Hello list,
could readers please respond with a short yes/no, if they would like
smartctl scripts that fix mismatching scterc/controller timeouts to
prevent data loss?
(Testing them only requires to unzip the file into /etc/udev/rules.d/)
And dear maintainers, would you consider shipping the
scripts?
(Fixing the timeouts seems important for all systems without a
default smartctl -l scterc timeout. And all redundancy controlling
packages (mdadm, lvm, btrfs, ...) may want to suggest or depend on the
smartctl package and set shorter timeouts, i.e. they call the
appropriate scripts depending on redundancy states. Thus smartctl seems
the right package to ship the scripts and default udev rule, and other
packages ship their device specific udev rules.)
Cheers,
Chris
PS: v1.01 has smartctl-default_test.rules for udev and is working well
here with drives that have scterc timeout default to "Disabled".
Am Fri, 27 Feb 2015 14:49:51 +0100
schrieb Chris [off-list ref]:
Hi there,
here are some "smartctl-timeouts" scripts that fix commonly
mismatching defaults with drives that have no error recovery timeout
configured, which has often lead to data loss. (See the README for
further information.)
The scripts are the result of discussions on the linux-raid
mailinglist, but it may be best to ship the scripts with smartctl
already (possibly together with a default "non-redundant-disk" udev
rule to ensure proper default timeouts), and let redundancy
controlling packages like mdadm, lvm, btrfs etc. ship their own udev
rules to call the scripts appropriately.
Please respond whether you'd like to ship the scripts together with
smartctl.
Cheers,
Chris