On Wed, 1 Feb 2023 at 17:13, emanuel stiebler via cctalk
<cctalk(a)classiccmp.org> wrote:
On 2023-02-01 10:56, Warner Losh wrote:
On Wed, Feb 1, 2023 at 1:41 AM emanuel stiebler via cctalk
<cctalk(a)classiccmp.org <mailto:cctalk@classiccmp.org>> wrote:
retension in case of power off.
If the power is applied all the time, the internal controller "can"
check the quality of the cells automatically (but this really
depends on
the controller, controller version, and the OS has to chose the right
strategy. And the controllers improved a lot lately)
The OS might not have a choice. All the SSDs that I've used in the
past decade at $WORK have not exposed any of this to the host, not
even enough stats to know when it's going on in real time... let alone
the ability to pause these operations for a little while until we're off
peak for the day...
But you should be able to choose (at least on controllers I know) if you
like to go for automatic or manual refresh.
If you go for the automatic, it could happen that the drive decides to
scan the drive, when your're busy and going nuts waiting for the drive
(you can also define on newer drives how many block to check per run)
If you're going for the manual refresh, just make sure you really run it
one day. But, you should be the one who knows, when your computer isn't
busy ...
That reminds me (looks at 43.5T of zfs pool that has not had a scrub
since 2021).
It can be nice to have a filesystem which handles redundancy and also
the option to occasionally read all the data, check end to end
checksums (in the unlikely case a device returns a successful read
with bad data), and fixup everything. Does not eliminate the need for
remote copies, but gives a little extra confidence that the master
copy is still what it should be :)
Its currently all on spinning rust, but I imagine zfs scrub should map
nicely onto refreshing SSDs with "read everything you care about,
write only on fixing up data" (in the unlikely event I can ever afford
sufficient SSD storage)
David