What to Look for in vmstat
Oracle tips by Burleson
As you can see, each dialect
of vmstat reports different information about the current status of the server.
Despite these dialect differences, there are only a small number of metrics that
are important for server monitoring. These metrics include:
r (runqueue) The
runqueue value shows the number of tasks executing and waiting for CPU
resources. When this number exceeds the number of CPUs on the server, a CPU
bottleneck exists, and some tasks are waiting for execution.
pi (page in) A
page-in operation occurs when the server is experiencing a shortage of RAM
memory. While all virtual memory servers will page out to the swap disk,
page-in operations show that the servers has exceeded the available RAM
storage. Any nonzero value for pi indicates excessive activity as RAM memory
contents are read in from the swap disk.
us (user CPU) This is
the amount of CPU that is servicing user tasks.
sy (system CPU) This
is the percentage of CPU being used to service system tasks.
id (idle) This
is the percentage of CPU that is idle.
wa (wait?IBM-AIX only) This
shows the percentage of CPU that is waiting on external operations such as
Note that all of the CPU metrics are expressed as
percentages. Hence, all of the CPU values (us + sy + id + wa) will always sum to
100. Now that we have a high-level understanding of the important vmstat data,
let's look into some methods for using vmstat to identify server problems.
CPU Bottlenecks with vmstat
Waiting CPU resources can be shown in UNIX
vmstat command output as the second column under the kthr (kernel thread
state change) heading. Tasks may be placed in the wait queue (b) if they are
waiting on a resource, while other tasks appear in the run queue (r) column. As
we see in Figure 5-1, server tasks are queued for execution by the server.
Figure 5-23: Tasks queuing for service by the CPUs
In short, the server is experiencing a CPU bottleneck
when r is greater than the number of CPUs on the server. To see the number of
CPUs on the server, you can use one of the following UNIX commands.
Display the Number of CPUs in IBM AIX and HP/UX
In AIX and HP/UX the
lsdev command can be used to see the number of CPUs on a server. This is
very important, because it shows the number of Parallel Query processes that can
be used on that server. That, in turn, limits the value that you can use
following the DEGREE keyword in a Parallel Query or DML statement. The following
example is taken from an AIX server, and shows that the server has four CPUs:
Display Number of CPUs in Solaris
In Solaris, the prsinfo
command can be used to count the number of CPUs on the processor. Here we see
that we have two CPUs on this server:
"Status of processor"|wc -l
Display Number of CPUs in Linux
To see the number of CPUs on
a Linux server, you can cat the /proc/cpuinfo file. In the example
here we see that our Linux server has four CPUs:
Remember that we need to
know the number of CPUs on our server because the vmstat runqueue value must
never exceed the number of CPUs. A runqueue value of 32 is perfectly acceptable
for a 36-CPU server, while a value of 32 would be a serious problem for a 24-CPU
In the following example, we run the vmstat utility. For
our purposes, we are interested in the first two columns: the run queue r, and
the kthr wait b column. In the next listing we see that there are an average of
about eight new tasks entering the run queue every five seconds (the r column),
while there are five other tasks that are waiting on resources (the b column).
Also, a nonzero value in the b column may indicate a bottleneck.
vmstat 5 5
kthr memory page faults cpu
----- ----------- ------------------------ ------------ -----------
r b avm fre re pi po fr sr cy in sy
cs us sy id wa
7 5 220214 141 0 0 0 42 53 0 1724 12381 2206 19
46 28 7
9 5 220933 195 0 0 1 216 290 0 1952 46118 2712 27
55 13 5
13 5 220646 452 0 0 1 33 54 0 2130 86185 3014 30
59 8 3
6 5 220228 672 0 0 0 0 0 0 1929 25068
2485 25 49 16 10
The rule for identifying a server with
CPU resource problems is quite simple. Whenever the value of the runqueue r
column exceeds the number of CPUs on the server, tasks are forced to wait for
execution. There are several solutions to
managing CPU overload, and these alternatives are presented in their order of
Add more processors (CPUs) to the server.
Load balance the system tasks by
rescheduling large batch tasks to execute during off-peak
Adjust the dispatching priorities (nice
values) of existing tasks.
To understand how
dispatching priorities work, we must remember that incoming tasks are placed in
the execution queue according to their nice value (see Figure 5-2). Here we see
that tasks with a low nice value are scheduled for execution above those tasks
with a higher nice value.
Figure 5-24: Tasks queued for execution according to their
We will go into detail about these resolutions later in
this chapter. Now that we can see when the CPUs are overloaded, let's look into
vmstat further and see how we can tell when the CPUs are running at full
This is an excerpt from "Oracle9i
High Performance tuning with STATSPACK" by Oracle Press.