Let ipv4-or-ipv6 prefer IPv6 instead of ipv4 #10500
Conversation
Previously the magical "ipv4-or-ipv6" CheckCommand template preferred the IPv4 address (`address` field of a Host object) to the IPv6 address (`address6` field of a Host object). This commit reverses the behaviour to favor IPv6 as it is common among different applications in order to further the transition to IPv6.
There was a problem hiding this comment.
As written over in the original PR, I am in favor for this change, reflecting the IETF's preference for IPv6 as also stated in RFC 8305. Since there are the $check_ipv4$ and $check_ipv6$ macros being set by multiple CheckCommands with an additional -4 or -6 argument, there should be no harm.
On the other hand, the current version works and just changing the behavior because one likes IPv6 better is a poor argument.
Thus, I would like to another opinion on this change.
|
IMHO there's no real benefit from this change. It just risks breaking existing configurations. If you want that behavior, you already have the option of setting the
No it doesn't as it requires that there was already an |
Previously the magical "ipv4-or-ipv6" CheckCommand template preferred the IPv4 address (
addressfield of a Host object) to the IPv6 address (address6field of a Host object).This commit reverses the behaviour to favor IPv6 as it is common among different applications in order to further the transition to IPv6.
This is a breakout PR related to #10494 where the idea originated.
Practically speaking this change will not affect many setups, since the
common enterprise network setup still relies on legacy IP only.
For some setups with dual-stack though this will change the behaviour quite
suddenly although the impact should be inconsiderable and go unnoticed
if the the network setup was done properly.