NAME
MakeRancidConf - Generate rancid Configuration
INTRODUCTION
This worker will generate a rancid configuration for all devices in Netdisco.
Optionally you can provide configuration to control the output, however the defaults are sane for rancid versions 3.x and will create one rancid group called default
which contains all devices. Those devices not discovered successfully within the past day will be marked as down
for rancid to skip. Configuration is saved to the ~/rancid subdirectory of Netdisco's home folder.
Note that this only generates the router.db files, you will still need to configure rancid's .cloginrc and schedule rancid-run
to run.
You could run this worker at 09:05 each day using the following configuration:
schedule:
makerancidconf:
when: '5 9 * * *'
Since MakeRancidConf is a worker module it can also be run via netdisco-do
:
~/bin/netdisco-do makerancidconf
Skipped devices and the reason for skipping them can be seen by using -D
:
~/bin/netdisco-do makerancidconf -D
CONFIGURATION
Here is a complete example of the configuration, which must be called rancid
. All keys are optional:
rancid:
rancid_cvsroot: '$ENV{NETDISCO_HOME}/rancid' # default
rancid_conf: '/etc/rancid' # default
down_age: '1 day' # default
delimiter: ';' # default
default_group: 'default' # default
excluded:
excludegroup1: 'host_group1_acl'
excludegroup2: 'host_group2_acl'
groups:
groupname1: 'host_group3_acl'
groupname2: 'host_group4_acl'
vendormap:
vname1: 'host_group5_acl'
vname2: 'host_group6_acl'
by_ip: 'host_group7_acl'
by_hostname: 'host_group8_acl'
Note that the default directory for writing files is not /var/lib/rancid so you may wish to set this in rancid_cvsroot
, (especially if migrating from the old netdisco-rancid-export
script).
Any values above that are a host group ACL will take either a single item or a list of network identifiers or device properties. See the ACL documentation wiki page for full details. We advise you to use the host_groups
setting and then refer to named entries in that, for example:
host_groups:
coredevices: '192.0.2.0/24'
edgedevices: '172.16.0.0/16'
grp-nxos: 'os:nx-os'
rancid:
groups:
core_devices: 'group:coredevices'
edge_devices: 'group:edgedevices'
vendormap:
cisco-nx: 'group:grp-nxos'
by_ip: 'any'
Do not forget that rancid also needs configuring when adding a new group, such as scheduling the group to run, adding it to rancid.conf, setting up the email config and creating the repository with rancid-cvs
.
rancid_conf
The location where the rancid configuration (rancid.types.base and rancid.types.conf) is installed. It will be used to check the existance of device types before exporting the devices to the rancid configuration. If no match is found the device will not be added to rancid.
rancid_cvsroot
The location to write rancid group configuration files (router.db) into. A subdirectory for each group will be created.
down_age
This should be the same or greater than the interval between regular discover jobs on your network. Devices which have not been discovered within this time will be marked as down
to rancid.
The format is any time interval known and understood by PostgreSQL, such as at https://www.postgresql.org/docs/10/static/functions-datetime.html.
delimiter
Set this to the delimiter character for your router.db entries if needed to be different from the default, the default is ;
.
default_group
Put devices into this group if they do not match any other groups defined.
excluded
This dictionary defines a list of devices that you do not wish to export to rancid configuration.
The value should be a Netdisco ACL to select devices in the Netdisco database.
groups
This dictionary maps rancid group names with configuration which will match devices in the Netdisco database.
The left hand side (key) should be the rancid group name, the right hand side (value) should be a Netdisco ACL to select devices in the Netdisco database.
vendormap
If the device vendor in Netdisco is not the same as the rancid vendor script or device type, configure a mapping here.
The left hand side (key) should be the rancid device type, the right hand side (value) should be a Netdisco ACL to select devices in the Netdisco database.
Note that vendors might have a large array of operating systems which require different rancid modules. Mapping operating systems to rancid device types is a good solution to use the correct device type. Example:
host_groups:
grp-ciscosb: 'os:ros'
rancid:
vendormap:
cisco-sb: 'group:grp-ciscosb'
by_ip
Netdisco ACL to select devices which will be written to the rancid config as an IP address, instead of the DNS FQDN or SNMP hostname.
by_hostname
Netdisco ACL to select devices which will have the unqualified hostname written to the rancid config. This is done simply by stripping the domain_suffix
configuration setting from the device FQDN.