Skip to content

P
Postgres FD Implementation
Commit c111a721 authored by Neil Conway's avatar Neil Conway
Browse files
Options

Update documentation for recent DTrace changes. Patch from Robert Lor.

parent 220db7cc
Show whitespace changes
Inline Side-by-side
Showing with 131 additions and 81 deletions
doc/src/sgml/installation.sgml
View file @ c111a721
<!-- $PostgreSQL: pgsql/doc/src/sgml/installation.sgml,v 1.304 2008/03/10 20:06:27 tgl Exp $ -->
<!-- $PostgreSQL: pgsql/doc/src/sgml/installation.sgml,v 1.305 2008/03/25 22:50:27 neilc Exp $ -->
<chapter id="installation">
<title><![%standalone-include[<productname>PostgreSQL</>]]>
......@@ -1193,8 +1193,8 @@ su - postgres
<primary>DTrace</primary>
</indexterm>
Compiles with support for the dynamic tracing tool DTrace.
Operating system support for DTrace is currently only
available in Solaris.
Operating system support for DTrace is currently available in
Solaris and Mac OS X Leopard.
</para>
<para>
......
doc/src/sgml/monitoring.sgml
View file @ c111a721
<!-- $PostgreSQL: pgsql/doc/src/sgml/monitoring.sgml,v 1.55 2007/11/28 15:42:31 petere Exp $ -->
<!-- $PostgreSQL: pgsql/doc/src/sgml/monitoring.sgml,v 1.56 2008/03/25 22:50:27 neilc Exp $ -->
<chapter id="monitoring">
<title>Monitoring Database Activity</title>
......@@ -947,26 +947,26 @@ SELECT pg_stat_get_backend_pid(s.backendid) AS procpid,
</para>
<para>
A number of trace points, often called probes, are already inserted
into the source code. By default these probes are disabled, and
the user needs to explicitly tell the configure script to make the
probes available in <productname>PostgreSQL</productname>.
A number of probes or trace points are already inserted
into the source code. By default these probes are not compiled into the
binary, and the user needs to explicitly tell the configure script to make
the probes available in <productname>PostgreSQL</productname>.
</para>
<para>
Currently, only the DTrace utility is supported, which is only available
on Solaris Express and Solaris 10+. It is expected that DTrace will
be available in the future on FreeBSD and Mac OS X.
Currently, only the DTrace utility is supported, which is available
on Solaris Express, Solaris 10, and Mac OS X Leopard. It is expected that
DTrace will be available in the future on FreeBSD.
Supporting other dynamic tracing utilities is theoretically possible by
changing the definitions for the <literal>PG_TRACE</> macros in
<filename>src/include/pg_trace.h</>.
changing the definitions for the macros in
<filename>src/include/utils/probes.h</>.
</para>
<sect2 id="compiling-for-trace">
<title>Compiling for Dynamic Tracing</title>
<para>
By default, trace points are disabled, so you will need to
By default, probes are not available, so you will need to
explicitly tell the configure script to make the probes available
in <productname>PostgreSQL</productname>. To include DTrace support
specify <option>--enable-dtrace</> to configure. See <xref
......@@ -975,16 +975,16 @@ SELECT pg_stat_get_backend_pid(s.backendid) AS procpid,
</sect2>
<sect2 id="trace-points">
<title>Built-in Trace Points</title>
<title>Built-in Probes</title>
<para>
A few standard trace points are provided in the source code
A few standard probes are provided in the source code
(of course, more can be added as needed for a particular problem).
These are shown in <xref linkend="trace-point-table">.
</para>
<table id="trace-point-table">
<title>Built-in Trace Points</title>
<title>Built-in Probes</title>
<tgroup cols="3">
<thead>
<row>
......@@ -996,67 +996,67 @@ SELECT pg_stat_get_backend_pid(s.backendid) AS procpid,
<tbody>
<row>
<entry>transaction__start</entry>
<entry>transaction-start</entry>
<entry>(int transactionId)</entry>
<entry>The start of a new transaction.</entry>
</row>
<row>
<entry>transaction__commit</entry>
<entry>transaction-commit</entry>
<entry>(int transactionId)</entry>
<entry>The successful completion of a transaction.</entry>
</row>
<row>
<entry>transaction__abort</entry>
<entry>transaction-abort</entry>
<entry>(int transactionId)</entry>
<entry>The unsuccessful completion of a transaction.</entry>
</row>
<row>
<entry>lwlock__acquire</entry>
<entry>lwlock-acquire</entry>
<entry>(int lockid, int mode)</entry>
<entry>An LWLock has been acquired.</entry>
</row>
<row>
<entry>lwlock__release</entry>
<entry>lwlock-release</entry>
<entry>(int lockid, int mode)</entry>
<entry>An LWLock has been released.</entry>
</row>
<row>
<entry>lwlock__startwait</entry>
<entry>lwlock-startwait</entry>
<entry>(int lockid, int mode)</entry>
<entry>An LWLock was not immediately available and a backend
has begun to wait for the lock to become available.
</entry>
</row>
<row>
<entry>lwlock__endwait</entry>
<entry>lwlock-endwait</entry>
<entry>(int lockid, int mode)</entry>
<entry>A backend has been released from its wait for an LWLock.
</entry>
</row>
<row>
<entry>lwlock__condacquire</entry>
<entry>lwlock-condacquire</entry>
<entry>(int lockid, int mode)</entry>
<entry>An LWLock was successfully acquired when the caller specified no
waiting.
</entry>
</row>
<row>
<entry>lwlock__condacquire__fail</entry>
<entry>lwlock-condacquire-fail</entry>
<entry>(int lockid, int mode)</entry>
<entry>An LWLock was not successfully acquired when the caller specified
no waiting.
</entry>
</row>
<row>
<entry>lock__startwait</entry>
<entry>lock-startwait</entry>
<entry>(int locktag_field2, int lockmode)</entry>
<entry>A request for a heavyweight lock (lmgr lock) has begun to wait
because the lock is not available.
</entry>
</row>
<row>
<entry>lock__endwait</entry>
<entry>lock-endwait</entry>
<entry>(int locktag_field2, int lockmode)</entry>
<entry>A request for a heavyweight lock (lmgr lock) has finished waiting
(i.e., has acquired the lock).
......@@ -1068,7 +1068,7 @@ SELECT pg_stat_get_backend_pid(s.backendid) AS procpid,
</sect2>
<sect2 id="using-trace-points">
<title>Using Trace Points</title>
<title>Using Probes</title>
<para>
The example below shows a DTrace script for analyzing transaction
......@@ -1096,11 +1096,9 @@ postgresql$1:::transaction-commit
self->ts=0;
}
</programlisting>
Note how the double underline in trace point names needs to
be replaced by a hyphen when using D script.
When executed, the example D script gives output such as:
<screen>
# ./txn_count.d `pgrep -n postgres`
# ./txn_count.d `pgrep -n postgres` or ./txn_count.d &lt;PID&gt;
^C
Start 71
......@@ -1109,7 +1107,7 @@ Total time (ns) 2312105013
</screen>
</para>
<para>
You should remember that trace programs need to be carefully written and
You should remember that DTrace scripts need to be carefully written and
debugged prior to their use, otherwise the trace information collected might
be meaningless. In most cases where problems are found it is the
instrumentation that is at fault, not the underlying system. When
......@@ -1119,30 +1117,86 @@ Total time (ns) 2312105013
</sect2>
<sect2 id="defining-trace-points">
<title>Defining Trace Points</title>
<title>Defining New Probes</title>
<para>
New trace points can be defined within the code wherever the developer
desires, though this will require a recompilation.
New probes can be defined within the code wherever the developer
desires, though this will require a recompilation. Below are the steps
for inserting new probes:
</para>
<procedure>
<step>
<para>
A trace point can be inserted by using one of the
trace macros. These are chosen according to how many variables will
be made available for inspection at that trace point. Tracing the
occurrence of an event can be achieved with a single line, using
just the trace point name, e.g.:
<programlisting>
PG_TRACE (my__new__trace__point);
</programlisting>
More complex trace points can be provided with one or more variables
for inspection by the dynamic tracing utility by using the
<literal>PG_TRACE</><replaceable>n</> macro that corresponds to the number
of parameters after the trace point name:
Decide on probe names and data to be made available through the probes
</para>
</step>
<step>
<para>
Add the probe definitons to <filename>src/backend/src/utils/probes.d</>
</para>
</step>
<step>
<para>
Insert a one-line probe macros at the desired locations in the source code
</para>
</step>
<step>
<para>
Recompile and verify that the new probes are available
</para>
</step>
</procedure>
<formalpara>
<title>Example:</title>
<para>
Here is an example of how you would add a probe to trace all new transactions by transaction ID.
</para>
</formalpara>
<procedure>
<step>
<para>
Name the probe transaction-start and give it a parameter of type integer (type of transaction ID)
</para>
</step>
<step>
<para>
Add <quote>probe transaction__start(int);</quote> to
<filename>src/backend/src/utils/probes.d</>, and it should look like the following:
<programlisting>
PG_TRACE3 (my__complex__event, varX, varY, varZ);
provider postgresql {
...
probe transaction__start(int);
...
};
</programlisting>
The definition of the transaction__start trace point is shown below:
Note the use of the double underline in the probe name. In the DTrace
script, the double underline needs to be replaced with a hyphen.
</para>
<para>
You should take care that the data types specified for the probe arguments
match the data types of the variables used in the macro. Otherwise, you
will get compilation errors.
</para>
</step>
<step>
<para>
At compile time, transaction__start is converted to a macro called
TRACE_POSTGRESQL_TRANSACTION_START, and it resides in
<filename>src/backend/src/utils/probes.h</>. Before recompiling, add
the single line macro to the appropriate location in the source code.
In this case, it looks like the following:
</para>
<para>
<programlisting>
static void
StartTransaction(void)
......@@ -1150,13 +1204,13 @@ StartTransaction(void)
...
/*
* generate a new transaction id
* Advertise it in the proc array. We assume assignment of
* LocalTransactionID is atomic, and the backendId should be set already.
*/
s->transactionId = GetNewTransactionId(false);
Assert(MyProc->backendId == vxid.backendId);
MyProc->lxid = vxid.localTransactionId;
XactLockTableInsert(s->transactionId);
PG_TRACE1(transaction__start, s->transactionId);
TRACE_POSTGRESQL_TRANSACTION_START(vxid.localTransactionId);
...
}
......@@ -1164,30 +1218,26 @@ StartTransaction(void)
Note how the transaction ID is made available to the dynamic tracing
utility.
</para>
</step>
<step>
<para>
The dynamic tracing utility might require you to further define these trace
points. For example, DTrace requires you to add new probes to the file
<filename>src/backend/utils/probes.d</> as shown here:
<programlisting>
provider postgresql {
...
probe transaction__start(int);
...
};
</programlisting>
After recompiling and running the new binary, check that your newly added
probe is available by executing the following DTrace command, and you
should see similar output.
<screen>
# dtrace -ln transaction-start
ID PROVIDER MODULE FUNCTION NAME
18705 postgresql49878 postgres StartTransactionCommand transaction-start
18755 postgresql49877 postgres StartTransactionCommand transaction-start
18805 postgresql49876 postgres StartTransactionCommand transaction-start
18855 postgresql49875 postgres StartTransactionCommand transaction-start
18986 postgresql49873 postgres StartTransactionCommand transaction-start
</screen>
</para>
</step>
</procedure>
<para>
You should take care that the data types specified for the probe arguments
match the data types of the variables used in the <literal>PG_TRACE</>
macro. This is not checked at compile time. You can check that your newly
added trace point is available by recompiling, then running the new binary,
and as root, executing a DTrace command such as:
<programlisting>
dtrace -l -n transaction-start
</programlisting>
</para>
</sect2>
</sect1>
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment