lib/snmp/mib.rb
#
# Copyright (c) 2004 David R. Halliday
# All rights reserved.
#
# This SNMP library is free software. Redistribution is permitted under the
# same terms and conditions as the standard Ruby distribution. See the
# COPYING file in the Ruby distribution for details.
#
require 'snmp/varbind'
require 'fileutils'
require 'yaml'
module SNMP
class MIB
DEFAULT_MIB_PATH = File.expand_path(
File.join(File.dirname(__FILE__), "..", "..", "data", "snmp", "mibs")
)
#:startdoc:
MODULE_EXT = 'yaml'
class ModuleNotLoadedError < RuntimeError; end
class << self
##
# Import an SMIv2 MIB file for later loading. A module only needs to
# be installed once.
#
# module_file - the filename of the module to be imported
# mib_dir - the output directory for the serialized MIB data
#
# NOTE: This implementation requires that the 'smidump' tool is available
# in the PATH. This tool can be obtained from the libsmi website at
# http://http://www.ibr.cs.tu-bs.de/projects/libsmi/ .
#
# ALSO NOTE: The file format in future releases is subject to
# change. For now, it is a simple YAML hash with the MIB symbol
# as the key and the OID as the value. These files could be
# generated manually if 'smidump' is not available.
#
# Here is an example of the contents of an output file:
#
# ---
# ipDefaultTTL: 1.3.6.1.2.1.4.2
# ipForwDatagrams: 1.3.6.1.2.1.4.6
# ipOutRequests: 1.3.6.1.2.1.4.10
# ipOutNoRoutes: 1.3.6.1.2.1.4.12
# ipReasmTimeout: 1.3.6.1.2.1.4.13
# icmpInDestUnreachs: 1.3.6.1.2.1.5.3
#
def import_module(module_file, mib_dir=DEFAULT_MIB_PATH)
raise "smidump tool must be installed" unless import_supported?
FileUtils.makedirs mib_dir
mib_hash = `smidump -f python #{module_file}`
mib = eval_mib_data(mib_hash)
if mib
module_name = mib["moduleName"]
raise "#{module_file}: invalid file format; no module name" unless module_name
if mib["nodes"]
oid_hash = {}
mib["nodes"].each { |key, value| oid_hash[key] = value["oid"] }
if mib["notifications"]
mib["notifications"].each { |key, value| oid_hash[key] = value["oid"] }
end
File.open(module_file_name(module_name, mib_dir), 'w') do |file|
YAML.dump(oid_hash, file)
file.puts
end
module_name
else
warn "*** No nodes defined in: #{module_file} ***"
nil
end
else
warn "*** Import failed for: #{module_file} ***"
nil
end
end
##
# Returns the full filename of the imported MIB file for the given
# module name.
#
def module_file_name(module_name, mib_dir=DEFAULT_MIB_PATH)
File.join(mib_dir, module_name + "." + MODULE_EXT)
end
##
# The MIB.import_module method is only supported if the external
# 'smidump' tool is available. This method returns true if a
# known version of the tool is available.
#
def import_supported?
`smidump --version` =~ /^smidump 0.4/ && $? == 0
end
##
# Returns a list of MIB modules that have been imported. All of
# the current IETF MIBs should be available from the default
# MIB directory.
#
# If a regex is provided, then the module names are matched
# against that pattern.
#
def list_imported(regex=//, mib_dir=DEFAULT_MIB_PATH)
list = []
Dir["#{mib_dir}/*.#{MODULE_EXT}"].each do |name|
module_name = File.basename(name, ".*")
list << module_name if module_name =~ regex
end
list
end
private
def eval_mib_data(mib_hash)
ruby_hash = mib_hash.
gsub(':', '=>'). # fix hash syntax
gsub('(', '[').gsub(')', ']'). # fix tuple syntax
sub('FILENAME =', 'filename ='). # get rid of constants
sub('MIB =', 'mib =')
mib = nil
eval(ruby_hash)
mib
end
end # class methods
def initialize
@by_name = {}
@by_module_by_name = {}
end
##
# Loads a module into this MIB. The module must be imported before it
# can be loaded. See MIB.import_module .
#
def load_module(module_name, mib_dir=DEFAULT_MIB_PATH)
oid_hash = nil
File.open(MIB.module_file_name(module_name, mib_dir)) do |file|
oid_hash = YAML.load(file.read)
end
@by_name.merge!(oid_hash) do |key, old, value|
warn "warning: overwriting old MIB name '#{key}'"
end
@by_module_by_name[module_name] = {}
@by_module_by_name[module_name].merge!(oid_hash)
end
##
# Returns a VarBindList for the provided list of objects. If a
# string is provided it is interpreted as a symbolic OID.
#
# This method accepts many different kinds of objects:
# - single string object IDs e.g. "1.3.6.1" or "IF-MIB::ifTable.1.1"
# - single ObjectId
# - list of string object IDs
# - list of ObjectIds
# - list of VarBinds
#
def varbind_list(object_list, option=:KeepValue)
vb_list = VarBindList.new
if object_list.respond_to? :to_str
vb_list << oid(object_list).to_varbind
elsif object_list.respond_to? :to_varbind
vb_list << apply_option(object_list.to_varbind, option)
else
object_list.each do |item|
if item.respond_to? :to_str
varbind = oid(item).to_varbind
else
varbind = item.to_varbind
end
vb_list << apply_option(varbind, option)
end
end
vb_list
end
def apply_option(varbind, option)
if option == :NullValue
varbind.value = Null
elsif option != :KeepValue
raise ArgumentError, "invalid option: #{option.to_s}", caller
end
varbind
end
private :apply_option
##
# Returns a VarBind object for the given name and value. The name
# can be a String, ObjectId, or anything that responds to :to_varbind.
#
# String names are in the format <ModuleName>::<NodeName>.<Index> with
# ModuleName and Index being optional.
#
def varbind(name, value=Null)
if name.respond_to? :to_str
vb = VarBind.new(oid(name), value)
else
vb = name.to_varbind
vb.value = value
end
vb
end
##
# Returns an ObjectId for the given name. Names are in the format
# <ModuleName>::<NodeName>.<Index> with ModuleName and Index being
# optional.
#
def oid(name)
module_parts = name.to_str.split("::")
if module_parts.length == 1
parse_oid(@by_name, name.to_str)
elsif module_parts.length == 2
module_name = module_parts[0]
oid = module_parts[1]
module_hash = @by_module_by_name[module_name]
if module_hash
parse_oid(module_hash, oid)
else
raise ModuleNotLoadedError, "module '#{module_name}' not loaded"
end
else
raise ArgumentError, "invalid format: #{name.to_str}"
end
end
def parse_oid(node_hash, name)
oid_parts = name.split(".")
first_part = oid_parts.shift
oid_string = node_hash[first_part]
if oid_string
oid_array = oid_string.split(".")
else
oid_array = [first_part]
end
oid_array.concat(oid_parts)
ObjectId.new(oid_array)
end
private :parse_oid
end
end # module SNMP