protocol.sgml

关系型数据库 Postgresql 6.5.2

<Chapter Id="protocol">
<DocInfo>
<Author>
<FirstName>Phil</FirstName>
<Surname>Thompson</Surname>
</Author>
<Date>1998-08-08</Date>
</DocInfo>
<Title>Frontend/Backend Protocol</Title>

<Para>
<Note>
<Para>
Written by <ULink url="mailto:[e-mail]">Phil Thompson</ULink>.
Updates for protocol 2.0 by <ULink url="mailto:[e-mail]">Tom Lane</ULink>.
</Para>
</Note>
</para>

<Para>
<ProductName>Postgres</ProductName> uses a message-based protocol for communication between frontends
and backends.  The protocol is implemented over <Acronym>TCP/IP</Acronym> and also on Unix sockets.
<ProductName>Postgres</ProductName> v6.3 introduced version numbers into the protocol.
This was done in such
a way as to still allow connections from earlier versions of frontends, but
this document does not cover the protocol used by those earlier versions.
</para>

<Para>
This document describes version 2.0 of the protocol, implemented in
<ProductName>Postgres</ProductName> v6.4 and later.
</para>

<Para>
Higher level features built on this protocol (for example, how <FileName>libpq</FileName> passes
certain environment variables after the connection is established)
are covered elsewhere.
</para>

<Sect1>
<Title>Overview</Title>

<Para>
The three major components are the frontend (running on the client) and the
postmaster and backend (running on the server).  The postmaster and backend
have different roles but may be implemented by the same executable.
</para>

<Para>
A frontend sends a startup packet to the postmaster.  This includes the names
of the user and the database the user wants to connect to.  The postmaster then
uses this, and the information in the pg_hba.conf(5) file to determine what
further authentication information it requires the frontend to send (if any)
and responds to the frontend accordingly.
</para>

<Para>
The frontend then sends any required authentication information.  Once the
postmaster validates this it responds to the frontend that it is authenticated
and hands over the connection to a backend.  The backend then sends a message
indicating successful startup (normal case) or failure (for example, an
invalid database name).
</para>

<Para>
Subsequent communications are query and result packets exchanged between the
frontend and the backend.  The postmaster takes no further part in ordinary
query/result communication.  (However, the postmaster is involved when the
frontend wishes to cancel a query currently being executed by its backend.
Further details about that appear below.)
</para>

<Para>
When the frontend wishes to disconnect it sends an appropriate packet and
closes the connection without waiting for a response for the backend.
</para>

<Para>
Packets are sent as a data stream.  The first byte determines what should be
expected in the rest of the packet.  The exception is packets sent from a
frontend to the postmaster, which comprise a packet length then the packet
itself.  The difference is historical.
</para>
</sect1>

<Sect1>
<Title>Protocol</Title>

<Para>
This section describes the message flow.  There are four different types of
flows depending on the state of the connection:
startup, query, function call, and termination.
There are also special provisions for notification responses and command
cancellation, which can occur at any time after the startup phase.
</para>


<Sect2>
<Title>Startup</Title>

<Para>
Startup is divided into an authentication phase and a backend startup phase.
</para>

<Para>
Initially, the frontend sends a StartupPacket.  The postmaster uses this info
and the contents of the pg_hba.conf(5) file to determine what authentication
method the frontend must use.  The postmaster then responds with one of the
following messages:
</para>

<Para>
<VariableList>
<VarListEntry>
<Term>
	ErrorResponse
</Term>
<ListItem>
<Para>
		The postmaster then immediately closes the connection.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	AuthenticationOk
</Term>
<ListItem>
<Para>
		The postmaster then hands over to the backend.  The postmaster
		takes no further part in the communication.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	AuthenticationKerberosV4
</Term>
<ListItem>
<Para>
		The frontend must then take part in a Kerberos V4
		authentication dialog (not described here) with the postmaster.
		If this is successful, the postmaster responds with an
		AuthenticationOk, otherwise it responds with an ErrorResponse.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	AuthenticationKerberosV5
</Term>
<ListItem>
<Para>
		The frontend must then take part in a Kerberos V5
		authentication dialog (not described here) with the postmaster.
		If this is successful, the postmaster responds with an
		AuthenticationOk, otherwise it responds with an ErrorResponse.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	AuthenticationUnencryptedPassword
</Term>
<ListItem>
<Para>
		The frontend must then send an UnencryptedPasswordPacket.
		If this is the correct password, the postmaster responds with
		an AuthenticationOk, otherwise it responds with an
		ErrorResponse.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	AuthenticationEncryptedPassword
</Term>
<ListItem>
<Para>
		The frontend must then send an EncryptedPasswordPacket.
		If this is the correct password, the postmaster responds with
		an AuthenticationOk, otherwise it responds with an
		ErrorResponse.
</Para>
</ListItem>
</VarListEntry>
</VariableList>
</Para>

<Para>
If the frontend does not support the authentication method requested by the
postmaster, then it should immediately close the connection.
</para>

<Para>
After sending AuthenticationOk, the postmaster attempts to launch a backend
process.  Since this might fail, or the backend might encounter a failure
during startup, the frontend must wait for the backend to acknowledge
successful startup.  The frontend should send no messages at this point.
The possible messages from the backend during this phase are:

<VariableList>
<VarListEntry>
<Term>
	BackendKeyData
</Term>
<ListItem>
<Para>
		This message is issued after successful backend startup.
		It provides secret-key data that the frontend must save
		if it wants to be able to issue cancel requests later.
		The frontend should not respond to this message, but should
		continue listening for a ReadyForQuery message.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	ReadyForQuery
</Term>
<ListItem>
<Para>
		Backend startup is successful.  The frontend may now issue
		query or function call messages.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	ErrorResponse
</Term>
<ListItem>
<Para>
		Backend startup failed.  The connection is closed after
		sending this message.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	NoticeResponse
</Term>
<ListItem>
<Para>
		A warning message has been issued.  The frontend should
		display the message but continue listening for ReadyForQuery
		or ErrorResponse.
</Para>
</ListItem>
</VarListEntry>
</VariableList>
</Para>

<Para>
The ReadyForQuery message is the same one that the backend will issue after
each query cycle.  Depending on the coding needs of the frontend, it is
reasonable to consider ReadyForQuery as starting a query cycle (and then
BackendKeyData indicates successful conclusion of the startup phase),
or to consider ReadyForQuery as ending the startup phase and each subsequent
query cycle.
</para>
</sect2>

<Sect2>
<Title>Query</Title>

<Para>
A Query cycle is initiated by the frontend sending a Query message to the
backend.  The backend then sends one or more response messages depending
on the contents of the query command string, and finally a ReadyForQuery
response message.  ReadyForQuery informs the frontend that it may safely
send a new query or function call.
</para>

<Para>
The possible response messages from the backend are:

<VariableList>
<VarListEntry>
<Term>
	CompletedResponse
</Term>
<ListItem>
<Para>
		An SQL command completed normally.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	CopyInResponse
</Term>
<ListItem>
<Para>
		The backend is ready to copy data from the frontend to a
		relation.  The frontend should then send a CopyDataRows
		message.  The backend will then respond with a
		CompletedResponse message with a tag of "COPY".
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	CopyOutResponse
</Term>
<ListItem>
<Para>
		The backend is ready to copy data from a relation to the
		frontend.  It then sends a CopyDataRows message, and then a
		CompletedResponse message with a tag of "COPY".
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	CursorResponse
</Term>
<ListItem>
<Para>
		The query was either an insert(l), delete(l), update(l),
		fetch(l) or a select(l) command.
                If the transaction has been
                aborted then the backend sends a CompletedResponse message with
                a tag of "*ABORT STATE*".  Otherwise the following responses
                are sent.
</Para>
<Para>
		For an insert(l) command, the backend then sends a
		CompletedResponse message with a tag of "INSERT <Replaceable>oid</Replaceable> <Replaceable>rows</Replaceable>"
		where <Replaceable>rows</Replaceable> is the number of rows inserted, and <Replaceable>oid</Replaceable> is the
		object ID of the inserted row if <Replaceable>rows</Replaceable> is 1, otherwise <Replaceable>oid</Replaceable>
		is 0.
</Para>
<Para>
		For a delete(l) command, the backend then sends a
		CompletedResponse message with a tag of "DELETE <Replaceable>rows</Replaceable>" where
		<Replaceable>rows</Replaceable> is the number of rows deleted.
</Para>
<Para>
		For an update(l) command, the backend then sends a
		CompletedResponse message with a tag of "UPDATE <Replaceable>rows</Replaceable>" where
		<Replaceable>rows</Replaceable> is the number of rows deleted.
</Para>
<Para>
		For a fetch(l) or select(l) command, the backend sends a
		RowDescription message.  This is then followed by an AsciiRow
		or BinaryRow message (depending on whether a binary cursor was
		specified) for each row being returned to the frontend.
		Finally, the backend sends a CompletedResponse message with a
		tag of "SELECT".
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	EmptyQueryResponse
</Term>
<ListItem>
<Para>
		An empty query string was recognized.  (The need to specially
		distinguish this case is historical.)
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	ErrorResponse
</Term>
<ListItem>
<Para>
		An error has occurred.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	ReadyForQuery
</Term>
<ListItem>
<Para>
		Processing of the query string is complete.  A separate
		message is sent to indicate this because the query string
		may contain multiple SQL commands.  (CompletedResponse marks
		the end of processing one SQL command, not the whole string.)
		ReadyForQuery will always be sent, whether processing
		terminates successfully or with an error.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	NoticeResponse
</Term>
<ListItem>
<Para>
		A warning message has been issued in relation to the query.
		Notices are in addition to other responses, ie. the backend
		will continue processing the command.
</Para>
</ListItem>
</VarListEntry>
</VariableList>
</Para>

<Para>
A frontend must be prepared to accept ErrorResponse and NoticeResponse
messages whenever it is expecting any other type of message.
</para>

<Para>
Actually, it is possible for NoticeResponse to arrive even when the frontend
is not expecting any kind of message, that is, the backend is nominally idle.
(In particular, the backend can be commanded to terminate by its postmaster.
In that case it will send a NoticeResponse before closing the connection.)
It is recommended that the frontend check for such asynchronous notices just
before issuing any new command.
</para>

<Para>
Also, if the frontend issues any listen(l) commands then it must be prepared
to accept NotificationResponse messages at any time; see below.
</para>
</sect2>

<Sect2>
<Title>Function Call</Title>

<Para>
A Function Call cycle is initiated by the frontend sending a FunctionCall
message to the backend.  The backend then sends one or more response messages
depending on the results of the function call, and finally a ReadyForQuery
response message.  ReadyForQuery informs the frontend that it may safely send
a new query or function call.
</para>

<Para>
The possible response messages from the backend are:

<VariableList>
<VarListEntry>
<Term>
	ErrorResponse
</Term>
<ListItem>
<Para>
		An error has occurred.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	FunctionResultResponse
</Term>
<ListItem>
<Para>
		The function call was executed and returned a result.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	FunctionVoidResponse
</Term>
<ListItem>
<Para>
		The function call was executed and returned no result.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	ReadyForQuery
</Term>
<ListItem>
<Para>
		Processing of the function call is complete.
		ReadyForQuery will always be sent, whether processing
		terminates successfully or with an error.
</Para>
</ListItem>
</VarListEntry>
<VarListEntry>
<Term>
	NoticeResponse
</Term>
<ListItem>
<Para>
		A warning message has been issued in relation to the function
		call.
		Notices are in addition to other responses, ie. the backend
		will continue processing the command.
</Para>
</ListItem>
</VarListEntry>
</VariableList>
</Para>

<Para>
A frontend must be prepared to accept ErrorResponse and NoticeResponse
messages whenever it is expecting any other type of message.  Also,
if it issues any listen(l) commands then it must be prepared to accept
NotificationResponse messages at any time; see below.
</para>
</sect2>

<Sect2>
<Title>Notification Responses</Title>

<Para>
If a frontend issues a listen(l) command, then the backend will send a
NotificationResponse message (not to be confused with NoticeResponse!)
whenever a notify(l) command is executed for the same notification name.
</para>

<Para>
Notification responses are permitted at any point in the protocol (after
startup), except within another backend message.  Thus, the frontend
must be prepared to recognize a NotificationResponse message whenever it is
expecting any message.  Indeed, it should be able to handle
NotificationResponse messages even when it is not engaged in a query.

<VariableList>
<VarListEntry>
<Term>
	NotificationResponse
</Term>
<ListItem>
<Para>
		A notify(l) command has been executed for a name for which
		a previous listen(l) command was executed.  Notifications
		may be sent at any time.
</Para>
</ListItem>
</VarListEntry>
</VariableList>
</Para>

<Para>
It may be worth pointing out that the names used in listen and notify
commands need not have anything to do with names of relations (tables)
in the SQL database.  Notification names are simply arbitrarily chosen
condition names.
</para>
</sect2>

<Sect2>
<Title>Cancelling Requests in Progress</Title>

<Para>
During the processing of a query, the frontend may request cancellation of the
query by sending an appropriate request to the postmaster.  The cancel request
is not sent directly to the backend for reasons of implementation efficiency:
we don't want to have the backend constantly checking for new input from
the frontend during query processing.  Cancel requests should be relatively
infrequent, so we make them slightly cumbersome in order to avoid a penalty
in the normal case.
</para>

<Para>
To issue a cancel request, the frontend opens a new connection to the
postmaster and sends a CancelRequest message, rather than the StartupPacket
message that would ordinarily be sent across a new connection.  The postmaster
will process this request and then close the connection.  For security
reasons, no direct reply is made to the cancel request message.
</para>

<Para>
A CancelRequest message will be ignored unless it contains the same key data
(PID and secret key) passed to the frontend during connection startup.  If the
request matches the PID and secret key for a currently executing backend, the
postmaster signals the backend to abort processing of the current query.
</para>

<Para>
The cancellation signal may or may not have any effect --- for example, if it
arrives after the backend has finished processing the query, then it will have
no effect.  If the cancellation is effective, it results in the current
command being terminated early with an error message.
</para>

<Para>
The upshot of all this is that for reasons of both security and efficiency,
the frontend has no direct way to tell whether a cancel request has succeeded.
It must continue to wait for the backend to respond to the query.  Issuing a
cancel simply improves the odds that the current query will finish soon,
and improves the odds that it will fail with an error message instead of
succeeding.
</para>

<Para>
Since the cancel request is sent to the postmaster and not across the
regular frontend/backend communication link, it is possible for the cancel
request to be issued by any process, not just the frontend whose query is
to be canceled.  This may have some benefits of flexibility in building
multiple-process applications.  It also introduces a security risk, in that
unauthorized persons might try to cancel queries.  The security risk is
addressed by requiring a dynamically generated secret key to be supplied
in cancel requests.
</para>
</sect2>

<Sect2>
<Title>Termination</Title>

<Para>
The normal, graceful termination procedure is that the frontend sends a
Terminate message and immediately closes the connection.  On receipt of the
message, the backend immediately closes the connection and terminates.
</para>

<Para>
An ungraceful termination may occur due to software failure (i.e., core dump)
at either end.  If either frontend or backend sees an unexpected closure of
the connection, it should clean up and terminate.  The frontend has the option
of launching a new backend by recontacting the postmaster, if it doesn't want
to terminate itself.
</para>
</sect2>
</sect1>

<Sect1>
<Title>Message Data Types</Title>

<Para>
This section describes the base data types used in messages.

<VariableList>
<VarListEntry>
<Term>
	Int<Replaceable>n</Replaceable>(<Replaceable>i</Replaceable>)
</Term>
<ListItem>
<Para>
		An <Replaceable>n</Replaceable> bit integer in network byte order.
                If <Replaceable>i</Replaceable> is specified it