[colug-432] failing hdd vs failing ssd

Chris Punches chris.punches at silogroup.org
Wed May 13 15:53:13 EDT 2020


between legacy and solid state* 

words are hard today

On Wed, 2020-05-13 at 15:46 -0400, Chris Punches wrote:
> Check out the attributes it reports when you do a smarctl test, like
> Wear_Leveling_Count, etc
> 
> https://www.techrepublic.com/article/how-to-check-ssd-health-in-linux/
> 
> I guess you could add a shim like some distros do for fsck on init
> and
> then pull the value, create an alert etc when certain values hit a
> threshold.  Not much really changed in that space between IDE and
> SDD.
> 
> -C
> 
> On Wed, 2020-05-13 at 12:21 -0700, Rick Hornsby wrote:
> > In most cases over the years when I’ve had an HDD fail, there were
> > early signs - often noises that weren’t normal coming from the
> > drive.
> > The noise usually doesn’t sound catastrophic, but rather the heads
> > can be heard moving around more than normal, or maybe the heads are
> > seeking, briefly parking, and then seeking (that’s a fairly obvious
> > sound). The problems might show up in the OS as an occasional I/O
> > error, or a corrupted file, but I can’t recall ever having any OS
> > warn me specifically that a drive was failing. Until, of course,
> > the
> > drive dies completely and becomes unusable to the OS.
> > 
> > I have a HDD that’s currently failing. It’s been making more noise
> > than normal for the last few days. Last night I ran smartctl -
> > mostly
> > to figure out which of the two HDDs might be a problem - and sure
> > enough it found some issues with one of the drives. Easy enough,
> > just
> > replace the drive and move on. SMART doesn’t catch every drive I’ve
> > had fail, but this time it did.
> > 
> > But it got me thinking about SSDs. There are no mechanical parts,
> > so
> > there’s nothing to hear. How do you know if an SSD is on the way
> > out,
> > other than running smartctl all the time and looking at the
> > results?
> > Do desktop OSs like Win/Mac/Linux have built-in facilities (I’m not
> > aware of any?) for notifying the user “hey, this drive is reporting
> > problems. It’s likely a sign of impending failure”? Would that even
> > work on an SSD?
> > 
> > I know both kinds of drives can remap bad sectors, and that they
> > are
> > designed with extra “space” to do just that. In my experience with
> > failing HDDs though, one bad sector tends to warn of, or worse
> > cascade into, more bad sectors rather quickly. I’m curious about
> > ways
> > we would know an SSD is going to kick it?
> > _______________________________________________
> > colug-432 mailing list
> > colug-432 at colug.net
> > http://lists.colug.net/mailman/listinfo/colug-432
> 
> _______________________________________________
> colug-432 mailing list
> colug-432 at colug.net
> http://lists.colug.net/mailman/listinfo/colug-432



More information about the colug-432 mailing list