<!-- $Header: /cvsroot/pgsql/doc/src/sgml/protocol.sgml,v 1.22 2001/11/21 05:53:41 thomas Exp $ -->

<!-- $Header: /cvsroot/pgsql/doc/src/sgml/protocol.sgml,v 1.23 2001/11/22 01:22:10 tgl Exp $ -->

<chapter id="protocol">

<title>Frontend/Backend Protocol</title>

<para>

A frontend opens a connection to the server and sends a start-up

packet. This includes the names of the user and the database the

packet. This includes the names of the user andofthe database the

user wants to connect to. The server then uses this, and the

information in the <filename>pg_hba.conf</filename> file to

determine what further authentication information it requires the

</para>

<para>

In order to serve multiple clients efficiently, the server would

normally create a new child process to handle each incoming

connection. However, this is not required. In the current

implementation, a new child process is created immediately after an

incoming connection is detected. In earlier versions of

<productname>PostgreSQL</>

(7.1 and earlier), the child process was created after sending the

authentication confirmation message.

In order to serve multiple clients efficiently, the server launches

a new <quote>backend</> process for each client. This is transparent

to the protocol, however. In the current implementation, a new child

process is created immediately after an incoming connection is detected.

</para>

<para>

When the frontend wishes to disconnect it sends an appropriate packet and

closes the connection without waiting for a responsefor the backend.

closes the connection without waiting for a responsefrom the backend.

</para>

<para>

<Term>CursorResponse</Term>

<ListItem>

<Para>

The query was either an <literal>INSERT</literal>,

<literal>UPDATE</literal>, <literal>DELETE</literal>,

<literal>FETCH</literal>, or a <literal>SELECT</literal>

command. If the transaction has been aborted then the backend

sends a CompletedResponse message with a tag of <literal>*ABORT

STATE*</literal>. Otherwise the following responses are sent.

</Para>

<Para>

For an <literal>INSERT</literal> command, the backend then

sends a CompletedResponse message with a tag of

<literal>INSERT <replaceable>oid</replaceable>

<replaceable>rows</replaceable></literal>, 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 <literal>DELETE</literal> command, the backend then

sends a CompletedResponse message with a tag of <literal>DELETE

<Replaceable>rows</Replaceable></literal> where

<Replaceable>rows</Replaceable> is the number of rows deleted.

</Para>

<Para>

For an <literal>UPDATE</literal> command, the backend then

sends a CompletedResponse message with a tag of <literal>UPDATE

<Replaceable>rows</Replaceable></literal> where

<Replaceable>rows</Replaceable> is the number of rows affected

by the update.

Beginning of the response to a <command>SELECT</command>,

<command>FETCH</command>, <command>INSERT</command>,

<command>UPDATE</command>, or <command>DELETE</command>

query. In the <command>FETCH</command> case the name of the

cursor being fetched from is included in the message. Otherwise

the message always mentions the <quote>blank</> cursor.

</Para>

</ListItem>

</VarListEntry>

<VarListEntry>

<Term>RowDescription</Term>

<ListItem>

<Para>

command, the backend sends a RowDescription message. This is

CompletedResponse message with a tag of <literal>SELECT</literal>.

a <command>SELECT</command> or <command>FETCH</command> query.

to the frontend.

</Para>

</ListItem>

</VarListEntry>

<Term>EmptyQueryResponse</Term>

<ListItem>

<Para>

An empty query string was recognized. (The need to specially

distinguish this case is historical.)

An empty query string was recognized.

</Para>

</ListItem>

</VarListEntry>

</VariableList>

</Para>

<Para>

The response to a <command>SELECT</> or <command>FETCH</> query

normally consists of CursorResponse, RowDescription, zero or more

AsciiRow or BinaryRow messages, and finally CompletedResponse.

<command>INSERT</command>, <command>UPDATE</command>, and

<command>DELETE</command> queries produce CursorResponse followed by

CompletedResponse.

<command>COPY</> to or from the frontend invokes special protocol

as mentioned above.

All other query types normally produce only

a CompletedResponse message.

</Para>

<Para>

Since a query string could contain several queries (separated by

semicolons), there might be several such response sequences before the

backend finishes processing the query string. ReadyForQuery is issued

when the entire string has been processed and the backend is ready to

accept a new query string.

</Para>

<Para>

If a completely empty (no contents other than whitespace) query string

is received, the response is EmptyQueryResponse followed by ReadyForQuery.

(The need to specially distinguish this case is historical.)

</Para>

<Para>

In the event of an error, ErrorResponse is issued followed by

ReadyForQuery. All further processing of the query string is aborted by

ErrorResponse (even if more queries remained in it). Note that this

may occur partway through the sequence of messages generated by an

individual query.

</Para>

<para>

A frontend must be prepared to accept ErrorResponse and

NoticeResponse messages whenever it is expecting any other type of

</para>

<Para>

Also, if the frontend issues any <literal>LISTEN</literal>

Also, if the frontend issues any <command>LISTEN</command>

commands then it must be prepared to accept NotificationResponse

messages at any time; see below.

</para>

<para>

Recommended practice is to code frontends in a state-machine style

that will accept any message type at any time that it could make sense,

rather than wiring in assumptions about the exact sequence of messages.

</para>

</sect2>

<Sect2>

<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 <literal>LISTEN</literal>

message. Also, if it issues any <command>LISTEN</command>

commands then it must be prepared to accept NotificationResponse

messages at any time; see below.

</para>

<title>Notification Responses</title>

<Para>

If a frontend issues a <literal>LISTEN</literal> command, then the

If a frontend issues a <command>LISTEN</command> command, then the

backend will send a NotificationResponse message (not to be

confused with NoticeResponse!) whenever a

<literal>NOTIFY</literal> command is executed for the same

<command>NOTIFY</command> command is executed for the same

notification name.

</para>

<Term>NotificationResponse</Term>

<ListItem>

<Para>

A <literal>NOTIFY</literal> command has been executed for a

name for which a previous <literal>LISTEN</literal> command

A <command>NOTIFY</command> command has been executed for a

name for which a previous <command>LISTEN</command> command

was executed. Notifications may be sent at any time.

</Para>

</ListItem>

same key data (PID and secret key) passed to the frontend during

connection start-up. If the request matches the PID and secret

key for a currently executing backend, the processing of the

current query is aborted. (In the existingimplemenation, this is

current query is aborted. (In the existingimplementation, this is

done by sending a special signal to the backend process that is

processing the query.)

</para>

by recontacting the server if it doesn't want to terminate

itself.

</para>

<para>

For either normal or abnormal termination, any open transaction is

rolled back, not committed. One should note however that if a

frontend disconnects while a query is being processed, the backend

will probably finish the query before noticing the disconnection.

If the query is outside any transaction block (<command>BEGIN</>

... <command>COMMIT</> sequence) then its results may be committed

before the disconnection is recognized.

</para>

</sect2>

<Sect2>

<Title>SSL Session Encryption</Title>

<Para>

Recent releases of <productname>PostgreSQL</> allow frontend/backend

communication to be encrypted using SSL. This provides communication

security in environments where attackers might be able to capture the

session traffic.

</para>

<para>

To initiate an SSL-encrypted connection, the frontend initially sends

an SSLRequest message rather than a StartupPacket. The server then

responds with a single byte containing <literal>Y</> or <literal>N</>,

indicating that it is willing or unwilling to perform SSL, respectively.

The frontend may close the connection at this point if it is dissatisfied

with the response. To continue after <literal>Y</>, perform an SSL

startup handshake (not described here, part of the SSL specification)

with the server. If this is successful, continue with

sending the usual StartupPacket. In this case the StartupPacket and

all subsequent data will be SSL-encrypted. To continue after

<literal>N</>, send the usual StartupPacket and proceed without

encryption.

</para>

<para>

The frontend should also be prepared to handle an ErrorMessage response

to SSLRequest from the server. This would only occur if the server

predates the addition of SSL support to <productname>PostgreSQL</>.

In this case the connection must be closed, but the frontend may choose

to open a fresh connection and proceed without requesting SSL.

</para>

<para>

An initial SSLRequest may also be used in a connection that is being

opened to send a CancelRequest message.

</para>

<para>

While the protocol itself does not provide a way for the server to

force SSL encryption, the administrator may configure the server to

reject unencrypted sessions as a byproduct of authentication checking.

</para>

</sect2>

</sect1>

</Term>

<ListItem>

<Para>

The command tag. This is usually (but not always) a single

word that identifies which SQL command was completed.

The command tag. This is usually a single

word that identifies which SQL command was completed.

</Para>

<Para>

For an <command>INSERT</command> command, the tag is

<literal>INSERT <replaceable>oid</replaceable>

<replaceable>rows</replaceable></literal>, 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 <command>DELETE</command> command, the tag is

<literal>DELETE <Replaceable>rows</Replaceable></literal> where

<Replaceable>rows</Replaceable> is the number of rows deleted.

</Para>

<Para>

For an <command>UPDATE</command> command, the tag is

<literal>UPDATE <Replaceable>rows</Replaceable></literal> where

<Replaceable>rows</Replaceable> is the number of rows updated.

</Para>

</ListItem>

</VarListEntry>

</Para>

</ListItem>

</VarListEntry>

<VarListEntry>

<Term>

SSLRequest (F)

</Term>

<ListItem>

<Para>

<VariableList>

<VarListEntry>

<Term>

Int32(8)

</Term>

<ListItem>

<Para>

The size of the packet in bytes.

</Para>

</ListItem>

</VarListEntry>

<VarListEntry>

<Term>

Int32(80877103)

</Term>

<ListItem>

<Para>

The SSL request code. The value is chosen to contain

<literal>1234</> in the most significant 16 bits, and <literal>5679</> in the

least 16 significant bits. (To avoid confusion, this code

must not be the same as any protocol version number.)

</Para>

</ListItem>

</VarListEntry>

</VariableList>

</Para>

</ListItem>

</VarListEntry>

<VarListEntry>

<Term>

StartupPacket (F)

<ListItem>

<Para>

The optional tty the backend should use for debugging messages.

(Currently, this field is unsupported and ignored.)

</Para>

</ListItem>

</VarListEntry>