Send controls to a SES device (via a SCSI SEND DIAGNOSTIC command)
or fetches status (via a SCSI RECEIVE DIAGNOSTIC RESULTS command).
The DEVICE should be a SES device which may be a dedicated
enclosure services processor (INQUIRY peripheral device type 0xd) or
attached to another type of SCSI device (e.g. a disk) in which case
the EncServ bit set in its INQUIRY response.
If no options are given (only the DEVICE argument) then all
diagnostic pages supported by the device (including SES pages)
The most recent reference for this utility is the draft SCSI Enclosure
Services 3 document T10/2149-D Revision 2 at http://ww.t10.org . Existing
standards for SES and SES-2 are ANSI INCITS 305-1998 and ANSI INCITS
Arguments to long options are mandatory for short options as well.
The options are arranged in alphabetical order based on the long
some control pages need byte 1 (i.e. the second byte) of the cdb set.
Only required in rare cases when the --control option is also set.
Default is 0; B1 is in decimal unless it is prefixed by '0x'
or '0X' (or has a trailing 'h' or 'H').
will send control information to the DEVICE via a SCSI SEND
DIAGNOSTIC command. Cannot give both this option and --status.
The Enclosure control, String Out, Threshold Out, Array control (obsolete
in SES-2) and Subenclosure String Out diagnostic pages can be set currently.
permits a string of comma separated (ASCII) hex bytes to be specified (limit
1024). A (single) space separated string of hex bytes is also allowed but
the list needs to be in quotes. This allows the parameters to a control
diagnostic page to be specified. The string given should not include the
first 4 bytes (i.e. page code and length). See next entry for using stdin.
reads a data string from stdin, limit 1024 bytes. Spaces, tabs and line feeds
additionally are permitted as separators.
cuts down on the amount of output from the enclosure status diagnostic
page. When this option is given, any line which has all its binary flags
cleared (i.e. 0) is filtered out (i.e. ignored). If a line has some other
value on it (e.g. a temperature) then it is output.
output the usage message then exit.
output the response in hexadecimal.
the outer levels of a status diagnostic page are decoded and printed out
but the innermost level (e.g. the element status descriptor) is output in
hex. Implemented for the more complex diagnostic pages.
list all known diagnostic page names and SES elements. DEVICE
is ignored and utility exits.
where PG is a page code. Assumed to be in decimal unless prefixed by
0x for hex. Valid range is 0 to 255 (0x0 to 0xff) inclusive. Default is
page_code 0 (i.e. "Supported diagnostic pages").
outputs the chosen status page in (ASCII) hex in a format suitable for
a later invocation using the --data= option. A status diagnostic
page less its first 4 bytes (page code and length) is output.
When used twice (e.g. -rr) outputs full diagnostic page in binary
will fetch status diagnostic page from the DEVICE via a SCSI RECEIVE
DIAGNOSTIC RESULTS command. If this option is not given and
--control is not given then --status is assumed.
increase the level of verbosity, (i.e. debug output).
print the version string and then exit.
Currently all status pages, control pages and element types defined in
SES-2 revision 17 (14th May 2007) are decoded.
This utility can be used to fetch arbitrary (i.e. non SES) diagnostic
pages (using the SCSI READ DIAGNOSTIC command). To this end the
--page=PG and --hex options would be appropriate. Arbitrary
diagnostic pages can be sent to a device with the sg_senddiag utility.
There is a related command set called SAF-TE (SCSI attached fault-tolerant
enclosure) for enclosure (including RAID) status and control. SCSI devices
that support SAF-TE report "Processor" peripheral device type (0x3) in their
INQUIRY response. See the sg_safte utility in this package or safte-monitor
on the internet.
These examples use Linux device names. For suitable device names in
other supported Operating Systems see the sg3_utils(8) man page.
To view the supported pages:
To view the configuration page:
sg_ses --page=1 /dev/sda
To view the status page:
sg_ses --page=2 /dev/sda
Changing a temperature threshold is possible, if a little awkward. The
current thresholds can be shown with:
sg_ses --page=5 /dev/sda
The threshold to be changed can be chosen. Then output the threshold page
in hex (suitable for editing) with:
sg_ses --page=5 --raw /dev/sda > t
Then with the aid of the SES-2 document (in revision 9: section 6.1.8)
use your favourite editor to change t. The change can be sent to the
sg_ses --control --page=5 --data=- /dev/sda < t
If the above is successful, the threshold should have been changed. To
sg_ses --page=5 /dev/sda
The exit status of sg_ses is 0 when it is successful. Otherwise see
the sg3_utils(8) man page.