I was wondering why /etc/runlevels/default/net.eth0 was a red link: uranus init.d # ls -l /etc/runlevels/default/net* lrwxrwxrwx 1 root root 20 Jan 13 21:58 /etc/runlevels/default/net.eth0 -> /etc/init.d/net.eth0 lrwxrwxrwx 1 root root 20 Jun 25 21:19 /etc/runlevels/default/netmount -> /etc/init.d/netmount uranus init.d # ls -l /etc/init.d/net* -rwxr-xr-x 1 root root 1122 Jan 19 14:09 /etc/init.d/net-check -rwxr-xr-x 1 root root 15143 Jun 26 00:13 /etc/init.d/net.lo -rwxr-xr-x 1 root root 2031 Jun 26 00:13 /etc/init.d/netmount lrwxrwxrwx 1 root root 6 Jan 13 20:53 /etc/init.d/net.peth0 -> net.lo lrwxrwxrwx 1 root root 6 Jan 13 21:05 /etc/init.d/net.ppeth0 -> net.lo -rwxr-xr-x 1 root root 6859 Jun 26 00:13 /etc/init.d/network uranus init.d # The bug is that xen installation renames init.d links, but not runlevel ones ... The bug is related with bug http://bugs.gentoo.org/show_bug.cgi?id=293717 Not, my /etc is messed up. And I have no clue how how I should fix it. Problem is that i don't know if /etc/conf.d/net should describe eth0 or peth0 . I confugred furst my system with network, and installed xen 2 weeks later; xen installation broke everything. config_eth0=( "192.168.0.52 netmask 255.255.255.0 broadcast 192.168.0.255" "192.168.0.80" "192.168.0.83" "2a01:e35:8ba8:e140::1234/64" ) After xen service, everything is broken, and, i have to fix everything manually (automated in /etc/init.d/net-check and /etc/conf.d/local.start After xen service, neither eth0 or peth0 are configured properly, so, for now, i have /etc/init.d/net-check doing this: checkconfig() { source /etc/conf_local IP_1="192.168.0.1" IP_2="72.14.221.104" IP_2="209.85.229.104" IP_2="${ADSL_GATEWAY}" # ebegin "exiting NOW ???" # return 0 ifconfig eth0 2>/dev/null | grep 192.168.0.52 >/dev/null && ebegin "eth0 is ok" || ( ifconfig eth0 192.168.0.52 ; ebegin "We fixed eth0" ; ) route -n 2>/dev/null | grep 192.168.0.1 >/dev/null && ebegin "Routes are fine" || ( route add default gw 192.168.0.1 ; ebegin "We fixed routes" ; ) ping -v -W5 -c3 ${IP_1} -i 0.2 >/dev/null && r1=0 || r1=1 if [ ${r1} -ne 0 ] ; then eerror "Could not ping ${IP_1} (should be gateway or Freebox)." /usr/bin/beep -f 100 -l 1000 -r 3 & return 1 fi ping -v -W5 -c3 ${IP_2} -i 0.2 >/dev/null && r2=0 || r2=1 if [ ${r2} -ne 0 ] ; then eerror "Could not ping ${IP_2} (should be google.fr)." /usr/bin/beep -f 100 -l 1000 -r 3 & return 1 fi return 0 } and in /etc/conf.d/local.start ip -6 addr add 2a01:e35:8ba8:e140::8080/64 dev eth0 /sbin/ifconfig eth0:80 192.168.0.80 Otherwise, the box don't have network at all. When i do these manipulations on peth0, in practice, connectivity does not "work", and network is not "usable". Dispite i am doing this on eth0, domUs work fine, and have usable network.
Created attachment 236635 [details] /tmp/emerge--info
Here is the console output; you will understand which case of net-check occurs: * Starting hddtemp daemon ... [ ok ] * Mounting network filesystems ... [ ok ] * Starting Xen control daemon ... [ ok ] * Starting sshd ... [ ok ] * Starting lighttpd ... [ ok ] * Starting Music Player Daemon ... [ ok ] /etc/init.d/munin-node: line 18: ebegin: command not found /etc/init.d/munin-node: line 20: eend: command not found * ERROR: munin-node failed to start * Checking network connectivity ... * We fixed eth0 ... * We fixed routes ... [ ok ] * Setting clock via the NTP client 'ntpdate' ... [ ok ] * Enabling numlock on ttys ... [ ok ] * Starting sensord ... [ ok ] * Starting S.M.A.R.T. monitoring daemon ... [ ok ] * Starting vixie-cron ... [ ok ] * Starting Xen domains from /etc/xen/auto * Creating screen session to hold domain consoles ... [ ok ] * Starting local ... If i don't do that and leave eth0 as xen left it, NTP fails, what can have many side effects (system date/time not corrected if BIOS time went wrong, other services not starting properly, exim turning mad, fsck, SQL ... ) ... that's why i had to make ntp depend on net-check.
Xen 4.1 in tree. Please test with it and reopen if it doesnt work
