GSP
Quick Navigator

Search Site

Unix VPS
A - Starter
B - Basic
C - Preferred
D - Commercial
MPS - Dedicated
Previous VPSs
* Sign Up! *

Support
Contact Us
Online Help
Handbooks
Domain Status
Man Pages

FAQ
Virtual Servers
Pricing
Billing
Technical

Network
Facilities
Connectivity
Topology Map

Miscellaneous
Server Agreement
Year 2038
Credits
 

USA Flag

 

 

Man Pages
GDNSD-PLUGIN-MULTIFO(8) gdnsd GDNSD-PLUGIN-MULTIFO(8)

gdnsd-plugin-multifo - gdnsd plugin for multi-address, all-active failover groups

Example plugin config:

  plugins => {
    multifo => {
      up_thresh => 0.3
      service_types => up,
      v4www => {
        lb01 => 192.0.2.200,
        lb02 => 192.0.2.201,
        lb03 => 192.0.2.202,
      }
      v6smtp => {
        service_types => [ smtp ],
        up_thresh => 0.1,
        lb01_v6 => 2001:DB8::1,
        lb02_v6 => 2001:DB8::2,
        lb03_v6 => 2001:DB8::3,
      }
      pubwww => {
        up_thresh => 0.5
        service_types => corpwww_type
        addrs_v4 => [ 192.0.2.100, 192.0.2.101, 192.0.2.102 ]
        addrs_v6 => {
          service_types => [ up ],
          up_thresh => 0.7
          lb01_v6 => 2001:DB8::1,
          lb02_v6 => 2001:DB8::2,
          lb03_v6 => 2001:DB8::3,
        }
      }
    }
  }

Example zonefile RRs:

  web4 180 DYNA multifo!v4www
  smtp 180 DYNA multifo!v6smtp
  www 180 DYNA multifo!pubwww

gdnsd-plugin-multifo is designed to do multi-address all-active failover grouping. Basically, for each configured resource name, you supply a labeled list of addresses. multifo monitors these addresses according to "service_types", and answers "DYNA" address queries using the non-"DOWN" subset. The core gdnsd code will round-robin rotate the records on the way out, as it does for all address RR-sets.

At the top level of the plugin's configuration stanza, three special parameters "up_thresh", "service_types", and "ignore_health" are supported. These set default per-resource options of the same name for any resources which do not define them explicitly.

The rest of the hash entries at the top level are the names of the resources you define. Each resource gets a configuration hash of its own for containing resource-specific parameters as well as the actual address data.

Within a resource, you have two basic options. You can either directly specify a set of "label => address" pairs which are all the same family (IPv4 or IPv6), or you can use the sub-stanzas "addrs_v4" and/or "addrs_v6" to specify one or both families in the same resource.

The "up_thresh", "service_types", and "ignore_health" parameters are inherited through every level, and can be overridden at any level (even per-address-family):

up_thresh
Floating point, default 0.5, range (0.0 - 1.0]. This configures the per-resource "up_thresh" threshold. More details in "UP THRESH" below.
service_types
Array of strings, or single string. Default "default". This sets the monitored service_types for this resource. If an array of more than one is provided, all will be monitored for each address, and the net monitored state will be the minimum (worst) of the set. See gdnsd.config(8) for more details on service_types.
ignore_health
Boolean, default false. If set to true, the health of individual addresses will not affect whether multifo adds them to the set of output addresses, but it will still be checked and used for the "up_thresh" calculation which is consumed by meta-plugins like geoip and metafo, which might use that information to fail over to a completely different datacenter as a result.

If you have no parameters (service_types, up_thresh, ignore_health) to configure in a given stanza (single-family direct resource config, or addrs_v[46]), and do not care about the descriptive per-address labels used in monitoring, you can replace the hash with an array of addresses. The labels will be generated for you as a series of integers starting with 1. For example, the following are equivalent:

   res1 => { addrs_v4 => [ 192.0.2.1, 192.0.2.2 ] }
   res1 => { addrs_v4 => { 1 => 192.0.2.1, 2 => 192.0.2.2 } }

All of the addresses for all of the resources are monitored using the per-address-family inherited "service_types" specified (default would be the static virtual monitor "up"). When the core daemon requests a lookup for address records of a given family on one of this plugin's resources, it goes through essentially the following process to determine the set of response addresses for that address family: 1) Add all non-DOWN addresses to the result set. 2) If the set of non-DOWN addresses fail the up_thresh check, add *all* addresses to the result set as a fallback. 3) If any address is in the DOWN state, cut the zonefile-specified TTL in half.

If "ignore_health" is true, all addresses are added to the result set regardless of health, but the up_thresh and TTL effects still happen, and the final resource-level state still reflects the overall state as it would without "ignore_health".

This process is repeated independently for each of the IPv4 and IPv6 address subsets, in the case that a resource has both address families configured (the TTL is only cut in half once of course). Details on the up_thresh check follow:

If there are not enough UP addresses to pass the threshold (per address family), all addresses (of a given address family) will be returned as a fallback.

The threshold is implemented mathematically as in the following pseudo-code "if(non_down >= ceil(thresh * total)) threshold_passed;". For example, if thresh is at the default value of 0.5, and there are 3 total IPv4 addresses, then 2 of them must be non-down to pass the threshold. The net result is that with the default threshold, the plugin will never return an isolated single address from a set of 3. It will either return all 3, or it will return 2/3 if a single address from the set has failed.

When the threshold check fails (and all addresses are returned) for either address family, resource-level total failure will also be signaled to any applicable upstream meta-plugins such as metafo or geoip.

General rules for the results of the up_thresh formula:

  • A threshold of 1.0 will only pass if all addresses are not-down. This is mostly pointless, you might as well not monitor anything and set up these addresses as a static set in a zonefile.
  • A threshold of 0.01 will pass even if only one address is alive and return just that one address, even if it's e.g. the only one left out of 40.
  • Because a threshold of 0.0 is illegal, if all addresses are down the threshold will always fail, returning all addresses.

Intermediate value examples: (threshold: non-down/total required to pass threshold):

  • 0.1: 1/1 1/2 1/3 1/4 1/5 1/6 1/7 1/8 2/16
  • 0.2: 1/1 1/2 1/3 1/4 1/5 2/6 2/7 2/8 4/16
  • 0.3: 1/1 1/2 1/3 2/4 2/5 2/6 3/7 3/8 5/16
  • 0.4: 1/1 1/2 2/3 2/4 2/5 3/6 3/7 4/8 7/16
  • 0.5: 1/1 1/2 2/3 2/4 3/5 3/6 4/7 4/8 8/16
  • 0.6: 1/1 2/2 2/3 3/4 3/5 4/6 5/7 5/8 10/16
  • 0.7: 1/1 2/2 3/3 3/4 4/5 5/6 5/7 6/8 12/16
  • 0.8: 1/1 2/2 3/3 4/4 4/5 5/6 6/7 7/8 13/16
  • 0.9: 1/1 2/2 3/3 4/4 5/5 6/6 7/7 8/8 15/16

gdnsd.config(5), gdnsd.zonefile(5), gdnsd(8), gdnsd-plugin-simplefo(8)

The gdnsd manual.

Copyright (c) 2012 Brandon L Black <blblack@gmail.com>

This file is part of gdnsd.

gdnsd is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

gdnsd is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with gdnsd. If not, see <http://www.gnu.org/licenses/>.

2022-04-09 gdnsd 3.7.0

Search for    or go to Top of page |  Section 8 |  Main Index

Powered by GSP Visit the GSP FreeBSD Man Page Interface.
Output converted with ManDoc.