Join us at engage.eucalyptus.com
Hi, again! I noticed today some strange behavior. I have 2 computers connecting to each other directly. One of them is euca-cluster (cloud controller + walrus + cluster controller + storage controller) and other is euca-node (node controller). In the current step i don't need any external connections, therefore i didn't set gateway parameter in /etc/network/interfaces for my only network interface:
auto eth0
iface eth0 inet static
address 192.168.33.33
netmask 255.255.255.0
network 192.168.33.0
broadcast 192.168.33.1
# gateway x.x.x.x - COMMENTED
After restarting network (sudo /etc/init.d/networking restart), i got this route table (route -n):
Destination Gateway Genmask Flags Metric Ref Use Iface
192.168.33.0 0.0.0.0 255.255.255.0 U O O O eth0
Then i rebooted, checked eucalyptus service status (sudo status eucalyptus) and got this:
eucalyptus stop/waiting
Then i decided set gateway parameter in /etc/network/interfaces:
auto eth0
iface eth0 inet static
address 192.168.33.33
netmask 255.255.255.0
network 192.168.33.0
broadcast 192.168.33.1
gateway 192.168.33.1
NOTE: I don't have any machine with ip-addr = 192.168.33.1. It's just fake address.
Route table for this configuration:
Destination Gateway Genmask Flags Metric Ref Use Iface
192.168.33.0 0.0.0.0 255.255.255.0 U O O O eth0
0.0.0.0 192.168.33.1 0.0.0.0 UG O O O eth0
After system rebooting eucalyptus service successfully started:
eucalyptus start/running, process XXXX
Answer:
So, in my opinion, default route have influence on successful running of eucalyptus service. Is it real so? Why? Is it depends on VNET_MODE?. The thing is, i don't have any real gateway-machine or switch. As i mentioned before both of my machines just connected directly each to other.
Thanks in advance!
P.S.
1. This behavior i have on both machines (euca-cluster and euca-node)
2. OS = Ubuntu 10.04 with last updates (29 Jule 2010)
3. eucalyptus 1.6.2 with last updates (29 Jule 2010) from ubuntu repo
4. VNET_MODE = "MANAGED-NOVLAN"