[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#686970: Inconsistent ipv6 handling of localhost in netbase and netcfg



Package: netbase,netcfg
Severity: normal
Tags: d-i ipv6

In #685362, a user complains that PostgreSQL doesn't listen on ipv6
localhost, unless he has "::1 localhost" in his /etc/hosts. I can
confirm that adding this fixes the problem for postgresql-9.1. Looking
through my systems, I'm seeing different defaults for "localhost". All
systems have

127.0.0.1	localhost

but then some have

::1     localhost ip6-localhost ip6-loopback

while some only have

::1     ip6-localhost ip6-loopback

Digging a bit further, I found that the "ipv6 localhost" variant is
installed by netbase's postinst:

create_hosts_file() {
  if [ -e /etc/hosts ]; then return 0; fi

    cat > /etc/hosts <<-EOF
            127.0.0.1       localhost
            ::1             localhost ip6-localhost ip6-loopback
[...]

And the "ipv4 only" variant is installed by netcfg's udeb:

#define IPV6_HOSTS \
"# The following lines are desirable for IPv6 capable hosts\n" \
"::1     ip6-localhost ip6-loopback\n" \

netcfg was actually using the "ipv6 localhost" variant until that got
reverted in 1.57, closing #595107.


I'm filing the bug on both netbase and netcfg so the maintainers can
make the behavior the same one way or the other.

From my point of view, localhost should be handled like any other
multi-homed host, that is, it should have both addresses. It might
have been different in 2010, wenn netcfg 1.57 was uploaded, but
nowadays applications really need to cope with ipv6 being present.

Christoph
-- 
cb@df7cb.de | http://www.df7cb.de/

Attachment: signature.asc
Description: Digital signature


Reply to: