Hp helion openstack

Download 467.98 Kb.

Page	1/8
Date	28.01.2017
Size	467.98 Kb.
	#8870

1 2 3 4 5 6 7 8

1.OpenStack Neutron

HP Helion OpenStack 1.1

May 2014

Tomáš Kubica, Solution architect

Document version 0.10 EN

HP HELION OPENSTACK

LAB GUIDE
networking deep dive

Obsah

1.OpenStack Neutron 3

2.Packet walkthrough 3

2.1.Troubleshooting methodology 3

2.2.Collecting information 3

2.3.VM to VM in single network on single compute node 4

2.3.1.Traffic leaving VM 5

2.3.2.Applying Security Group 6

2.3.3.Traffic after Security Groups are applied 6

2.3.4.Entering vSwitch br-int 7

2.3.5.From br-int vSwitch to destination instance 8

2.4.VM to VM in single network on two compute nodes 9

2.4.1.Entering vSwitch br-int 9

2.4.2.Setting up tunnels and sending to underlay 9

2.4.3.OpenFlow rules in br-tun 10

2.4.4.Going out of compute node 11

2.4.5.Receiving traffic on second compute node 12

2.4.6.Entering vSwitch br-tun 12

2.4.7.OpenFlow rules in receiving br-tun 13

2.4.8.Receiving br-int vSwitch 14

2.4.9.Final step 15

2.5.North-South with Floating IP 15

2.5.1.Traffic leaving VM 16

2.5.2.Entering vSwitch br-int 16

2.5.3.OpenFlow rules in br-int 17

2.5.4.Router 18

2.5.5.Floating IP name space 21

2.5.6.Traffic leaving virtual world 21

2.6.East-West routing 22

2.6.1.Traffic leaving VM 23

2.6.2.Entering vSwitch br-int 23

2.6.3.Router 24

2.6.4.Back from router to br-int 26

2.6.5.Going out of compute node 26

2.6.6.Receiving on second compute node 28

2.6.7.OpenFlow rules in receiving br-tun vSwitch 28

2.6.8.OpenFlow rules in receiving br-int 29

2.6.9.Final step 30

2.6.10.And what about way back? 31

2.7.North-South routing with SNAT 31

2.7.1.Traffic leaving VM 32

2.7.2.Entering vSwitch br-int 33

2.7.3.Router 33

2.7.4.Back from router to br-int 35

2.7.5.Going out of compute node 35

2.7.6.Receiving in Network Node 37

2.7.7.OpenFlow rules in Network Node br-tun 38

2.7.8.OpenFlow rules Network Node br-int 38

2.7.9.SNAT namespace 39

2.7.10.Final step 41

2.8.DHCP traffic 42

2.8.1.Traffic leaving VM 43

2.8.2.Spoofing protection 44

2.8.3.Entering vSwitch br-int 44

2.8.4.Going out of Compute Node 45

2.8.5.Receiving in Network Node 46

2.8.6.OpenFlow rules in Network Node br-tun vSwitch 47

2.8.7.OpenFlow rules Network Node br-int vSwitch 48

2.8.8.DHCP server 49

3.Monitoring and troubleshooting 50

3.1.Mirroring VM traffic to centralized analyzer analyzátoru (RSPAN) 50

3.2.Flow monitoring 51

3.2.1.NetFlow 51

3.2.2.sFlow 55

1.OpenStack Neutron

…

2.Packet walkthrough

2.1.Troubleshooting methodology

…

2.2.Collecting information

For troubleshooting it is essential to start by collecting information regarding instances, compute nodes, MAC and IP addresses, various IDs such as tenant ID, instance ID etc.

Get your project tenant ID

keystone tenant-get mujprojekt

+-------------+----------------------------------+

| Property | Value |

+-------------+----------------------------------+

| description | |

| enabled | True |

| id | baa7096fe1d54571900c3758397e0939 |

| name | mujprojekt |

+-------------+----------------------------------+

On what physical compute nodes your instances are running and what is its local VM name (you need to run this OpenStack CLI command using admin account)?

nova list --all-tenants 1 --tenant baa7096fe1d54571900c3758397e0939 --fields name,OS-EXT-SRV-ATTR:host,OS-EXT-SRV-ATTR:instance_name

+--------------------------------------+------------+-------------------------------------+--------------------------------+

+--------------------------------------+------------+-------------------------------------+--------------------------------+

+--------------------------------------+------------+-------------------------------------+--------------------------------+

Let’s look inside of compute node – first we need to know its IP address

nova hypervisor-list

+----+-----------------------------------------------+

| ID | Hypervisor hostname |

+----+-----------------------------------------------+

| 1 | overcloud-novacompute0-vli5de2egecg.novalocal |

| 2 | overcloud-novacompute1-c4ia2jfbd75d.novalocal |

+----+-----------------------------------------------+

nova hypervisor-show overcloud-novacompute0-vli5de2egecg.novalocal | grep host_ip

| host_ip | 10.0.10.14

Log into compute node. You need to have valid certificate for that, typically you can do this from your Seed VM

root@hLinux:~# ssh heat-admin@10.0.10.14

Linux overcloud-novacompute0-vli5de2egecg 3.14.29-4-amd64-hlinux #hlinux1 SMP Mon Feb 9 20:32:22 UTC 2015 x86_64

The programs included with the hLinux system are free software; the exact

license terms for each program are described in the individual files in

/usr/share/doc/*/copyright.

Last login: Mon May 4 13:31:09 2015 from 10.0.10.2

$ sudo -i

root@overcloud-novacompute0-vli5de2egecg:~#

Let’s see running VMs

root@overcloud-novacompute0-vli5de2egecg:~# virsh list

Id Name State

----------------------------------------------------

5 instance-00000055 running

6 instance-00000056 running

74 instance-000000bd running

79 instance-000000c8 running

96 instance-000000e2 running

104 instance-000000eb running

105 instance-000000ed running

106 instance-000000ee running

107 instance-000000ef running

124 instance-00000108 running

125 instance-00000109 running

126 instance-0000010a running

127 instance-0000010b running

128 instance-0000010c running

129 instance-0000010d running

130 instance-0000010f running

131 instance-0000010e running

We are especially interested in our particular VM with local name instance-0000010d. Use dumpxml to get all information, but we are specifically looking for its virtual NIC (tap interface).

root@overcloud-novacompute0-vli5de2egecg:~# virsh dumpxml instance-0000010d | grep -A 7 "

Download 467.98 Kb.

Share with your friends:

1 2 3 4 5 6 7 8