Comsumer External HDD Lifespan

[sczze]

I have a seagte backup plus portable drive that is a few years old. It's still quite new with <100 power on hours.

I am wondering on average how long do these consumer drives last? How much can I depend on them if I decide to plug them in 24/7?

[AbydosOne]

5 minutes ago, sczze said:

I have a seagte backup plus portable drive that is a few years old. It's still quite new with <100 power on hours.

I am wondering on average how long do these consumer drives last? How much can I depend on them if I decide to plug them in 24/7?

Well, I have one external going on 7 years (HGST), and another going on 9 (LaCie/Seagate). My parents have one that's at least 12 (Verbatim/Seagate), I think...

 

Power-on hours don't mean much in terms of drive life. Power-on count is probably a better indicator (I've had drives in my server with POH above 20K and POC around 100 that are still going strong).

 

On the flip-side, infant mortality is a thing, so basically don't count on any hard drive to turn on the next time you need to use it. Keep backups. Keep backups of backups.

[Matty2048]

i wouldnt count on it lasting longer than 15 years but lifetime varies alot.

[Thirdgen89GTA]

18 minutes ago, sczze said:

I have a seagte backup plus portable drive that is a few years old. It's still quite new with <100 power on hours.

I am wondering on average how long do these consumer drives last? How much can I depend on them if I decide to plug them in 24/7?

Like all spinning rust, you have about 2 major opportunities for failure.

 

Early failure due to MFG defect, usually happens early in the life cycle, probably 3 to 6 months out of the box.

 

After that, the drive is likely normal and will provide reliable usage for the first 4-6 years of its life. Beyond that, every day is a blessing and the chances of data corruption rise dramatically.  You should never consider single drive storage as "reliable" for important information.  Doesn't matter if its family photos, financial records, etc.

 

If the "data", whatever that may be, is important to you, you should be be storing it on some kind of Mirror, or Raid setup, or be paying for some kind of cloud storage solution where they deal with redundancy for you.

 

Not saying you need some kind of beastly home server.  But a simple disk mirror either via computer, or NAS like device is fine.

[sczze]

4 minutes ago, Thirdgen89GTA said:

Like all spinning rust, you have about 2 major opportunities for failure.

 

Early failure due to MFG defect, usually happens early in the life cycle, probably 3 to 6 months out of the box.

 

After that, the drive is likely normal and will provide reliable usage for the first 4-6 years of its life. Beyond that, every day is a blessing and the chances of data corruption rise dramatically.  You should never consider single drive storage as "reliable" for important information.  Doesn't matter if its family photos, financial records, etc.

 

If the "data", whatever that may be, is important to you, you should be be storing it on some kind of Mirror, or Raid setup, or be paying for some kind of cloud storage solution where they deal with redundancy for you.

 

Not saying you need some kind of beastly home server.  But a simple disk mirror either via computer, or NAS like device is fine.

When a data corruption occurs, what happens? I have not experienced data corruptions before (at least I did not know that that I did). Most of the time I have been using SSD but now I create more content. So I decided I will archive my work with HDD. (before that i only use hdd for backup). When data corruption occurs, will it repare itself (I read somewhere that i would do that but I dont know how true) or will there be an error message. 

And also if im not wrong, data corruptions happens more frequently with HDD as compared to SSD am I right?

 

Thank you so much!

[Thirdgen89GTA]

You won't notice data corruption unit you either read, or write to the bad sectors.  Typically the system locks up while attempting to read/write the sectors because the I/O request sent to the disk is not being responded to.  

 

Think of the Computer as the boss, and the corrupted disk as a lazy employee.  The boss asks the lazy employee for the report, the lazy employee can't provide the report.  So the lazy employee scrambles like mad to fix, create, correct the report.  Mean while the boss is waiting, wondering where the report is.

 

If the disk error correction works, then after a delay, the boss gets the report.

 

If the disk error cannot be corrected, the lazy employee tells the boss sorry I can't give you the report.

 

Then if you have mirrored array, you have 1 boss and two employees doing the exact same jobs.  When one employee is lazy, the other provides the report while the lazy employee bumbles around.

 

Raid 5 would be 5 employee's working on a project, and each employee does their own work + a portion of each other employee's work.  When one employee is lazy, the other 4 employee's pick up the 5th employee's slack and provide a finished project.

 

[sczze]

4 hours ago, Thirdgen89GTA said:

Then if you have mirrored array, you have 1 boss and two employees doing the exact same jobs.  When one employee is lazy, the other provides the report while the lazy employee bumbles around.

Wait so if there is a bad sector, the computer will use data from the other drive(s) in the raid array. But will the sector on the bad drive remain bad? Or will the computer use the data from the other drive to repair the bad drive?

[Thirdgen89GTA]

11 hours ago, sczze said:

Wait so if there is a bad sector, the computer will use data from the other drive(s) in the raid array. But will the sector on the bad drive remain bad? Or will the computer use the data from the other drive to repair the bad drive?

Yes, because one of the drives will have failed to respond in time or returned data that is different from the other drive. When the bad disk fails to read the sector, it will mark the sector bad and attempt to relocate the data to a reserve sector.

 

The parent RAID controller, be it Hardware, or Software will also conduct a checksum test comparing the data returned from each drive.

 

Filesystems like ZFS are insanely robust, even on single drive solutions.  When you add in Mirrored, or RaidZ# setups..etc.