[smartmontools-support] False alarms after hot-swapping drives

[Brent Howards]

Hello—

I have a number of drives specified in my smartd.conf by path, for example:
/dev/disk/by-path/pci-0000:02:00.0-sas-exp0x5001b3d25c10803f-phy8-lun-0 -a
-s L/../../1/04

About 20 minutes hot-swapping out two hard drives (currently inactive and
not part of any array) for drives of a different make/model, I received
emails that included the following (one for each of the paths that had had
its drive swapped):

"""
The following warning/error was logged by the smartd daemon:

Device: /dev/disk/by-path/pci-0000:02:00.0-sas-exp0x5001b3d25c10803f-phy8-lun-0
[SAT], Failed SMART usage Attribute: 5 Reallocated_Sector_Ct.

Device info:
WDC WD6001FZWX-00A2VA0, S/N:, WWN:, FW:01.01A01, 6.00 TB
"""

The device infos in the emails corresponded to the previous drives in the
slots, rather than the ones currently (which are Seagate 8TB Archive
Drives). Running "smartctl -a" did not show any problems under
Reallocated_Sector_Ct for either Seagate drive. I also tried running
"smartctl -a" on the previously inserted drive (after connecting it
externally) and that also indicated no issues with Reallocated_Sector_Ct.

smartmontools persisted in alerting me to these phantom errors via my
syslog for the next several hours, however, until I restarted it at which
point there were no more messages.

Is this a known problem with hot-swapping drives or is there something more
worrisome going on?

Thank you,
Brent

[Christian Franke]

Post by Brent Howards
...
About 20 minutes hot-swapping out two hard drives (currently inactive
and not part of any array) for drives of a different make/model, I
received emails that included the following (one for each of the paths
...
smartmontools persisted in alerting me to these phantom errors via my
syslog for the next several hours, however, until I restarted it at
which point there were no more messages.
Is this a known problem with hot-swapping drives or is there something
more worrisome going on?

Hot swapping is not supported yet:
https://www.smartmontools.org/ticket/695

Could possibly be fixed by checking device identity before each check
(enabled by e.g. a new -d hotswap directive) and trigger a rescan if
changed.

Thanks,
Christian