This warning is normal for compute nodes that are slow to boot an instance. As it states, it is "Retrying". I would wait a while and see if the compute service will finish booting the instance. You can also watch it progress by using the compute console. If you are running this compute node inside virtualbox or on some other nested virtualization (without vmx/smx support) scenario nova can take up to 18-20 minutes to fully boot an instance. If the instance does not finish booting and this times out, the load balancer will go into an ERROR state. If that is the case, we will have to debug nova and the image you are using. Michael On Tue, Dec 10, 2019 at 2:21 AM Ignazio Cassano <ignaziocassano@gmail.com> wrote:
Hello, I've just installed stein on centos 7 with octavia. I created the amphora image based on centos , but when a load balance is created the amphora instance does note have any proces on port 9443 and the worker reports:
2019-12-10 08:40:14.456 3710 WARNING octavia.amphorae.drivers.haproxy.rest_api_driver [-] Could not connect to instance. Retrying.: ConnectTimeout: HTTPSConnectionPool(host='10.102.143.94', port=9443): Max retries exceeded with url: // (Caused by ConnectTimeoutError(<urllib3.connection.VerifiedHTTPSConnection object at 0x7f4a04de0d10>, 'Connection to 10.102.143.94 timed out. (connect timeout=10.0)'))
Network connectivity and security groups are ok. I connected on the instance via ssh and I cannot see any listening 9443 port. Anyone can help me, please ? Ignazio