Really excited to see more NVMe support; my experience of vendor-supplied tools and reporting forces me to rely on smartmontools, which has been quietly reliable for as long as I can remember.
Has anyone ever had SMART stats actually usefully predict the death of a drive? The vast majority of uncontrolled (ie, not death by age and natural causes) drive failures are crib deaths with no warning I feel like.
I had some SSDs which reported that they have reached the maximum amount of data that may be written.
I have replaced them promptly, without waiting to see if they will really die.
I also had some new HDDs which have failed the SMART self-tests, so I was able to request replacements. After having once problems with a new HDD, I have always run all the SMART self-tests on all new HDDs, before storing any valuable data on them, and that has proved to be a good policy.
It is, indeed. E.g. by looking on remapped sectors on traditional HDD you may see trend that drive will die soon. Selftests are only way to check surface before you will hit it with real files...
On SSD the most important is wear out and usage data. It may, from one side, indicate that drive needs to be replaces soon and from other - if you see that wear out is too fast - optimise your software/os to prevent that.
- The docker image used for CI and release builds is now based on
Debian 12 instead of Ubuntu 18.04.
- macOS: CI and release builds are now generated for the x86_64 and arm64
targets. 32 bit platforms will require to be compiled from the source.
- smartctl '-t short', '-t long' and '-X': NVMe support.
- smartctl '-l selftest': NVMe support.
- smartctl '-l farm': Prints Seagate's vendor-specific Field Access
Reliability Metrics (FARM) log for ATA and SCSI drives.
- smartctl '-l error': Now also prints an error message for each entry
of NVMe error information log.
- smartctl '-l genstats': Prints SCSI General statistics and performance
log page.
- smartctl '-i' and '--identify': ACS-4/5/6 enhancements.
- smartctl '-c': Added NVMe 2.0 capability flags.
- smartctl '-g security': Added 'ata_security.master_password_id'
to JSON output. Plaintext output shows Master Password ID if set
to a non-default value.
- smartctl '-q noserial': Now also suppresses the output of NVMe Namespace
IEEE EUI-64.
- smartctl '-j': '-l error -l selftest' JSON output for NVMe devices.
- smartctl '-j': Avoid invalid UTF-8 sequences in JSON/YAML strings.
- smartctl '-j': Fixed a bogus exception during SCSI JSON output.
- smartctl '-j': Renamed JSON element 'scsi_temperature' back to
'temperature' (regression).
- smartctl '-a': Now suggests '-x' for ATA devices because '-a' only
provides legacy SMART information.
- smartd: Now detects accidental use of smartd_warning script as
'-M exec' parameter.
- smartd: No longer writes the 'Copyright...' line to syslog.
- smartd.conf '-M always': Sends reminder emails without any delay.
- smartd.conf '-M diminishing': Limited email delay to 32 days.
- ATA: Fixed decoding of extended self-test log on big endian hosts.
- ATA: Enhanced LBA range for device types '-d jmb39x-q,...' and
'-d jms56x,...' from 33-62 to 1-255.
- ATA: Device type '-d intelliprop,N' now fails with a deprecation message.
Added '-d intelliprop,N,force' flag to use it anyway.
- ATA/USB: Device type '-d usbasm1352r,N' for ASMedia ASM1352R USB to SATA
RAID bridges
- SCSI: Fixed possible corruption issue with the Error Counter and
Non medium Error log pages.
- SCSI: Added more "Informational Exceptions" strings.
- SCSI: Added initial support for REPORT SUPPORTED OPERATION command.
- SCSI: Initial rework of SCSI debug output.
- NVMe: Added error messages for NVMe status values.
- NVMe: Fixed crash after read of error information log on big endian hosts.
- HDD, SSD and USB additions to drive database.
- update-smart-drivedb: Fixed syntax for 'sed' versions which require
';' before '}' or do not support ';' at all.
- configure: Default for '--with-nvme-devicescan' is now 'yes' also on
Darwin and FreeBSD. It is still 'no' on NetBSD only.
- configure: Defines '_FORTIFY_SOURCE=3' if supported and not predefined.
- configure: No longer fails if libsystemd-dev is installed and
'LDFLAGS=-static' is used.
- Compile fix for systems without legacy 'getdtablesize()'.
- Pre-releases from SVN snapshots now show "pre-VERSION" in version
information and 'smartctl.pre_release=true' in JSON output.
- Linux: Device type '-d sssraid' for 3SNIC RAID controllers.
- Linux: Device type '-d marvell' now fails with a deprecation message.
Added '-d marvell,force' flag to use it anyway.
- Linux: The generic SCSI code now defaults to SG_IO_V3 and does no
longer fall back to the deprecated SCSI_IOCTL_SEND_COMMAND
(but this ioctl is still used for '-d 3ware' and '-d marvell,force').
- Linux smartd: Now prevents systemd unit startup timeout when many
devices are registered and then initially checked.
- Linux smartd: Systemd no longer reports a service failure if no device
is present and a '-q *nodev0*' option is used.
- Windows: IOCTL_STORAGE_PROTOCOL_COMMAND variant for NVMe self-tests.
- Windows: Installer now defaults to 64-bit executables.
