smbstat
— report
SMB server statistics
smbstat |
[-ctu ] [-r
[-anz ]] [interval] |
The smbstat
command shows statistical
information for the SMB server, including any or all of the following four
categories: counters, throughput,
utilization,
requests.
By default, smbstat
shows
throughput and utilization.
-c
- Display counters. The columns shown are:
- nbt
- NetBIOS connections
- tcp
- TCP connections
- users
- logged on users
- trees
- share connections
- files
- open files and directories
- pipes
- open named pipes
-r
- Display
request
data, one row for each SMB command. The columns shown are, for each
request type:
- (name)
- command name
- code
- command code
- %
- % of requests that fall in this row
- rbytes/s
- received bytes per second
- tbytes/s
- transmitted bytes per second
- reqs/s
- requests per second
- rt-mean
- response time average
- rt-stddev
- response time standard deviation
-t
- Display throughput. The columns shown are:
- rbytes/s
- received bytes per second
- tbytes/s
- transmitted bytes per second
- reqs/s
- requests per second
- reads/s
- number of read requests per second
- writes/s
- number of write requests per second
-u
- Display utilization. The columns shown are:
- wcnt
- average number of waiting requests
- rcnt
- average number of running requests
- wtime
- average wait time per request
- rtime
- average run time per request
- w%
- % of time there were waiting requests
- r%
- % of time there were running requests
- u%
- utilization, computed as
rcnt/max_workers
- sat
- has the server been "saturated" (u% at 100)
- usr%
- % of CPU time in user space
- sys%
- % of CPU time in kernel
- idle%
- % of CPU time spent idle
The -r
option supports additional
modifiers including:
-a
- show "all" request types (including unsupported ones)
-n
- "name" order (sort by request name)
-z
- suppress "zero" count rows (types for which none were
received)
- interval
- When interval is specified,
smbstat
runs in a loop, printing the requested
output every interval seconds. When no
interval is specified, the statistics presented are
based on running averages accumulated since the system started. The first
output shows the same cumulative statistics one would see without the
interval specified, and subsequent outputs represent
the activity in the interval that just finished.