Deploying OpenStack Error and Solutions

1. Nova image-show error $ yum downgrade PYTHON-URLLIB3 version 1.10 yum downgrade python-requests version 2.7
2, HTTP start error cp/usr/share/keystone/keystone.wsgi/var/www/cgi-bin/keystone/admin cp/usr/share/keystone/ Keystone.wsgi/var/www/cgi-bin/keystone/main httpd
3, neutron newspaper timeout/etc/neutron/neutron.conf there will be a configuration item "Rpc_response_timeout", which is used to configure the RPC time-out, the default is 60s, Therefore, the timeout exception is caused. Workaround for setting rpc_response_timeout=180 on controller
4, rabbitmq error 406 Delete openstack user, re-establish a rabbitmqctl delete_user openstack rabbitmqctl add_user OpenStack Rabbit_pass Rabbitmqctl set_permissions OpenStack ". *" ". *" ". *"
5. Remove the Zombie instance http://www.bubuko.com/infodetail-644952.html
6, Cinder create a cloud drive has been displayed in the creation, view debug to catch 8776 and then it won't move. Cinder profile is not associated to Ceph
7, dashboard can be normal landing, can do a series of operations, but in the deletion operation will be reported as follows error resolution: In/etc/openstack-dashboard/local_settings import OS after adding the relevant module three line import SYS Reload (SYS) sys.setdefaultencoding (' utf-8 ') then restart httpd and memcached
8, Dashbaord session expires can not login, error in/etc/openstack-dashboard/local.settings plus a line auth_user_model= ' Openstack_auth. User '
9, Fwaas installed after the database does not have a corresponding table, resulting in the inability to create a firewall su-s/bin/sh-c "Neutron-db-manage--service Fwaas--service_plugins Openstack-neutron-fwaas Upgrade Head "neutron
10. Image of Raw format modify instance Password network node: Yum install cloud-init compute node: vim/etc/nova/nova.conf plus force_config_drive=true/always reference file when creating host: #cloud-config
Ssh_pwauth:true
disable_root:0
User:root
Password:abc123
Chpasswd:expire:false
11, RAW format image address centos7:http://cloud.centos.org/centos/7/images/centos6:http://cloud.centos.org/centos/6/images/ Remove root ubuntu:http://cloud-images.ubuntu.com/daily/server conversion image format for users in clond.conf: http://docs.openstack.org/ Image-guide/convert-images.html
12, RABBITMQ error 2016-08-12 23:13:26.072 3683 error oslo_messaging._drivers.impl_rabbit [-] AMQP server on 172.16.0.1:5672 i S unreachable: [Errno 104] Connection reset by peer. Trying again in 1 seconds: Config file inside rabbit-hosts Remember to add "s"
13, neutron 9696 port and lb problem: lb needs to be tuned into 7 Layer mode, session remains consistent with OpenStack
14, HA environment to view the VNC interface failed to refresh a few to good, and then refresh and not good in all controllers on the/ETC/NOVA/NOVA.CONF default module add: Memcached_servers = 172.16.0.1:11211,172.16.0.2:11211,172.16.0.3:11211 If not good: check Netstat-lntap |grep memcache Ensure that you are listening to 0.0.0.0 : 11211, if the listener is 127.0.0.1, you need to modify:/etc/sysconfig/memcached options= "-l 127.0.0.1,::1" to options= ""
15. Modify the OpenStack default quota modify the quota option in the Nova.conf file on the controller network related options in the Controller neutron.conf
16, Cinder Storage node log error: (Operationalerror) (1040, ' Too many connections ') none none increase MySQL maximum number of connections
17, Ceilometer meter-list Error: The service catalog is empty modify environment variables: Vim ceilometer-openrc.sh unset os_project_domain_id unset OS_USER_DOMAIN_ID Export os_project_name=admin Export os_tenant_name=admin export os_username=admin export OS_PASSWORD =admin@1233 Export os_auth_url=http://172.16.0.100:35357/v2.0 export os_image_api_version=2 export OS_VOLUME_API_ version=2
18, RABBITMQ start error error:node with name "Rabbit" already running PS aux | grep Erl kill-9 The process
19. RABBITMQ Optimization command Rabbitmqctl stop rabbitmq-server-detached
20, create an instance always show Scheduling,nova-schedul log always dispatched: There is a compute node is broken, but dashboard does not close this bad compute node, first shut it down, and then restart the compute node on the Nova-compute service
21, Cinder-volumes cloud hard disk Dashboard and commands are deleted: Enter the database: >update volumes SET attach_status= ' detached ', status= "available" WHERE id= ' 647d95ff-851d-415c-aba6-9419e5541758 ';
22. View image error: The resource could not being found. (HTTP 404) Deferring Keystone Exception:the plugin Token_endpoint could not being found delete current environment variable: unset os_token os_url re-created:. Admin-openrc
23, through the Libvirt-set-admin-password way to modify the instance login password: Install related software (compute + control): Libvirt 1.2.16+ host Python-libvirt 1.2.16+ host Qemu-guest-agent 2.3+ Virtual machine internal HTTPS://LAUNCHPAD.NET/UBUNTU/+SOURCE/QEMU virtual machine image In addition to the installation of qemu-guest-agent, the image itself will be set hw_qemu_ Guest_agent Properties: JSON file Import metadata file under control node/etc/glance/metadefs directory: Compute-libvirt-image.json # glance-manage Db_load_ Metadefs Import all default metadata files Select Update metadata at the Mirror: libvirt Driver options for Images The value of hw_qemu_guest_agent is yes. In control node: # Nova List # Nova Root-password instance name New Password:again:
24. Use OvS security Group Policy does not take effect: in the compute node in the Openvswitch_agent.ini file of the [Securitygroup] module, add:: Enable_ipset = True Enable_security_group = True firewall_driver = Neutron.agent.linux.iptables_firewall. Ovshybridiptablesfirewalldriver Open iptables Service for all compute nodes
25, dashboard above adjust cloud host size does not take effect, compute node Error: The reason: OpenStack Cloud Host Configuration type modification, actually equivalent to do a cloud host in different host migration, so need to be in the relevant migration cloud host for no password access, Because OpenStack is managed by the Nova component to manage the cloud host, it requires no password access to the Nova user. Need to configure SSH password-free access for each node
26. Change the cloud host password Nova meta hostname set changepasswd=true admin_pass= password
27. Cloud Host Hot Migration configuration: All COMPUTE nodes: Vim/etc/nova/nova.conf added under [libvirt]: Live_migration_flag = vir_migrate_undefine_source,vir_ Migrate_peer2peer,vir_migrate_live,vir_migrate_persist_dest,vir_migrate_tunnelled vim/etc/libvirt/libvirtd.conf Added: Listen_tls = 0 listen_tcp = 1 auth_tcp = "none" Tcp_port = "16509" listen_addr = "Native IP" Vim/etc/sysconfig/libvirtd added : libvirtd_args= "--listen" # Service Openstack-nova-compute Restart # service LIBVIRTD Restart # Netstat-lntp|grep 16509 as And the port was successful.
28, configure the cloud host instance CPU model and physical machine consistent: Modify all COMPUTE nodes: # vim/etc/nova/nova.conf Add Cpu_mode=host-passthrough under [Libvirt] if you want to specify a model: Cpu_mode =custom cpu_model= Model (supported model list file:/usr/share/libvirt/cpu_map.xml) Note: The model is compatible with the physical machine, otherwise the cloud host system will not be