Problems and Solutions for openstack icehouse 3 node deployment

Shortly after I first came into contact with openstack, I encountered some problems when implementing 3-node deployment by referring to the official documentation, mainly focusing on the compute node. Fortunately, I had more than a decade of O & M experience to help me solve the problem one by one. The solution is marked in red.

System Environment: centos 6.5 64-bit

IP addresses of each node: configured according to the IP addresses in the official documentation.

Official documents: http://docs.openstack.org/icehouse/install-guide/install/zypper/content/

Log Record Date:

Problem section:

Controller node

Install the dashboard

Updatethe allowed_hosts in local_settings.py to include theaddresses you wish to access the dashboard from.

Edit/etc/openstack-dashboard/local_settings:

ALLOWED_HOSTS= [‘localhost‘, ‘my-desktop‘]

The actual configuration is as follows:

ALLOWED_HOSTS = [‘10.0.0.11‘, ‘0.0.0.0‘]

 

This configuration is estimated to be a list of Python, that is, the hosts that allows access to the dashboard,

According to the configuration in the official document, the dashboard page reports an error and cannot be enabled. As shown in:

650) This. width = 650; "src =" http://s3.51cto.com/wyfs02/M02/39/52/wKiom1O4pRLQTYwsAAF9xiLiSTA828.jpg "Title =" qq 40706091743.jpg "alt =" wkiom1o4prlqtywsaaf9425ista828.jpg "/>

Besides this issue, check Apache logs/var/log/httpd/error_log.

650) This. width = 650; "src =" http://s3.51cto.com/wyfs02/M01/39/52/wKiom1O4pabAbXvkAAMOBEtXqXs112.jpg "Title =" qq 40706092024.jpg "alt =" wkiom1o4pababxvkaamobetxqxs112.jpg "/>

Compute Node

Question 1:

Openstack networking (neutron)

Configure compute Node

To install the networkingcomponents

yum install openstack-neutron-ml2 openstack-neutron-openvswitch

It is still missing:

Openstack-Nova-compute

The complete content is:

yum install openstack-neutron-ml2openstack-neutron-openvswitch openstack-nova-compute -y

Question 2:

To configurecompute to use networking

By default, mostdistributions configure compute to use legacy networking. You must reconfigurecompute to manage networks through networking.

• Run the following commands:

ReplaceNeutron_passWith the password you chose for the neutron user in the identity service.

# openstack-config --set/etc/nova/nova.conf DEFAULT  network_api_class nova.network.neutronv2.api.API# openstack-config --set/etc/nova/nova.conf DEFAULT  neutron_url http://controller:9696# openstack-config --set /etc/nova/nova.confDEFAULT  neutron_auth_strategy keystone# openstack-config --set/etc/nova/nova.conf DEFAULT  neutron_admin_tenant_name service# openstack-config --set/etc/nova/nova.conf DEFAULT  neutron_admin_username neutron# openstack-config --set/etc/nova/nova.conf DEFAULT  neutron_admin_password NEUTRON_PASS# openstack-config --set/etc/nova/nova.conf DEFAULT  neutron_admin_auth_url http://controller:35357/v2.0# openstack-config --set/etc/nova/nova.conf DEFAULT  linuxnet_interface_driver nova.network.linux_net.LinuxOVSInterfaceDriver# openstack-config --set/etc/nova/nova.conf DEFAULT  firewall_driver nova.virt.firewall.NoopFirewallDriver# openstack-config --set/etc/nova/nova.conf DEFAULT  security_group_api neutron

It is still missing:

openstack-config --set /etc/nova/nova.conf DEFAULT qpid_hostname controlleropenstack-config --set /etc/nova/nova.conf DEFAULT rpc_backend qpidopenstack-config --set /etc/nova/nova.conf DEFAULT glance_host controlleropenstack-config --set /etc/nova/nova.conf DEFAULT auth_strategy keystoneopenstack-config --set /etc/nova/nova.conf DEFAULT novncproxy_base_url http://10.0.0.11:6080/vnc_auto.htmlopenstack-config --set /etc/nova/nova.conf DEFAULT vncserver_proxyclient_address 10.0.0.31openstack-config --set /etc/nova/nova.conf DEFAULT vncserver_listen 0.0.0.0 chkconfig openstack-nova-compute on

Cause analysis:

Check the compute log/var/log/NOVA/compute. log to find the following problems.

1. the compute node cannot communicate with the controller node, as shown in

650) This. width = 650; "src =" http://s3.51cto.com/wyfs02/M02/39/52/wKiom1O4s0DR5CLDAAP5R7MLZGY847.jpg "Title =" qq 40706101408.jpg "alt =" wkiom1o4s0dr5cldaap5r7mlzgy847.jpg "/>

Check the/etc/NOVA. conf configuration file and confirm that the default configuration is rpc_backend = rabbit, qpid_hostname = localhost.

Rpc_backend = qpid

Qpid_hostname = Controller

2. the compute node cannot get the imported image from the controller node, as shown in

Controller node log/var/log/glance/API. Log

650) This. width = 650; "src =" http://s3.51cto.com/wyfs02/M01/39/53/wKiom1O4t6PBpELSAAQTMRxiQ3g742.jpg "Title =" qq 40706103653.jpg "alt =" wkiom1o4t6pbpelsaaqtmrxiq1_742.jpg "/>

Compute node log/var/log/NOVA/compute. Log

650) This. width = 650; "src =" http://s3.51cto.com/wyfs02/M02/39/53/wKioL1O4uBPjm0tFAABf-N_e5GM177.jpg "Title =" qq 40706101167.jpg "alt =" wKioL1O4uBPjm0tFAABf-N_e5GM177.jpg "/>

View the/etc/NOVA. conf configuration file and confirm that the default configuration is glance_host = $ my_ip and my_ip = 10.0.0.1.

Glance_host = Controller

After the modification, the image still cannot be obtained. Analyze/var/log/glance/API. log again and confirm that the token information is not included in the get operation. View the/etc/NOVA. conf configuration file and confirm that the default configuration is auth_strategy = noauth.

Auth_strategy = keystone

3. You cannot open the VM console on the dashboard on the controller node. No error information is reported when you view logs. The front-end reports the following error:

650) This. width = 650; "src =" http://s3.51cto.com/wyfs02/M01/39/53/wKiom1O4vCeCNuyPAAFBNkf-rdk401.jpg "Title =" qq 40706105547.jpg "alt =" wKiom1O4vCeCNuyPAAFBNkf-rdk401.jpg "/>

Solved this problem after searching for the official help Area

Https://ask.openstack.org/en/question/520/vnc-console-in-dashboard-fails-to-connect-ot-server-code-1006/

Edit the/etc/NOVA. conf configuration file of the compute Node

Novncproxy_base_url = http: // 10.0.0.11: 6080/vnc_auto.html

Vncserver_proxyclient_address = 10.0.0.31

Vncserver_listen = 0.0.0.0

After the configuration is modified, restart the openstack-Nova-compute service and run the netstat command to check the port 5900 status. This is the port on which the console needs to access compute.
650) This. width = 650; "src =" http://s3.51cto.com/wyfs02/M00/39/53/wKiom1O4wqLwgqMjAAVJajDZD8w286.jpg "Title =" qq 407061120.0.jpg "alt =" wkiom1o4wqlwgqmjaavjdzd8w286.jpg "/>

Question 3:

Run the openstack-Nova-compute startup command. The front-end does not have any error information and the service can be started normally.

/etc/init.d/openstack-nova-compute start

Check the openstack-Nova-compute status later and you will find that the prompt message is that the process does not exist. However, if the PID file exists, you must delete the PID file before starting the service again.

/etc/init.d/openstack-nova-compute statusrm -f /var/run/nova/nova-compute.pid

Solution:

This problem is not solved by checking compute logs/var/log/NOVA/compute. log.

650) This. width = 650; "src =" http://s3.51cto.com/wyfs02/M01/39/52/wKioL1O4rvfQYVirAARrwWnwvRE956.jpg "Title =" qq 40706100059.jpg "alt =" wkiol1o4rvfqyviraarrwwnwvre956.jpg "/>

The libmongod settings are not described in the official documentation. In actual use, you need to configure libmongod as follows.

· Edit the cgroup_device_acl array in the/etc/libvirt/qemu. conf file:

cgroup_device_acl = ["/dev/null", "/dev/full","/dev/zero","/dev/random", "/dev/urandom","/dev/ptmx", "/dev/kvm","/dev/kqemu","/dev/rtc","/dev/hpet","/dev/net/tun"]

· Enable Live migration by updating/etc/libvirt/lib1_d. conf file:

listen_tls = 0listen_tcp = 1auth_tcp = "none"

Before starting openstack-Nova-compute, you must start the libmongod service and set libmongod to start the service.

/etc/init.d/libvirtd startchkconfig libvirtd on

This article from "lan2003" blog, please be sure to keep this source http://lan2003.blog.51cto.com/252098/1434981