I decided to install a handy utility called gsmartcontrol on my Fedora 27 laptop. After successfully installing it, I get this prompt when trying to run it from the GNOME application menu:
Authentication Required
Authentication is required to run GSmartControl Administrator
Password:
Root password does not work, nor does that of the current user (who is in the sudoer file.) Who the heck is Administrator? This ain’t Windows NT…
Thanks for the suggestions. Here are the results; unfortunately neither of them worked.
# gsmartcontrol
Unable to init server: Could not connect: Connection refused
(gsmartcontrol:16690): Gtk-WARNING **: cannot open display:
$ gksu gsmartcontrol
ksh: gksu: not found [No such file or directory]
Searching the repo for gksu suggested an alternative called beesu, so I installed that. However, I get exactly the same error that @Dynamic_Gravity 's suggestion generated.
So that gave me another idea. I logged out and for my X session selected “GNOME Classic” instead of “GNOME”. And gsmartcontrol fired right up, after feeding it my password.
Now I need to figure out how severe an IDNF SMART error is…
Any SMART errors at all, I immediately replace the drive. Served me well so far as every single drive has subsequently failed in a matter of months at most.
Ive had the opposite experince. every drive ive had with smart errors has continued to work for many years. Only dead drive ive had is when the controller died and could not longer be written to and would randomlly drop out.
This is actually a good question. I’ve heard some people say that anything other than reallocated/pending sectors can be ignored. Some people even say that if reallocated sectors is only a couple of sectors, you can ignore it.
Other people like @Ruffalo think that any error calls for replacement. Personally, when I’ve been, um, economically challenged, I’ve used a drive with a single reallocated sector for more than a year.
SMART, to my knowledge, is a predictive failure mechanism. Meaning that a failure is coming and the drive should be replaced.
I know that where I work, standard practice is to replace the drive if we get that but we’re doing work for a bulk customer who bought a massive extended warranty.
Reallocation is part of how mechanical drives work. Theres usually a threshold to this though. If you’re throwing out drives with a couple reallocated sectors, you’re wasting money really. Other metrics do matter though, I had an HGST drive fail smart on spin retry. Luckily within warranty.
Yes, that’s a logic trap. I learned my lesson when I tried ignoring the errors and a couple drives died.
Nobody throws out drives with “a couple reallocated sectors”. You don’t look at the raw SMART data, you get an alert saying the number of reallocated sectors (or whatever) passed a set threshold, which indicates the drive is likely to fail.
Well, that’s the thing. I’m not getting an alert that the drive is failing. However, when I look in the drive’s onboard error log I see several “ID Not Found” (IDNF) errors. Some googling indicates that this is a corrupt sector. However its been about 400 hours of uptime since the last error occurred.