Synology Issue

How many programmers does it take to change a lightbulb...

They can't, it's a hardware problem!

It sounds either like an unexpected hardware error on HDD3, or that the SMART stats were so far out of parameter that although the disk was usable it was in the imminent catastrophic failure zone; but these are just conjecture.

You're right about not using desktop drives for NAS, they just aren't designed for that type of duty cycle and hence have high failure rates.

Hi Huge,

I was just VERY glad that things are recovering and that I, hopefully, won't need to rebuild!

M

I was reading an article somewhere of a data centre that used 3tb drives and had a 47% failure rate after 18 months or something. Our IT guy certainly thinks 2tb is the sweet point between size and reliability.

I'm replacing with 4TB WD Red ....I'll let you know if I have issues!

M

I have only ever bought one other type of manufacturer's hard drives - a Samsung 1TB when I was building a desktop PC, but it failed from out the box......I swear by Western Digital, have only ever had one issue, a failed interface board on a portable HD, now located in an 'Icy Box' enclosure. The Western Digital hard drive count in our house is up at about 7 x 3" of varying capacity, and about 11 or so x 2.5" of varying capacities with not a single HD failure.

As for the NAS, our two bay Sonology has WD green drives, left over from PC builds and stores Music only.

The four bay Sonology has 4 x 4TB WD Red NAS specific drives, for all our movies, photo's and backup of other files.

Haven't an an issue yet, and would really recommend Sonology for their user interface and 'apps'.

screen shot

There are a few www reports on HDD reliability worth a look ...

...  search  "Hard Drive Reliability"

A report by ExtremeTech concerns the two makes mentioned by Mr Underhill - Seagate & WD

Seagate were bottom of the class & WD significantly better,  Hitachi were the best,  but only marginally better than WD

While in the ExtremeTech report, follow the link to Backblaze who compiled the report from there own back up service data base 

One last observation from me:

When I manually ran SMART quick tests on the HDDs one other disk was reporting issues; which I also replaced. I am therefore assuming that the NAS doesn't do this automatically.

The Synology Storage Manager HDD/SSD section has a 'Test Scheduler' tab that allows you to automate the SMART testing; which I now do nightly.

.....and I have added a Hot Spare.

M

Nightly !!!!   isn't that a little OTT ???

I thought it was recommended to run SMART tests aprx every 1000 op hours,  that about once a year for peeps like us.  Plus I understand it can be a risk of corruption is the disk has an error

What does the installed & always running background check system do ?   Its on the widgets pop-ups that are available on the www.web page front screen.This tells you the disks are running OK & seems effective as I had a minor error once that it shouted up a warning on (only needed a reboot) so I would have thought the SMART test was almost redundant.  

Hi Mike,

The SMART quick test shouldn't stress the disk, and I think the general advice is to run it weekly. As I still have some Seagate drives I think the daily is fine ....until they are all replaced.

M

Yes - that's definitely excessive.

One thing that has interested me out of this is that the reliability of SSDs has now far exceeded HDDs, even for MLC SSDs (which some years ago had very short life expectancy), and even under benign conditions for the HDD.  In my view only argument against SSDs is now simply the cost per GB.

Mike, Mr Underhill,

The quick SMART test imposes a relatively small degree of stress on the disk, but there is still some stress (i.e. a bit above normal operation), and that caries a very small risk of inducing a failure condition, so you're both right!

If you suspect a significant risk of imminent catastrophic failure (sudden death syndrome) then once a day could be justified, but that would suggest something more than just a suspect disk type.  For instance it could be appropriate for desktop drives being used as a temporary replacement for NAS drives in a hardware RAID NAS used for 24x7 availability.

If you have reason to suspect some unreliability in one or more disks, then  I'd agree with one a week being a reasonable period.

In more normal circumstances (e.g. drives designed for NAS use, operating in their normal working envelope), then once a month should be sufficient (i.e. once per 720 hours of the disk being powered, either operating or standby as we are taking NAS drives here).

If you don't use RAID the smart check is almost redundant () as you'd have to go to your backup to restore the data anyway (all it does is to allow you time to order a replacement disk).

If you use RAID then you can set the SMART check frequency to 'a bit more often than the time in which it's exceptionally unlikely that you'll get two (or more) disk failures'; the 'bit' being the length of time it takes to get a replacement disk in place.

A useful read -  http://forum.synology.com/wiki..._repair_basic_faults

Mike,

For NAS drives like the WD Reds that are designed to manage their own sleep cycles (rather than having the OS do it), how do you determine the 'Operating Time' as opposed to their 'Powered Time'?

Hi Huge,  I don't - simple'z - as the russian speaking Namibian suricates say 

I have WD Reds & was not aware they were designed as you indicate - but whatever I don't bother with SMART tests,  I rely on the Synology installed disk check that tells me its OK everytime I get into the web-page & if it does have an error it sends me an e-mail.  That happened once when it failed to boot up properly after a few weeks holiday;  Synology advised to shut down (the blue switch) & restart - problem solved.   So I wonder why we need a SMART test 

WD Reds use a variable platen speed (I believe it was something like 0, 3600 to 6800 rpm) to optimise power consumption, life and availability; they are designed for 24x7 uptime.

Mike-B,

I am doing the SMART tests following my last HDD failure, onceI had recovered I decided to do manual SMART checks across all the remaining disks, this then reported another Seagate had potential issues.

Whilst I still have the Seagate drives in the NAS I will do the checks, once they have all died I will revisit the topic.

M