Update ovsdb specifications based on suggestions from Brandon Heller.

[openvswitch] / ovsdb / SPECS

         ===================================================
          Open vSwitch Configuration Database Specification
         ===================================================

Basic Notation
--------------

The descriptions below use the following shorthand notations for JSON
values.  Additional notation is presented later.

<string>

    A JSON string.  Any Unicode string is allowed, as specified by RFC
    4627.  Implementations may disallowed null bytes.

<id>

    A JSON string matching [a-zA-Z_][a-zA-Z0-9_]*.

    <id>s that begin with _ are reserved to the implementation and may
    not be used by the user.

<boolean>

    A JSON true or false value.

<number>

    A JSON number.

<integer>

    A JSON number with an integer value, within a certain range
    (currently -2**63...+2**63-1).

<value>

    Any JSON value.

Schema Format
-------------

An Open vSwitch configuration database consists of a set of tables,
each of which has a number of columns and zero or more rows.  A schema
is represented by <database-schema>, as described below.

<database-schema>

    A JSON object with the following members:

        "name": <id>                            required
        "comment": <string>                     optional
        "tables": {<id>: <table-schema>, ...}   required

    The "name" identifies the database as a whole.  The "comment"
    optionally provides more information about the database.  The
    value of "tables" is a JSON object whose names are table names and
    whose values are <table-schema>s.

<table-schema>

    A JSON object with the following members:

        "comment": <string>                       optional
        "columns": {<id>: <column-schema>, ...}   required

    The "comment" optionally provides information about this table for
    a human reader.  The value of "columns" is a JSON object whose
    names are column names and whose values are <column-schema>s.

    Every table has the following columns whose definitions are not
    included in the schema:

        "_uuid": This column, which contains exactly one UUID value,
        is initialized to a random value by the database engine when
        it creates a row.  It is read-only, and its value never
        changes during the lifetime of a row.

        "_version": Like "_uuid", this column contains exactly one
        UUID value, initialized to a random value by the database
        engine when it creates a row, and it is read-only.  However,
        its value changes to a new random value whenever any other
        field in the row changes.  Furthermore, its value is
        ephemeral: when the database is closed and reopened, or when
        the database process is stopped and then started again, each
        "_version" also changes to a new random value.

<column-schema>

    A JSON object with the following members:

        "comment": <string>                       optional
        "type": <type>                            required
        "ephemeral": <boolean>                    optional

    The "comment" optionally provides information about this column
    for a human reader.  The "type" specifies the type of data stored
    in this column.  If "ephemeral" is specified as true, then this
    column's values are not guaranteed to be durable; they may be lost
    when the database restarts.

<type>

    The type of a database column.  Either an <atomic-type> or a JSON
    object that describes the type of a database column, with the
    following members:

        "key": <atomic-type>               required
        "value": <atomic-type>             optional
        "min": <integer>                   optional
        "max": <integer> or "unlimited"    optional

    If "min" or "max" is not specified, each defaults to 1.  If "max"
    is specified as "unlimited", then there is no specified maximum
    number of elements, although the implementation will enforce some
    limit.  After considering defaults, "min" must be at least 0,
    "max" must be at least 1, and "max" must be greater than or equal
    to "min".

    If "min" and "max" are both 1 and "value" is not specified, the
    type is the scalar type specified by "key".

    If "min" is not 1 or "max" is not 1, or both, and "value" is not
    specified, the type is a set of scalar type "key".

    If "value" is specified, the type is a map from type "key" to type
    "value".

<atomic-type>

    One of the strings "integer", "real", "boolean", "string", or
    "uuid", representing the specified scalar type.

Wire Protocol
-------------

The database wire protocol is implemented in JSON-RPC 1.0.  We
encourage use of JSON-RPC over stream connections instead of JSON-RPC
over HTTP, for these reasons:

    * JSON-RPC is a peer-to-peer protocol, but HTTP is a client-server
      protocol, which is a poor match.  Thus, JSON-RPC over HTTP
      requires the client to periodically poll the server to receive
      server requests.

    * HTTP is more complicated than stream connections and doesn't
      provide any corresponding advantage.

    * The JSON-RPC specification for HTTP transport is incomplete.

The database wire protocol consists of the following JSON-RPC methods:

get_schema
..........

Request object members:

    "method": "get_schema"            required
    "params": []                      required
    "id": any JSON value except null  required

Response object members:

    "result": <database-schema>
    "error": null
    "id": same "id" as request

This operation retrieves a <database-schema> that describes the
hosted database.

transact
........

Request object members:

    "method": "transact"              required
    "params": [<operation>*]          required
    "id": any JSON value except null  required

Response object members:

    "result": [<object>*]
    "error": null
    "id": same "id" as request

The "params" array for this method consists of zero or more JSON
objects, each of which represents a single database operation.  The
"Operations" section below describes the valid operations.

The value of "id" must be unique among all in-flight transactions
within the current JSON-RPC session.  Otherwise, the server may return
a JSON-RPC error.

The database server executes each of the specified operations in the
specified order, except that if an operation fails, then the remaining
operations are not executed.

The set of operations is executed as a single atomic, consistent,
isolated transaction.  The transaction is committed only if every
operation succeeds.  Durability of the commit is not guaranteed unless
the "commit" operation, with "durable" set to true, is included in the
operation set (see below).

Regardless of whether errors occur, the response is always a JSON-RPC
response with null "error" and a "result" member that is an array with
the same number of elements as "params".  Each element of the "result"
array corresponds to the same element of the "params" array.  The
"result" array elements may be interpreted as follows:

    - A JSON object that does not contain an "error" member indicates
      that the operation completed successfully.  The specific members
      of the object are specified below in the descriptions of
      individual operations.  Some operations do not produce any
      results, in which case the object will have no members.

    - A JSON object that contains an "error" member indicates that the
      operation completed with an error.  The value of the "error"
      member is a short string, specified in this document, that
      broadly indicates the class of the error.  Besides the ones
      listed for a specific operation, any operation may result in one
      the following "error"s:

      "error": "resources exhausted"

          The operation or the transaction requires more resources
          (memory, disk, CPU, etc.) than are currently available to
          the database server.

      "error": "syntax error"

          The operation is not specified correctly: a required request
          object member is missing, an unknown or unsupported request
          object member is present, the operation attempts to act on a
          table that does not exist, the operation modifies a
          read-only table column, etc.

      Database implementations may use "error" strings not specified
      in this document to indicate errors that do not fit into any of
      the specified categories.

      Optionally, the object may include a "details" member, whose
      value is a string that describes the error in more detail for
      the benefit of a human user or administrator.  The object may
      also have other members that describe the error in more detail.
      This document does not specify the names or values of these
      members.

    - A JSON null value indicates that the operation was not attempted
      because a prior operation failed.

In general, "result" contains some number of successful results,
possibly followed by an error, in turn followed by enough JSON null
values to match the number of elements in "params".  There is one
exception: if all of the operations succeed, but the results cannot be
committed (e.g. due to I/O errors), then "result" will have one more
element than "params", with the additional element describing the
error.

If "params" contains one or more "wait" operations, then the
transaction may take an arbitrary amount of time to complete.  The
database implementation must be capable of accepting, executing, and
replying to other transactions and other JSON-RPC requests while a
transaction or transactions containing "wait" operations are
outstanding on the same or different JSON-RPC sessions.

The section "Notation for the Wire Protocol" below describes
additional notation for use with the wire protocol.  After that, the
"Operations" section describes each operation.

cancel
......

Request object members:

    "method": "cancel"                              required
    "params": [the "id" for an outstanding request] required
    "id": null                                      required

Response object members:

    <no response>

This JSON-RPC notification instructs the database server to
immediately complete or cancel the "transact" request whose "id" is
the same as the notification's "params" value.  

If the "transact" request can be completed immediately, then the
server sends a response in the form described for "transact", above.
Otherwise, the server sends a JSON-RPC error response of the following
form:

    "result": null
    "error": "canceled"
    "id": the request "id" member

The "cancel" notification itself has no reply.

monitor
.......

Request object members:

    "method": "monitor"                        required
    "params": [<value>, <monitor-requests>]    required
    "id": any JSON value except null           required

<monitor-requests> is an object that maps from a table name to a
<monitor-request>.

Each <monitor-request> is an object with the following members:

    "columns": [<column>*]            optional
    "select": <monitor-select>        optional

<monitor-select> is an object with the following members:

    "initial": <boolean>              optional
    "insert": <boolean>               optional
    "delete": <boolean>               optional
    "modify": <boolean>               optional

Response object members:

    "result": <table-updates>
    "error": null
    "id": same "id" as request

This JSON-RPC request enables a client to replicate tables or subsets
of tables.  Each <monitor-request> specifies a table to be replicated.
The JSON-RPC response to the "monitor" includes the initial contents
of each table.  Afterward, when changes to those tables are committed,
the changes are automatically sent to the client using the "update"
monitor notification.  This monitoring persists until the JSON-RPC
session terminates or until the client sends a "monitor_cancel"
JSON-RPC request.

Each <monitor-request> describes how to monitor a table:

    The circumstances in which an "update" notification is sent for a
    row within the table are determined by <monitor-select>:

        If "initial" is omitted or true, every row in the table is
        sent as part of the reply to the "monitor" request.

        If "insert" is omitted or true, "update" notifications are
        sent for rows newly inserted into the table.

        If "delete" is omitted or true, "update" notifications are
        sent for rows deleted from the table.

        If "modify" is omitted or true, "update" notifications are
        sent whenever when a row in the table is modified.

    The "columns" member specifies the columns whose values are
    monitored.  If "columns" is omitted, all columns in the table,
    except for "_uuid", are monitored.

The "result" in the JSON-RPC response to the "monitor" request is a
<table-updates> object (see below) that contains the contents of the
tables for which "initial" rows are selected.  If no tables' initial
contents are requested, then "result" is an empty object.

update
......

Notification object members:

    "method": "update"
    "params": [<value>, <table-updates>]
    "id": null

The <value> in "params" is the same as the value passed as the <value>
in "params" for the "monitor" request.

<table-updates> is an object that maps from a table name to a
<table-update>.

A <table-update> is an object that maps from the row's UUID (as a
36-byte string) to a <row-update> object.

A <row-update> is an object with the following members:

    "old": <row>         present for "delete" and "modify" updates
    "new": <row>         present for "initial", "insert", and "modify" updates

This JSON-RPC notification is sent from the server to the client to
tell it about changes to a monitored table (or the initial state of a
modified table).  Each table in which one or more rows has changed (or
whose initial view is being presented) is represented in "updates".
Each row that has changed (or whose initial view is being presented)
is represented in its <table-update> as a member with its name taken
from the row's _uuid member.  The corresponding value is a
<row-update>:

    The "old" member is present for "delete" and "modify" updates.
    For "delete" updates, each monitored column is included.  For
    "modify" updates, the prior value of each monitored column whose
    value has changed is included (monitored columns that have not
    changed are represented in "new").

    The "new" member is present for "initial", "insert", and "modify"
    updates.  For "initial" and "insert" updates, each monitored
    column is included.  For "modify" updates, the new value of each
    monitored column is included.

monitor_cancel
..............

Request object members:

    "method": "monitor_cancel"                              required
    "params": [<value>]                                     required
    "id": any JSON value except null                        required

Response object members:

    "result": {}
    "error": null
    "id": the request "id" member

Cancels the ongoing table monitor request, identified by the <value>
in "params" matching the <value> in "params" for an ongoing "monitor"
request.  No more "update" messages will be sent for this table
monitor.

echo
....

Request object members:

    "method": "echo"                                required
    "params": JSON array with any contents          required
    "id": <value>                                   required

Response object members:

    "result": same as "params"
    "error": null
    "id": the request "id" member

Both the JSON-RPC client and the server must implement this request.

This JSON-RPC request and response can be used to implement connection
keepalives, by allowing the server to check that the client is still
there or vice versa.


Notation for the Wire Protocol
------------------------------

<table>

    An <id> that names a table.

<column>

    An <id> that names a table column.

<row>

    A JSON object that describes a table row or a subset of a table
    row.  Each member is the name of a table column paired with the
    <value> of that column.

<value>

    A JSON value that represents the value of a column in a table row,
    one of <atom>, a <set>, or a <map>.

<atom>

    A JSON value that represents a scalar value for a column, one of
    <string>, <number>, <boolean>, <uuid>, <named-uuid>.

<set>

    A 2-element JSON array that represents a database set value.  The
    first element of the array must be the string "set" and the second
    element must be an array of zero or more <atom>s giving the values
    in the set.  All of the <atom>s must have the same type.

<map>

    A 2-element JSON array that represents a database map value.  The
    first element of the array must be the string "map" and the second
    element must be an array of zero or more <pair>s giving the values
    in the map.  All of the <pair>s must have the same key and value
    types.

    (JSON objects are not used to represent <map> because JSON only
    allows string names in an object.)

<pair>

    A 2-element JSON array that represents a pair within a database
    map.  The first element is an <atom> that represents the key, the
    second element is an <atom> that represents the value.

<uuid>

    A 2-element JSON array that represents a UUID.  The first element
    of the array must be the string "uuid" and the second element must
    be a 36-character string giving the UUID in the format described
    by RFC 4122.  For example, the following <uuid> represents the
    UUID 550e8400-e29b-41d4-a716-446655440000:

        ["uuid", "550e8400-e29b-41d4-a716-446655440000"]

<named-uuid>

    A 2-element JSON array that represents the UUID of a row inserted
    in a previous "insert" operation within the same transaction.  The
    first element of the array must be the string "named-uuid" and the
    second element must be the string specified on this "insert"
    operation's "uuid-name" or on a preceding "insert" within the same
    transaction.  For example, if this or a previous "insert"
    operation specified a "uuid-name" of "myrow", the following
    <named-uuid> represents the UUID created by that operation:

        ["named-uuid", "myrow"]

    A <named-uuid> may be used anywhere a <uuid> is valid.

<condition>

    A 3-element JSON array of the form [<column>, <function>,
    <value>] that represents a test on a column value.

    Except as otherwise specified below, <value> must have the same
    type as <column>.

    The meaning depends on the type of <column>:

        integer
        real

            <function> must be "<", "<=", "==", "!=", ">=", ">",
            "includes", or "excludes".

            The test is true if the column's value satisfies the
            relation <function> <value>, e.g. if the column has value
            1 and <value> is 2, the test is true if <function> is "<",
            "<=" or "!=", but not otherwise.

            "includes" is equivalent to "=="; "excludes" is equivalent
            to "!=".

        boolean
        string
        uuid

            <function> must be "!=", "==", "includes", or "excludes".

            If <function> is "==" or "includes", the test is true if
            the column's value equals <value>.  If <function> is "!="
            or "excludes", the test is inverted.

        set
        map

            <function> must be "!=", "==", "includes", or "excludes".

            If <function> is "==", the test is true if the column's
            value contains exactly the same values (for sets) or pairs
            (for maps).  If <function> is "!=", the test is inverted.

            If <function> is "includes", the test is true if the
            column's value contains all of the values (for sets) or
            pairs (for maps) in <value>.  The column's value may also
            contain other values or pairs.

            If <function> is "excludes", the test is true if the
            column's value does not contain any of the values (for
            sets) or pairs (for maps) in <value>.  The column's value
            may contain other values or pairs not in <value>.

            If <function> is "includes" or "excludes", then the
            required type of <value> is slightly relaxed, in that it
            may have fewer than the minimum number of elements
            specified by the column's type.  If <function> is
            "excludes", then the required type is additionally relaxed
            in that <value> may have more than the maximum number of
            elements specified by the column's type.

<function>

    One of "<", "<=", "==", "!=", ">=", ">", "includes", "excludes".

Operations
----------

Each of the available operations is described below.

insert
......

Request object members:

    "op": "insert"          required
    "table": <table>        required
    "row": <row>            required
    "uuid-name": <string>   optional

Result object members:

    "uuid": <uuid>

Semantics:

    Inserts "row" into "table".  If "row" does not specify values
    for all the columns in "table", those columns receive default
    values.

    The new row receives a new, randomly generated UUID, which is
    returned as the "uuid" member of the result.  If "uuid-name" is
    supplied, then the UUID is made available under that name to this
    operation and later operations within the same transaction.

select
......

Request object members:

    "op": "select"                required
    "table": <table>              required
    "where": [<condition>*]       required
    "columns": [<column>*]        optional

Result object members:

    "rows": [<row>*]

Semantics:

    Searches "table" for rows that match all the conditions specified
    in "where".  If "where" is an empty array, every row in "table" is
    selected.

    The "rows" member of the result is an array of objects.  Each
    object corresponds to a matching row, with each column
    specified in "columns" as a member, the column's name as the
    member name and its value as the member value.  If "columns"
    is not specified, all the table's columns are included.  If
    two rows of the result have the same values for all included
    columns, only one copy of that row is included in "rows".
    Specifying "_uuid" within "columns" will avoid dropping
    duplicates, since every row has a unique UUID.

    The ordering of rows within "rows" is unspecified.

update
......

Request object members:

    "op": "update"                required
    "table": <table>              required
    "where": [<condition>*]       required
    "row": <row>                  required

Result object members:

    "count": <integer>

Semantics:

    Updates rows in a table.

    Searches "table" for rows that match all the conditions
    specified in "where".  For each matching row, changes the
    value of each column specified in "row" to the value for that
    column specified in "row".

    The "_uuid" and "_version" columns of a table may not be directly
    updated with this operation.  Columns designated read-only in the 
    schema also may not be updated.

    The "count" member of the result specifies the number of rows
    that matched.

delete
......

Request object members:

    "op": "delete"                required
    "table": <table>              required
    "where": [<condition>*]       required

Result object members:

    "count": <integer>

Semantics:

    Deletes all the rows from "table" that match all the conditions
    specified in "where".

    The "count" member of the result specifies the number of deleted
    rows.

wait
....

Request object members:

    "op": "wait"                        required
    "timeout": <integer>                optional
    "table": <table>                    required
    "where": [<condition>*]             required
    "columns": [<column>*]              required
    "until": "==" or "!="               required
    "rows": [<row>*]                    required

Result object members:

    none

Semantics:

    Waits until a condition becomes true.

    If "until" is "==", checks whether the query on "table" specified
    by "where" and "columns", which is evaluated in the same way as
    specified for "select", returns the result set specified by
    "rows".  If it does, then the operation completes successfully.
    Otherwise, the entire transaction rolls back.  It is automatically
    restarted later, after a change in the database makes it possible
    for the operation to succeed.  The client will not receive a
    response until the operation permanently succeeds or fails.
    
    If "until" is "!=", the sense of the test is negated.  That is, as
    long as the query on "table" specified by "where" and "columns"
    returns "rows", the transaction will be rolled back and restarted
    later.

    If "timeout" is specified, then the transaction aborts after the
    specified number of milliseconds.  The transaction is guaranteed
    to be attempted at least once before it aborts.  A "timeout" of 0
    will abort the transaction on the first mismatch.

Errors:

    "error": "not supported"

        One or more of the columns in this table do not support
        triggers.  This error will not occur if "timeout" is 0.

    "error": "timed out"

        The "timeout" was reached before the transaction was able to
        complete.

commit
......

Request object members:

    "op": "commit"                      required
    "durable": <boolean>                required

Result object members:

    none

Semantics:

    If "durable" is specified as true, then the transaction, if it
    commits, will be stored durably (to disk) before the reply is sent
    to the client.

Errors:

    "error": "not supported"

        When "durable" is true, this database implementation does not
        support durable commits.

abort
.....

Request object members:

    "op": "abort"                      required

Result object members:

    (never succeeds)

Semantics:

    Aborts the transaction with an error.  This may be useful for
    testing.

Errors:

    "error": "aborted"

        This operation always fails with this error.