weekly fstrim cron job reports errors

[Nathaniel W. Turner]

Every week I get an email from cron with error output from the /etc/cron.weekly/fstrim job:

/etc/cron.weekly/fstrim:
 HDIO_DRIVE_CMD(identify) failed: Invalid argument
device /dev/md0 is not an Intel or Samsung drive
SG_IO: bad/missing sense data, sb[]: 70 00 02 00 00 00 00 0a 00 00 00 00 3a 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
SG_IO: bad/missing sense data, sb[]: 70 00 02 00 00 00 00 0a 00 00 00 00 3a 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
device /dev/sdf1 is not an Intel or Samsung drive

It's true that those block devices are not Intel nor Samsung SSDs (though /dev/sdc is an Intel SSD), but I don't think fstrim-all should be printing a warning in this case.

ProblemType: Bug
DistroRelease: Ubuntu 14.04
Package: util-linux 2.20.1-5.1ubuntu14
ProcVersionSignature: Ubuntu 3.13.0-8.28-generic 3.13.2
Uname: Linux 3.13.0-8-generic x86_64
NonfreeKernelModules: nvidia
ApportVersion: 2.13.2-0ubuntu4
Architecture: amd64
CurrentDesktop: Unity
Date: Mon Feb 17 12:23:49 2014
EcryptfsInUse: Yes
SourcePackage: util-linux
UpgradeStatus: Upgraded to trusty on 2014-01-23 (25 days ago)

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[latimerio (fomember)]

This bug is not a duplicate of 1276941 as the error messages are totally different (although the root cause may be the same)
I installed 14.04LTS on a virtual machine and now I am getting the error from /etc/cron.weekly/fstrim

SG_IO: bad/missing sense data, sb[]: 70 00 05 00 00 00 00 0a 00 00 00 00 20 00 00 c0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

There is no SSD in my VM thus fstrim should not be running at all on my virtual disk.

[latimerio (fomember)]

Just found that the problem is possibly not originating from fstrim.
As hdparm -I /dev/sda reports the error
 SG_IO: bad/missing sense data, sb[]: 70 00 05 00 ... I suppose fstrim is just passing the error.

I doubt that it is a real error as my ubunt is a VM with vmfs disk.
Thus fstrim should be silent about the missing sense data at least when the disk has no TRIM capability at all.

[latimerio (fomember)]

Digging deeper into it I found a line in /sbin/fstrim-all saying:

HDPARM="`hdparm -I $REALDEV`" 2>/dev/null || continue

I guess it should read instead:
HDPARM="`hdparm -I $REALDEV 2>/dev/null`" || continue

The 2>/dev/null only makes sense to me if it ignores the error of hdparm not the possible error of setting the HDPARM variable.