Hello! I have been trying to run Alpine (edge) containers on Debian (bookworm unstable, kernel 6.1.0, lxd/lxc version 5.0.2), but every time I run lxc stop or lxc restart, the host’s network adapter goes down, and my SSH connection hangs.
Based on this discussion, I’ve tried statically defining the bridge.hwaddr for the lxdbr0 bridge config
config:
bridge.hwaddr: 02:3a:f6:d5:08:f1
ipv4.address: 10.128.37.1/24
ipv4.nat: "true"
ipv6.address: fd42:fb55:60e2:5f53::1/64
ipv6.nat: "true"
description: ""
name: lxdbr0
type: bridge
used_by:
- /1.0/profiles/default
- /1.0/instances/still-elk
- /1.0/instances/above-cockatoo
[truncating instance list for brevity]
managed: true
status: Created
locations:
- none
but to no avail.
Based on this discussion, I disabled the Docker userlandproxy, and added the ENV{INTERFACE}=="veth*", ENV{NM_UNMANAGED}="1" udev rule.
After running lxc reboot on one of the containers, the network froze, and I used Linode’s lish to check the kernel logs, which show this, but I couldn’t see anything that stood out to me.
[ 103.846824] veth755f10f4: renamed from physpVCCoj
[ 103.854632] lxdbr0: port 1(vethc5d309a8) entered disabled state
[ 104.019509] device vethc5d309a8 left promiscuous mode
[ 104.020028] lxdbr0: port 1(vethc5d309a8) entered disabled state
[ 104.749945] kauditd_printk_skb: 7 callbacks suppressed
[ 104.749949] audit: type=1400 audit(1677465441.806:61): apparmor="STATUS" operation="profile_remove" profile="unconfined" name="lxd-above-cockatoo_</var/lib/lxd>" pid=6774 comm="apparmor_parser"
[ 104.782461] lxdbr0: port 1(vethe3381d02) entered blocking state
[ 104.783109] lxdbr0: port 1(vethe3381d02) entered disabled state
[ 104.783699] device vethe3381d02 entered promiscuous mode
[ 104.784269] lxdbr0: port 1(vethe3381d02) entered blocking state
[ 104.784840] lxdbr0: port 1(vethe3381d02) entered forwarding state
[ 104.862468] lxdbr0: port 1(vethe3381d02) entered disabled state
[ 104.872434] audit: type=1400 audit(1677465441.930:62): apparmor="STATUS" operation="profile_load" profile="unconfined" name="lxd-above-cockatoo_</var/lib/lxd>" pid=6807 comm="apparmor_parser"
[ 104.927807] physBKzS9V: renamed from veth911ecca6
[ 104.938745] eth0: renamed from physBKzS9V
[ 104.951534] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[ 104.952262] lxdbr0: port 1(vethe3381d02) entered blocking state
[ 104.952934] lxdbr0: port 1(vethe3381d02) entered forwarding state
[ 104.972410] Not activating Mandatory Access Control as /sbin/tomoyo-init does not exist.
[ 105.116836] audit: type=1400 audit(1677465442.174:63): apparmor="DENIED" operation="mount" info="failed flags match" error=-13 profile="lxd-above-cockatoo_</var/lib/lxd>" name="/dev/" pid=7052 comm="busybox" flags="rw, nosuid, noexec, remount, silent"
[ 105.120142] audit: type=1400 audit(1677465442.182:64): apparmor="DENIED" operation="mount" info="failed flags match" error=-13 profile="lxd-above-cockatoo_</var/lib/lxd>" name="/dev/" pid=7052 comm="busybox" flags="ro, nosuid, noexec, remount, silent"
Even though “link becomes ready” is shown, I am still unable to ssh back in.
I have no problem starting containers, or using the network within the containers once they are up.
My apologies if this post is a bit unwieldy, I’m still new to the forum!