root@helen:~# uname -a
Linux helen 2.6.28-17-generic #58-Ubuntu SMP Tue Dec 1 21:27:25 UTC 2009 x86_64 GNU/Linux
This box has two Ethernet interfaces. The onboard nic - eth0 - connects to a Ciena 3290 and the Internet.
A cable cut in our upstream infrastructure yesterday resulted in eth0 and eth2 going down, eth0 coming up but eth2 remaining down until a reboot.
Feb 18 10:26:27 helen kernel: [3251918.453137] r8169: eth0: link down
Feb 18 10:26:30 helen kernel: [3251921.244552] r8169: eth0: link up
Feb 18 10:26:31 helen kernel: [3251922.156904] eth2: link down
Feb 18 21:50:17 helen syslogd 1.5.0#5ubuntu3: restart.
Feb 18 21:50:18 helen kernel: Inspecting /boot/System.map-2.6.28-17-generic
The upstream provider says the Ciena 3290 did not get a mac address from eth0 (ARP failure?)
On reboot dmesg says eth2 above is now eth1.
root@helen:~# dmesg | grep eth
[*** 1.329422] Driver 'sd' needs updating - please use bus_type methods
iface eth2 inet static
*** address 172.17.100.5
*** netmask 255.255.255.252
*** up ip route add 18.104.22.168/26 via 172.17.100.6
*** up ip route add 10.4.0.0/24 via 172.17.100.6
*** up ip route add 10.5.0.0/24 via 172.17.100.6
1) would an upgrade to 9.10 help?
2) is there a simple watchdog that would test upstream connectivity periodically and bounce the interface is there is none
3) same for downstream?
4) can the link down for eth2 be prevented?
5) is there a way to refresh detection of eth2?
Seems like Hotel California. I am stuck in Networking 101.
Rev. Jim Tarvid, PCA
ubuntu-server mailing list
More info: https://wiki.ubuntu.com/ServerTeam