Macro 32 Ramblings

Here’s some advice if you notice your system may be having a performance problem. Run the ls of your /proc filesystem. This will tell you how many active processes are running on your system.

If you’re dealing with a large process count, consider running either command with a short number of iterations, otherwise your monitoring attempts will only contribute to the overall load. I would argue that in a lot of cases, people running Top or Prstat in a background window are actually contributing to any performance problems that they eventually discover through the use of those tools.

http://developers.sun.com/solaris/articles/prstat.html

Topping top in Solaris 8 with prstat
By Tom Kincaid and Andrei Dorofeev, March 2001

Introduction
Different views of the system

Finding processes that are using up the CPU
Determining User Consumption
Looking at processor sets

Identifying virtual memory usage

An in-depth look at individual processes

Focusing on specific processes
Redirecting the output of prstat
Tracking down resource leaks
Getting the Resource Statistics for Each Thread Within a Process
Getting the micro statistics on a process

Other Solaris Performance Tools
Conclusion

Introduction

This article describes how the prstat utility can be used to provide views of a system’s activity and resource consumption. In addition, identifying activity and resource consumption of individual processes and threads is discussed. The advantages of prstat over top are identified. Examples of commonly encountered performance problems, and solutions that utilize prstat, are included.

This article will help:

1. Solaris developers understand how their applications are consuming system resources and how the applications are spending time. With this understanding, developers can identify and correct resource leaks in their applications and gain an understanding of how their applications can be changed to perform better on the Solaris platform.
2. Solaris system administrators understand how prstat can be used to identify the resources consumption taking place on their systems. With this understanding of resource consumption, system administrators can identify system performance problems and correct them.

Different Views of the System

When users see a noticeable drop in system performance, it is almost always because a resource is being exhausted. The resources most commonly exhausted are CPU, memory, and disks. The Solaris operating environment has many tools for identifying which resources is being overtaxed. Some common Solaris tools for monitoring a system’s usage of memory, CPU, and disks are vmstat, mpstat, and iostat respectively. In addition, the Solaris performance tool perfmeter can be used to graphically monitor system resource utilization on systems that have the appropriate graphic ability.

Note: On some Solaris servers, windowing systems are not available. For this reason, the subject of viewing resource utilization graphically will not be presented in any detail in this article.

Once you have identified which resource is being exhausted, you can use prstat to identify which processes causing it.

Finding Processes that are Using Up the CPU

Suppose you are using a Solaris system to write a document in StarOffice and you notice that it is taking a long time to process large spreadsheet calculations and other tasks.

If you suspect that the system is behaving poorly because the CPU resources are being overtaxed, a quick way to get some kernel statistics on CPU usage is to use the mpstat command.

The mpstat 5 5 command will print the CPU statistics 5 times at 5 second intervals. The following is a sample output of this command.

kincaid/tartan 49 $ mpstat 5 5

kincaid/tartan 50 $

In the output sample above, 4 of the 5 samples have CPU 0 with a combined user time and sytem time at 100 and idle time at 0 (column headings usr, sys, idl). This indicates that the CPU is completely consumed on this system.

The following is the perfmeter output for CPU, page, and disks.

After gathering the data from perfmeter and mpstat, which indicates that the system CPU resources are overtaxed, you can use prstat to identify which processes are consuming the CPU resources. The prstat -s cpu -n 5 command is used to list the five processes that are consuming the most CPU resources. The -s cpu flag tells prstat to sort the output by CPU usage. The -n 5 flag tells prstat to restrict the output to the top five processes.

kincaid/tartan 43 $ prstat -s cpu -n 5

In the example output, there is a process named cpuhog that is consuming the majority (67 %) of the CPU cycles. If this process is killed using the kill -9 13974 command, and the mpstat command is subsequently repeated, the output shows that the CPU is idle the majority of the time. The office application that was tested becomes very responsive and the spreadsheet calculations complete faster.

Note: The -s cpu option is the default setting for prstat. Therefore, if the intent is to sort output by CPU usage, specifying the -s cpu option is not necessary. For the purpose of this article, the -s cpu option is set to distinguish it from other ways of sorting the output produced by prstat.

mpstat 5 5

kincaid/tartan 53 $

Determining User Consumption

Occassionally there are many small processes, each of which consume a small piece of the CPU. On a system such as a computing server that is shared by many users, prstat can be used to determine which user (as opposed to which processes) is consuming the most resources. If the user consuming the most resources on the system can be identified, it is possible to move at least part of the work to another machine. To have prstat report statistics about resource consumption by user, add the -a option to the prstat command line.

Adding the -a option to any prstat command will identify how many processes each user is using, what percent of the CPUs, and how much memory, they are using on a system, as shown above. The command prstat -s cpu -a -n 8 asks for the top 8 processes consuming the CPU and a list of resource consumption statistics for each user.

The output below shows that user larry is consuming the most CPU resources.

kincaid/tartan 43 $ prstat -s cpu -a -n 8

kincaid/tartan 44 $

Looking at Processor Sets

The Solaris operating environment is used to host applications on multi-processor machines. Frequently, system administrators create what is known as a processor set. A processor set is a group of CPUs to which tasks can be assigned. For instance, consider a machine containing four CPUs that is being used to run a web server and an application server. One approach to partitioning the system CPU resources between these two tasks is to create two processors with two CPUs each, and to bind the web server to one processor set and the application server to the other processor set. This avoids having one server consume all the CPU resources on the machine.

Included below is a set of commands you can use to create two processor sets.

# psrset -c 0
created processor set 1
processor 0: was not assigned, now 1

# psrset -a 1 1
processor 1: was not assigned, now 1

# psrset -c 2
created processor set 2
processor 2: was not assigned, now 2

# psrset
user processor set 1: processors 0 1
user processor set 2: processor 2

After creating processor sets, prstat commands can be restricted to retrieve the process statistics for processes bound to a specific processor set. This is accomplished with the -C option. For example, prstat -C 1 -s cpu will report all the process activity for processes bound to processor set 1, and sort the results by CPU usage. This is extremely useful for identifying which processes are running on what processor sets.

Evaluating the CPU column for processor sets is also useful for determining how busy a processor set is, which then determines what processor set to assign future computing tasks to.

Note: The CPU column of prstat always reports the percentage of system CPU resources a process is consuming and not the percentage of CPU resources of a processor or a processor set, even if the -C option is specified on the command line. For example, if there is a two-processor set on a four-processor system and a prstat -C is executed on the processor set, since the processor set has 50% of the system’s CPUs, the total percentages of the CPU column will not exceed 50%.

Identifying Virtual Memory Usage

When the active computing tasks on a system require memory resources that exceed the physical memory available on the machine, the system moves units of memory called pages to disk to make memory available for an active section of a program. This is known as paging. When a system starts paging, there will be a significant drop in performance. A good tool for determining if a system is paging is sar. The sar -g command will yield the paging statistics for a given system.

A common cause for system paging is a process or group of processes that are using a majority of the system’s memory. prstat is an excellent tool for identifying which processes are consuming the majority of a system’s memory. Use prstat -s size, which is similar to the previous command, but which sorts prstat output by size instead of by CPU usage.

The following output illustrates sar -g for a system that is paging at a very high rate.

sar -g 5 5

SunOS tartan 5.8 Generic_108528-01 sun4u 02/12/01

Once it has been determined that the system performance drop off is a result of heavy paging activity, the next step is to determine which processes have introduced the increase. Also, any time scanning occurs (as indicated by the column pgscan/s in the above output) there is a memory shortage on the system. It is not easy to identify all the reasons for paging, but identifying the processes that are consuming the most virtual memory is a good start. To view the process consuming the most virtual memory, use the command prstat with the -s size option. The command prstat -s size -n 5 provides the top five processes on a system in terms of virtual memory consumption. Included below is the output from the prstat -s size -n 5 command on the system on which the above sar command was run.

tartan% prstat -s size -n 5

One process in the example is using over 1000 megabytes of virtual memory. The system only has 1 gigabyte of physical memory total. The process with ID 21307, memhog, is most likely the process that is slowing down the system.

After the kill -9 21307 command is issued to terminate the process on the system, the performance returns to normal and repeating the sar command shows that all paging and scanning have ceased, as shown in the following output.

tartan% sar -g 5 5

SunOS tartan 5.8 Generic_108528-01 sun4u 02/12/01

An In-depth Look at Individual Processes

The previous section decribed how prstat can be used to identify processes that are causing system performance issues. This section describes how to examine the details of individual processes using prstat.

The techniques presented in this section are useful for debugging certain classes of bugs and problems frequently encountered when developing or running server applications.

prstat, like top, periodically updates the screen display with a new set of statistics for processes. There are two command line options that can be used to focus in on specific processes and then monitor the process for a longer period.

Focusing on Specific Processes

To limit the statistics reported by prstat to a specific process or a set of processes, use the -p option followed by the list of process IDs. Process IDs can be obtained with the ps -ef command. This can be a useful way to gather information on a running process that you suspect has some performance issues.

Redirecting the Output of prstat

By default, prstat will update a section of a terminal screen each time it reports new statistics about the processes it is monitoring. However, if the objective is to determine if a process is leaking resources or to learn how the process behaves over time, it is useful to have multiple sets of data reports from prstat for comparison.

If you redirect the output of prstat to a file, each set of statistics produced by prstat will be preserved in the file.

Tracking Down Resource Leaks

Consider the following scenario. Suppose you suspect that the server application received from a software vendor contains a slow memory leak. To illustrate the problem to the vendor, run a prstat command that is limited to the process ID of the server application and gathers the statistics every 15 seconds. The following code example shows how you can do this.

kincaid/tartan 74 $ server &

[1] 2423
kincaid/tartan 75 $ 
kincaid/tartan 56 $ prstat -p 2443 15 > server.out &
kincaid/tartan 57 $ tail -f server.out
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  2443 kincaid    32M 1008K sleep   58    0   0:00.00 0.0% server/1
Total: 1 processes, 1 lwps, load averages: 0.00, 0.01, 0.01
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  2443 kincaid    47M 1248K sleep   58    0   0:00.00 0.0% server/1
Total: 1 processes, 1 lwps, load averages: 0.00, 0.01, 0.01
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  2443 kincaid    62M 1488K sleep   58    0   0:00.00 0.0% server/1
Total: 1 processes, 1 lwps, load averages: 0.00, 0.01, 0.01
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  2443 kincaid    77M 1728K sleep   58    0   0:00.00 0.0% server/1
Total: 1 processes, 1 lwps, load averages: 0.00, 0.01, 0.01
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  2443 kincaid    92M 1968K sleep   58    0   0:00.00 0.0% server/1
Total: 1 processes, 1 lwps, load averages: 0.00, 0.01, 0.01
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  2443 kincaid   107M 2208K sleep   58    0   0:00.00 0.0% server/1

From the data gathered by prstat, it is clear that the process is growing by roughly 15 meg. every 15 seconds. In addition, the resident set size of the process is growing by roughly 40K every 15 seconds as well. While there may be explanations for this other than a memory leak in the server application, the data, like that shown in the code example, should raise strong suspicions about there being a memory leak in the application.

Here is an example of a how prstat can be used to observe a Java server application leaking threads. prstat always lists the number of lwps (threads) in each process.

kincaid/tartan 47 $ prstat -p 3221 > threads.out &

[1] 3235
kincaid/tartan 48 $ tail -f threads.out	
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  3221 kincaid    29M   11M run     58    0   0:00.00  32% java/43
Total: 1 processes, 43 lwps, load averages: 1.29, 1.20, 1.18
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  3221 kincaid    29M   11M sleep   58    0   0:00.00  36% java/45
Total: 1 processes, 45 lwps, load averages: 1.44, 1.23, 1.19
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  3221 kincaid    29M   11M sleep   58    0   0:00.00  38% java/48
Total: 1 processes, 48 lwps, load averages: 1.43, 1.23, 1.19
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  3221 kincaid    29M   11M run     44    0   0:00.00  45% java/50
Total: 1 processes, 50 lwps, load averages: 1.65, 1.29, 1.21
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  3221 kincaid    29M   11M sleep   58    0   0:00.00  52% java/52
Total: 1 processes, 52 lwps, load averages: 2.03, 1.37, 1.24
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP       
  3221 kincaid    29M   11M run     48    0   0:00.00  55% java/55

This output shows the number of lwps increasing over time. Hence, it is possible that the server application is leaking threads.

Getting the Resource Statistics for Each Thread Within a Process.

You can also use prstat to find other resource leaks. A very nice feature of prstat is that by using the -L switch, prstat will report statistics for each thread of a process.

This has a variety of uses, including:

• The ability to see how the balance of the work is being distributed across the pool by viewing the CPU usage of each thread.
• The option to further narrow down resource leaks to individual threads and not just a process. For example you can determine which thread is leaking memory.

To illustrate, the following output shows the resource statistics for each thread of a server application.

^C
kincaid/tartan 49 $ 
kincaid/tartan 49 $ prstat -L -p 3295


kincaid/tartan 50 $


Getting the micro statistics on a process

To examine further what a process is doing, you can use the -m option to have prstat print out the micro statistics of the process. When you specify the -m option, a different set of columns is presented by prstat that represent how the process spent its time since the last sample was taken. The following table lists the micro statistic column headings and their meanings.

Following is sample output from using the -m option on the server application in the previous section.

prstat -m -p 3295
     PID USERNAME USR SYS TRP TFL DFL LCK SLP LAT VCX ICX SCL SIG PROCESS/NLWP  
  3295 kincaid  0.8 0.0 0.0 0.0 0.0 0.0  99 0.0  22   4  25   0 java/24

Explaining the meaning and implications of all these columns is beyond the scope of this paper. However, as you become familiar with operating system concepts and Solaris internals, the ability to monitor the micro statistics of an individual process or a group of processes becomes a very powerful tool, especially when trying to identify performance-related problems or issues. In addition, the -m option can be used with the -L option so the micro statistics of each thread of a process can be monitored.

Other Solaris Performance Tools

As mentioned previously, prstat can be used with other Solaris performance monitoring tools to strengthen a developer’s arsenal for debugging and diagnosing system performance issues. Some of the other Solaris tools that are worth reading about are: sar, iostat, netstat, and mpstat. All of these are great tools for identifying system performance issues. prstat is a nice complement to these tools because it helps further identify the processes and threads that may be responsible for system performance problems.

Conclusion

prstat is a great addition to the Solaris tool set. It is standard, beginning with Solaris 2.8. Developers no longer need to track down a version of top with each new release of Solaris. prstat has all of the most commonly used features of top plus several very powerful features not found in top.

System Administration Commands                         prstat(1M)



NAME
     prstat - report active process statistics

SYNOPSIS
     prstat [-acJLmRtTv] [-C psrsetlist] [-j projlist] [-k  task-
     list]  [-n ntop[,nbottom]] [-p pidlist] [-P cpulist] [-s key
     | -S key ] [-u euidlist] [-U uidlist] [-z  zoneidlist]  [-Z]
     [interval [count]]

DESCRIPTION
     The prstat utility iteratively examines all active processes
     on  the  system and reports statistics based on the selected
     output mode and sort order. prstat provides options to exam-
     ine  only processes matching specified PIDs, UIDs, zone IDs,
     CPU IDs, and processor set IDs.

     The -j, -k, -C, -p, -P, -u, -U, and -z options accept  lists
     as  arguments.  Items  in  a list can be either separated by
     commas or enclosed in quotes  and  separated  by  commas  or
     spaces.

     If you  do  not  specify  an  option,  prstat  examines  all
     processes and reports statistics sorted by CPU usage.

OPTIONS
     The following options are supported:

     -a

         Report information about processes and  users.  In  this
         mode  prstat  displays  separate reports about processes
         and users at the same time.



     -c

         Print new reports  below  previous  reports  instead  of
         overprinting them.



     -C psrsetlist

         Report only processes or lwps that are bound to  proces-
         sor  sets in the given list. Each processor set is iden-
         tified by an integer as reported by psrset(1M). The load
         averages  displayed  are the sum of the load averages of
         the specified processor sets (see  pset_getloadavg(3C)).
         Processes  with one or more LWPs bound to processor sets
         in the given list are reported even when the  -L  option
         is not used.



SunOS 5.10          Last change: 13 Sep 2005                    1






System Administration Commands                         prstat(1M)



     -j projlist

         Report only processes or lwps whose project ID is in the
         given list. Each project ID can be specified as either a
         project name or a numerical project ID.  See project(4).



     -J

         Report information about processes and projects. In this
         mode  prstat  displays  separate reports about processes
         and projects at the same time.



     -k tasklist

         Report only processes or lwps whose task ID is in  task-
         list.



     -L

         Report statistics for each light-weight  process  (LWP).
         By  default,  prstat reports only the number of LWPs for
         each process.



     -m

         Report microstate  process  accounting  information.  In
         addition to all fields listed in -v mode, this mode also
         includes the percentage of time the  process  has  spent
         processing  system  traps,  text  page faults, data page
         faults, waiting for  user  locks  and  waiting  for  CPU
         (latency time).



     -n ntop[,nbottom]

         Restrict number  of  output  lines.  The  ntop  argument
         determines  how  many lines of process or lwp statistics
         are reported, and the nbottom  argument  determines  how
         many  lines  of  user,  task, or projects statistics are
         reported if the -a, -t, -T, or -J options are specified.
         By default, prstat displays as many lines of output that
         fit in a window or terminal. When  you  specify  the  -c
         option  or  direct  the  output  to  a file, the default



SunOS 5.10          Last change: 13 Sep 2005                    2






System Administration Commands                         prstat(1M)



         values for ntop and nbottom are 15 and 5.



     -p pidlist

         Report only processes whose process ID is in  the  given
         list.



     -P cpulist

         Report only processes or lwps which have  most  recently
         executed on a CPU in the given list. Each CPU is identi-
         fied by an integer as reported by psrinfo(1M).



     -R

         Put prstat in the real time scheduling class. When  this
         option  is  used,  prstat  is  given priority over time-
         sharing and interactive processes. This option is avail-
         able only for superuser.



     -s key

         Sort output lines (that is, processes, lwps,  or  users)
         by  key in descending order. Only one key can be used as
         an argument.

         There are five possible key values:


         cpu

             Sort by process CPU usage. This is the default.




         pri

             Sort by process priority.



         rss




SunOS 5.10          Last change: 13 Sep 2005                    3






System Administration Commands                         prstat(1M)



             Sort by resident set size.



         size

             Sort by size of process image.



         time

             Sort by process execution time.



     -S key

         Sort output lines by key in  ascending  order.  Possible
         key values are the same as for the -s option. See -s.



     -t

         Report total usage summary for each  user.  The  summary
         includes  the total number of processes or LWPs owned by
         the user, total size of process images,  total  resident
         set  size, total cpu time, and percentages of recent cpu
         time and system memory.



     -T

         Report information about processes and  tasks.  In  this
         mode  prstat  displays  separate reports about processes
         and tasks at the same time.



     -u euidlist

         Report only processes whose effective user ID is in  the
         given  list.  Each  user ID may be specified as either a
         login name or a numerical user ID.



     -U uidlist

         Report only processes whose real user ID is in the given



SunOS 5.10          Last change: 13 Sep 2005                    4






System Administration Commands                         prstat(1M)



         list.  Each  user  ID may be specified as either a login
         name or a numerical user ID.



     -v

         Report  verbose  process  usage.  This   output   format
         includes the percentage of time the process has spent in
         user  mode,  in  system  mode,  and  sleeping.  It  also
         includes the number of voluntary and involuntary context
         switches,  system  calls  and  the  number  of   signals
         received.  Statistics  that  are not reported are marked
         with the - sign.



     -z zoneidlist

         Report only processes or LWPs whose zone ID  is  in  the
         given  list.  Each  zone ID can be specified as either a
         zone name or a numerical zone ID. See zones(5).



     -Z

         Report information about processes and  zones.  In  this
         mode,  prstat  displays separate reports about processes
         and zones at the same time.



OUTPUT
     The following list defines the column headings and the mean-
     ings of a prstat report:

     PID

         The process ID of the process.



     USERNAME

         The real user (login) name or real user ID.



     SIZE

         The total virtual memory size of the process,  including



SunOS 5.10          Last change: 13 Sep 2005                    5






System Administration Commands                         prstat(1M)



         all  mapped  files  and devices, in kilobytes (K), mega-
         bytes (M), or gigabytes (G).



     RSS

         The resident set size of the process (RSS), in kilobytes
         (K),  megabytes  (M), or gigabytes (G). The RSS value is
         an estimate provided by proc(4) that might underestimate
         the actual resident set size. Users who want to get more
         accurate usage information for capacity planning  should
         use the -x option to pmap(1) instead.



     STATE

         The state of the process:


         cpuN

             Process is running on CPU N.




         sleep

             Sleeping: process is waiting for an  event  to  com-
             plete.



         run

             Runnable: process in on run queue.



         zombie

             Zombie state:  process  terminated  and  parent  not
             waiting.



         stop

             Process is stopped.




SunOS 5.10          Last change: 13 Sep 2005                    6






System Administration Commands                         prstat(1M)



     PRI

         The priority of the process. Larger numbers mean  higher
         priority.



     NICE

         Nice value used in priority computation. Only  processes
         in certain scheduling classes have a nice value.



     TIME

         The cumulative execution time for the process.



     CPU

         The percentage of recent CPU time used by  the  process.
         If executing in a non-global zone and the pools facility
         is active, the percentage will be that of the processors
         in  the  processor  set  in use by the pool to which the
         zone is bound.



     PROCESS

         The name of the process (name of executed file).



     LWPID

         The lwp ID of the lwp being reported.



     NLWP

         The number of lwps in the process.



     With the some options, in addition to a number of the column
     headings shown above, there are:





SunOS 5.10          Last change: 13 Sep 2005                    7






System Administration Commands                         prstat(1M)



     NPROC           Number of processes in a  specified  collec-
                     tion.



     MEMORY          Percentage of memory  used  by  a  specified
                     collection of processes.



     The following columns are displayed when the -v or -m option
     is specified

     USR             The percentage of time the process has spent
                     in user mode.



     SYS             The percentage of time the process has spent
                     in system mode.



     TRP             The percentage of time the process has spent
                     in processing system traps.



     TFL             The percentage of time the process has spent
                     processing text page faults.



     DFL             The percentage of time the process has spent
                     processing data page faults.



     LCK             The percentage of time the process has spent
                     waiting for user locks.



     SLP             The percentage of time the process has spent
                     sleeping.



     LAT             The percentage of time the process has spent
                     waiting for CPU.





SunOS 5.10          Last change: 13 Sep 2005                    8






System Administration Commands                         prstat(1M)



     VCX             The number of voluntary context switches.



     ICX             The number of involuntary context switches.



     SCL             The number of system calls.



     SIG             The number of signals received.



     Under the -L option, one line is printed for each lwp in the
     process  and  some  reporting fields show the values for the
     lwp, not the process.

OPERANDS
     The following operands are supported:

     count           Specifies  the  number  of  times  that  the
                     statistics  are repeated. By default, prstat
                     reports statistics until a termination  sig-
                     nal is received.



     interval        Specifies the sampling interval in  seconds;
                     the default interval is 5 seconds.



EXAMPLES
     Example  1:  Reporting  the  Five  Most  Active   Super-User
     Processes

     The following command reports the five  most  active  super-
     user processes running on CPU1 and CPU2:

     example% prstat -u root -n 5 -P 1,2 1 1

     PID   USERNAME  SIZE   RSS STATE  PRI  NICE      TIME  CPU PROCESS/LWP
      306   root     3024K 1448K sleep   58    0   0:00.00 0.3% sendmail/1
      102   root     1600K  592K sleep   59    0   0:00.00 0.1% in.rdisc/1
      250   root     1000K  552K sleep   58    0   0:00.00 0.0% utmpd/1
      288   root     1720K 1032K sleep   58    0   0:00.00 0.0% sac/1
        1   root      744K  168K sleep   58    0   0:00.00 0.0% init/1
     TOTAL:       25, load averages:  0.05, 0.08, 0.12




SunOS 5.10          Last change: 13 Sep 2005                    9






System Administration Commands                         prstat(1M)



     Example 2: Displaying Verbose Process Usage Information

     The following command displays verbose process usage  infor-
     mation  about processes with lowest resident set sizes owned
     by users root and john.

     example% prstat -S rss -n 5 -vc -u root,john

      PID USERNAME USR SYS TRP TFL DFL LCK SLP LAT VCX ICX SCL SIG PROCESS/LWP
        1 root     0.0 0.0  -   -   -   -  100  -    0   0   0   0 init/1
      102 root     0.0 0.0  -   -   -   -  100  -    0   0   3   0 in.rdisc/1
      250 root     0.0 0.0  -   -   -   -  100  -    0   0   0   0 utmpd/1
     1185 john     0.0 0.0  -   -   -   -  100  -    0   0   0   0 csh/1
      240 root     0.0 0.0  -   -   -   -  100  -    0   0   0   0 powerd/4
      TOTAL:       71, load averages:  0.02, 0.04, 0.08


EXIT STATUS
     The following exit values are returned:

     0               Successful completion.



     1               An error occurred.



ATTRIBUTES
     See attributes(5) for descriptions of the  following  attri-
     butes:

     ____________________________________________________________
    | ATTRIBUTE TYPE              | ATTRIBUTE VALUE             |
    |_____________________________|_____________________________|
    | Availability                | SUNWcsu                     |
    |_____________________________|_____________________________|


SEE ALSO
     proc(1),       psrinfo(1M),       psrset(1M),       sar(1M),
     pset_getloadavg(3C),   proc(4),  project(4),  attributes(5),
     zones(5)

NOTES
     The snapshot of system usage displayed  by  prstat  is  true
     only  for  a split-second, and it may not be accurate by the
     time it is displayed.  When  the  -m  option  is  specified,
     prstat  tries to turn on microstate accounting for each pro-
     cess; the original state is restored when prstat exits.  See
     proc(4)  for  additional  information  about  the microstate
     accounting facility.



SunOS 5.10          Last change: 13 Sep 2005                   10






System Administration Commands                         prstat(1M)



     The total memory size reported in the SIZE and  RSS  columns
     for  groups  of  processes  can  sometimes  overestimate the
     actual amount of memory used by processes with shared memory
     segments.