On Sun, 19 May 2024 at 05:06, Stefanos Harhalakis <[email protected]> wrote:
> See attached patch for matching NVMe devices too in smartd logs Thanks - yes i'd noticed that the rules for smartd need an update and this is on the radar - however, the update to the names of the .state files i was not aware of! I think the first part of the patch is not quite right, per the code i think it should be: # from https://sources.debian.org/src/smartmontools/7.4-2/smartd.cpp/#L5829 ... Monitoring [[:digit:]]+ ATA/SATA, [[:digit:]]+ SCSI/SAS and [[:digit:]]+ NVMe devices It would be really helpful if you could: 1/ send some sample log output logs, something like: journalctl -t smartd | awk '{$1=$2=$3=$4="";$5="X";print}' | sort -u would be helpful I'm especially unsure about the bits of rules that start "Device:" as i could not work out what format is expected -- the current rules have Device: /dev/[^[:space:]]+( \[[_/[:alnum:][:space:]]+\])?( \[SAT\])? but im not sure if the optional groups are up-to-date, -- i couldnt find what bit of code determines those. 2/ the attached rules (untested) are what i'm wokring on for smartd - any testing of these would be great (I generalised the name of the .state file - again i dont know where smartd sets this bit, so hard to know what is expected!) (these are completely untested, so may be completely broken)
smartd.rules
Description: Binary data
