Re: [gentoo-user] Hard drive error from SMART

[Dale]

Rich Freeman wrote:
> On Sat, Apr 16, 2022 at 10:59 AM Dale <rdalek1967@×××××.com> wrote:
>> I have googled and can not find a way to reset udev and it naming
>> drives.  I may have to rework some things since the drive kept the sdk
>> instead of switching to sdd when I made the physical change.  Thing is,
>> I suspect it will when I reboot the next time.
> IMO it is best to make that not matter.  If you're referencing drives
> by letter in configuration files, you're just asking for some change
> to re-order things and cause problems.
>
> You're using LVM, so all the drives should be assembled based on their
> embedded metadata.  It is fine to reference whatever temporary device
> name you're using when running pvmove/pvcreate since that doesn't
> really get stored anywhere. If you are directly mounting anything
> without using LVM then it is best to use labels/uuids/etc to identify
> partitions.

I have to use sd** when using cryptsetup to decrypt the drive.  I
haven't found a way around that that is easier yet.  My command was
something like cryptsetup open /dev/sdk1 <name> and then it asks for the
password.  After that, I use UUID and a entry in fstab to mount.  If
there is a easier way, I'm open to it.  I have three external drives and
as long as I only power them up one at a time, they all used sdk.  Now
they use sdd and I keep trying to type in sdk, from habit.  :/

My next project, find a good external drive enclosure like the three I
got now.  They no longer available tho.  I like them because they have a
fan, a eSATA port and a nifty display to let me know things are
working.  Really a good price for the features.  I don't like USB
connected drives.  Long story.


>> It also triggered
>> messages from SMART too.  It got upset that it couldn't find sdd anymore.
> That is typical when hotswapping.  I believe smartd only scans drives
> at startup, and of course if a drive does go offline it isn't a bad
> thing that it is noisy about it.  From a quick read of the manpage
> SIGHUP might or might not get it to rescan the drives, and if not you
> can just restart it.  The daemon works by polling so if there are any
> pending issues they should still get picked up after restarting the
> daemon.
>


Yea, it is a good thing.  I just disabled it for sdd, enabled for the
new sdk and restarted the service.  It was happy then but getting a
email from SMART always makes my heart beat a few extra beats and
sometimes causes me to swallow big too.  It's rarely good news.  Maybe
the next reboot will sort things out.  Then I get to switch everything
back to the old way again.  :/

Dale

:-)  :-)