iostat - report I/O statistics
/usr/bin/iostat [-cCdDeEiImMnpPfFrstxXYz] [-l n] [-T u | d]
[ disk]... [interval [count]]
The
iostat utility iteratively reports terminal, disk, and tape
I/O activity, as well as
CPU utilization. The first line of
output is for all time since boot; each subsequent line is for the prior
interval only.
To compute this information, the kernel maintains a number of counters. For each
disk, the kernel counts reads, writes, bytes read, and bytes written. The
kernel also takes hi-res time stamps at queue entry and exit points, which
allows it to keep track of the residence time and cumulative residence-length
product for each queue. Using these values,
iostat produces highly
accurate measures of throughput, utilization, queue lengths, transaction rates
and service time. For terminals collectively, the kernel simply counts the
number of input and output characters.
During execution of the kernel status command, the
state of the system
can change. If relevant, a state change message is included in the
iostat output, in one of the following forms:
<<device added: sd0>>
<<device removed: sd0>>
<<partition added: sd0,a>>
<<partition removed: sd0,a>>
<<NFS mounted: nfs1>>
<<NFS unmounted: nfs1>>
<<multi-path added: ssd4>>
<<multi-path removed: ssd4>>
<<controller added: c1>>
<<controller removed: c1>>
<<processors added: 1, 3>>
<<processors removed: 1, 3>>
Note that the
names printed in these state change messages are affected
by the
-n and
-m options as appropriate.
For more general system statistics, use
sar(1),
sar(1M), or
vmstat(1M).
The output of the
iostat utility includes the following information.
device
name of the disk
r/s
reads per second
w/s
writes per second
kr/s
kilobytes read per second
The average I/O size during the interval can be computed from
kr/s
divided by
r/s.
kilobytes written per second
The average I/O size during the interval can be computed from
kw/s
divided by
w/s.
average number of transactions waiting for service (queue
length)
This is the number of I/O operations held in the device driver queue waiting for
acceptance by the device.
average number of transactions actively being serviced
(removed from the queue but not yet completed)
This is the number of I/O operations accepted, but not yet serviced, by the
device.
average response time of transactions, in milliseconds
The
svc_t output reports the overall
response time, rather than
the
service time, of a device. The overall time includes the time that
transactions are in queue and the time that transactions are being serviced.
The time spent in queue is shown with the
-x option in the
wsvc_t output column. The time spent servicing transactions is the true
service time. Service time is also shown with the
-x option and appears
in the
asvc_t output column of the same report.
percent of time there are transactions waiting for
service (queue non-empty)
%b
percent of time the disk is busy (transactions in
progress)
wsvc_t
average service time in wait queue, in milliseconds
asvc_t
average service time of active transactions, in
milliseconds
The following options are supported:
-c
Report the percentage of time the system has spent in
user mode, in system mode, in dtrace probes, and idling. See the NOTES section
and mpstat(1m) for more information.
-C
When the -x option is also selected, report
extended disk statistics aggregated by controller id.
-d
For each disk, report the number of kilobytes transferred
per second, the number of transfers per second, and the average service time
in milliseconds.
-D
For each disk, report the reads per second, writes per
second, and percentage disk utilization.
-e
Display device error summary statistics. The total
errors, hard errors, soft errors, and transport errors are displayed.
-E
Display all device error statistics.
-i
In -E output, display the Device ID instead
of the Serial No. The Device Id is a unique identifier
registered by a driver through ddi_devid_register(9F).
-I
Report the counts in each interval, rather than rates
(where applicable).
-l n
Limit the number of disks included in the report to
n; the disk limit defaults to 4 for -d and -D, and
unlimited for -x. Note: disks explicitly requested (see disk
below) are not subject to this disk limit.
-m
Report file system mount points. This option is most
useful if the -P or -p option is also specified or used in
conjunction with -Xn or -en. The -m option is useful only
if the mount point is actually listed in the output. This option can only be
used in conjunction with the -n option.
-M
Display data throughput in MB/sec instead of
KB/sec.
-n
Display names in descriptive format. For example,
cXtYdZ,
rmt/N,
server:/export/path.
By default, disks are identified by instance names such as
ssd23 or
md301. Combining the
-n option with the
-x option causes
disk names to display in the
cXtYdZsN format which is more easily
associated with physical hardware characteristics. The
cXtYdZsN format
is particularly useful in FibreChannel (FC) environments where the FC World
Wide Name appears in the
t field.
For each disk, report per-partition statistics in
addition to per-device statistics.
-P
For each disk, report per-partition statistics only, no
per-device statistics.
-f
Report ZFS-level statistics for ZFS pool and individual
vdevs in addition to per-device statistics.
-F
Report ZFS pool and individual physical vdevs statistics
only, no per-device statistics.
-r
Display data in a comma-separated format.
-s
Suppress messages related to state changes.
-t
Report the number of characters read and written to
terminals per second.
-T u | d
Display a time stamp.
Specify
u for a printed representation of the internal representation of
time. See
time(2). Specify
d for standard date format. See
date(1).
For disks under scsi_vhci(7D) control, in addition
to disk lun statistics, also report statistics for
lun.controller.
-x
Report extended disk statistics. By default, disks are
identified by instance names such as
ssd23 or
md301. Combining
the
x option with the
-n option causes disk names to display in
the
cXtYdZsN format, more easily associated with physical hardware
characteristics. Using the
cXtYdZsN format is particularly helpful in
the FibreChannel environments where the FC World Wide Name appears in the
t field.
If no output display is requested (no
-x,
-e,
-E),
-x is implied.
For disks under
scsi_vhci(7D) control, in addition
to disk
lun statistics, also report statistics for
lun.
targetport and
lun.
targetport.
controller.
In
-n (descriptive) mode the
targetport is shown in using the
target-port property of the path. Without
-n the
targetport is shown using the shorter
port-id. All target ports
with the same
target-port property value share the same
port-id.
The
target-port-to-
port-id association does not persist across
reboot.
If no output display is requested (no
-x,
-e,
-E),
-x is implied.
Do not print lines whose underlying data values are all
zeros.
The option set
-xcnCXTdz interval is particularly useful for
determining whether disk I/O problems exist and for identifying problems.
The following operands are supported:
count
Display only count reports.
disk
Explicitly specify the disks to be reported; in addition
to any explicit disks, any active disks up to the disk limit (see -l
above) will also be reported.
interval
Report once each interval seconds.
Example 1 Using
iostat to Generate User and System Operation
Statistics
The following command displays two reports of extended device statistics,
aggregated by
controller id, for user (
us) and system
(
sy) operations. Because the
-n option is used with the
-x option, devices are identified by controller names.
example% iostat -xcnCXTdz 5
Mon Nov 24 14:58:36 2003
cpu
us sy dt id
14 31 0 20
extended device statistics
r/s w/s kr/s kw wait actv wsvc_t asvc_t %w %b device
3.8 29.9 145.8 44.0 0.0 0.2 0.1 6.4 0 5 c0
666.3 814.8 12577.6 17591.1 91.3 82.3 61.6 55.6 0 2 c12
180.0 234.6 4401.1 5712.6 0.0 147.7 0.0 356.3 0 98 d10
Mon Nov 24 14:58:41 2003
cpu
us sy dt id
11 31 1 22
extended device statistics
r/s w/s kr/s kw wait actv wsvc_t asvc_t %w %b device
0.8 41.0 5.2 20.5 0.0 0.2 0.2 4.4 0 6 c0
565.3 581.7 8573.2 10458.9 0.0 26.6 0.0 23.2 0 3 c12
106.5 81.3 3393.2 1948.6 0.0 5.7 0.0 30.1 0 99 d10
Example 2 Using
iostat to Generate TTY Statistics
The following command displays two reports on the activity of five disks in
different modes of operation. Because the
-x option is used, disks are
identified by instance names.
example% iostat -x tc 5 2
extended device statistics tty cpu
device r/s w/s kr/s kw/s wait actv svc_t %w %b tin tout us sy dt id
sd0 0.4 0.3 10.4 8.0 0.0 0.0 36.9 0 1 0 10 0 0 0 99
sd1 0.0 0.0 0.3 0.4 0.0 0.0 35.0 0 0
sd6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0
nfs1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0
nfs2 0.0 0.0 0.0 0.1 0.0 0.0 35.6 0 0
extended device statistics tty cpu
device r/s w/s kr/s kw/s wait actv svc_t %w %b tin tout us sy dt id
sd0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 0 155 0 0 0 100
sd1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0
sd6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0
nfs1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0
nfs2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0
Example 3 Using
iostat to Generate Partition and Device Statistics
The following command generates partition and device statistics for each disk.
Because the
-n option is used with the
-x option, disks are
identified by controller names.
example% iostat -xnp
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.4 0.3 10.4 7.9 0.0 0.0 0.0 36.9 0 1 c0t0d0
0.3 0.3 9.0 7.3 0.0 0.0 0.0 37.2 0 1 c0t0d0s0
0.0 0.0 0.1 0.5 0.0 0.0 0.0 34.0 0 0 c0t0d0s1
0.0 0.0 0.0 0.1 0.0 0.0 0.6 35.0 0 0 fuji:/export/home/user3
Example 4 Show Translation from Instance Name to Descriptive Name
The following example illustrates the use of
iostat to translate a
specific instance name to a descriptive name.
example% iostat -xn sd1
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 c8t1d0
Example 5 Show Target Port and Controller Activity for a Specific Disk
In the following example, there are four controllers, all connected to the same
target port.
# iostat -Y ssd22
extended device statistics
device r/s w/s kr/s kw/s wait actv svc_t %w %b
ssd22 0.2 0.0 1.5 0.0 0.0 0.0 0.7 0 0
ssd22.t2 0.2 0.0 1.5 0.0 0.0 0.0 0.0 0 0
ssd22.t2.fp0 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0 0
ssd22.t2.fp1 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0 0
ssd22.t2.fp2 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0 0
ssd22.t2.fp3 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0 0
See
attributes(5) for descriptions of the following attributes:
| ATTRIBUTE TYPE |
ATTRIBUTE VALUE |
|
| Interface Stability |
See below. |
Invocation is evolving. Human readable output is unstable.
date(1),
sar(1),
sar(1M),
mpstat(1M),
vmstat(1M),
time(2),
attributes(5),
scsi_vhci(7D)
The sum of
CPU utilization might vary slightly from 100 because of
rounding errors in the production of a percentage figure.
The
svc_t response time is not particularly significant when the
I/0 (
r/s+
w/s) rates are under 0.5 per second. Harmless
spikes are fairly normal in such cases.
The
mpstat utility reports the same
dt,
usr, and
sys
statistics. See
mpstat(1M) for more information.
When executed in a
zone and if the pools facility is active,
iostat(1M) will only provide information for those processors in the
processor set of the pool to which the
zone is bound.