[one-users] Strange Networking Problems
Markus Hubig
mhubig at imko.de
Thu Nov 15 04:44:30 PST 2012
Hi @all,
I'm experencing a strage networking behavor here. Every time I login to a VM
from my workstation, my SSH session freezes regularlily for a couple of
seconds, the it starts to work again. Also if I try to ping something I have
a huge package drop.
| # ping puppet.imko
| PING puppet.imko (192.168.1.60) 56(84) bytes of data.
| From jenkins.imko (192.168.1.70) icmp_seq=1 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=2 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=3 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=4 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=5 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=6 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=7 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=8 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=9 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=10 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=11 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=12 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=13 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=14 Destination Host Unreachable
| From jenkins.imko (192.168.1.70) icmp_seq=15 Destination Host Unreachable
| 64 bytes from cc00.imko (192.168.1.60): icmp_req=16 ttl=64 time=1.26 ms
| 64 bytes from cc00.imko (192.168.1.60): icmp_req=17 ttl=64 time=0.504 ms
| 64 bytes from cc00.imko (192.168.1.60): icmp_req=18 ttl=64 time=0.487 ms
| 64 bytes from cc00.imko (192.168.1.60): icmp_req=19 ttl=64 time=0.464 ms
| 64 bytes from cc00.imko (192.168.1.60): icmp_req=20 ttl=64 time=0.541 ms
| 64 bytes from cc00.imko (192.168.1.60): icmp_req=21 ttl=64 time=0.498 ms
| 64 bytes from cc00.imko (192.168.1.60): icmp_req=22 ttl=64 time=0.440 ms
| 64 bytes from cc00.imko (192.168.1.60): icmp_req=23 ttl=64 time=0.534 ms
I suspect that I have some kind of a network missconfiguration, but I can't
find the problem... ;-(
I set up a CloudController running one with two network connections eth0 and
eth1. eth0 (192.168.1.60) is intended to connect to the company LAN and eth1
(172.16.1.10) is the infrastrukture network intendet to connect to the
computing nodes and the storage servers (glusterfs). The computing nodes and
the glusterfs nodes have only ips in the 172.168.1.0/24 network to separate
them from the company lan (192.168.1.0/24).
The computing nodes have 4 network ports which I bond together to incrase the
thouput (as seen on the image). bond0 is added to the bridge br0 and has no ip
address, bond1 has an address from the 172.168.1.0/24 network.
The ifconfig output on node01 shows that only bond1 has an ip address. And the
routing table looks like this:
| Destination Gateway Genmask Flags Metric Ref Use Iface
| 0.0.0.0 172.16.1.10 0.0.0.0 UG 100 0 0 bond1
| 172.16.1.0 0.0.0.0 255.255.255.0 U 0 0 0 bond1
The /etc/network/interfaces on node01 looks like this:
| # The loopback network interface
| auto lo
| iface lo inet loopback
|
| # The eth0 interface
| auto eth0
| iface eth0 inet manual
| bond-master bond0
|
| # The eth1 interface
| auto eth1
| iface eth1 inet manual
| bond-master bond0
|
| # The eth2 interface
| auto eth2
| iface eth2 inet manual
| bond-master bond1
|
| # The eth3 interface
| auto eth3
| iface eth3 inet manual
| bond-master bond1
|
| # The public network interface
| auto bond0
| iface bond0 inet manual
| # Bind this interfaces
| bond-slaves none
| # Balance-RR configuration
| bond_mode balance-rr
| bond_miimon 100
| bond_updelay 200
| bond_downdelay 200
|
| # The storage network interface
| auto bond1
| iface bond1 inet static
| # Static assign the IP, netmask, default gateway.
| address 172.16.1.21
| netmask 255.255.255.0
| gateway 172.16.1.10
| # Bind this interfaces
| bond-slaves none
| # Balance RR configuration
| bond_mode balance-rr
| bond_miimon 100
| bond_updelay 200
| bond_downdelay 200
|
| # The public bridge interface
| auto br0
| iface br0 inet static
| # Static assign the IP, netmask, default gateway.
| address 0.0.0.0
| # Bind one or more interfaces to the bridge.
| bridge_ports bond0
| # Tune the bridge for a single interface.
| bridge_stp off
| bridge_fd 0
| bridge_maxwait 0
+------------------------------+
| Cloud Controller |
| |
| eth0 eth1 |
| 192.168.1.60 172.16.1.10 |
+-------+------------+---------+
| |
+-------------+------------+-------------+ +---------------+
| Switch +---+ Workstation |
+--+----------+------------+----------+--+ | 192.168.1.101 |
| | | | +---------------+
| | | |
| | | |
+----+----------+------------+----------+----+
| eth0 eth1 eth2 eth3 |
| | | | | |
| | | | | |
| +--+----------+--+ +--+----------+--+ |
| | bond0 | | bond1 | |
| | (no ip) | | (172.168.1.20) | |
| +--------+-------+ +----------------+ |
| | |
| +--------+-----------+ |
| | | | +--------------+ |
| | bond0 vnet0--+--+ eth0 (vm0) | |
| | | | 192.168.1.20 | |
| | | +--------------+ |
| | | |
| | br0 | +--------------+ |
| | (no IP) vnet1--+--+ eth0 (vm1) | |
| | | | 192.168.1.21 | |
| | | +--------------+ |
| +--------------------+ |
| node01 |
+--------------------------------------------+
As far as I know the nodes don't need a special route for the VM Network.
And maybe the problem is from the bonding configuration. Or maby I need to
enable STP on the bridge ... but I don't unterstand the problem yet.
I could really need some help here!
Cheers, Markus
More information about the Users
mailing list