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

Re: Interface configuration - inet6 issues



>>>>> Axton  <axton.grams@gmail.com> writes:

	(Please also consider joining the news:comp.os.linux.networking
	and news:alt.os.linux.debian Usenet newsgroups.  In particular,
	these are available via the free http://Aioe.org/ service, as
	well as through http://groups.google.com/.)

 > To add to this, the following results in a /etc/network/run/ifstate
 > that does not show eth1:

	There's a generic reason of ifstate not being updated after an
	ifup(8) invocation due to a failed pre-up or post-up command.
	Thus, I'd try to debug this problem by running ifup(8) with
	--verbose, like:

# ifup -v eth1

	Such a command would likely to point out the failed command.

	Please note that it may be necessary to manually deconfigure the
	now-misconfigured interface before invoking ifup(8) (as in:
	# ip addr del.)

 > root@ntp01:/# cat /etc/network/interfaces
 > auto lo
 > iface lo inet loopback

 > auto eth1

 > iface eth1 inet static
 > address 10.0.4.240
 > netmask 255.255.252.0
 > broadcast 10.0.7.255

	There's little point in specifying a broadcast address
	coinciding with the one that'd be the default for this
	configuration.

 > gateway 10.0.4.1

 > iface eth1 inet6 static
 > address 2001:740:7063:10::240
 > netmask 64
 > gateway 2001:740:7063:10::10
 > pre-up echo 0 > /proc/sys/net/ipv6/conf/eth1/autoconf
 > pre-up echo 0 > /proc/sys/net/ipv6/conf/eth1/accept_ra

	Why not to use # sysctl -w here instead?  Also, using explicit
	interface names in these commands may lead to hard to debug
	issues when fragments are moved or copied between different
	iface stanzas, especially if the interface names look similar.

	Thus:

    pre-up sysctl -w net.ipv6.conf."$IFACE".accept_ra=0

 > root@ntp01:/# cat /etc/network/run/ifstate
 > lo=lo

 > The interfaces are configured as follows with the above
 > configuration:

	Then it's my guess that the interface wasn't correctly
	deconfigured before invoking ifup(8).  (It might've been a
	failed pre-up command, but then the interface wouldn't have been
	assigned the addresses.)

[…]

-- 
FSF associate member #7257


Reply to: