• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

NAS drives stress testing

Liver

Supreme [H]ardness
Joined
Oct 24, 2005
Messages
6,300
I've decided on replacing my NAS, and I am finally getting around to it. For my needs a Synology 1522+ will do just fine.

I haven't decided on the drives yet, but that isn't (necessarily) my question. Once I decide on the drives, is there a way I can briefly stress test them to weed out any early failures? Is that even possible?

Basically, what tests would you run on a new (or new to you) drive before putting data on it?

I only have a NAS and MBP and an older Windows laptop.
 
When I bought my 4x16TB WD Golds for my DS920+ I just threw each one in an external USB 3.0 HD enclosure and ran Short, Long and Extended SMART tests in HDSentinel noting their SMART Data out of the box and after all the tests.
 
I have been looking at OSX alternatives, and DriveDX seems to be good. Any other recommendations?

I have used terminal commands before, but it’s not something I am comfortable with. Its a copy and paste thing for me.
 
There's no good way to stress test a hard drive. Just make sure your data is backup and use any of the raid levels that can tolerate at least a drive failure (RAID 1 / 10 / 5 / 6) and call it a day. I've build at least 5-6 enterprise NAS within the last 3-4 years (12 to 24 drives, RAID 6 or 10) and never feel the need to stress test them.
 
There's no good way to stress test a hard drive. Just make sure your data is backup and use any of the raid levels that can tolerate at least a drive failure (RAID 1 / 10 / 5 / 6) and call it a day. I've build at least 5-6 enterprise NAS within the last 3-4 years (12 to 24 drives, RAID 6 or 10) and never feel the need to stress test them.
+1. Failure will occur. Have a strategy.

For me, I classify my data in various tiers based on criticality. I use cloud plus backup for my most critical, for example. I also use my old NAS for "quasi-critical" stuff. And then I just have stuff that sits on RAID and I'm fine if it goes - but I have hard drives on hand for rapid restoration.
 
I’m replacing my existing NAS so I can keep the old NAS as another back up. I can put the stuff I really can’t afford to lose on it (have it on both of the servers).

I guess I don’t need to stress test the drives, based on the opinions here. Thank you.

Looking at different raid set ups, which would be the best for data safety as well as maximizing overall pooled drive size? I’m not that concerned with speed (within reason).

I’m getting a Synology 1821+ with 8 bays. I planned on starting with 4 16Tb drives, but I can do 6 drives if needed to get a different raid set up.

I’m researching raid 6 vs 10.

Thanks.
 
I’m replacing my existing NAS so I can keep the old NAS as another back up. I can put the stuff I really can’t afford to lose on it (have it on both of the servers).

I guess I don’t need to stress test the drives, based on the opinions here. Thank you.

Looking at different raid set ups, which would be the best for data safety as well as maximizing overall pooled drive size? I’m not that concerned with speed (within reason).

I’m getting a Synology 1821+ with 8 bays. I planned on starting with 4 16Tb drives, but I can do 6 drives if needed to get a different raid set up.

I’m researching raid 6 vs 10.

Thanks.
I’m a big fan of RAID6. Research the pros and cons and see if it’s right for you. I’m a Synology guy, too. Love them!

Go 6 drives and RAID6 - best for prosumer, IMO. In case of vacation, etc - plus drives are so big rebuilds take a long time. That’s the riskiest time for your array.
 
Last edited:
If it is for home use and uptime is not important... You might try SHR (Synology Hybrid Raid) with 1 drive redundancy (basically a RAID 5). It will only allow 1 drive failure and the write performance will be terrible but should not be slower than 100 MB/s and you only waste 1 drive's worth of space on parity. Personally, I'd go with 6 drives and just do RAID 6.
 
If it is for home use and uptime is not important... You might try SHR (Synology Hybrid Raid) with 1 drive redundancy (basically a RAID 5). It will only allow 1 drive failure and the write performance will be terrible but should not be slower than 100 MB/s and you only waste 1 drive's worth of space on parity. Personally, I'd go with 6 drives and just do RAID 6.
SHR is cool too as you can use all kinds of drives. It’s like hobo RAID, lol.
 
I was sorta in this predicament when doing my 4x16TB in Synology DS920+ (4-bay). I originally went with single 32TB SHR-2 array, but then decided (and after many questions to SamirD ...thanks for all the help!) that was too risky if I had a drive go bad as a rebuild might kill another drive. So for my use (music, movies, pics, files, PC backups, etc) I wiped and started over with 2x 16TB RAID1 arrays instead. Who knows if I'll regret not staying with SHR-2 lol.
 
I was sorta in this predicament when doing my 4x16TB in Synology DS920+ (4-bay). I originally went with single 32TB SHR-2 array, but then decided (and after many questions to SamirD ...thanks for all the help!) that was too risky if I had a drive go bad as a rebuild might kill another drive. So for my use (music, movies, pics, files, PC backups, etc) I wiped and started over with 2x 16TB RAID1 arrays instead. Who knows if I'll regret not staying with SHR-2 lol.
Solid choice. The redundancy is worth it. It's also why I buy 8 bays. Allows you to balance redundancy with capacity. It costs $$$ but they've lasted me years and years.

Built this guy about a year ago (DS1821+):

Screenshot 2023-07-07 at 8.55.48 PM.png
 
That’s exactly what I’ll be doing.

Syn 1821+ with 16 Tb drives.

Currently I have a Syn with 20Tb. 15Tb useable and I’ve used 13Tb of it. I’ve had it for almost 10 years. As you can see, I’m not a power data user.

After doing a bunch of research I had decided on a 1621+ because of the imbedded Ryzen 1500b chip instead of the Ryzen 1600. I then saw the 1821+ had eight bays instead of 6 and was only $100 more. Yea, that’s how they upsell you. It’s just $100 more.

Realistically I only need 2-4 bays, but why spend that much to be limited by drives? The 1522+ with 5 bays is $700 and the Ryzen 1600. For an extra $300 (almost 50% more, but still only $300) I can get a much more capable 1821+
 
There's no good way to stress test a hard drive. Just make sure your data is backup and use any of the raid levels that can tolerate at least a drive failure (RAID 1 / 10 / 5 / 6) and call it a day. I've build at least 5-6 enterprise NAS within the last 3-4 years (12 to 24 drives, RAID 6 or 10) and never feel the need to stress test them.

yes and no, when you buy several drives from the same batches, even with those raid levels if still using hardware raid, you could have multiple drives die at the same time, yes, unlikely, but still possible. a stress test off the bat, can quickly rule out a bad drive right away at least and save headaches for when you got everything built and data on it to have a drive reporting as bad a week later..
 
yes and no, when you buy several drives from the same batches, even with those raid levels if still using hardware raid, you could have multiple drives die at the same time, yes, unlikely, but still possible. a stress test off the bat, can quickly rule out a bad drive right away at least and save headaches for when you got everything built and data on it to have a drive reporting as bad a week later..
Everytime you build a RAID volume, you basically run a full scan of all of the sector on all of the disks. And with parity based RAID, you supposed to run a disk scrubbing once in awhile to minimize URE. So tell me again, why and how would you “stress” test every single disk before they’re actually put into an array? Do you know how long it would take to run a stress test on 48 x 18 TB HDD?
 
Everytime you build a RAID volume, you basically run a full scan of all of the sector on all of the disks. And with parity based RAID, you supposed to run a disk scrubbing once in awhile to minimize URE. So tell me again, why and how would you “stress” test every single disk before they’re actually put into an array? Do you know how long it would take to run a stress test on 48 x 18 TB HDD?
A long time. My 12TB Exos took a bit over a full day to run a long DST per drive.

But the OP is interested in a 5 bay NAS. So, maybe running a long DST on each drive wouldn't be so bad if they have the time.
 
A long time. My 12TB Exos took a bit over a full day to run a long DST per drive.

But the OP is interested in a 5 bay NAS. So, maybe running a long DST on each drive wouldn't be so bad if they have the time.

I have plenty of time. I do NOT even know how to run a DST on OSX.

Anyways. With prime days the DS923+ was on sale. While not unit I wanted, it was literally less than 1/2 the price of the DS1821+ I wanted. It was $480 vs $1000.

Yes, there are compromising differences.

I also wanted 16Tb drives. Found IronWolf Pro 16Tb for $260 a piece on Adorama. On Amazon the Toshiba N300 Pro 14Tb were $240 a piece.

Instead of spending $2050 I am spending $1440 (before taxes and free shipping).

Yea it’s a compromise but it’s $600.

I upgraded the memory to 32 gig and upgraded the network card. By the time I need more storage I’m sure I’ll need another server anyways.
 
I have plenty of time. I do NOT even know how to run a DST on OSX.

...
If it's anything like a QNAP NAS, you use the local IP address of the NAS on your web browser to login, and there is a way to run the drive self test from the drive/storage section. OSX (or Windows) has nothing to do with it other than to provide a web browser.
 
Last edited:
  • Like
Reactions: Liver
like this
Whenever you set them up just use data verification.
I've actually had brand new drives throw an error and go offline.
If you skip this you will be sorry.
 
I got the drives today and I put them all in the NAS. I am running and extended SMART test. Is that all I should do? Or is there another setting on the NAS to look at?
 
Last edited:
I got the drives today and I put them all in the NAS. I am running and extended SMART test. Is that all I should do? Or is there another setting on the NAS to look at?
That should be fine.
This is how they came from Amazon. Pitiful.

View attachment 584376

You're lucky it's in a box!
I've had them shipped in the static bag only loose in one of those bags!

Instant return request in that case.

I generally get mine at Microcenter or if I need bulk Provantage. They know how to ship them properly.
 
I will typically boot up a parted magic live cd in an old system attached to an 8-bay sata enclsoure and then run all the smart tests (short, transport, and long). If it fails any of these, it's just a trigger for a warranty, which is what I'm trying to figure out--is a drive that was shipped doa. The long test will need to run overnight typically, but what's great about parted magic is that it can run several tests in parallel, so I can put in say, 5x 16TB drives for a long test and by the am all 5 will have completed. (I also put in a 4 bay 2.5" in one of the drive bays so now I can test up to 12x drives at the same time--saves a lot of time. :))
 
An extensive surface test can be done with WD data lifeguard. It is on the free Hirens USB bootstick together with Windows 10PE and many other tools. I use it whenever my ZFS server complains about too many or checksum errors to verify disk state especially as WD data lifeguard can repair some problems or give a final disk fail.

To use it on a laptop, you need an Sata->USB bay/adapter

https://www.hirensbootcd.org/
 
Last edited:
I use ZFS for more than 15 years with hundreds of disks. I have never tested disks prior use with ZFS as ZFS will detect and report real problems more accurat during use than smart or badblocks prior use. Only care about redundancy (allow disks to fail), hotspares for immediate replacement and email alerts in case of problems.

I test disks only if I asume there are bad ones or when ZFS detects problems (Intensive disk tests that are worth the effort run several days per disk, smart is often no more than a prediction like high blood pressure for your health. You can die soon or in 30 years with it.On

just to add
With a homeserver you normally use disks until they fail either when ZFS throw them out due "too many errors" or on Solaris when the fault management service fmd retires a device with errors to avoid further problems. As all data are distributed over all disks, ZFS will report problems immediatly (beside hot spare disks)

On a production system, situation is different. All disks are regularly replaced every 3-5 years. If a smart error occors that may indicate a higher propability of an earlier failure, you want to replace on a lower activity time. Not because it has failed but because on a failure you must expect a reduced performance during replace/resilvering that you want to avoid if possible.
 
Last edited:
An extensive surface test can be done with WD data lifeguard. It is on the free Hirens USB bootstick together with Windows 10PE and many other tools. I use it whenever my ZFS server complains about too many or checksum errors to verify disk state especially as WD data lifeguard can repair some problems or give a final disk fail.

To use it on a laptop, you need an Sata->USB bay/adapter

https://www.hirensbootcd.org/
Unless you've got an older laptop with an esata port ;) Then you just need an esata dock. :)
 
Back
Top