Saturday, March 25, 2017

Excon error in fog.log ManageIQ Euwe and Openstack Newton Provider Error in refreshing the OpenStack Provider status in ManageIQ Providers Error::Socket: getaddrinfo: Name or service not known (SocketError)


[----] E, [2017-03-25T06:38:18.640082 #3557:7bd130] ERROR -- : excon.error     #<Excon::Error::Socket: getaddrinfo: Name or service not known (SocketError)>This also shows up in the OpenStack Provider in the WebUI of the ManageIQ as Socket: getaddrinfo: Name or service not known (SocketError)>


I was using the Newton Installation on the OpenStack and had to use the IP addresses of the controller and the compute nodes for openstack.

Deployment of OpenStack Newton was using the packstack and had to use IP addresses for the controller nodes and the compute nodes owing to a bug in PackStack for Newton. 

The below link can be informative for the packstack issue in Newton OpenStack




But with that as I had the ManageIQ appliance of VMware and had set it to an IP address in the same subnet as the OpenStack controller, I was successfully able to add the OpenStack provider under cloud providers in the ManageIQ.

But post that each attempt of having the synchronizing the current status of the provider threw me the error "getaddrinfo: Name or service not known (SocketError)" in the ManageIQ UI as well as in the appliance's /var/log/manageiq/fog.log

The WorkAround:

The workaround was to have an entry for the OpenStack controller IP address and FQDN entry in the  /etc/hosts file of the ManageIQ appliance.

Please note: I was not using any DNS service at the time, and for host name resolution by the ManageIQ appliance I had to rely on /etc/hosts on the ManageIQ server. Not sure if there had been a DNS server to resolve the IP address of the controller to the ManageIQ host this problem would have arose or not.



OpenStack Newton Pakcstack bug on Neutron Manifest run Wrong number of arguments given (1 for 2) in openstacksetup.log

The error appearing in the openstack-setup.log

PuppetError: Error appeared during Puppet run: controller1.example.com_network.pp
Error: member(): Wrong number of arguments given (1 for 2) at /var/tmp/packstack/6b760137219c43bc8e7f4365e9778689/modules/packstack/manifests/neutron/ovs_agent.pp:29 on node controller1.example.com
You will find full trace in log /var/tmp/packstack/20170325-051853-oQnZA6/manifests/controller1.example.com_network.pp.log

2017-03-25 05:21:52::INFO::shell::94::root:: [compute1.example.com] Executing script:
rm -rf /var/tmp/packstack/7ee6b5f014df46f982a179d372186bc8
2017-03-25 05:21:52::INFO::shell::94::root:: [controller1.example.com] Executing script:
rm -rf /var/tmp/packstack/6b760137219c43bc8e7f4365e9778689


The workaround is to replace all the HOST parameters in the packstack answers file with the IP addresses instead of the hostnames

as below

[root@controller1 ~]# cat answers1.txt | grep 192.168.17
CONFIG_CONTROLLER_HOST=192.168.17.52
CONFIG_COMPUTE_HOSTS=192.168.17.61
CONFIG_NETWORK_HOSTS=192.168.17.52
CONFIG_SSL_CERT_SUBJECT_CN=192.168.17.52
CONFIG_SSL_CERT_SUBJECT_MAIL=admin@192.168.17.52
CONFIG_AMQP_HOST=192.168.17.52
CONFIG_MARIADB_HOST=192.168.17.52
CONFIG_MONGODB_HOST=192.168.17.52
CONFIG_REDIS_HOST=192.168.17.52
[root@controller1 ~]#



Relevant Redhat Bugzilla


https://bugs.launchpad.net/packstack/+bug/1635133