Created attachment 30641 [details] Scary! DeviceKit-disks-008-1.fc12.x86_64 libatasmart-0.16-1.fc12.x86_64 Says 2 million sectors are failing, and that the temperature is zero.
Please attach out skdump binary blob (option --save). Thanks.
Hmm.. 2 million sectors of 512 bytes a pop gives.. 1TB.. which is much bigger than the size of the disk. Perhaps libatasmart could be smart about this (e.g. decide that the parsing of the attribute failed) - and then possibly complain on stderr (or, better, syslog) so the owner can file a bug.
(In reply to comment #2) > Hmm.. 2 million sectors of 512 bytes a pop gives.. 1TB.. which is much bigger > than the size of the disk. MATH FAIL... 2 million times 512 bytes is roughly 1GB. Never mind the suggestion in comment 2 then.
Also, for comparison, please attach smartctl -a output. Thanks.
Created attachment 30644 [details] skdump /dev/sda output
Created attachment 30645 [details] smartctl -a output
This is the 64GB SSD in a gen1 macbook air. It's as new, and (if I can read the libatasmart output correctly) has only been turned on for 9 hours when I uploaded the original screenshot.
(In reply to comment #7) > This is the 64GB SSD in a gen1 macbook air. It's as new, and (if I can read the > libatasmart output correctly) has only been turned on for 9 hours when I > uploaded the original screenshot. Attribute five (which is interpreted as "reallocated sector count") seems to be constant and smartctl(8) seems to agree with libatasmart. Might be worth checking if a disk firmware upgrade from the vendor is available, and if so, fixes this issue. Otherwise I guess we can get Lennart to blacklist things.
(In reply to comment #2) > Perhaps libatasmart could be smart about this (e.g. decide that the parsing of > the attribute failed) - and then possibly complain on stderr (or, better, > syslog) so the owner can file a bug. We actually do something like this. But this raw value is still in range, so we don't complain.
I have now blacklisted the two bogus attributes for this drive in exactly this firmware in the hope that Apple fixes this in a later firmware.
http://git.0pointer.de/?p=libatasmart.git;a=patch;h=12d2dcdb7176cfe71489954a93e6028ffef259c1
(In reply to comment #0) > Created an attachment (id=30641) [details] > Scary! Btw, I wonder why it reads "60 GB Hard Disk" and not "60 GB Solid-State Disk" like with my disk http://people.freedesktop.org/~david/f12-smart.png What's the content of /sys/block/sda/queue/rotational (should be 0 but is probably 1). If 1, please attach the output of 'hdparm -I /dev/sda' (mine says: Nominal Media Rotation Rate: Solid State Device). Hmm, could be that the IDENTIFY data from the drive is just lying - 1st gen Macbook Air sounds a bit vintage - but you say the disk is brand new? Does this mean 'never used' instead? Thanks,
I use MacOS X so often that I didn't realise it didn't run the latest software/wasn't updated. I'll update the machine and report back. As for vintage, the machine was released early 2008: http://en.wikipedia.org/wiki/MacBook_Air#Specifications The machine is a refurb, never used.
Snow Leopard with all the software/firmware updates installed doesn't update the firmware of the SSD, so no changes there. MacOS X claims the device has passed SMART tests just fine as well, so it probably ignores that value the same way the libatasmart patch does. Thanks for the fix Lennart, could you please make sure it gets into F12, so it doesn't complain about my setup on startup?
(In reply to comment #14) > Snow Leopard with all the software/firmware updates installed doesn't update > the firmware of the SSD, so no changes there. OK - I guess this makes sense since Apple don't worry about other OSes than their own on their hardware. > MacOS X claims the device has passed SMART tests just fine as well, so it > probably ignores that value the same way the libatasmart patch does. libatasmart/DKD/gnome-disk-utility is more thorough than most other SMART stacks / user-interfaces (including both smartctl and Apple's stack) insofar that we warn the user if there's "many" bad sectors instead of blindly trusting the self-assessment done by the disk. ("many" depends on a number of things - we don't warn if there's only one - but we do warn if you have ten or hundreds - IIRC it depends on the size of the disk - Lennart knows.)
Shall I file a separate bug about the hard disk not showing up as SSD?
(In reply to comment #16) > Shall I file a separate bug about the hard disk not showing up as SSD? Please provide the info requested in comment 12 ;-)
(In reply to comment #14) > Thanks for the fix Lennart, could you please make sure it gets into F12, so it > doesn't complain about my setup on startup? Sure. will do. will combine this with some other patches (idle wakeup stuff). Those are a bit more dangerous so I am not sure I will ask for breaking F12 freeze for this, but rather push it into F12 after the release via the usual karma stuff.
FYI: http://thread.gmane.org/gmane.linux.utilities.smartmontools/5252 and: https://bugs.launchpad.net/ubuntu/+source/gnome-disk-utility/+bug/438136/comments/4 Maybe the firmware check needs to be updated.
(In reply to comment #19) > FYI: > http://thread.gmane.org/gmane.linux.utilities.smartmontools/5252 > and: > https://bugs.launchpad.net/ubuntu/+source/gnome-disk-utility/+bug/438136/comments/4 > > Maybe the firmware check needs to be updated. I modified the blacklist entry now slightly so that it covers more firmware versions: http://git.0pointer.de/?p=libatasmart.git;a=patch;h=4a6caeca4e8e9585a86470c11d2e75d5c1368d59
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.