This module implements versions 1, 2c and 3 of the Simple Network Management Protocol (SNMP).
Read this in other languages: English, 简体ä¸ć–‡.
This module is installed using node package manager (npm):
npm install net-snmp
It is loaded using the require()
function:
var snmp = require ("net-snmp");
Sessions to remote hosts can then be created and used to perform SNMP requests and send SNMP traps or informs:
var session = snmp.createSession ("127.0.0.1", "public");
var oids = ["1.3.6.1.2.1.1.5.0", "1.3.6.1.2.1.1.6.0"];
session.get (oids, function (error, varbinds) {
if (error) {
console.error (error);
} else {
for (var i = 0; i < varbinds.length; i++) {
if (snmp.isVarbindError (varbinds[i])) {
console.error (snmp.varbindError (varbinds[i]));
} else {
console.log (varbinds[i].oid + " = " + varbinds[i].value);
}
}
}
session.close ();
});
session.trap (snmp.TrapType.LinkDown, function (error) {
if (error) {
console.error (error);
}
});
RFC 3413 describes five types of SNMP applications:
- Command Generator Applications — which initiate read or write requests
- Command Responder Applications — which respond to received read or write requests
- Notification Originator Applications — which generate notifications (traps or informs)
- Notification Receiver Applications — which receive notifications (traps or informs)
- Proxy Forwarder Applications — which forward SNMP messages
This library provides support for all of the above applications, with the documentation for each shown in this table:
Application | Common Use | Documentation |
---|---|---|
Command Generator | NMS / SNMP tools | Application: Command & Notification Generator |
Command Responder | SNMP agents | Application: SNMP Agent |
Notification Originator | SNMP agents / NMS-to-NMS notifications | Application: Command & Notification Generator |
Notification Receiver | NMS | Application: Notification Receiver |
Proxy Forwarder | SNMP agents | Agent Forwarder Module |
- Support for all SNMP versions: SNMPv1, SNMPv2c and SNMPv3
- SNMPv3 message authentication using MD5 or SHA, and privacy using DES or AES encryption
- Community-based and user-based authorization
- SNMP initiator for all relevant protocol operations: Get, GetNext, GetBulk, Set, Trap, Inform
- Convenience methods for MIB "walking", subtree collection, table and table column collection
- SNMPv3 context support
- Notification receiver for traps and informs
- MIB parsing and MIB module store
- Translation between numeric and named OIDs
- SNMP agent with MIB management for both scalar and tabular data
- Agent table index support for non-integer keys, foreign keys, composite keys and table augmentation
- Agent support for "RowStatus" protocol-based creation and deletion of table rows
- Agent support for these MIB constraints: MAX-ACCESS, integer ranges, string sizes, integer enumerations
- SNMP proxy forwarder for agent
- AgentX subagent
- IPv4 and IPv6
This module aims to be fully compliant with the following RFCs:
- 1098 - A Simple Network Management Protocol (version 1)
- 1155 - Structure and Identification of Management Information
- 2571 - Agent Extensibility (AgentX) Protocol Version 1
- 2578 - Structure of Management Information Version 2 (SMIv2)
- 3413 - Simple Network Management Protocol (SNMP) Applications
- 3414 - User-based Security Model (USM) for version 3 of the Simple Network Management Protocol (SNMPv3)
- 3416 - Version 2 of the Protocol Operations for the Simple Network Management Protocol (SNMP)
- 3417 - Transport Mappings for the Simple Network Management Protocol (SNMP)
- 3826 - The Advanced Encryption Standard (AES) Cipher Algorithm in the SNMP User-based Security Model
The following sections describe constants exported and used by this module.
These constants are used to specify which of version supported by this module should be used.
This object contains constants for all valid values the error-status field in
response PDUs can hold. If when parsing a PDU the error-index field contains
a value not defined in this object the constant snmp.ErrorStatus.GeneralError
will be used instead of the value in the error-status field. The following
constants are defined in this object:
NoError
TooBig
NoSuchName
BadValue
ReadOnly
GeneralError
NoAccess
WrongType
WrongLength
WrongEncoding
WrongValue
NoCreation
InconsistentValue
ResourceUnavailable
CommitFailed
UndoFailed
AuthorizationError
NotWritable
InconsistentName
This object contains constants used to specify syntax for varbind objects, e.g.:
var varbind = {
oid: "1.3.6.1.2.1.1.4.0",
type: snmp.ObjectType.OctetString,
value: "user.name@domain.name"
};
The following constants are defined in this object:
Boolean
Integer
OctetString
Null
OID
IpAddress
Counter
Gauge
TimeTicks
Opaque
Integer32
Counter32
Gauge32
Unsigned32
Counter64
NoSuchObject
NoSuchInstance
EndOfMibView
This object contains constants used to specify a type of SNMP trap. These
constants are passed to the trap()
and inform()
methods exposed by the
Session
class. The following constants are defined in this object:
ColdStart
WarmStart
LinkDown
LinkUp
AuthenticationFailure
EgpNeighborLoss
EnterpriseSpecific
This object contains constants used to identify the SNMP PDU types specified in RFC 3416. The values, along with their numeric codes, are:
160 - GetRequest
161 - GetNextRequest
162 - GetResponse
163 - SetRequest
164 - Trap
165 - GetBulkRequest
166 - InformRequest
167 - TrapV2
168 - Report
This object contains constants to specify the security of an SNMPv3 message as per RFC 3414:
noAuthNoPriv
- for no message authentication or encryptionauthNoPriv
- for message authentication and no encryptionauthPriv
- for message authentication and encryption
This object contains constants to select a supported digest algorithm for SNMPv3 messages that require authentication:
md5
- for HMAC-MD5 message authenticationsha
- for HMAC-SHA-1 message authenticationsha224
- for HMAC-SHA-224 message authenticationsha256
- for HMAC-SHA-256 message authenticationsha384
- for HMAC-SHA-384 message authenticationsha512
- for HMAC-SHA-512 message authentication
MD5 and SHA (actually SHA-1) are the hash algorithms specified in the original SNMPv3 User-Based Security Model RFC (RFC 3414); the other four were added later in RFC 7860.
This object contains constants to select a supported encryption algorithm for SNMPv3 messages that require privacy:
des
- for DES encryption (CBC-DES)aes
- for 128-bit AES encryption (CFB-AES-128)aes256b
- for 256-bit AES encryption (CFB-AES-256) with "Blumenthal" key localizationaes256r
- for 256-bit AES encryption (CFB-AES-256) with "Reeder" key localization
DES is the sole encryption algorithm specified in the original SNMPv3 User-Based Security Model RFC (RFC 3414); 128-bit AES for SNMPv3 was added later in RFC 3826. 256-bit AES has not been standardized, and as such comes with two varieties of key localization. Cisco and a number of other vendors commonly use the "Reeder" key localization variant. Other encryption algorithms are not supported.
The Agent Extensibility (AgentX) Protocol specifies these PDUs in RFC 2741:
1 - Open
2 - Close
3 - Register
4 - Unregister
5 - Get
6 - GetNext
7 - GetBulk
8 - TestSet
9 - CommitSet
10 - UndoSet
11 - CleanupSet
12 - Notify
13 - Ping
14 - IndexAllocate
15 - IndexDeallocate
16 - AddAgentCaps
17 - RemoveAgentCaps
18 - Response
None
- no access control for defined communities and usersSimple
- simple access control of levels "read-only" or "read-write" for defined communites and users
None
- no access granted to the community or userReadOnly
- read-only access granted to the community or userReadWrite
- read-write access granted to the community or user
0 - not-accessible
1 - accessible-for-notify
2 - read-only
3 - read-write
4 - read-create
Status values
1 - active
2 - notInService
3 - notReady
Actions
4 - createAndGo
5 - createAndWait
6 - destroy
1 - EIp4AddressSize
2 - EUnknownObjectType
3 - EUnknownPduType
4 - ECouldNotDecrypt
5 - EAuthFailure
6 - EReqResOidNoMatch
- `7 - (no longer used)
8 - EOutOfOrder
9 - EVersionNoMatch
10 - ECommunityNoMatch
11 - EUnexpectedReport
12 - EResponseNotHandled
13 - EUnexpectedResponse
oid - oid
path - path
module - module
Some parts of this module accept simple OID strings, e.g.:
var oid = "1.3.6.1.2.1.1.5.0";
Other parts take an OID string, it's type and value. This is collectively referred to as a varbind, and is specified as an object, e.g.:
var varbind = {
oid: "1.3.6.1.2.1.1.5.0",
type: snmp.ObjectType.OctetString,
value: new Buffer ("host1")
};
The type
parameter is one of the constants defined in the snmp.ObjectType
object.
The JavaScript true
and false
keywords are used for the values of varbinds
with type Boolean
.
All integer based types are specified as expected (this includes Integer
,
Counter
, Gauge
, TimeTicks
, Integer32
, Counter32
, Gauge32
, and
Unsigned32
), e.g. -128
or 100
.
Since JavaScript does not offer full 64 bit integer support objects with type
Counter64
cannot be supported in the same way as other integer types,
instead Node.js Buffer
objects are used. Users are responsible for
producing (i.e. for set()
requests) and consuming (i.e. the varbinds passed
to callback functions) Buffer
objects. That is, this module does not work
with 64 bit integers, it simply treats them as opaque Buffer
objects.
Dotted decimal strings are used for the values of varbinds with type OID
,
e.g. 1.3.6.1.2.1.1.5.0
.
Dotted quad formatted strings are used for the values of varbinds with type
IpAddress
, e.g. 192.168.1.1
.
Node.js Buffer
objects are used for the values of varbinds with
type Opaque
and OctetString
. For varbinds with type OctetString
this
module will accept JavaScript strings, but will always give back Buffer
objects.
The NoSuchObject
, NoSuchInstance
and EndOfMibView
types are used to
indicate an error condition. Currently there is no reason for users of this
module to to build varbinds using these types.
Most of the request methods exposed by this module require a mandatory callback function. This function is called once a request has been processed. This could be because an error occurred when processing the request, a trap has been dispatched or a successful response was received.
The first parameter to every callback is an error object. In the case no error occurred this parameter will be "null" indicating no error, e.g.:
function responseCb (error, varbinds) {
if (error) {
console.error (error);
} else {
// no error, do something with varbinds
}
}
When defined, the error
parameter is always an instance of the Error
class,
or a sub-class described in one of the sub-sections contained in this section.
The semantics of error handling is slightly different between SNMP version 1 and subsequent versions 2c and 3. In SNMP version 1 if an error occurs when calculating the value for one OID the request as a whole will fail, i.e. no OIDs will have a value.
This failure manifests itself within the error-status and error-index fields
of the response. When the error-status field in the response is non-zero,
i.e. not snmp.ErrorStatus.NoError
the callback
will be called with error
defined detailing the error.
Requests made with SNMP version 1 can simply assume all OIDs have a value when
no error object is passed to the callback
, i.e.:
var oids = ["1.3.6.1.2.1.1.5.0", "1.3.6.1.2.1.1.6.0"];
session.get (oids, function (error, varbinds) {
if (error) {
console.error (error.toString ());
} else {
var sysName = varbinds[0].value; // this WILL have a value
}
});
In SNMP versions 2c and 3, instead of using the error-status and error-index
fields of the response to signal an error, the value for the varbind placed in the
response for an OID will have an object syntax describing an error. The
error-status and error-index fields of the response will indicate the request
was successul, i.e. snmp.ErrorStatus.NoError
.
This changes the way in which error checking is performed in the callback
.
When using SNMP version 2c each varbind must be checked to see if its value
was computed and returned successfully:
var oids = ["1.3.6.1.2.1.1.5.0", "1.3.6.1.2.1.1.6.0"];
session.get (oids, function (error, varbinds) {
if (error) {
console.error (error.toString ());
} else {
if (varbinds[0].type != snmp.ErrorStatus.NoSuchObject
&& varbinds[0].type != snmp.ErrorStatus.NoSuchInstance
&& varbinds[0].type != snmp.ErrorStatus.EndOfMibView) {
var sysName = varbinds[0].value;
} else {
console.error (snmp.ObjectType[varbinds[0].type] + ": "
+ varbinds[0].oid);
}
}
});
This module exports two functions and promotes a specifc pattern to make error
checking a little simpler. Firstly, regardless of version in use varbinds can
always be checked. This results in a generic callback
that can be used for
both versions.
The isVarbindError()
function can be used to determine if a varbind has an
error condition. This function takes a single varbind
parameter and returns
true
if the varbind has an error condition, otherwise false
. The exported
varbindError()
function can then be used to obtain the error string
describing the error, which will include the OID for the varbind:
session.get (oids, function (error, varbinds) {
if (error) {
console.error (error.toString ());
} else {
if (snmp.isVarbindError (varbinds[0])) {
console.error (snmp.varbindError (varbinds[0]));
} else {
var sysName = varbinds[0].value;
}
}
});
If the varbindError
function is called with a varbind for which
isVarbindError
would return false, the string NotAnError
will be returned
appended with the related OID.
The sections following defines the error classes used by this module.
This error indicates a remote host failed to process a request. The exposed
message
attribute will contain a detailed error message. This error also
exposes a status
attribute which contains the error-index value from a
response. This will be one of the constants defined in the
snmp.ErrorStatus
object.
This error indicates a failure to render a request message before it could be
sent. The error can also indicate that a parameter provided was invalid.
The exposed message
attribute will contain a detailed error message.
This error states that no response was received for a particular request. The
exposed message
attribute will contain the value Request timed out
.
This error indicates a failure to parse a response message. The
exposed message
attribute will contain a detailed error message, and
as a sub-class of Error, its toString()
method will yield that
message
attribute.
An error of this class will always additionally include a code
attribute (one of the values in ResponseInvalidCode
); and in some
cases, also have an info
attribute which provides code
-specific
supplemental information. An authentication error, for example -- code
ResponseInvalidCode.EAuthFailure
-- will contain a map in info
with the attempted authentication data which failed to authenticate.
If a receiver or an agent receives a packet it is unable to decode,
then it will produce a ProcessingError
containing:
rinfo
information on the origin of the packet,- a
buffer
containing the packet contents, and - an
error
containing the original error encountered during processing.
This library provides a Session
class to provide support for building
"Command Generator" and "Notification Originator" SNMP applications.
All SNMP requests are made using an instance of the Session
class. This
module exports two functions that are used to create instances of the
Session
class:
createSession()
- for v1 and v2c sessionscreateV3Session()
- for v3 sessions
The createSession()
function instantiates and returns an instance of the
Session
class for SNMPv1 or SNMPv2c:
// Default options
var options = {
port: 161,
retries: 1,
timeout: 5000,
backoff: 1.0,
transport: "udp4",
trapPort: 162,
version: snmp.Version1,
backwardsGetNexts: true,
reportOidMismatchErrors: false,
idBitsSize: 32
};
var session = snmp.createSession ("127.0.0.1", "public", options);
The optional target
parameter defaults to 127.0.0.1
. The optional
community
parameter defaults to public
. The optional options
parameter
is an object, and can contain the following items:
port
- UDP port to send requests too, defaults to161
retries
- Number of times to re-send a request, defaults to1
sourceAddress
- IP address from which SNMP requests should originate, there is no default for this option, the operating system will select an appropriate source address when the SNMP request is sentsourcePort
- UDP port from which SNMP requests should originate, defaults to an ephemeral port selected by the operation systemtimeout
- Number of milliseconds to wait for a response before re-trying or failing, defaults to5000
backoff
- The factor by which to increase thetimeout
for every retry, defaults to1
for no increasetransport
- Specify the transport to use, can be eitherudp4
orudp6
, defaults toudp4
trapPort
- UDP port to send traps and informs too, defaults to162
version
- Eithersnmp.Version1
orsnmp.Version2c
, defaults tosnmp.Version1
backwardsGetNexts
- boolean to allow GetNext operations to retrieve lexicographically preceding OIDs, defaults totrue
reportOidMismatchErrors
- boolean to allow error reporting of OID mismatches between requests and responses, defaults tofalse
idBitsSize
- Either16
or32
, defaults to32
. Used to reduce the size of the generated id for compatibility with some older devices.
When a session has been finished with it should be closed:
session.close ();
The createV3Session()
function instantiates and returns an instance of the
same Session
class as createSession()
, only instead initialized for SNMPv3:
// Default options for v3
var options = {
port: 161,
retries: 1,
timeout: 5000,
transport: "udp4",
trapPort: 162,
version: snmp.Version3,
engineID: "8000B98380XXXXXXXXXXXXXXXXXXXXXXXX", // where the X's are random hex digits
backwardsGetNexts: true,
reportOidMismatchErrors: false,
idBitsSize: 32,
context: ""
};
// Example user
var user = {
name: "blinkybill",
level: snmp.SecurityLevel.authPriv,
authProtocol: snmp.AuthProtocols.sha,
authKey: "madeahash",
privProtocol: snmp.PrivProtocols.des,
privKey: "privycouncil"
};
var session = snmp.createV3Session ("127.0.0.1", user, options);
The target
and user
parameters are mandatory. The optional options
parameter
has the same meaning as for the createSession()
call. The one additional field
in the options parameter is the context
field, which adds an SNMPv3 context to
the session.
The user
object must contain a name
and level
field. The level
field can
take these values from the snmp.SecurityLevel
object:
snmp.SecurityLevel.noAuthNoPriv
- for no message authentication or encryptionsnmp.SecurityLevel.authNoPriv
- for message authentication and no encryptionsnmp.SecurityLevel.authPriv
- for message authentication and encryption
The meaning of these are as per RFC3414. If the level
supplied is authNoPriv
or
authPriv
, then the authProtocol
and authKey
fields must also be present. The
authProtocol
field can take values from the snmp.AuthProtocols
object:
snmp.AuthProtocols.md5
- for MD5 message authenticationsnmp.AuthProtocols.sha
- for SHA message authentication
If the level
supplied is authPriv
, then the privProtocol
and privKey
fields
must also be present. The privProtocol
field can take values from the
snmp.PrivProtocols
object:
snmp.PrivProtocols.des
- for DES encryptionsnmp.PrivProtocols.aes
- for AES encryption
Once a v3 session is created, the same set of session
methods are available as
for v1 and v2c.
The close
event is emitted by the session when the session's underlying UDP
socket is closed.
No arguments are passed to the callback.
Before this event is emitted all outstanding requests are cancelled, resulting
in the failure of each outstanding request. The error passed back through to
each request will be an instance of the Error
class with the errors
message
attribute set to Socket forcibly closed
.
The following example prints a message to the console when a session's underlying UDP socket is closed:
session.on ("close", function () {
console.log ("socket closed");
});
The error
event is emitted by the session when the session's underlying UDP
socket emits an error.
The following arguments will be passed to the callback
function:
error
- An instance of theError
class, the exposedmessage
attribute will contain a detailed error message.
The following example prints a message to the console when an error occurs with a session's underlying UDP socket, the session is then closed:
session.on ("error", function (error) {
console.log (error.toString ());
session.close ();
});
The close()
method closes the sessions underlying UDP socket. This will
result in the close
event being emitted by the sessions underlying UDP
socket which is passed through to the session, resulting in the session also
emitting a close
event.
The following example closes a sessions underlying UDP socket:
session.close ();
The get()
method fetches the value for one or more OIDs.
The oids
parameter is an array of OID strings. The callback
function is
called once the request is complete. The following arguments will be passed
to the callback
function:
error
- Instance of theError
class or a sub-class, ornull
if no error occurredvarbinds
- Array of varbinds, will not be provided if an error occurred
The varbind in position N in the varbinds
array will correspond to the OID
in position N in the oids
array in the request.
Each varbind must be checked for an error condition using the
snmp.isVarbindError()
function when using SNMP version 2c.
The following example fetches values for the sysName (1.3.6.1.2.1.1.5.0
) and
sysLocation (1.3.6.1.2.1.1.6.0
) OIDs:
var oids = ["1.3.6.1.2.1.1.5.0", "1.3.6.1.2.1.1.6.0"];
session.get (oids, function (error, varbinds) {
if (error) {
console.error (error.toString ());
} else {
for (var i = 0; i < varbinds.length; i++) {
// for version 1 we can assume all OIDs were successful
console.log (varbinds[i].oid + "|" + varbinds[i].value);
// for version 2c we must check each OID for an error condition
if (snmp.isVarbindError (varbinds[i]))
console.error (snmp.varbindError (varbinds[i]));
else
console.log (varbinds[i].oid + "|" + varbinds[i].value);
}
}
});
The getBulk()
method fetches the value for the OIDs lexicographically
following one or more OIDs in the MIB tree.
The oids
parameter is an array of OID strings. The optional nonRepeaters
parameter specifies the number of OIDs in the oids
parameter for which only
1 varbind should be returned, and defaults to 0
. For each remaining OID
in the oids
parameter the optional maxRepetitions
parameter specifies how
many OIDs lexicographically following an OID for which varbinds should be
fetched, and defaults to 20
.
The callback
function is called once the request is complete. The following
arguments will be passed to the callback
function:
error
- Instance of theError
class or a sub-class, ornull
if no error occurredvarbinds
- Array of varbinds, will not be provided if an error occurred
The varbind in position N in the varbinds
array will correspond to the OID
in position N in the oids
array in the request.
For for the first nonRepeaters
items in varbinds
each item will be a
single varbind. For all remaining items in varbinds
each item will be an
array of varbinds - this makes it easy to tie response varbinds with requested
OIDs since response varbinds are grouped and placed in the same position in
varbinds
.
Each varbind must be checked for an error condition using the
snmp.isVarbindError()
function when using SNMP version 2c.
The following example fetches values for the OIDs following the sysContact
(1.3.6.1.2.1.1.4.0
) and sysName (1.3.6.1.2.1.1.5.0
) OIDs, and up to the
first 20 OIDs in the ifDescr (1.3.6.1.2.1.2.2.1.2
) and ifType
(1.3.6.1.2.1.2.2.1.3
) columns from the ifTable (1.3.6.1.2.1.2.2
) table:
var oids = [
"1.3.6.1.2.1.1.4.0",
"1.3.6.1.2.1.1.5.0",
"1.3.6.1.2.1.2.2.1.2",
"1.3.6.1.2.1.2.2.1.3"
];
var nonRepeaters = 2;
session.getBulk (oids, nonRepeaters, function (error, varbinds) {
if (error) {
console.error (error.toString ());
} else {
// step through the non-repeaters which are single varbinds
for (var i = 0; i < nonRepeaters; i++) {
if (i >= varbinds.length)
break;
if (snmp.isVarbindError (varbinds[i]))
console.error (snmp.varbindError (varbinds[i]));
else
console.log (varbinds[i].oid + "|" + varbinds[i].value);
}
// then step through the repeaters which are varbind arrays
for (var i = nonRepeaters; i < varbinds.length; i++) {
for (var j = 0; j < varbinds[i].length; j++) {
if (snmp.isVarbindError (varbinds[i][j]))
console.error (snmp.varbindError (varbinds[i][j]));
else
console.log (varbinds[i][j].oid + "|"
+ varbinds[i][j].value);
}
}
}
});
The getNext()
method fetches the value for the OIDs lexicographically
following one or more OIDs in the MIB tree.
The oids
parameter is an array of OID strings. The callback
function is
called once the request is complete. The following arguments will be passed
to the callback
function:
error
- Instance of theError
class or a sub-class, ornull
if no error occurredvarbinds
- Array of varbinds, will not be provided if an error occurred
The varbind in position N in the varbinds
array will correspond to the OID
in position N in the oids
array in the request.
Each varbind must be checked for an error condition using the
snmp.isVarbindError()
function when using SNMP version 2c.
The following example fetches values for the next OIDs following the
sysObjectID (1.3.6.1.2.1.1.1.0
) and sysName (1.3.6.1.2.1.1.4.0
) OIDs:
var oids = [
"1.3.6.1.2.1.1.1.0",
"1.3.6.1.2.1.1.4.0"
];
session.getNext (oids, function (error, varbinds) {
if (error) {
console.error (error.toString ());
} else {
for (var i = 0; i < varbinds.length; i++) {
// for version 1 we can assume all OIDs were successful
console.log (varbinds[i].oid + "|" + varbinds[i].value);
// for version 2c we must check each OID for an error condition
if (snmp.isVarbindError (varbinds[i]))
console.error (snmp.varbindError (varbinds[i]));
else
console.log (varbinds[i].oid + "|" + varbinds[i].value);
}
}
});
The inform()
method sends a SNMP inform.
The typeOrOid
parameter can be one of two types; one of the constants
defined in the snmp.TrapType
object (excluding the
snmp.TrapType.EnterpriseSpecific
constant), or an OID string.
The first varbind to be placed in the request message will be for the
sysUptime.0
OID (1.3.6.1.2.1.1.3.0
). The value for this varbind will
be the value returned by the process.uptime ()
function multiplied by 100
(this can be overridden by providing upTime
in the optional options
parameter, as documented below).
This will be followed by a second varbind for the snmpTrapOID.0
OID
(1.3.6.1.6.3.1.1.4.1.0
). The value for this will depend on the typeOrOid
parameter. If a constant is specified the trap OID for the constant will be
used as supplied for the varbinds value, otherwise the OID string specified
will be used as is for the value of the varbind.
The optional varbinds
parameter is an array of varbinds to include in the
inform request, and defaults to the empty array []
.
The optional options
parameter is an object, and can contain the following
items:
upTime
- Value of thesysUptime.0
OID (1.3.6.1.2.1.1.3.0
) in the inform, defaults to the value returned by theprocess.uptime ()
function multiplied by 100
The callback
function is called once a response to the inform request has
been received, or an error occurred. The following arguments will be passed
to the callback
function:
error
- Instance of theError
class or a sub-class, ornull
if no error occurredvarbinds
- Array of varbinds, will not be provided if an error occurred
The varbind in position N in the varbinds
array will correspond to the
varbind in position N in the varbinds
array in the request. The remote host
should echo back varbinds and their values as specified in the request, and
the varbinds
array will contain each varbind as sent back by the remote host.
Normally there is no reason to use the contents of the varbinds
parameter
since the varbinds are as they were sent in the request.
The following example sends a generic cold-start inform to a remote host, it does not include any varbinds:
session.inform (snmp.TrapType.ColdStart, function (error) {
if (error)
console.error (error);
});
The following example sends an enterprise specific inform to a remote host, and includes two enterprise specific varbinds:
var informOid = "1.3.6.1.4.1.2000.1";
var varbinds = [
{
oid: "1.3.6.1.4.1.2000.2",
type: snmp.ObjectType.OctetString,
value: "Periodic hardware self-check"
},
{
oid: "1.3.6.1.4.1.2000.3",
type: snmp.ObjectType.OctetString,
value: "hardware-ok"
}
];
// Override sysUpTime, specfiying it as 10 seconds...
var options = {upTime: 1000};
session.inform (informOid, varbinds, options, function (error) {
if (error)
console.error (error);
});
The set()
method sets the value of one or more OIDs.
The varbinds
parameter is an array of varbind objects. The callback
function is called once the request is complete. The following arguments will
be passed to the callback
function:
error
- Instance of theError
class or a sub-class, ornull
if no error occurredvarbinds
- Array of varbinds, will not be provided if an error occurred
The varbind in position N in the varbinds
array will correspond to the
varbind in position N in the varbinds
array in the request. The remote host
should echo back varbinds and their values as specified in the request unless
an error occurred. The varbinds
array will contain each varbind as sent
back by the remote host.
Each varbind must be checked for an error condition using the
snmp.isVarbindError()
function when using SNMP version 2c.
The following example sets the value of the sysName (1.3.6.1.2.1.1.4.0
) and
sysLocation (1.3.6.1.2.1.1.6.0
) OIDs:
var varbinds = [
{
oid: "1.3.6.1.2.1.1.5.0",
type: snmp.ObjectType.OctetString,
value: "host1"
}, {
oid: "1.3.6.1.2.1.1.6.0",
type: snmp.ObjectType.OctetString,
value: "somewhere"
}
];
session.set (varbinds, function (error, varbinds) {
if (error) {
console.error (error.toString ());
} else {
for (var i = 0; i < varbinds.length; i++) {
// for version 1 we can assume all OIDs were successful
console.log (varbinds[i].oid + "|" + varbinds[i].value);
// for version 2c we must check each OID for an error condition
if (snmp.isVarbindError (varbinds[i]))
console.error (snmp.varbindError (varbinds[i]));
else
console.log (varbinds[i].oid + "|" + varbinds[i].value);
}
}
});
The subtree()
method fetches the value for all OIDs lexicographically
following a specified OID in the MIB tree which have the specified OID as
their base. For example, the OIDs sysName (1.3.6.1.2.1.1.5.0
) and
sysLocation (1.3.6.1.2.1.1.6.0
) both have the same base system
(1.3.6.1.2.1.1
) OID.
For SNMP version 1 repeated get()
calls are made until the one of the
returned OIDs does not use the specified OID as its base. For SNMP version
2c repeated getBulk()
calls are made until the one of the returned OIDs
does no used the specified OID as its base.
The oid
parameter is an OID string. The optional maxRepetitions
parameter
is passed to getBulk()
requests when SNMP version 2c is being used.
This method will not call a single callback once all OID values are fetched.
Instead the feedCallback
function will be called each time a response is
received from the remote host. The following arguments will be passed to the
feedCallback
function:
varbinds
- Array of varbinds, and will contain at least one varbind
Each varbind must be checked for an error condition using the
snmp.isVarbindError()
function when using SNMP version 2c.
Once at least one of the returned OIDs does not use the specified OID as its
base, or an error has occurred, the doneCallback
function will be called.
The following arguments will be passed to the doneCallback
function:
error
- Instance of theError
class or a sub-class, ornull
if no error occurred
Once the doneCallback
function has been called the request is complete and
the feedCallback
function will no longer be called.
If the feedCallback
function returns a true
value when called no more
get()
or getBulk()
method calls will be made and the doneCallback
will
be called.
The following example fetches all OIDS under the system (1.3.6.1.2.1.1
) OID:
var oid = "1.3.6.1.2.1.1";
function doneCb (error) {
if (error)
console.error (error.toString ());
}
function feedCb (varbinds) {
for (var i = 0; i < varbinds.length; i++) {
if (snmp.isVarbindError (varbinds[i]))
console.error (snmp.varbindError (varbinds[i]));
else
console.log (varbinds[i].oid + "|" + varbinds[i].value);
}
}
var maxRepetitions = 20;
// The maxRepetitions argument is optional, and will be ignored unless using
// SNMP verison 2c
session.subtree (oid, maxRepetitions, feedCb, doneCb);
The table()
method fetches the value for all OIDs lexicographically
following a specified OID in the MIB tree which have the specified OID as
their base, much like the subtree()
method.
This method is designed to fetch conceptual tables, for example the ifTable
(1.3.6.1.2.1.2.2
) table. The values for returned varbinds will be structured
into objects to represent conceptual rows. Each row is then placed into an
object with the rows index being the key, e.g.:
var table = {
// Rows keyed by ifIndex (1 and 2 are shown)
1: {
// ifDescr (column 2) and ifType (columnd 3) are shown
2: "interface-1",
3: 6,
...
},
2: {
2: "interface-2",
3: 6,
...
},
...
}
Internally this method calls the subtree()
method to obtain the subtree of
the specified table.
The oid
parameter is an OID string. If an OID string is passed which does
not represent a table the resulting object produced to hold table data will be
empty, i.e. it will contain no indexes and rows. The optional
maxRepetitions
parameter is passed to the subtree()
request.
The callback
function will be called once the entire table has been fetched.
The following arguments will be passed to the callback
function:
error
- Instance of theError
class or a sub-class, ornull
if no error occurredtable
- Object containing object references representing conceptual rows keyed by index (e.g. for the ifTable table rows are keyed by ifIndex), each row object will contain values keyed by column number, will not be provided if an error occurred
If an error occurs with any varbind returned by subtree()
no table will be
passed to the callback
function. The reason for failure, and the related
OID string (as returned from a call to the snmp.varbindError()
function),
will be passed to the callback
function in the error
argument as an
instance of the RequestFailedError
class.
The following example fetches the ifTable (1.3.6.1.2.1.2.2
) table:
var oid = "1.3.6.1.2.1.2.2";
function sortInt (a, b) {
if (a > b)
return 1;
else if (b > a)
return -1;
else
return 0;
}
function responseCb (error, table) {
if (error) {
console.error (error.toString ());
} else {
// This code is purely used to print rows out in index order,
// ifIndex's are integers so we'll sort them numerically using
// the sortInt() function above
var indexes = [];
for (index in table)
indexes.push (parseInt (index));
indexes.sort (sortInt);
// Use the sorted indexes we've calculated to walk through each
// row in order
for (var i = 0; i < indexes.length; i++) {
// Like indexes we sort by column, so use the same trick here,
// some rows may not have the same columns as other rows, so
// we calculate this per row
var columns = [];
for (column in table[indexes[i]])
columns.push (parseInt (column));
columns.sort (sortInt);
// Print index, then each column indented under the index
console.log ("row for index = " + indexes[i]);
for (var j = 0; j < columns.length; j++) {
console.log (" column " + columns[j] + " = "
+ table[indexes[i]][columns[j]]);
}
}
}
}
var maxRepetitions = 20;
// The maxRepetitions argument is optional, and will be ignored unless using
// SNMP verison 2c
session.table (oid, maxRepetitions, responseCb);
The tableColumns()
method implements the same interface as the table()
method. However, only the columns specified in the columns
parameter will
be in the resulting table.
This method should be used when only selected columns are required, and
will be many times faster than the table()
method since a much smaller
amount of data will be fetched.
The following example fetches the ifTable (1.3.6.1.2.1.2.2
) table, and
specifies that only the ifDescr (1.3.6.1.2.1.2.2.1.2
) and ifPhysAddress
(1.3.6.1.2.1.2.2.1.6
) columns should actually be fetched:
var oid = "1.3.6.1.2.1.2.2";
var columns = [2, 6];
function sortInt (a, b) {
if (a > b)
return 1;
else if (b > a)
return -1;
else
return 0;
}
function responseCb (error, table) {
if (error) {
console.error (error.toString ());
} else {
// This code is purely used to print rows out in index order,
// ifIndex's are integers so we'll sort them numerically using
// the sortInt() function above
var indexes = [];
for (index in table)
indexes.push (parseInt (index));
indexes.sort (sortInt);
// Use the sorted indexes we've calculated to walk through each
// row in order
for (var i = 0; i < indexes.length; i++) {
// Like indexes we sort by column, so use the same trick here,
// some rows may not have the same columns as other rows, so
// we calculate this per row
var columns = [];
for (column in table[indexes[i]])
columns.push (parseInt (column));
columns.sort (sortInt);
// Print index, then each column indented under the index
console.log ("row for index = " + indexes[i]);
for (var j = 0; j < columns.length; j++) {
console.log (" column " + columns[j] + " = "
+ table[indexes[i]][columns[j]]);
}
}
}
}
var maxRepetitions = 20;
// The maxRepetitions argument is optional, and will be ignored unless using
// SNMP verison 2c
session.tableColumns (oid, columns, maxRepetitions, responseCb);
The trap()
method sends a SNMP trap.
The typeOrOid
parameter can be one of two types; one of the constants
defined in the snmp.TrapType
object (excluding the
snmp.TrapType.EnterpriseSpecific
constant), or an OID string.
For SNMP version 1 when a constant is specified the following fields are set in the trap:
- The enterprise field is set to the OID
1.3.6.1.4.1
- The generic-trap field is set to the constant specified
- The specific-trap field is set to 0
When an OID string is specified the following fields are set in the trap:
- The final decimal is stripped from the OID string and set in the specific-trap field
- The remaining OID string is set in the enterprise field
- The generic-trap field is set to the constant
snmp.TrapType.EnterpriseSpecific
In both cases the time-stamp field in the trap PDU is set to the value
returned by the process.uptime ()
function multiplied by 100
.
SNMP version 2c messages are quite different in comparison with version 1.
The version 2c trap has a much simpler format, simply a sequence of varbinds.
The first varbind to be placed in the trap message will be for the
sysUptime.0
OID (1.3.6.1.2.1.1.3.0
). The value for this varbind will
be the value returned by the process.uptime ()
function multiplied by 100
(this can be overridden by providing upTime
in the optional options
parameter, as documented below).
This will be followed by a second varbind for the snmpTrapOID.0
OID
(1.3.6.1.6.3.1.1.4.1.0
). The value for this will depend on the typeOrOid
parameter. If a constant is specified the trap OID for the constant
will be used as supplied for the varbinds value, otherwise the OID string
specified will be used as is for the value of the varbind.
The optional varbinds
parameter is an array of varbinds to include in the
trap, and defaults to the empty array []
.
The optional agentAddrOrOptions
parameter can be one of two types; one is
the IP address used to populate the agent-addr field for SNMP version 1 type
traps, and defaults to 127.0.0.1
, or an object, and can contain the
following items:
agentAddr
- IP address used to populate the agent-addr field for SNMP version 1 type traps, and defaults to127.0.0.1
upTime
- Value of thesysUptime.0
OID (1.3.6.1.2.1.1.3.0
) in the trap, defaults to the value returned by theprocess.uptime ()
function multiplied by 100
NOTE When using SNMP version 2c the agentAddr
parameter is ignored if
specified since version 2c trap messages do not have an agent-addr field.
The callback
function is called once the trap has been sent, or an error
occurred. The following arguments will be passed to the callback
function:
error
- Instance of theError
class or a sub-class, ornull
if no error occurred
The following example sends an enterprise specific trap to a remote host using
a SNMP version 1 trap, and includes the sysName (1.3.6.1.2.1.1.5.0
) varbind
in the trap. Before the trap is sent the agentAddr
field is calculated using
DNS to resolve the hostname of the local host:
var enterpriseOid = "1.3.6.1.4.1.2000.1"; // made up, but it may be valid
var varbinds = [
{
oid: "1.3.6.1.2.1.1.5.0",
type: snmp.ObjectType.OctetString,
value: "host1"
}
];
dns.lookup (os.hostname (), function (error, agentAddress) {
if (error) {
console.error (error);
} else {
// Override sysUpTime, specfiying it as 10 seconds...
var options = {agentAddr: agentAddress, upTime: 1000};
session.trap (enterpriseOid, varbinds, agentAddress,
function (error) {
if (error)
console.error (error);
});
}
});
The following example sends a generic link-down trap to a remote host using a
SNMP version 1 trap, it does not include any varbinds or specify the
agentAddr
parameter:
session.trap (snmp.TrapType.LinkDown, function (error) {
if (error)
console.error (error);
});
The following example sends an enterprise specific trap to a remote host using a SNMP version 2c trap, and includes two enterprise specific varbinds:
var trapOid = "1.3.6.1.4.1.2000.1";
var varbinds = [
{
oid: "1.3.6.1.4.1.2000.2",
type: snmp.ObjectType.OctetString,
value: "Hardware health status changed"
},
{
oid: "1.3.6.1.4.1.2000.3",
type: snmp.ObjectType.OctetString,
value: "status-error"
}
];
// version 2c should have been specified when creating the session
session.trap (trapOid, varbinds, function (error) {
if (error)
console.error (error);
});
The walk()
method fetches the value for all OIDs lexicographically following
a specified OID in the MIB tree.
For SNMP version 1 repeated get()
calls are made until the end of the MIB
tree is reached. For SNMP version 2c repeated getBulk()
calls are made
until the end of the MIB tree is reached.
The oid
parameter is an OID string. The optional maxRepetitions
parameter
is passed to getBulk()
requests when SNMP version 2c is being used.
This method will not call a single callback once all OID values are fetched.
Instead the feedCallback
function will be called each time a response is
received from the remote host. The following arguments will be passed to the
feedCallback
function:
varbinds
- Array of varbinds, and will contain at least one varbind
Each varbind must be checked for an error condition using the
snmp.isVarbindError()
function when using SNMP version 2c.
Once the end of the MIB tree has been reached, or an error has occurred, the
doneCallback
function will be called. The following arguments will be
passed to the doneCallback
function:
error
- Instance of theError
class or a sub-class, ornull
if no error occurred
Once the doneCallback
function has been called the request is complete and
the feedCallback
function will no longer be called.
If the feedCallback
function returns a true
value when called no more
get()
or getBulk()
method calls will be made and the doneCallback
will
be called.
The following example walks to the end of the MIB tree starting from the
ifTable (1.3.6.1.2.1.2.2
) OID:
var oid = "1.3.6.1.2.1.2.2";
function doneCb (error) {
if (error)
console.error (error.toString ());
}
function feedCb (varbinds) {
for (var i = 0; i < varbinds.length; i++) {
if (snmp.isVarbindError (varbinds[i]))
console.error (snmp.varbindError (varbinds[i]));
else
console.log (varbinds[i].oid + "|" + varbinds[i].value);
}
}
var maxRepetitions = 20;
// The maxRepetitions argument is optional, and will be ignored unless using
// SNMP verison 2c
session.walk (oid, maxRepetitions, feedCb, doneCb);
RFC 3413 classifies a "Notification Receiver" SNMP application that receives "Notification-Class" PDUs. Notifications include both SNMP traps and informs. This library is able to receive all types of notification PDU:
Trap-PDU
(original v1 trap PDUs, which are now considered obselete)Trapv2-PDU
(unacknowledged notifications)InformRequest-PDU
(same format asTrapv2-PDU
but with message acknowledgement)
The library provides a Receiver
class for receiving SNMP notifications. This
module exports the createReceiver()
function, which creates a new Receiver
instance.
The receiver creates an Authorizer
instance to control incoming access. More
detail on this is found below in the Authorizer Module section
below.
The createReceiver()
function instantiates and returns an instance of the Receiver
class:
// Default options
var options = {
port: 162,
disableAuthorization: false,
includeAuthentication: false,
accessControlModelType: snmp.AccessControlModelType.None,
engineID: "8000B98380XXXXXXXXXXXXXXXXXXXXXXXX", // where the X's are random hex digits
address: null,
transport: "udp4"
};
var callback = function (error, notification) {
if ( error ) {
console.error (error);
} else {
console.log (JSON.stringify(notification, null, 2));
}
};
receiver = snmp.createReceiver (options, callback);
The options
and callback
parameters are mandatory. The options
parameter is
an object, possibly empty, and can contain the following fields:
port
- the port to listen for notifications on - defaults to 162. Note that binding to port 162 on some systems requires the receiver process to be run with administrative privilege. If this is not possible then choose a port greater than 1024.disableAuthorization
- disables local authorization for all community-based notifications received and for those user-based notifications received with no message authentication or privacy (noAuthNoPriv) - defaults to falseengineID
- the engineID used for SNMPv3 communications, given as a hex string - defaults to a system-generated engineID containing elements of randomtransport
- the transport family to use - defaults toudp4
address
- the IP address to bind to - default tonull
, which means bind to all IP addressesincludeAuthentication
- adds the community (v1/2c) or user name (v3) information to the notification callback - defaults tofalse
sockets
- an array of objects containing triples oftransport
,address
andport
that can be used to specify multiple socket listeners. This option overrides any individualtransport
,address
andport
options.
The callback
parameter is a callback function of the form
function (error, notification)
. On an error condition, the notification
parameter is set to null
. On successful reception of a notification, the error
parameter is set to null
, and the notification
parameter is set as an object
with the notification PDU details in the pdu
field and the sender socket details
in the rinfo
field. For example:
{
"pdu": {
"type": 166,
"id": 45385686,
"varbinds": [
{
"oid": "1.3.6.1.2.1.1.3.0",
"type": 67,
"value": 5
},
{
"oid": "1.3.6.1.6.3.1.1.4.1.0",
"type": 6,
"value": "1.3.6.1.6.3.1.1.5.2"
}
],
"scoped": false
},
"rinfo": {
"address": "127.0.0.1",
"family": "IPv4",
"port": 43162,
"size": 72
}
}
Returns the receiver's Authorizer
instance, used to control access
to the receiver. See the Authorizer
section for further details.
Closes the receiver's listening socket, ending the operation of the receiver.
The SNMP agent responds to all four "request class" PDUs relevant to a Command Responder application:
- GetRequest - request exactly matched OID instances
- GetNextRequest - request lexicographically "next" OID instances in the MIB tree
- GetBulkRequest - request a series of "next" OID instances in the MIB tree
- SetRequest - set values for specified OIDs
The agent sends a GetResponse PDU to all four request PDU types, in conformance with RFC 3416.
The agent - like the notification receiver - maintains an Authorizer
instance
to control access to the agent, details of which are in the Authorizer Module
section below.
The central data structure that the agent maintains is a Mib
instance, the API of which is
detailed in the Mib Module section below. The agent allows the MIB to be queried
and manipulated through the API, as well as queried and manipulated through the SNMP interface with
the above four request-class PDUs.
The agent also supports SNMP proxy forwarder applications with its singleton Forwarder
instance,
which is documented in the Forwarder Module section below.
The createAgent()
function instantiates and returns an instance of the Agent
class:
// Default options
var options = {
port: 161,
disableAuthorization: false,
accessControlModelType: snmp.AccessControlModelType.None,
engineID: "8000B98380XXXXXXXXXXXXXXXXXXXXXXXX", // where the X's are random hex digits
address: null,
transport: "udp4"
};
var callback = function (error, data) {
if ( error ) {
console.error (error);
} else {
console.log (JSON.stringify(data, null, 2));
}
};
agent = snmp.createAgent (options, callback);
The options
and callback
parameters are mandatory. The options
parameter is
an object, possibly empty, and can contain the following fields:
port
- the port for the agent to listen on - defaults to 161. Note that binding to port 161 on some systems requires the receiver process to be run with administrative privilege. If this is not possible, then choose a port greater than 1024.disableAuthorization
- disables local authorization for all community-based notifications received and for those user-based notifications received with no message authentication or privacy (noAuthNoPriv) - defaults to falseaccessControlModelType
- specifies which access control model to use. Defaults tosnmp.AccessControlModelType.None
, but can be set tosnmp.AccessControlModelType.Simple
for further access control capabilities. See theAuthorization
class description for more information.engineID
- the engineID used for SNMPv3 communications, given as a hex string - defaults to a system-generated engineID containing elements of randomtransport
- the transport family to use - defaults toudp4
address
- the IP address to bind to - default tonull
, which means bind to all IP addressessockets
- an array of objects containing triples oftransport
,address
andport
that can be used to specify multiple socket listeners. This option overrides any individualtransport
,address
andport
options.
The mib
parameter is optional, and sets the agent's singleton Mib
instance.
If not supplied, the agent creates itself a new empty Mib
singleton. If supplied,
the Mib
instance needs to be created and populated as per the Mib Module
section below.
Returns the agent's singleton Authorizer
instance, used to control access
to the agent. See the Authorizer
section for further details.
Returns the agent's singleton Mib
instance, which holds all of the management data
for the agent.
Sets the agent's singleton Mib
instance to the supplied one. The agent discards
its existing Mib
instance.
Returns the agent's singleton Forwarder
instance, which holds a list of registered
proxies that specify context-based forwarding to remote hosts.
Closes the agent's listening socket, ending the operation of the agent.
Both the receiver and agent maintain an singleton Authorizer
instance, which is
responsible for maintaining an authorization list of SNMP communities (for v1 and
v2c notifications) and also an authorization list of SNMP users (for v3 notifications).
These lists are used to authorize notification access to the receiver, and to store
security protocol and key settings. RFC 3414 terms the user list as the the
"usmUserTable" stored in the receiver's "Local Configuration Database".
If a v1 or v2c notification is received with a community that is not in the
receiver's community authorization list, the receiver will not accept the notification,
instead returning a error of class RequestFailedError
to the supplied callback
function. Similarly, if a v3 notification is received with a user whose name is
not in the receiver's user authorization list, the receiver will return a
RequestFailedError
. If the disableAuthorization
option is supplied for the
receiver on start-up, then these local authorization list checks are disabled for
community notifications and noAuthNoPriv user notifications. Note that even with
this setting, the user list is still checked for authNoPriv and authPriv notifications,
as the library still requires access to the correct keys for the message authentication
and encryption operations, and these keys are stored against a user in the user
authorization list.
The API allows the receiver's / agent's community authorization and user authorization lists to be managed with adds, queries and deletes.
For an agent, there is a further optional access control check, that can limit the
access for a given community or user according to the AccessControlModelType
supplied
as an option to the agent. The default model type is snmp.AccessControlModelType.None
,
which means that - after the authorization list checks described in the preceding paragraphs -
there is no further access control restrictions i.e. all requests are granted access by
the agent. A second access control model type snmp.AccessControlModelType.Simple
can
be selected, which creates a SimpleAccessControlModel
object that can be manipulated
to specify that a community or user has one of three levels of access to agent information:
- none
- read-only
- read-write
More information on how to configure access with the SimpleAccessControlModel
class is
provided below under the description of that class.
The authorizer instance can be obtained by using the getAuthorizer()
call, for both the receiver and the agent. For example:
receiver.getAuthorizer ().getCommunities ();
Adds a community string to the receiver's community authorization list. Does nothing if the community is already in the list, ensuring there is only one occurence of any given community string in the list.
Returns a community string if it is stored in the receiver's community authorization
list, otherwise returns null
.
Returns the receiver's community authorization list.
Deletes a community string from the receiver's community authorization list. Does nothing if the community is not in the list.
Adds a user to the receiver's user authorization list. If a user of the same name
is in the list, this call deletes the existing user, and replaces it with the supplied
user, ensuring that only one user with a given name will exist in the list. The user
object is in the same format as that used for the session.createV3Session()
call.
var user = {
name: "elsa",
level: snmp.SecurityLevel.authPriv,
authProtocol: snmp.AuthProtocols.sha,
authKey: "imlettingitgo",
privProtocol: snmp.PrivProtocols.des,
privKey: "intotheunknown"
};
receiver.getAuthorizer ().addUser (elsa);
Returns a user object if a user with the supplied name is stored in the receiver's
user authorization list, otherwise returns null
.
Returns the receiver's user authorization list.
Deletes a user from the receiver's user authorization list. Does nothing if the user with the supplied name is not in the list.
Returns the snmp.AccessControlModelType
of this authorizer, which is one of:
snmp.AccessControlModelType.None
snmp.AccessControlModelType.Simple
Returns the access control model object:
- for a type of
snmp.AccessControlModelType.None
- returns null (as the access control check returns positive every time) - for a type of
snmp.AccessControlModelType.Simple
- returns aSimpleAccessControlModel
object
The SimpleAccessControlModel
class can be optionally selected as the access control model used by an Agent
. The
SimpleAccessControlModel
provides basic three-level access control for a given community or user.
The access levels are specified in the snmp.AccessLevel constant:
snmp.AccessLevel.None
- no access is granted to the community or usersnmp.AccessLevel.ReadOnly
- access is granted for the community or user for Get, GetNext and GetBulk requests but not Set requestssnmp.AccessLevel.ReadWrite
- access is granted for the community or user for Get, GetNext, GetBulk and Set requests
The SimpleAccessControlModel
is not created via a direct API call, but is created internally by an Agent
's Authorizer
singleton.
So an agent's access control model can be accessed with:
var acm = agent.getAuthorizer ().getAccessControlModel ();
Note that any community or user that is used in any of the API calls in this section must first be created in the agent's Authorizer
,
otherwise the agent will fail the initial community/user list check that the authorizer performs.
When using the Simple Access Control Model, the default access level for a newly created community or user in the
Authorizer
is read-only.
Example use:
var agent = snmp.createAgent({
accessControlModelType: snmp.AccessControlModelType.Simple
}, function (error, data) {
// null callback for example brevity
});
var authorizer = agent.getAuthorizer ();
authorizer.addCommunity ("public");
authorizer.addCommunity ("private");
authorizer.addUser ({
name: "fred",
level: snmp.SecurityLevel.noAuthNoPriv
});
var acm = authorizer.getAccessControlModel ();
// Since read-only is the default, explicitly setting read-only access is not required - just shown here as an example
acm.setCommunityAccess ("public", snmp.AccessLevel.ReadOnly);
acm.setCommunityAccess ("private", snmp.AccessLevel.ReadWrite);
acm.setUserAccess ("fred", snmp.AccessLevel.ReadWrite);
Grant the given community the given access level.
Remove all access for the given community.
Return the access level for the given community.
Return a list of all community access control entries defined by this access control model.
Grant the given user the given access level.
Remove all access for the given user.
Return the access level for the given user.
Return a list of all user access control entries defined by this access control model.
An Agent
instance, when created, in turn creates an instance of the Mib
class.
An agent always has one and only one Mib
instance. The agent's Mib
instance is
accessed through the agent.getMib ()
call.
The MIB is a tree structure that holds management information. Information is "addressed" in the tree by a series of integers, which form an Object ID (OID) from the root of the tree down.
There are only two kinds of data structures that hold data in a MIB:
-
scalar data - the scalar variable is stored at a node in the MIB tree, and the value of the variable is a single child node of the scalar variable node, always with address "0". For example, the sysDescr scalar variable is located at "1.3.6.1.2.1.1.1". The value of the sysDescr variable is stored at "1.3.6.1.2.1.1.1.0"
1.3.6.1.2.1.1.1 <= sysDescr (scalar variable) 1.3.6.1.2.1.1.1.0 = OctetString: MyAwesomeHost <= sysDescr.0 (scalar variable value)
-
table data - the SNMP table stores data in columns and rows. Typically, if a table is stored at a node in the MIB, it has an "entry" object addressed as "1" directly below the table OID. Directly below the "entry" is a list of columns, which are typically numbered from "1" upwards. Directly below each column are a series of rows. In the simplest case a row is "indexed" by a single column in the table, but a row index can be a series of columns, or columns that give multiple integers (e.g. an IPv4 address has four integers to its index), or both. Here is an example of the hierarchy of an SNMP table for part of the ifTable:
1.3.6.1.2.1.2.2 <= ifTable (table) 1.3.6.1.2.1.2.2.1 <= ifEntry (table entry) 1.3.6.1.2.1.2.2.1.1 <= ifIndex (column 1) 1.3.6.1.2.1.2.2.1.1.1 = Integer: 1 <= ifIndex row 1 value = 1 1.3.6.1.2.1.2.2.1.1.2 = Integer: 2 <= ifIndex row 2 value = 2
On creation, an Agent
instance creates a singleton instance of the Mib
module. You can
then register a "provider" to the agent's Mib
instance that gives an interface to either a scalar
data instance, or a table.
var myScalarProvider = {
name: "sysDescr",
type: snmp.MibProviderType.Scalar,
oid: "1.3.6.1.2.1.1.1",
scalarType: snmp.ObjectType.OctetString,
maxAccess: snmp.MaxAccess["read-write"],
handler: function (mibRequest) {
// e.g. can update the MIB data before responding to the request here
mibRequest.done ();
}
};
var mib = agent.getMib ();
mib.registerProvider (myScalarProvider);
mib.setScalarValue ("sysDescr", "MyAwesomeHost");
This code first gives the definition of a scalar "provider". A further explanation of
these fields is given in the mib.registerProvider()
section. Importantly, the name
field is the unique identifier of the provider, and is used to select the specific
provider in subsequent API calls.
The registerProvider()
call adds the provider to the list of providers that the MIB holds.
Note that this call does not add the "oid" node to the MIB tree. The first call of
setScalarValue()
will add the instance OID "1.3.6.1.2.1.1.1.0" to the MIB tree,
along with its value.
At this point, the agent will serve up the value of this MIB node when the instance OID "1.3.6.1.2.1.1.1.0" is queried via SNMP.
A table provider has a similar definition:
var myTableProvider = {
name: "smallIfTable",
type: snmp.MibProviderType.Table,
oid: "1.3.6.1.2.1.2.2.1",
maxAccess: snmp.MaxAccess["not-accessible"],
tableColumns: [
{
number: 1,
name: "ifIndex",
type: snmp.ObjectType.Integer,
maxAccess: snmp.MaxAccess["read-only"]
},
{
number: 2,
name: "ifDescr",
type: snmp.ObjectType.OctetString,
maxAccess: snmp.MaxAccess["read-write"],
},
{
number: 3,
name: "ifType",
type: snmp.ObjectType.Integer,
maxAccess: snmp.MaxAccess["read-only"],
constraints: {
enumeration: {
"1": "goodif",
"2": "averageif",
"3": "badif"
}
}
}
],
tableIndex: [
{
columnName: "ifIndex"
}
]
};
var mib = agent.getMib ();
mib.registerProvider (myTableProvider);
mib.addTableRow ("smallIfTable", [1, "eth0", 6]);
Here, the provider definition takes two additions fields: tableColumns
for the column defintions,
and tableIndex
for the columns used for row indexes. In the example the tableIndex
is the
ifIndex
column. The mib.registerProvider()
section has further details on the fields that
make up the provider definition.
The oid
must be that of the "table entry" node, not its parent "table" node e.g. for
ifTable
, the oid
in the provider is "1.3.6.1.2.1.2.2.1" (the OID for ifEntry
).
Note that there is no handler
callback function in this particular example, so any interaction
is directly between SNMP requests and MIB values with no other intervention.
Three types of constraints are supported: enumerations, integer
ranges, and string sizes. These can be specified in a handler's
constraints
map, with keys enumeration
, ranges
, or sizes
.
Any SetRequest protocol operations are checked against the defined constraints, and are not actioned if the value in the SetRequest would violate the constraints e.g. the value is not a member of the defined enumeration.
The MIB parser converts definitions such as this to enumeration
constraints (see RFC 2578 Section 7.1.1):
SYNTAX INTEGER { cont(0), alt(1) }
It converts definitions such as these to ranges
constraints (see RFC 2578 Appendix A):
SYNTAX Integer32 (172..184)
And it converts definitions like these to sizes
constraints (see RFC 2578 Appendix A):
SYNTAX OCTET STRING (SIZE (0..31))
Enumerations identify each of the valid values of an object of type Integer, like this:
constraints: {
enumeration: {
"1": "goodif",
"2": "averageif",
"3": "badif"
}
Ranges are used in Integer types, to limit the object's allowable
values. They are specified using an array of maps. Each map optionally
contains min
and/or max
values, specifying a single range.
Mutliple ranges allow the union of values specified by those ranges.
Specifying only min
in a range allows all values greater than or
equal to the specified one to be valid. Specifying only max
in a
range allows all values less than or equal to the specified one to be
valid. This example shows that any value between 1 and 3 (inclusive)
or 5 or greater is allowed, i.e., all integers greater than or equal
to 1, except 4:
constraints: {
ranges: [
{ min: 1, max: 3 },
{ min: 5 }
]
},
Sizes are used to limit the lengths of strings. The syntax is similar
to ranges
, allowing multiple ranges of sizes each potentially
providing min
andmax
values. A constraint that allows a string to
be any length 1 or greater, except length 4, would look like this:
constraints: {
sizes: [
{ min: 1, max: 3 },
{ min: 5 }
]
}
The createMib()
function instantiates and returns an instance of the
Mib
class. The new Mib does not have any nodes (except for a single
root node) and does not have any registered providers.
Note that this is only usable for an agent, not an AgentX subagent. Since an agent instanciates
a Mib
instance on creation, this call is not needed in many scenarios. Two scenarios where it
might be useful are:
- where you want to pre-populate a
Mib
instance with providers and scalar/tabular data before creating theAgent
instance itself. - where you want to swap out an agent's existing
Mib
instance for an entirely new one.
Registers a provider definition with the MIB. Doesn't add anything to the MIB tree.
A provider definition has these fields:
name
(mandatory) - the name of the provider, which serves as a unique key to reference the provider for getting and setting valuestype
(mandatory) - must be eithersnmp.MibProviderType.Scalar
orsnmp.MibProviderType.Table
(mandatory)oid
(mandatory) - the OID where the provider is registered in the MIB tree. Note that this is not the "instance node" (the ".0" node), but the one above it. In this case, the provider registers at "1.3.6.1.2.1.1.1", to provide the value at "1.3.6.1.2.1.1.1.0".scalarType
(mandatory for scalar types) - only relevant to scalar provider type, this give the type of the variable, selected fromsnmp.ObjectType
tableColumns
(mandatory for table types) - gives any array of column definition objects for the table. Each column object must have a uniquenumber
, aname
, atype
fromsnmp.ObjectType
, and amaxAccess
value fromsnmp.MaxAccess
. A column object with typeObjectType.Integer
can optionally contain aconstraints
object, the format and meaning of which is identical to that defined on a single scalar provider (see the "Constraints" section above for further details on this).tableIndex
(optional for table types) - gives an array of index entry objects used for row indexes. Use a single-element array for a single-column index, and multiple values for a composite index. An index entry object has acolumnName
field, and if the entry is in another provider's table, then include aforeign
field with the name of the foreign table's provider. If thetableAugments
field is absent,tableIndex
is mandatory.tableAugments
(optional for table types) - gives the name of another registered provider that this table "augments". This means that the index information is taken from the given provider's table, and doesn't exist in the local table's column definitions. If thetableIndex
field is absent,tableAugments
is mandatory i.e. one oftableIndex
andtableAugments
needs to be present to define the table index.maxAccess
(mandatory) - specifies the maximum allowed access level provided by this provider. The allowable values are the numeric values from the MaxAccess export. If amaxAccess
value is specified, aget
request to the agent will return anoAccess
error ifmaxAccess
is not at least "read-only" (2).maxAccess
must be at least "read-write" (3) for aset
request to succeed.defVal
(optional) - the default value to assign for scalar objects automatically created, whenmaxAccess
is set to "read-create" (4). Note that table columns can specify suchdefVal
default values in an identical way, to be used when a new row is to be automatically created, except that these are stored under the column object definition for each column. SeeAutomatic creation of objects
, below, for details.handler
(optional) - an optional callback function, which is called before the request to the MIB is made. This could update the MIB value(s) handled by this provider. If not given, the values are simply returned from (or set in) the MIB without any other processing. The callback function takes aMibRequest
instance, which has adone()
function. This must be called when finished processing the request. To signal an error, give a single error object in the form of{errorStatus: <status>}
, where<status>
is a value from ErrorStatus e.g.{errorStatus: snmp.ErrorStatus.GeneralError}
. TheMibRequest
also has anoid
field with the instance OID being operated on, and anoperation
field with the request type fromsnmp.PduType
. If theMibRequest
is for aSetRequest
PDU, then variablessetValue
andsetType
contain the value and type received in theSetRequest
varbind.constraints
(optional for scalar types) - an optional object to specify constraints for integer-based enumerated types, integer range restrictions and string size restrictions. Note that table columns can specify suchconstraints
in an identical way, except that these are stored under the column object definition for each column. See the "Constraints" section above for further details.
After registering the provider with the MIB, the provider is referenced by its name
in other API calls.
While this call registers the provider to the MIB, it does not alter the MIB tree.
Convenience method to register an array of providers in one call. Simply calls registerProvider()
for each provider definition in the array.
Unregisters a provider from the MIB. This also deletes all MIB nodes from the provider's oid
down
the tree. It will also do upstream MIB tree pruning of any interior MIB nodes that only existed for
the MIB tree to reach the provider oid
node.
Returns an object of provider definitions registered with the MIB, indexed by provider name.
Returns a single registered provider object for the given name.
Retrieves the value from a scalar provider.
Sets the value for a scalar provider. If this is the first time the scalar is set since the provider has registered with the MIB, it will also add the instance (".0") node and all required ancestors to the MIB tree.
Adds a table row - in the form of an array of values - to a table provider. If
the table is empty, this instantiates the provider's oid
node and ancestors,
its columns, before adding the row of values. Note that the row is an array of
elements in the order of the table columns. If the table has any foreign index
columns (i.e. those not belonging to this table), then values for these must be
included the at the start of the row array, in the order they appear in the
MIB INDEX clause.
Returns a list of column definition objects for the provider.
Returns a two-dimensional array of the table data. If byRow
is false (the default),
then the table data is given in a list of column arrays i.e. by column. If byRow
is true
, then the data is instead a list of row arrays. If includeInstances
is
true
, then, for the column view there will be an extra first column with instance
index information. If includeInstances
is true
for the row view, then there is
an addition element at the start of each row with index information.
Returns a single column of table data for the given column number. If includeInstances
is true
, then two arrays are returned: the first with instance index information,
and the second with the column data.
Returns a single row of table data for the given row index. The row index is an array of index values built from the node immediately under the column down to the node at the end of the row instance, which will be a leaf node in the MIB tree. Ultimately, non-integer values need to be converted to a sequence of integers that form the instance part of the OID. Here are the details of the conversions from index values to row instance OID sequences:
- ObjectType.Integer - single integer
- ObjectType.OctetString - a sequence of integer ASCII values
- ObjectType.OID - the exact sequence of integers in the OID
- ObjectType.IpAddress - a sequence of the four integers in the IP address
Returns a single cell value from the column and row specified. The row index array is specified
in the same way as for the getTableRowCells()
call.
Sets a single cell value at the column and row specified. The row index array is specified
in the same way as for the getTableRowCells()
call.
Deletes a table row at the row index specified. The row index array is specified
in the same way as for the getTableRowCells()
call. If this was the last row in the table,
the table is pruned from the MIB, although the provider still remains registered with the MIB.
Meaning that on the addition of another row, the table will be instantiated again.
Adds a default value, called defVal
, to a scalar provider. This
default value will be used for automatic creation of the scalar's
object instance, when its maxAccess
value is "read-create". This
method is of primary usefulness when providers are automatically
created, e.g., via store.getProvidersForModule
. See Automatic creation of objects
for details.
Add default values, called defVal
, to each table column in a table
provider. These default values will be used for automatic creation of
a table row. defaultValues
must be an array of values of length
equal to the length of the tableColumns array in the provider. When a
specific column need not be given a default value, that element of the
array should be set to undefined
. This method is of primary
usefulness when providers are automatically created, e.g., via
store.getProvidersForModule
. See Automatic creation of objects
for
details.
Dumps the MIB in text format. The options
object controls the display of the dump with these
options fields (all are booleans that default to true
):
leavesOnly
- don't show interior nodes separately - only as prefix parts of leaf nodes (instance nodes)showProviders
- show nodes where providers are attached to the MIBshowTypes
- show instance value typesshowValues
- show instance values
For example:
mib.dump ();
produces this sort of output:
1.3.6.1.2.1.1.1 [Scalar: sysDescr]
1.3.6.1.2.1.1.1.0 = OctetString: Rage inside the machine!
1.3.6.1.2.1.2.2.1 [Table: ifTable]
1.3.6.1.2.1.2.2.1.1.1 = Integer: 1
1.3.6.1.2.1.2.2.1.1.2 = Integer: 2
1.3.6.1.2.1.2.2.1.2.1 = OctetString: lo
1.3.6.1.2.1.2.2.1.2.2 = OctetString: eth0
1.3.6.1.2.1.2.2.1.3.1 = Integer: 24
1.3.6.1.2.1.2.2.1.3.2 = Integer: 6
When a provider's maxAccess
is set to "read-create" (4), then an
agent request to access the object's instance will result in the
instance being automatically created, if defVal
is also defined in
the provider. The new instance's value will be set to the default
value specified in defVal
. If defVal
is not specified in the
provider, then the instance will not, by default, be automatically
created.
The default handling of instance creation can be overridden by
providing a handler in a provider, called, createHandler
. The
handler is passed a createRequest
object, containing a singe
field provider
- the provider for the scalar. The
method must return either the value to be assigned to the
newly-created instance; or undefined
to indicate that the instance
should not be created.
An example handler method, accomplishing the default behavior, looks like this:
function scalarReadCreateHandler (createRequest) {
let provider = createRequest.provider;
// If there's a default value specified...
if ( typeof provider.defVal != "undefined" ) {
// ... then use it
return provider.defVal;
}
// We don't have enough information to auto-create the scalar
return undefined;
}
Automatic instance creation of table rows can be disabled entirely by
setting createHandler
to null.
Table rows may be added to a table, or deleted from it, if the table
has a column defined with rowStatus: true
in the provider.
The semantics of adding and deleting rows is described beginning on
page 5 of RFC 2579, and in
SNMPv2-TC.mib.
The row status column is typically referred to, simply, as the Status
column.
When a row does not exist and its Status column's value is set to
"createAndGo" (4) or "createAndWait" (5), the specified row will be
created, by default, using the default values specified in each
non-index and non-Status column's defVal
member. If defVal
is not
specified in any column other than index or Status columns, the row
will not be automatically created.
The default handling of row creation can be overridden by providing a
handler in a provider, called, createHandler
. The handler is
passed a createRequest
object with three fields:
provider
- the provider for the tableaction
- the action invoking the row creation: one of "createAndGo" or "createAndWait"row
- an array of columns forming the table index, where each element of the array is an index into thetableColumns
array of the provider
The handler must return either an array of column values for the new
row, with exactly one value corresponding to each column specified in
tableColumns
; or undefined
to indicate that the row should not be
created.
An example handler method, accomplishing the default behavior, looks like this:
function tableRowStatusHandler(createRequest) {
let provider = createRequest.provider;
let action = createRequest.action;
let row = createRequest.row;
let values = [];
let missingDefVal = false;
let rowIndexValues = Array.isArray( row ) ? row.slice(0) : [ row ];
const tc = provider.tableColumns;
tc.forEach(
(columnInfo, index) => {
let entries;
// Index columns get successive values from the rowIndexValues array.
// RowStatus columns get either "active" or "notInService" values.
// Every other column requires a defVal.
entries = provider.tableIndex.filter( entry => columnInfo.number === entry.columnNumber );
if (entries.length > 0 ) {
// It's an index column. Use the next index value
values.push(rowIndexValues.shift());
} else if ( columnInfo.rowStatus ) {
// It's the RowStatus column. Replace the action with the appropriate state
values.push( RowStatus[action] );
} else if ( "defVal" in columnInfo] ) {
// Neither index nor RowStatus column, so use the default value
values.push( columnInfo.defVal );
} else {
// Default value was required but not found
console.log("No defVal defined for column:", columnInfo);
missingDefVal = true;
values.push( undefined ); // just for debugging; never gets returned
}
}
);
// If a default value was missing, we can't auto-create the table row.
// Otherwise, we're good to go: give 'em the column values.
return missingDefVal ? undefined : values;
}
Automatic instance creation of table rows can be disabled entirely by
setting createHandler
to null.
When a MIB is read from a file using ModuleStore
's loadFromFile
method, and the providers for that module automatically created via a
call to the store's getProvidersForModule
method, default values
specified as DEFVAL
in the MIB are mapped to defVal
within the
provider, both from scalar definitions and from table columns
definitions.
If the MIB files do not contain some or all of the default values
needed for automatic creation of scalar objects or table rows, the
methods Mib.setScalarDefaultValue
and Mib.setTableRowDefaultValues
may be used to conveniently add defaults after the MIB files are
loaded.
The library supports MIB parsing by providing an interface to a ModuleStore
instance into which
you can load MIB modules from files, and fetch the resulting JSON MIB module representations.
Additionally, once a MIB is loaded into the module store, you can produce a list of MIB "provider"
definitions that an Agent
can register (see the Agent
documentation for more details), so
that you can start manipulating all the values defined in your MIB file right away.
// Create a module store, load a MIB module, and fetch its JSON representation
var store = snmp.createModuleStore ();
store.loadFromFile ("/path/to/your/mibs/SNMPv2-MIB.mib");
var jsonModule = store.getModule ("SNMPv2-MIB");
// Fetch MIB providers, create an agent, and register the providers with your agent
var providers = store.getProvidersForModule ("SNMPv2-MIB");
// Not recommended - but authorization and callback turned off for example brevity
var agent = snmp.createAgent ({disableAuthorization: true}, function (error, data) {});
var mib = agent.getMib ();
mib.registerProviders (providers);
// Start manipulating the MIB through the registered providers using the `Mib` API calls
mib.setScalarValue ("sysDescr", "The most powerful system you can think of");
mib.setScalarValue ("sysName", "multiplied-by-six");
mib.addTableRow ("sysOREntry", [1, "1.3.6.1.4.1.47491.42.43.44.45", "I've dreamed up this MIB", 20]);
Then hit those bad boys with your favourite SNMP tools (or library ;-), e.g.
snmpwalk -v 2c -c public localhost 1.3.6.1
Meaning you can get right to the implementation of your MIB functionality with a minimum of boilerplate code.
Creates a new ModuleStore
instance, which comes pre-loaded with some "base" MIB modules that
that provide MIB definitions that other MIB modules commonly refer to ("import"). The list of
pre-loaded "base" modules is:
- RFC1155-SMI
- RFC1158-MIB
- RFC-1212
- RFC1213-MIB
- SNMPv2-SMI
- SNMPv2-CONF
- SNMPv2-TC
- SNMPv2-MIB
Loads all MIB modules in the given file into the module store. By convention, there is
typically only a single MIB module per file, but there can be multiple module definitions
stored in a single file. Loaded MIB modules are then referred to by this API by their
MIB module name, not the source file name. The MIB module name is the name preceding the
DEFINITIONS ::= BEGIN
in the MIB file, and is often the very first thing present in
a MIB file.
Note that if your MIB depends on ("imports") definitions from other MIB files, these must be loaded first e.g. the popular IF-MIB uses definitions from the IANAifType-MIB, which therefore must be loaded first. These dependencies are listed in the IMPORTS section of a MIB module, usually near the top of a MIB file. The pre-loaded "base" MIB modules contain many of the commonly used imports.
Retrieves the named MIB module from the store as a JSON object.
Retrieves all MIB modules from the store. If the includeBase
boolean is set to true,
then the base MIB modules are included in the list. The modules are returned as a single
JSON "object of objects", keyed on the module name, with the values being entire JSON
module represenations.
Retrieves a list of the names of all MIB modules loaded in the store. If the includeBase
boolean is set to true, then the base MIB modules names are included in the list.
Returns an array of Mib
"provider" definitions corresponding to all scalar and table instance
objects contained in the named MIB module. The list of provider definitions are then
ready to be registered to an agent's MIB by using the agent.getMib().registerProviders()
call.
Takes an OID in one of the three supported formats (the library automatically detects the given OID's format):
- OidFormat.oid - canonical (numerical) OID format e.g. 1.3.6.1.2.1.1.1
- OidFormat.path - named OID path format e.g. iso.org.dod.internet.mgmt.mib-2.system.sysDescr
- OidFormat.module - module-qualified format e.g. SNMPv2-MIB::sysDescr
Returns the given OID translated to the provided destination format - also one of the above three formats.
For example:
var numericOid = store.translate ('SNMPv2-MIB::sysDescr', snmp.OidFormat.oid);
=> '1.3.6.1.2.1.1.1'
var moduleQualifiedName = store.translate ('1.3.6.1.2.1.1.1', snmp.OidFormat.module);
=> 'SNMPv2-MIB::sysDescr'
var namedOid = store.translate ('1.3.6.1.2.1.1.1', snmp.OidFormat.path);
=> 'iso.org.dod.internet.mgmt.mib-2.system.sysDescr'
An Agent
instance, when created, in turn creates an instance of the Forwarder
class.
There is no direct API call to create a Forwarder
instance; this creation is the
responsibility of the agent. An agent always has one and only one Forwarder
instance.
The agent's Forwarder
instance is accessed through the agent.getForwarder ()
call.
A Forwader
is what RFC 3413 terms a "Proxy Forwarder Application". It maintains a list
of "proxy" entries, each of which configures a named SNMPv3 context name to enable access
to a given target host with the given user credentials. The Forwarder
supports proxying
of SNMPv3 sessions only.
var forwarder = agent.getForwarder ();
forwarder.addProxy({
context: "slatescontext",
host: "bedrock",
user: {
name: "slate",
level: snmp.SecurityLevel.authNoPriv,
authProtocol: snmp.AuthProtocols.sha,
authKey: "quarryandgravel"
}
});
Now requests to the agent with the context "slatescontext" supplied will be forwarded to host "bedrock", with the supplied credentials for user "slate".
You can query the proxy with a local agent user (added with the agent's Authorizer
instance).
Assuming your proxy runs on localhost, port 161, you could add local user "fred", and access the proxy
with the new "fred" user.
var authorizer = agent.getAuthorizer();
authorizer.addUser ({
name: "fred",
level: snmp.SecurityLevel.noAuthNoPriv
});
// Test access using Net-SNMP tools (-n is the context option):
snmpget -v 3 -u fred -l noAuthNoPriv -n slatescontext localhost 1.3.6.1.2.1.1.1.0
This proxies requests through to "bedrock" as per the proxy definition.
Adds a new proxy to the forwarder. The proxy is an object with these fields.
context
(mandatory) - the name of the SNMPv3 context for this proxy entry. This is the unique key for proxy entries i.e. there cannot be two proxies with the same context name.transport
(optional) - specifies the transport to use to reach the remote target. Can be eitherudp4
orudp6
, defaults toudp4
.target
(mandatory) - the remote host that will receive proxied requests.port
(optional) - the port of the SNMP agent on the remote host. Defaults to 161.user
(mandatory) - the SNMPv3 user. The format for the user is described in thecreateV3Session()
call documentation.
Delete the proxy for the given context from the forwarder.
Returns the forwarder's proxy for the given context.
Returns an object containing a list of all registered proxies, keyed by context name.
Prints a dump of all proxy definitions to the console.
The AgentX subagent implements the functionality specified in RFC 2741 to become a "subagent" of an AgentX "master agent". The goal of AgentX is to extend the functionality of an existing "master" SNMP agent by a separate "subagent" registering parts of the MIB tree that it would like to manage for the master agent.
The AgentX subagent supports the generation of all but two of the "administrative" PDU types, all of which are sent from the subagent to the master agent:
- Open PDU - opens a new session with a master agent
- Close PDU - closes an existing session with the master agent
- Register PDU - registers a MIB region to control with the master agent
- Unregister PDU - unregisters a previously registered MIB region with the master agent
- Notify PDU - sends a notification to the master agent
- Ping PDU - sends a "ping" to confirm the master agent is still available
- AddAgentCaps PDU - adds an agent capability to the master agent's sysORTable
- RemoveAgentCaps PDU - remove a previously added agent capability from the master agent's sysORTable
The two unsupported "administrative" PDU types are:
- IndexAllocate PDU - request allocation of an index from a table whose index is managed by a master agent
- IndexDeallocate PDU - request deallocation of a previously allocated index from a master agent's table
These are unsupported as they do not fit the current MIB provider registration model, which only supports registering scalars and entire tables. These could be supported in the future by further generalizing the registration model to support table row registration.
The subagent responds to all "request processing" PDU types relevant to a Command Responder application, which are received from the master agent:
- Get PDU - requests exactly matched OID instances
- GetNext PDU - requests lexicographically "next" OID instances in the MIB tree
- GetBulk PDU - requests a series of "next" OID instances in the MIB tree
- TestSet PDU - tests a list of "set" operations to be committed as a single transaction
- CommitSet PDU - commits a list of "set" operations as a single transaction
- UndoSet PDU - undoes a list of "set" operations as a single transaction
- CleanupSet PDU - ends a "set" transaction
As per RFC 2741, all of these except the CleanupSet PDU return a Response PDU to the master agent.
Like the SNMP agent, the AgentX subagent maintains is a Mib
instance, the API of which is
detailed in the Mib Module section above. The subagent allows the MIB to be queried
and manipulated through the API, as well as queried and manipulated through the AgentX interface with
the above "request processing" PDUs (which are produced by the master agent when its SNMP interface
is invoked).
It is important that MIB providers are registered using the subagent's subagent.registerProvider ()
call (outlined below), and not using subagent.getMib ().registerProvider ()
, as the subagent needs
to both register the provider on its internal Mib
object, and send a Register PDU to the master
agent for the provider's MIB region. The latter step is skipped if registering the provider directly
on the MIB object.
The createSubagent ()
function instantiates and returns an instance of the Subagent
class:
// Default options
var options = {
master: "localhost",
masterPort: 705,
timeout: 0,
description: "Node net-snmp AgentX sub-agent"
};
subagent = snmp.createSubagent (options);
The options
parameter is a mandatory object, possibly empty, and can contain the following fields:
master
- the host name or IP address of the master agent, which the subagent connects to.masterPort
- the TCP port for the subagent to connect to the master agent on - defaults to 705.timeout
- set the session-wide timeout on the master agent - defaults to 0, which means no session-wide timeout is set.description
- a textual description of the subagent.
Returns the agent's singleton Mib
instance, which is automatically created on creation
of the subagent, and which holds all of the management data for the subagent.
Sends an Open
PDU to the master agent to open a new session, invoking the callback on
response from the master.
Sends a Close
PDU to the master agent to close the subagent's session to the master,
invoking the callback on response from the master.
See the Mib
class registerProvider()
call for the definition of a provider. The format
and meaning of the provider
object is the same for this call. This sends a Register
PDU to
the master to register a region of the MIB for which the master will send "request processing"
PDUs to the subagent. The supplied callback
is used only once, on reception of the
subsequent Response
PDU from the master to the Register
PDU. This is not to be confused
with the handler
optional callback on the provider definition, which is invoked for any
"request processing" PDU received by the subagent for MIB objects in the registered MIB region.
Unregisters a previously registered MIB region by the supplied name of the provider. Sends
an Unregister
PDU to the master agent to do this. The supplied callback
is used only
once, on reception of the subsequent Response
PDU from the master to the Unregister
PDU.
Convenience method to register an array of providers in one call. Simply calls registerProvider()
for each provider definition in the array. The callback
function is called once for each
provider registered.
Returns an object of provider definitions registered with the MIB, indexed by provider name.
Returns a single registered provider object for the given name.
Adds an agent capability - consisting of oid
and descr
- to the master agent's sysORTable.
Sends an AddAgentCaps
PDU to the master to do this. The supplied callback
is called on
reception of the subsequent Response
PDU from the master to the AddAgentCaps
PDU.
Remove an previously added capability from the master agent's sysORTable. Sends a RemoveAgentCaps
PDU to the master to do this. The supplied callback
is called on reception of the subsequent
Response
PDU from the master to the RemoveAgentCaps
PDU.
Sends a notification to the master agent using a Notify
PDU. The notification takes the same
form as outlined in the session.inform()
section above and also in RFC 2741 Section 6.2.10,
which is creating two varbinds that are always included in the notification:
- sysUptime.0 (1.3.6.1.2.1.1.3.0) - containing the subagent's uptime
- snmpTrapOID.0 (1.3.6.1.6.3.1.1.4.1.0) - containing the supplied OID (or supplied
snmp.TrapType
value)
The optional varbinds
list is an additional list of varbind objects to append to the above
two varbinds. The supplied callback
is called on reception of the subsequent
Response
PDU from the master to the Notify
PDU.
Sends a "ping" to the master agent using a Ping
PDU, to confirm that the master agent is still
responsive. The supplied callback
is called on reception of the subsequent
Response
PDU from the master to the Ping
PDU.
The close
event is emitted by the subagent when its underlying TCP socket is closed.
No arguments are passed to the callback.
The following example prints a message to the console when a subagent's underlying TCP socket is closed:
subagent.on ("close", function () {
console.log ("Subagent socket closed");
});
The error
event is emitted by the subagent when its underlying TCP socket emits an error.
The following argument will be passed to the callback
function:
error
- An instance of theError
class, the exposedmessage
attribute will contain a detailed error message.
The following example prints a message to the console when an error occurs with a subagent's underlying TCP socket, and the subagent is then closed:
subagent.on ("error", function (error) {
console.log (error.toString ());
subagent.close ();
});
Example programs are included under the module's example
directory.
- Initial release including only SNMP version 1 support
- Implement SNMP version 2c support
- Correct name used in example
require()
call to include this module
- Implement
subtree()
,table()
andwalk()
methods - Support IPv6 (added
transport
option to thecreateSession()
function) - Re-order some methods in README.md
- Fix some typos and grammar errors in README.md
- Example
snmp-table
program hadsnmp-subtree
in its usage message - Implement example
snmp-tail
program to constantly poll for an OIDs value - Add note to README.md about the ability to stop the
walk()
andsubtree()
methods by returningtrue
- Fix incorrect usage of the term "NPM" in README.md, should be "npm"
- The
transport
option tocreateSession()
was not used
- Implement
tableColumns()
method - Added example program
snmp-table-columns.js
- Correct name of the
table
parameter to thetable()
callback - Slight OID comparison performance enhancement
- Use MIT license instead of GPL
- Added the example program
cisco-device-inventory.js
- Receive
Trap failed: TypeError: value is out of bounds
when sending traps using SNMP version 2c
- Corrected a few instances of the parameter named
requestCallback
to some methods in the README.md file which should have beenfeedCallback
- Null type is used for varbinds with a 0 value
- Correct instances of snmp.Type to snmp.ObjectType in the README.md file
- Error handler in the
dgram.send()
callback in thesend()
method was creating a new instance of theError
class from theerror
parameter, but it was already an instance of theError
class (thanks Ray Solomon) - Add stack traces to Error classes exported by this module (thanks Ray Solomon)
- Allow users to specify
0
retries when creating a session (thanks Ray Solomon) - Update the list of SNMP version 1 related RFCs we adhere to in the
Standards Compliance
section of the README.md file
- Add
sourceAddress
andsourcePort
optional options to theSession
classescreateSession()
method, which can be used to control from which IP address and port messages should be sent - Allow users to specify sysUpTime for SNMP traps and informs
- The
agentAddr
attribute is not used when passed in theoptions
object to thetrap()
method
- Not catching error events for the UDP socket returned from the
dgram.createSocket()
function - Some request methods do not copy arguments which results in sometimes unexpected behaviour
- Use a single UDP socket for all requests in a single SNMP session
- Use a try/catch block in the timer callback in the
Session.send()
method - The
Session
can now emit anerror
event to catch errors in a sessions underlying UDP socket - The
Session
can now emit aclose
event to catch close events from a sessions underlying UDP socket, which results in the cancellation of all outstanding requests - Added a
close()
method toSession
to close a sessions underlying UDP socket, which results aclose
event - Signed integers are treated as unsigned integers when parsing response messages
- Host repository on GitHub
- When parsing an invalid response an exception in message parsing does not interupt response processing
- Incorrectly passing
req
object in call toreq.responseCb
when handling errors during response processing
- Address a number of issues detected with the Mocha test suite by a user
- Correct reference to non-existant
req
variable in theSession
objects constructor (should bethis
)
- Correct argument number and names to the
snmp.createSession()
function - Add missing braces to an example in the README.md file
- Remove 64bit integer check to ensure a maximum of 8 bytes are given in send and received messages
- Replace asn1 dependancy with asn1-ber
- Add support of 16bit ids to help interoperate with older devices (added the
idBitsSize
option to thecreateSession()
function - Add note to README.md that sessions should be closed when done with
- Set NoSpaceships Ltd to be the owner and maintainer
- Remove redundant sections from README.md
- Add SNMPv3 support
- Add trap and inform receiver
- Add CONTRIBUTING.md guidelines
- Add SNMPv3 context to Session class
- Add IPv6 option for tests
- Add SNMP agent
- Add MIB parser and module store
- Add proxy forwarder to agent
- Add AES-128 encryption
- Add non-integer, composite key, foreign key and augmented table index handling
- Update CONTRIBUTING.md and parser example
- Add backoff option
- Fix agent crash with unexpected GetNext start OID
- Fix double report PDU time synchronisation handling
- Fix agent handling of GetNext from off-tree OID
- Handle periodic report PDUs on a long running session
- Fix OID and namespace calculations in MIB parser
- Fix Windows absolute path for reading MIB files
- Improve SNMPv3 error messages
- Receiver close fix and receiver example fix
- Add backwardsGetNexts option for handling of errant GetNexts
- Add AgentX subagent
- Fix backwardsGetNexts session option and fix null MIB entry reading
- Add missing agent.close() API call
- Add set value to MibRequest and fix backwardsGetNexts
- Improve socket error handling
- Add agent support for handling short OIDs and noSuchInstance
- Fix async mibRequest handler
- Add support for zero-index rows in agent tables
- Fix GetBulk async mibRequest handling
- Add MIB create, add MIB setting for agent, and fix MIB error response crash
- Fix AgentX subagent noSuchInstance crash
- Declare variables to fix transpile errors
- MIB getobject callback convention update
- Fix columnNumber check in getColumnProvider
- Fix parsing of iso.org
- Add revisions/descriptions MIB parsing
- Fix double callback invocation on callback error
- Add eslint rules and conformance, fix AgentX subagent Unregister
- Add Travis CI configuration
- Add simple access control model for agent
- Add MIB integer enumeration constraints for providers and SetRequests
- Fix MIB parsing of files leading with a comment
- Add bind address support for agent
- Fix getBulk documentation errors
- Add syntax definitions to README code blocks
- Fix providers for MIB table at end of MIB file
- Add README.cn.md for Chinese language
- Add support for BER long-form length encoding
- Add message security level checks against user security level
- Fix UNITS key recognition in MIB parser
- Add MAX-ACCESS provider and agent support
- Fix error indexing for failed varbinds from agent
- Fix agent hang on GetBulk with non-repeaters greater than varbind length
- Add agent errorStatus signalling support
- Prevent non-accessible index objects from being columns in table rows
- Fix MIB file reading from relative paths
- Fix MIB parsing of tab characters
- Fix MIB parsing of quoted unmatched brackets
- Add RowStatus support to agent tables
- Fix scalar default createHandler
- Add row index, row and column to agent table callback info
- Fix agent scalar read-create and set
- Fix agent state for row deletion to empty table
- Fix return from agent set outside of constraints to WrongValue
- Add range and size constraints support to MIB variables
- Add range and size constraints to MIB parsing and provider generation
- Add range and size constraints documentation
- Add column position to agent varbind callback
- Add 256-bit AES encryption
- Add oldValue to agent callback info and convert buffers to strings
- Add error codes to ResponseInvalidError
- Add documentation for ResponseInvalidError error codes
- Add engineID option to v3 session
- Fix MIB parsing of sized integers without whitespace
- Fix MIB table index handling of Buffer type
- Fix error with empty varbind array in walk
- Align accessible-for-notify row cell behaviour with not-accessible behaviour
- Add missing return in getbulk feedCb callback non-repeaters error condition
- Fix GetNext OID calculation for off-tree OIDs
- Fix handing of null varbinds in walk
- Fix processing of negative integers larger than 32 bits
- Add calculated key cache to remove authNoPriv and authPriv performance bottleneck
- Add v3 context to non-initial PDUs
- Add option for receiver to receive client authentication identity
- Fix logic for v3 time synchronization requirement
- Ignore mismatched returned OIDs by default
- Add SHA-2 authentication support (SHA-224, SHA-256, SHA-384, SHA-512)
- Fix DES decrypt corruption issue
- Improve getBulk response handling
- Fix 32-bit unsigned integer writing and add integer range checking
- Add bit string type to varbind reading
- Fix PDU error status field writing
- Fix incorrect user level assignment
- Fix IpAddress byte array set request handling
- Add ProcessingError to handle agent/receiver decode failures
- Fix MIB parsing of unclosed brackets in comments
- Fix MIB parsing of non-comments in descriptions and unmatched quotations
- Add AgentX subagent "error" and "close" events
- Fix syntax constraints tokenization and fix applying of size constraints
- Normalize whitespace parsing for OBJECT IDENTIFIER value
- Add type constraint support for textual conventions
- Add tolerance for reading malformed 32-bit unsigned integers
- Fix subtree callback termination
- Add SMIv1 integer enumeration support
- Add numeric/named OID translate function
- Fix table column type in provider definition with column type constraints
- Add MIB parsing support for OID values with intermediate entries in list
- Add MIB parsing support for final member of OID value list containing identifier/descriptor pair
- Add MIB parsing support for middle member of OID value list containing identifier only
- Fix parent object retrieval to respect MIB module imports
- Add error status and index handling to AgentX subagent
- Add provider to MIB request
- Add multiple socket listener support for agent and receiver
- Fix SNMPv1 session walk infinite loop condition
Copyright (c) 2020 Mark Abrahams mark@abrahams.co.nz
Copyright (c) 2018 NoSpaceships Ltd hello@nospaceships.com
Copyright (c) 2013 Stephen Vickers stephen.vickers.sv@gmail.com
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.