| opendevreview | Bodo Petermann proposed openstack/octavia master: Health sender: support to send via TCP https://review.opendev.org/c/openstack/octavia/+/852269 | 10:37 |
|---|---|---|
| opendevreview | Bodo Petermann proposed openstack/octavia master: Health sender: support to send via TCP https://review.opendev.org/c/openstack/octavia/+/852269 | 10:38 |
| wncslln | morning folks o/ | 14:01 |
| wncslln | im configuring the tenant network of octavia | 14:02 |
| wncslln | i ran reconfigure to it and o-hm0 interface was created, but not get ip address | 14:02 |
| wncslln | and the interface state is UNKNOWN | 14:03 |
| wncslln | there is a known issue? | 14:03 |
| tweining | Hi wncslln . What do you mean with reconfigure? Are you using a specific deployment tool to install and configure octavia? | 14:07 |
| wncslln | tweining: im using kolla-ansible | 14:07 |
| tweining | unfortuntly, I have no experience with kolla-ansible. johnsom might be able to help you, but you'll have to wait for a couple of hour because he is in US west coast time zone. | 14:10 |
| wncslln | is all good to me. during this time, i will make some tests in my environment | 14:11 |
| wncslln | thanks tweining! | 14:11 |
| tweining | np. I'd say it is probably rather a configuration issue than an issue with octavia directly | 14:11 |
| wncslln | yeah. i noticed that the created ports in openstack has different mac address than node interface | 14:12 |
| wncslln | i tried to delete this interface to recreate it, but i could not | 14:13 |
| wncslln | tweining: i can configure the network and interface gets an ip, but when i create a LB, the amphora stuck in BOOTING state | 14:59 |
| wncslln | log show the following message: | 14:59 |
| wncslln | 2024-02-14 14:58:29.683 20 WARNING octavia.amphorae.drivers.haproxy.rest_api_driver [-] Could not connect to instance. Retrying.: requests.exceptions.ConnectionError: HTTPSConnectionPool(host='10.5.2.162', port=9443): Max retries exceeded with url: // (Caused by NewConnectionError('<urllib3.connection.HTTPSConnection object at 0x7f7542483f70>: Failed to establish a new connection: [Errno 113] No route to host')) | 14:59 |
| tweining | okay, that means you cannot ping that ip either? | 15:01 |
| johnsom | That means we are waiting on nova to finish booting the VM instance. The LB should go ACTIVE in 30-45 seconds | 15:02 |
| johnsom | You will get a few of those retrying messages while we wait on nova | 15:03 |
| johnsom | Though if I remember the boot sequence right, if the vm comes up and you still see these, there is a networking problem with the lb-mgmt-net | 15:05 |
| wncslln | the instance status is ACTIVE and the amphora still BOOTING | 15:07 |
| wncslln | i should be able to ping the instance from anywhere? | 15:07 |
| wncslln | johnsom: i just configure a tenant network with autoconfigure on. what kind of things in mgmt-network could be causing this? | 15:10 |
| tweining | ping should work from the namespace of o-hm0 I think | 15:10 |
| tweining | I haven't had to debug these kind of issues for a long time | 15:11 |
| wncslln | yeah... i cant ping from node with o-hm0 interface | 15:12 |
| tweining | then it's obviously a problem with the network config of o-hm0. maybe the default route is missing? | 15:15 |
| wncslln | maybe a firewall problem, i will investigate more | 15:17 |
| wncslln | thanks for the help tweining | 15:17 |
| johnsom | You cannot ping an amphora. Next step is look in the amphora console log in nova, did cloud-init get an IP address from neutron? | 15:17 |
| johnsom | Yes, check the security groups kolla created | 15:18 |
| wncslln | johnsom: yeah, I already check this. cloud init shows a table with ip | 15:18 |
| wncslln | gets ans IP from subnetwork created in management-net | 15:18 |
| johnsom | Ok, good, so you can test with ssh or curl | 15:18 |
| johnsom | Next on my list would be checking the security group got setup | 15:20 |
| wncslln | for some how, the instance and the amphora was deleted o_o | 15:22 |
| opendevreview | Bodo Petermann proposed openstack/octavia master: Health sender: support to send via TCP https://review.opendev.org/c/openstack/octavia/+/852269 | 15:26 |
| wncslln | johnsom: i think is all good with secgroups. there is a rule to SSH, ICMP and 9443 port | 15:27 |
| johnsom | wncslln Yes, the controller will automatically clean up the resources in use if the provisioning was not successful. It does this after the retry timeout settings in the octavia.conf. That way you don't end up with a bunch of non-functional provisioned VMs in your cloud. | 15:52 |
| wncslln | johnsom: in fact, there was an interface with same IP from management-network, i think octavia tries to connect with amphora through this interface | 15:56 |
| wncslln | i just set to down and works perfectly! | 15:56 |
| wncslln | loadbalancer was created | 15:57 |
| wncslln | thanks al ot johnsom, tweining | 15:57 |
| johnsom | Ah, so someone neutron issued an IP twice? maybe the DHCP range on the subnet is off | 15:57 |
| tweining | I'm glad it works now | 15:57 |
| johnsom | Awesome, glad you are up and running. | 15:57 |
| opendevreview | Tom Weininger proposed openstack/octavia master: Integrate mypy type checker https://review.opendev.org/c/openstack/octavia/+/879749 | 16:43 |
| opendevreview | Elod Illes proposed openstack/python-octaviaclient master: DNM: gate health test https://review.opendev.org/c/openstack/python-octaviaclient/+/909017 | 16:58 |
| *** jph4 is now known as jph | 18:32 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!