This is quite long but is necessary to understand the non-problem.
I stumbled onto something very fishy... we have many offices around the
world, that connect to the central office trought VPNs. All the offices
are equipped with a dedicated firewall/vpn machine that is basically a
sealed-pre-installed Linux server. We can login remotely to debug or
check if there are problems in the local lan using ssh trought the
internet or by a dial-up modem.
Last week one of our office went off the net completely. We could ssh
or dial-in into the firewall, but from the firewall to the rest of the
lan there was nothing. At the phone with one of the guy over there that
play 'techie' I had him pinging the firewall (success), loggin into the
firewall and pinging the internal network (no luck), then reboot the
main server (inside the lan, no differences).
But what struck me was this: if I started tcpdump on the LAN interface
of the firewall, the one connected to the main switch inside the lan,
the firewall was rebooting every time. Just run tcpdum -i eth0 and BUM!
there goes the firewall...
I diagonsticated a busted switch and had the local techie run for the
first computer shop in the area. Yesterday they finally got a new switch,
replaced the old one (not so old actually) and everything was nice and
tosty again.
But.... how can a busted firewall make an entire machine reboot??
Anybody had ever saw something like this?
Davide
--
In Tennessee, it is illegal to shoot any game other than whales from a
moving automobile.
|
Similar Threads
Linear Mode
