20042010 Ericsson AB. All Rights Reserved. The contents of this file are subject to the Erlang Public License, Version 1.1, (the "License"); you may not use this file except in compliance with the License. You should have received a copy of the Erlang Public License along with this software. If not, it can be retrieved online at http://www.erlang.org/. Software distributed under the License is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License for the specific language governing rights and limitations under the License. snmpm snmpm.xml
snmpm Interface functions to the SNMP toolkit manager

The module snmpm contains interface functions to the SNMP manager.

Common Data Types

The following data types are used in the functions below:

monitor() -> Ref Monitor the snmp manager Ref = reference()

Monitor the SNMP manager. In case of a crash, the calling (monitoring) process will get a 'DOWN' message (see the erlang module for more info).

demonitor(Ref) -> void() Turn off monitoring of the snmp manager Ref = reference()

Turn off monitoring of the SNMP manager.

notify_started(Timeout) -> Pid Request to be notified when manager started Timeout = integer() Pid = pid()

Request a notification (message) when the SNMP manager has started.

The Timeout is the time the request is valid. The value has to be greater then zero.

The Pid is the process handling the supervision of the SNMP manager start. When the manager has started a completion message will be sent to the client from this process: {snmpm_started, Pid}. If the SNMP manager was not started in time, a timeout message will be sent to the client: {snmpm_start_timeout, Pid}.

A client application that is dependent on the SNMP manager will use this function in order to be notified of when the manager has started. There are two situations when this is useful:

During the start of a system, when a client application could start prior to the SNMP manager but is dependent upon it, and therefor has to wait for it to start.

When the SNMP manager has crashed, the dependent client application has to wait for the SNMP manager to be restarted before it can reconnect.

The function returns the pid() of a handler process, that does the supervision on behalf of the client application. Note that the client application is linked to this handler.

This function is used in conjunction with the monitor function.

 cancel_notify_started(Pid) -> void() Cancel request to be notified when manager started Pid = pid()

Cancel a previous request to be notified of SNMP manager start.

 register_user(Id, Module, Data) -> ok | {error, Reason} register_user(Id, Module, Data, DefaultAgentConfig) -> ok | {error, Reason} Register a user of the manager Id = term() Module = snmpm_user() Data = term() DefaultAgentConfig = [default_agent_config()] default_agent_config() = {Item, Val} Item = community | timeout | max_message_size | version | sec_model | sec_name | sec_level Val = term() Reason = term() snmpm_user() = Module implementing the snmpm_user behaviour

Register the manager entity (=user) responsible for specific agent(s).

Module is the callback module (snmpm_user behaviour) which will be called whenever something happens (detected agent, incoming reply or incoming trap/notification). Note that this could have already been done as a consequence of the node config. (see users.conf).

The argument DefaultAgentConfig is used as default values when this user register agents.

The type of Val depends on Item:

register_user_monitor(Id, Module, Data) -> ok | {error, Reason} register_user_monitor(Id, Module, Data, DefaultAgentConfig) -> ok | {error, Reason} Register a monitored user of the manager Id = term() Module = snmpm_user() DefaultAgentConfig = [default_agent_config()] default_agent_config() = {Item, Val} Item = community | timeout | max_message_size | version | sec_model | sec_name | sec_level Val = term() Data = term() Reason = term() snmpm_user() = Module implementing the snmpm_user behaviour

Register the monitored manager entity (=user) responsible for specific agent(s).

The process performing the registration will be monitored. Which means that if that process should die, all agents registered by that user process will be unregistered. All outstanding requests will be canceled.

Module is the callback module (snmpm_user behaviour) which will be called whenever something happens (detected agent, incoming reply or incoming trap/notification). Note that this could have already been done as a consequence of the node config. (see users.conf).

The argument DefaultAgentConfig is used as default values when this user register agents.

The type of Val depends on Item:

unregister_user(Id) -> ok | {error, Reason} Unregister the user Id = term()

Unregister the user.

 which_users() -> Users Get a list of all users Users = [UserId] UserId = term()

Get a list of the identities of all registered users.

 register_agent(UserId, TargetName, Config) -> ok | {error, Reason} Register this agent UserId = term() Addr = ip_address() TargetName = target_name() Config = [agent_config()] agent_config() = {Item, Val} Item = engine_id | address | port | community | timeout | max_message_size | version | sec_model | sec_name | sec_level Val = term() Reason = term()

Explicitly instruct the manager to handle this agent, with UserId as the responsible user.

Called to instruct the manager that this agent shall be handled. This function is used when the user knows in advance which agents the manager shall handle. Note that there is an alternate way to do the same thing: Add the agent to the manager config files (see agents.conf).

TargetName is a non-empty string, uniquely identifying the agent.

The type of Val depends on Item:

Note that if no Port is given, the default value is used.

 unregister_agent(UserId, TargetName) -> ok | {error, Reason} Unregister the user UserId = term() TargetName = target_name()

Unregister the agent.

 agent_info(TargetName, Item) -> {ok, Val} | {error, Reason} Retrieve agent config TargetName = target_name() Item = atom() Reason = term()

Retrieve agent config.

 update_agent_info(UserId, TargetName, Item, Val) -> ok | {error, Reason} Update agent config UserId = term() TargetName = target_name() Item = atom() Val = term() Reason = term()

Update agent config.

 which_agents() -> Agents which_agents(UserId) -> Agents List the registered agents UserId = term() Agents = [TargetName] TargetName = target_name()

Get a list of all registered agents or all agents registered by a specific user.

 register_usm_user(EngineID, UserName, Conf) -> ok | {error, Reason} Register this USM user EngineID = string() UserName = string() Conf = [usm_config()] usm_config() = {Item, Val} Item = sec_name | auth | auth_key | priv | priv_key Val = term() Reason = term()

Explicitly instruct the manager to handle this USM user. Note that there is an alternate way to do the same thing: Add the usm user to the manager config files (see usm.conf).

The type of Val depends on Item:

unregister_usm_user(EngineID, UserName) -> ok | {error, Reason} Unregister this USM user EngineID = string() UserName = string() Reason = term()

Unregister this USM user.

 usm_user_info(EngineID, UserName, Item) -> {ok, Val} | {error, Reason} Retrieve usm user config EngineID = string() UsmName = string() Item = sec_name | auth | auth_key | priv | priv_key Reason = term()

Retrieve usm user config.

 update_usm_user_info(EngineID, UserName, Item, Val) -> ok | {error, Reason} Update agent config EngineID = string() UsmName = string() Item = sec_name | auth | auth_key | priv | priv_key Val = term() Reason = term()

Update usm user config.

 which_usm_users() -> UsmUsers List all the registered usm users UsmUsers = [{EngineID,UserName}] EngineID = string() UsmName = string()

Get a list of all registered usm users.

 which_usm_users(EngineID) -> UsmUsers List the registered usm users UsmUsers = [UserName] UserName = string()

Get a list of all registered usm users with engine-id EngineID.

 sync_get(UserId, TargetName, Oids) -> {ok, SnmpReply, Remaining} | {error, Reason} sync_get(UserId, TargetName, Oids, SendOpts) -> {ok, SnmpReply, Remaining} | {error, Reason} Synchronous get-request UserId = term() TargetName = target_name() Oids = [oid()] SendOpts = send_opts() send_opts() = [send_opt()] send_opt() = {context, string()} | {timeout, pos_integer()} | {extra, term()} | {community, community()} | {sec_model, sec_model()} | {sec_name, string()} | {sec_level, sec_level()} | {max_message_size, pos_integer()} SnmpReply = snmp_reply() Remaining = integer() Reason = {send_failed, ReqId, ActualReason} | {invalid_sec_info, SecInfo, SnmpInfo} | term() ReqId = term() ActualReason = term() SecInfo = [sec_info()] sec_info() = {sec_tag(), ExpectedValue, ReceivedValue} sec_tag() = atom() ExpectedValue = ReceivedValue = term() SnmpInfo = term()

Synchronous get-request.

Remaining is the remaining time of the given (or default) timeout time.

When Reason is {send_failed, ...} it means that the net_if process failed to send the message. This could happen because of any number of reasons, i.e. encoding error. ActualReason is the actual reason in this case.

The send option extra specifies an opaque data structure passed on to the net-if process. The net-if process included in this application makes no use of this info, so the only use for it in such a option (when using the built in net-if) would be tracing.

Some of the send options (community, sec_model, sec_name, sec_level and max_message_size) are override options. That is, for this request, they override any configuration done when the agent was registered.

For SnmpInfo, see the user callback function handle_report.

 async_get(UserId, TargetName, Oids) -> {ok, ReqId} | {error, Reason} async_get(UserId, TargetName, Oids, SendOpts) -> {ok, ReqId} | {error, Reason} Asynchronous get-request UserId = term() TargetName = target_name() Oids = [oid()] SendOpts = send_opts() send_opts() = [send_opt()] send_opt() = {context, string()} | {timeout, pos_integer()} | {extra, term()} | {community, community()} | {sec_model, sec_model()} | {sec_name, string()} | {sec_level, sec_level()} | {max_message_size, pos_integer()} ReqId = term() Reason = term()

Asynchronous get-request.

The reply, if it arrives, will be delivered to the user through a call to the snmpm_user callback function handle_pdu.

The send option timeout specifies for how long the request is valid (after which the manager is free to delete it).

The send option extra specifies an opaque data structure passed on to the net-if process. The net-if process included in this application makes no use of this info, so the only use for it in such a configuration (when using the built in net-if) would be tracing.

Some of the send options (community, sec_model, sec_name, sec_level and max_message_size) are override options. That is, for this request, they override any configuration done when the agent was registered.

sync_get_next(UserId, TargetName, Oids) -> {ok, SnmpReply, Remaining} | {error, Reason} sync_get_next(UserId, TargetName, Oids, SendOpts) -> {ok, SnmpReply, Remaining} | {error, Reason} Synchronous get-next-request UserId = term() TargetName = target_name() Oids = [oid()] SendOpts = send_opts() send_opts() = [send_opt()] send_opt() = {context, string()} | {timeout, pos_integer()} | {extra, term()} | {community, community()} | {sec_model, sec_model()} | {sec_name, string()} | {sec_level, sec_level()} | {max_message_size, pos_integer()} SnmpReply = snmp_reply() Remaining = integer() Reason = {send_failed, ReqId, ActualReason} | {invalid_sec_info, SecInfo, SnmpInfo} | term() ReqId = term() ActualReason = term() SecInfo = [sec_info()] sec_info() = {sec_tag(), ExpectedValue, ReceivedValue} sec_tag() = atom() ExpectedValue = ReceivedValue = term() SnmpInfo = term()

Synchronous get-next-request.

Remaining is the remaining time of the given (or default) timeout time.

When Reason is {send_failed, ...} it means that the net_if process failed to send the message. This could happen because of any number of reasons, i.e. encoding error. ActualReason is the actual reason in this case.

The send option extra specifies an opaque data structure passed on to the net-if process. The net-if process included in this application makes no use of this info, so the only use for it in such a configuration (when using the built in net-if) would be tracing.

Some of the send options (community, sec_model, sec_name, sec_level and max_message_size) are override options. That is, for this request, they override any configuration done when the agent was registered.

For SnmpInfo, see the user callback function handle_report.

 async_get_next(UserId, TargetName, Oids) -> {ok, ReqId} | {error, Reason} async_get_next(UserId, TargetName, Oids, SendOpts) -> {ok, ReqId} | {error, Reason} Asynchronous get-next-request UserId = term() TargetName = target_name() Oids = [oid()] send_opt() = {context, string()} | {timeout, pos_integer()} | {extra, term()} | {community, community()} | {sec_model, sec_model()} | {sec_name, string()} | {sec_level, sec_level()} | {max_message_size, pos_integer()} ReqId = integer() Reason = term()

Asynchronous get-next-request.

The reply will be delivered to the user through a call to the snmpm_user callback function handle_pdu.

The send option timeout specifies for how long the request is valid (after which the manager is free to delete it).

The send option extra specifies an opaque data structure passed on to the net-if process. The net-if process included in this application makes no use of this info, so the only use for it in such a configuration (when using the built in net-if) would be tracing.

Some of the send options (community, sec_model, sec_name, sec_level and max_message_size) are override options. That is, for this request, they override any configuration done when the agent was registered.

sync_set(UserId, TargetName, VarsAndVals) -> {ok, SnmpReply, Remaining} | {error, Reason} sync_set(UserId, TargetName, VarsAndVals, SendOpts) -> {ok, SnmpReply, Remaining} | {error, Reason} Synchronous set-request UserId = term() TargetName = target_name() VarsAndVals = vars_and_vals() SendOpts = send_opts() send_opts() = [send_opt()] send_opt() = {context, string()} | {timeout, pos_integer()} | {extra, term()} | {community, community()} | {sec_model, sec_model()} | {sec_name, string()} | {sec_level, sec_level()} | {max_message_size, pos_integer()} SnmpReply = snmp_reply() Remaining = integer() Reason = {send_failed, ReqId, ActualReason} | {invalid_sec_info, SecInfo, SnmpInfo} | term() ReqId = term() ActualReason = term() SecInfo = [sec_info()] sec_info() = {sec_tag(), ExpectedValue, ReceivedValue} sec_tag() = atom() ExpectedValue = ReceivedValue = term() SnmpInfo = term()

Synchronous set-request.

Remaining is the remaining time of the given (or default) timeout time.

When Reason is {send_failed, ...} it means that the net_if process failed to send the message. This could happen because of any number of reasons, i.e. encoding error. ActualReason is the actual reason in this case.

When var_and_val() is {oid(), value()}, the manager makes an educated guess based on the loaded mibs.

The send option extra specifies an opaque data structure passed on to the net-if process. The net-if process included in this application makes no use of this info, so the only use for it in such a configuration (when using the built in net-if) would be tracing.

Some of the send options (community, sec_model, sec_name, sec_level and max_message_size) are override options. That is, for this request, they override any configuration done when the agent was registered.

For SnmpInfo, see the user callback function handle_report.

 async_set(UserId, TargetName, VarsAndVals) -> {ok, ReqId} | {error, Reason} async_set(UserId, TargetName, VarsAndVals, SendOpts) -> {ok, ReqId} | {error, Reason} Asynchronous set-request UserId = term() TargetName = target_name() VarsAndVals = vars_and_vals() SendOpts = send_opts() send_opts() = [send_opt()] send_opt() = {context, string()} | {timeout, pos_integer()} | {extra, term()} | {community, community()} | {sec_model, sec_model()} | {sec_name, string()} | {sec_level, sec_level()} | {max_message_size, pos_integer()} ReqId = term() Reason = term()

Asynchronous set-request.

The reply will be delivered to the user through a call to the snmpm_user callback function handle_pdu.

The send option timeout specifies for how long the request is valid (after which the manager is free to delete it).

When var_and_val() is {oid(), value()}, the manager makes an educated guess based on the loaded mibs.

The send option extra specifies an opaque data structure passed on to the net-if process. The net-if process included in this application makes no use of this info, so the only use for it in such a configuration (when using the built in net-if) would be tracing.

Some of the send options (community, sec_model, sec_name, sec_level and max_message_size) are override options. That is, for this request, they override any configuration done when the agent was registered.

sync_get_bulk(UserId, TragetName, NonRep, MaxRep, Oids) -> {ok, SnmpReply, Remaining} | {error, Reason} sync_get_bulk(UserId, TragetName, NonRep, MaxRep, Oids, SendOpts) -> {ok, SnmpReply, Remaining} | {error, Reason} Synchronous get-bulk-request UserId = term() TargetName = target_name() NonRep = integer() MaxRep = integer() Oids = [oid()] SendOpts = send_opts() send_opts() = [send_opt()] send_opt() = {context, string()} | {timeout, pos_integer()} | {extra, term()} | {community, community()} | {sec_model, sec_model()} | {sec_name, string()} | {sec_level, sec_level()} | {max_message_size, pos_integer()} SnmpReply = snmp_reply() Remaining = integer() Reason = {send_failed, ReqId, ActualReason} | {invalid_sec_info, SecInfo, SnmpInfo} | term() ReqId = term() ActualReason = term() SecInfo = [sec_info()] sec_info() = {sec_tag(), ExpectedValue, ReceivedValue} sec_tag() = atom() ExpectedValue = ReceivedValue = term() SnmpInfo = term()

Synchronous get-bulk-request (See RFC1905).

Remaining is the remaining time of the given (or default) timeout time.

When Reason is {send_failed, ...} it means that the net_if process failed to send the message. This could happen because of any number of reasons, i.e. encoding error. ActualReason is the actual reason in this case.

The send option extra specifies an opaque data structure passed on to the net-if process. The net-if process included in this application makes no use of this info, so the only use for it in such a configuration (when using the built in net-if) would be tracing.

Some of the send options (community, sec_model, sec_name, sec_level and max_message_size) are override options. That is, for this request, they override any configuration done when the agent was registered.

For SnmpInfo, see the user callback function handle_report.

 async_get_bulk(UserId, TargetName, NonRep, MaxRep, Oids) -> {ok, ReqId} | {error, Reason} async_get_bulk(UserId, TargetName, NonRep, MaxRep, Oids, SendOpts) -> {ok, ReqId} | {error, Reason} Asynchronous get-bulk-request UserId = term() TargetName = target_name() NonRep = integer() MaxRep = integer() Oids = [oid()] SendOpts = send_opts() send_opts() = [send_opt()] send_opt() = {context, string()} | {timeout, pos_integer()} | {extra, term()} | {community, community()} | {sec_model, sec_model()} | {sec_name, string()} | {sec_level, sec_level()} | {max_message_size, pos_integer()} ReqId = integer() Reason = term()

Asynchronous get-bulk-request (See RFC1905).

The reply will be delivered to the user through a call to the snmpm_user callback function handle_pdu.

The send option timeout specifies for how long the request is valid (after which the manager is free to delete it).

The send option extra specifies an opaque data structure passed on to the net-if process. The net-if process included in this application makes no use of this info, so the only use for it in such a configuration (when using the built in net-if) would be tracing.

Some of the send options (community, sec_model, sec_name, sec_level and max_message_size) are override options. That is, for this request, they override any configuration done when the agent was registered.

cancel_async_request(UserId, ReqId) -> ok | {error, Reason} Cancel a asynchronous request UserId = term() ReqId = term() Reason = term()

Cancel a previous asynchronous request.

 log_to_txt(LogDir, Mibs) log_to_txt(LogDir, Mibs, OutFile) -> ok | {error, Reason} log_to_txt(LogDir, Mibs, OutFile, LogName) -> ok | {error, Reason} log_to_txt(LogDir, Mibs, OutFile, LogName, LogFile) -> ok | {error, Reason} log_to_txt(LogDir, Mibs, OutFile, LogName, LogFile, Start) -> ok | {error, Reason} log_to_txt(LogDir, Mibs, OutFile, LogName, LogFile, Start, Stop) -> ok | {error, Reason} Convert an Audit Trail Log to text format LogDir = string() Mibs = [MibName] MibName = string() OutFile = string() LogName = string() LogFile = string() Start = Stop = null | datetime() | {local_time,datetime()} | {universal_time,datetime()} Reason = disk_log_open_error() | file_open_error() | term() disk_log_open_error() = {LogName, term()} file_open_error() = {OutFile, term()}

Converts an Audit Trail Log to a readable text file. OutFile defaults to "./snmpm_log.txt". LogName defaults to "snmpm_log". LogFile defaults to "snmpm.log". See snmp:log_to_txt for more info.

 change_log_size(NewSize) -> ok | {error, Reason} Change the size of the Audit Trail Log NewSize = {MaxBytes, MaxFiles} MaxBytes = integer() MaxFiles = integer() Reason = term()

Changes the log size of the Audit Trail Log. The application must be configured to use the audit trail log function. Please refer to disk_log(3) in Kernel Reference Manual for a description of how to change the log size.

The change is permanent, as long as the log is not deleted. That means, the log size is remembered across reboots.

 set_log_type(NewType) -> {ok, OldType} | {error, Reason} Change the Audit Trail Log type NewType = OldType = atl_type() Reason = term()

Changes the run-time Audit Trail log type.

Note that this has no effect on the application configuration as defined by configuration files, so a node restart will revert the config to whatever is in those files.

This function is primarily useful in testing/debugging scenarios.

load_mib(Mib) -> ok | {error, Reason} Load a MIB into the manager Mib = MibName MibName = string() Reason = term()

Load a Mib into the manager. The MibName is the name of the Mib, including the path to where the compiled mib is found. For example,

Dir = code:priv_dir(my_app) ++ "/mibs/", snmpm:load_mib(Dir ++ "MY-MIB"). unload_mib(Mib) -> ok | {error, Reason} Unload a MIB from the manager Mib = MibName MibName = string() Reason = term()

Unload a Mib from the manager. The MibName is the name of the Mib, including the path to where the compiled mib is found. For example,

Dir = code:priv_dir(my_app) ++ "/mibs/", snmpm:unload_mib(Dir ++ "MY-MIB"). which_mibs() -> Mibs Which mibs are loaded into the manager Mibs = [{MibName, MibFile}] MibName = atom() MibFile = string()

Get a list of all the mib's loaded into the manager.

 name_to_oid(Name) -> {ok, Oids} | {error, Reason} Get all the possible oid's for an alias-name Name = atom() Oids = [oid()]

Transform a alias-name to its oid.

Note that an alias-name is only unique within the mib, so when loading several mib's into a manager, there might be several instances of the same aliasname.

 oid_to_name(Oid) -> {ok, Name} | {error, Reason} Get the alias-name of the oid Oid = oid() Name = atom() Reason = term()

Transform a oid to its aliasname.

 oid_to_type(Oid) -> {ok, Type} | {error, Reason} Get the type of the the oid Oid = oid() Type = atom() Reason = term()

Retreive the type (asn1 bertype) of an oid.

 backup(BackupDir) -> ok | {error, Reason} Backup manager data BackupDir = string()

Backup persistent data handled by the manager.

BackupDir cannot be identical to DbDir.

info() -> [{Key, Value}] Return information about the manager Key = atom() Value = term()

Returns a list (a dictionary) containing information about the manager. Information includes statistics counters, miscellaneous info about each process (e.g. memory allocation), and so on.

 verbosity(Ref, Verbosity) -> void() Assign a new verbosity for the process Ref = server | config | net_if | note_store | all Verbosity = verbosity() verbosity() = silence | info | log | debug | trace

Sets verbosity for the designated process. For the lowest verbosity silence, nothing is printed. The higher the verbosity, the more is printed.

 format_reason(Reason) -> string() format_reason(Prefix, Reason) -> string() Assign a new verbosity for the process Reason = term() Prefix = integer() | string()

This utility function is used to create a formatted (pretty printable) string of the error reason received from either:

The Reason returned value if any of the sync/async get/get-next/set/get-bulk functions returns {error, Reason}

The Reason parameter in the handle_error user callback function.

Prefix should either be an indention string (e.g. a list of spaces) or a positive integer (which will be used to create the indention string of that length).