.
The best way to run drives is 24/7. They don't like being power cycled. It's the same for electronics too - doesn't like being power cycled.
 
I run around a dozen assorted WD drives 24x7 and usually change them out at around 4 to 5 years at which point I often use them as part of my off site backup set for another couple of years. I have had only a couple of failures but I do ensure that the drives don't unload/load the heads too often which was an issue with some WD drives as after 8 seconds of being idle they would unload the heads only for some software to trigger them to reload potentially several times a minute.

This could cause stupidly high counts as seen in SMART data and being over the expected maximums for these within a few years. I have no idea if this would increase the chance of failure but its certainly not ideal so I check all new WD drives are set for around 300 seconds idle timeout (search for WDIDLE3.exe or similar) and make sure no software needlessly keeps waking up the drives if asleep.
 
Also check the RMA drives carefully. I RMAd a WD drive a couple of years ago and received a drive in turn that failed the fitness test. That one went back and the next one failed as well. Ended up buying a new drive that has been fine in a 4 disk array for the last 4 years.
 
I've mostly been running generic Seagate 500GB drives in mine - they spin down during the hours I'm not active and mostly spun up otherwise - seem to hit reallocated sectors starting to happen at 5.7 years on time on the dot but none died totally yet though one randomly stopped being visible for a short while but I think I might have moved it slightly while changing another drive.
 
OverDrive said:
I installed it and had the NAS repair the RAID 1 configuration - all was well...so I thought.
Click to expand...
It may well have been that the rebuild operation put extra strain on the other disk and caused it to fail - not uncommon.

Undesirable said:
I'm not paying another 200 squid for 6tb of backup. Raid will have to do.
Click to expand...
Only you can put a value on your data ;)
 
You must log in or register to reply here.
Back
Top Bottom