"
# The loopback network interface
auto lo
iface lo inet loopback
# The primary network interface
#allow-hotplug eth0
#iface eth0 inet dhcp
iface eth0 inet manual
auto br0
iface br0 inet dhcp
bridge_ports eth0
bridge_stp off
bridge_waitport 0
bridge_fd 0
bridge_maxwait 0
#bridge_hw
"
I noticed two systems running Jessie. The setup for Station 1 with the
above interfaces works without issue, but Station 2 currently has a
problem with "iface br0 inet static"
Station 2 I'll be referring to even though I've managed a workaround. I
show the workaround I use. I'll also give a little note here before
diving into this..
A little story here before setting up a bridge device, since my time
with testing Jessie few weeks ago I noticed that systemd(or "something"
with networking, I'm making sure I'm not using NM with these server
machines), the "allow-hotplug eth0" is necessary for eth0 to work
properly on bootup. In other words,
[Tested on Station 1]
a)
"allow-hotplug eth0
iface eth0 inet dhcp
"
, or
b)
""allow-hotplug eth0
iface eth0 inet dhcp
"
is not good enough, but
"allow-hotplug eth0
auto eth0
iface eth0 inet dhcp
"
gets eth0 working with dhcp as intended
I'm assuming a) is not an issue with NM because NM would require the
ifupdown plugins with a managed= directive in networkmanager.conf..
Substracting NM, now looking at setting up bridge br0 on the Station2's
"inet br0 static" configuration.. You are right allow-hotplug is not
needed for br0(and neither should it be with inet br0 dhcp, which I
approve as correct on Station 1's example) but there's something broken
to having bridging working properly with "iface eth0 static"(Station 2)
which is I'm bothering to throw in allow-hotplug with br0.
^ If you can explain to me why "allow-hotplug" is needed for the
non-bridge eth0 setup that I'm interested after, so I can only speculate
it would also be needed without harm for trying with br0 as well. But
here again I know "iface br0 static" on station 1 does not need it.
The station 2 configuration file I have for its interfaces:
"allow-hotplug eth0
#auto eth0
iface eth0 inet manual
#pre-up /sbin/ip link set eth0 address 70:54:d2:19:69:f7
up /sbin/ip link set eth0 up
pre-down /sbin/ip addr flush dev eth0
down /sbin/ip link set eth0 down
#allow-hotplug br0
auto br0
iface br0 inet static
bridge_ports eth0
bridge_ ..
address <>
netmask <>
network <>192.95.29.0
broadcast <>
gateway <>
"
The workaround I use is with /etc/rc.local
"
/sbin/ifup eth0
/sbin/ip link set eth0 up
/sbin/ifup br0
exit 0
"
And yes, you'll just say it all has to be set properly in the interfaces
and one should be using "auto br0" to have it load automatically on
boot-up. Unfortunately this is not happening otherwise I would of solved
it quite easily. The Station 2 is a completely new install of Debian
Jessie on a "remote" system. It was actually hard for me to diagnose
remotely it so I got leased a KVM-over-ip for a day and I know what I'm
filing here is as accurate as I can possibly describe because I really
wish somebody can fix this with due all respect (if you don't mind me
saying because I really want to get to why this is happening! :).
I went to setup a KVM on a local station(Station 1) here from scratch
and tested "iface br0 inet dhcp" without issues.. And not to add to any
confusion here, I went ahead with another test on Station 1 using "iface
br0 inet static" and it definitely works.. so why oh why is it not
working on Station 2?
They both are loading the "bridge" module (and needless to say I don't
even think they need it? I do know iproute2/"ip bridge" adds more port
features to the traditional bridge devices.)
So what is the problem with bridge-utils? Can it be using a problematic
probe tool causing it not to work with e1000e nic interfaces? A few days
ago I installed "ethtool" to see if anything of ifup's scripts was
problematic without it. Could it be a "hardware" issue? I don't think so
because I can just issue "ifup eth0, ifup br0" from /etc/rc.local and it
works properly (why ethtool? because I noticed one of the ifup scripts
was scanning for ethtool and thought it can ameliorate something to do
with Station'2 particular e1000e driver. It generates the file
/etc/network/if-up.d/ethtool)
The system on Station 2 was setup from scratch by me with a debian
jessie installer. I'm sure something is broken somewhere but I don't
know where to look..
Also if anyone knows as to why
"allow-hotplug eth0" is needed for an eth0-only connection, I'm
interested on this as well, I suppose if it is not needed then I would
have to file a bugreport about this because over here when testing
things I actually needed to use it.
If this sounds complicating then I apologize, I've been on this over the
weekend and right now br0 is being loaded up properly with that rc.local
workaround when I know this is not the "perfect fix" so there's
definitely a bug somewhere. Let me me know what I should try next and
I'll do my best to see if I can manage things without too much disruption..
thanks
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
