On bootup smartd reports FailedReadSmartSelfTestLog but there is no hardware problem

[Richard]

SYMPTOM:
smartd reports it cannot read the smart self-test log for two Intel ssd drives.
Problems are reported for both drives, indicating that this is a software problem.

AFFECTS:
The issue started after upgrading to Ubuntu 15.04.
The issue did not happen on any Ubuntu release up to and including 14.10.

IMPACT:
errors written to syslog on bootup
emails sent to root on bootup

DIAGNOSE:
The log can be read wihtout issuesa when the computer is running:
sudo smartctl --log=selftest /dev/sda

Perhaps the drive is accessed before it is ready or other similar problem preventing the smartd commands from succeeding.

Emails are sent to root:
This message was generated by the smartd daemon running on:

   host name: c505
   DNS domain: [Empty]

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

Device: /dev/sda [SAT], Read SMART Self-Test Log Failed

Device info:
INTEL SSDSA2MH080G1GC, S/N:CVEM838000PY080DGN, WWN:5-001517-0ed62633c, FW:045C8820, 80.0 GB

For details see host's SYSLOG.

You can also use the smartctl utility for further investigation.
The original message about this issue was sent at Sat Jun 20 06:48:32 2015 PDT
Another message will be sent in 24 hours if the problem persists.

On bootup, the log displays:

Jul 3 21:47:26 c505 kernel: [1128701.864062] ata4.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
Jul 3 21:47:26 c505 kernel: [1128701.864069] ata4.00: failed command: SMART
Jul 3 21:47:26 c505 kernel: [1128701.864075] ata4.00: cmd b0/d5:01:06:4f:c2/00:00:00:00:00/00 tag 2 pio 512 in
Jul 3 21:47:26 c505 kernel: [1128701.864075] res 40/00:ff:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
Jul 3 21:47:26 c505 kernel: [1128701.864079] ata4.00: status: { DRDY }
Jul 3 21:47:26 c505 kernel: [1128701.864083] ata4: hard resetting link
Jul 3 21:47:27 c505 kernel: [1128702.356041] ata4: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Jul 3 21:47:27 c505 kernel: [1128702.356534] ata4.00: configured for UDMA/133
Jul 3 21:47:27 c505 kernel: [1128702.356552] ata4: EH complete
Jul 3 21:47:27 c505 kernel: [1128702.356782] ata4.00: Enabling discard_zeroes_data

ProblemType: Bug
DistroRelease: Ubuntu 15.04
Package: smartmontools 6.3+svn4002-2
ProcVersionSignature: Ubuntu 3.19.0-22.22-generic 3.19.8-ckt1
Uname: Linux 3.19.0-22-generic x86_64
NonfreeKernelModules: nvidia
ApportVersion: 2.17.2-0ubuntu1.1
Architecture: amd64
CurrentDesktop: GNOME-Flashback:Unity:
Date: Sat Jul 4 12:03:27 2015
ProcEnviron:
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: smartmontools
UpgradeStatus: Upgraded to vivid on 2015-06-20 (14 days ago)

[Christian Franke (christian-franke)]

> ... kernel: [1128701.864075] ata4.00: cmd b0/d5:01:06:4f:c2/...
> ... kernel: [1128701.864075] res ... Emask 0x4 (timeout)

This ATA cmd is a correct SMART READ LOG 0x06 (self-test log) command. If the command fails with "timeout", this is possibly not a smartmontools issue.

> Device info:
> INTEL SSDSA2MH080G1GC, ...

Some older Intel SSDs have known issues with the SMART READ LOG 0x00 (log directory) command: After this command was completed, the drive may hang. Smartd reads the log directory to check which log is supported, see https://www.smartmontools.org/ticket/89.

This very old X18/X25-M G1 series SSD may be affected or not. Please try whether the problem persists if '-F nologdir' is added to the smartd.conf configuration line.

[Christian Franke (christian-franke)]

This is possibly not the Intel firmware bug mentioned in my previous comment but an older one:

X18/X25-M/V G2 series with Firmware before 2CV102J8 may hang if self-test log is read. G1 series may also be affected. Smartctl and smartd print a related warning for G2 series but not for G1.

Please update the SSD firmware if possible.

See also this ticket from 2010:
"smartd Read SMART Self Test Log Failed on Intel X25-M SSD":
https://bugs.launchpad.net/ubuntu/+source/smartmontools/+bug/597518

[Richard (ismail-a)]

for the 34 nm, Intel has an update.
For the 50 nm, there is no fix from Intel.

[Robie Basak (racb)]

Christian, thank you for looking after smartmontools bugs in Ubuntu. Do you consider this a valid bug for which we should track a fix in upstream and Ubuntu?

[Christian Franke (christian-franke)]

It is possibly is a bug in the Intel X25 G1 firmware Intel did never fix (a similar bug in G2 was fixed). But I don't remember any related report since the G1 series was introduced in 2009(!).

It is possibly not a smartmontools bug, except if the problem did not occur with older versions.

Please test if '-F nologdir' changes anything. If not, test whether another problem occurs if '-l selftest' is not specified: Don't use '-a' in smartd.conf, use '-H -f -t -l error' instead.

All that could be done in smartmontools is to add another -F ( --firmwarebug) option to suppress accesses to self-test log. If desired, please create a ticket upstream. But I'm not sure whether this enhancement is worth the (non-trivial) effort for a single SSD series introduced 6+ years ago.

[Richard (ismail-a)]

I updated the G2 34 nm drive, works fine.

G1 has the problem and Intel has decided no fix.

It is true that these drives are 7 years old, however the wear indicator says there's another 14 years in them.
They sure were expensive but the only product tested to be reliable at the time.