| *** prithiv has joined #openstack-astara | 00:08 | |
| *** prithiv has quit IRC | 01:31 | |
| *** shashank_hegde has quit IRC | 01:41 | |
| *** leonstack1 has joined #openstack-astara | 03:32 | |
| *** leonstack has quit IRC | 03:33 | |
| stupidnic | Okay. I am still having some sort of DHCP issue | 03:45 |
|---|---|---|
| stupidnic | There is some sort of issue where the appliance's dnsmasq.d/ interface entry isn't being updated with a newly created instance's MAC address, and so it doesn't respond to DHCP requests | 03:57 |
| stupidnic | If I restart orchestrator the entries appear | 03:57 |
| stupidnic | Then I run into my next problem which is the metadata proxy is returning 500 | 03:58 |
| elo | about the earlier issue. There is a patch for the DHCP issue | 04:06 |
| elo | https://review.openstack.org/#/c/266586/ | 04:07 |
| elo | anything in logs for the astra-metadata-proxy service? | 04:11 |
| stupidnic | elo: I wasn't expecting you here | 04:13 |
| stupidnic | elo: thanks for peeking in | 04:13 |
| stupidnic | elo: the DHCP issue I am seeing might be related to something else... Here is what I have been able to determine so far | 04:14 |
| stupidnic | basically if I create a brand new router for a tenant the instance boots, but then I see this in the astara orchestrator logs: | 04:14 |
| stupidnic | Instance has booted, attempting initial config | 04:15 |
| stupidnic | I worked around the DHCP traffic issues with security group rules... not ideal but whatever | 04:15 |
| stupidnic | I will check into that patch though | 04:16 |
| elo | the patch I mentioned should resolve this issue. | 04:16 |
| stupidnic | which? | 04:17 |
| elo | the security group adding | 04:17 |
| stupidnic | Okay cool | 04:17 |
| stupidnic | The issue I am currently seeing is that when a router is newly created... orchestrator can't talk to it | 04:17 |
| stupidnic | it tries to send the config, but I think it has the wrong IP address | 04:18 |
| stupidnic | it eventually recycles the router and when the router is built a second time, it can talk to it no problem and updates the config which then updates the dnsmasq.d/eth2 file with the DHCP information for the instance | 04:19 |
| stupidnic | (eth2 being the internal network interface on the router) | 04:19 |
| stupidnic | I can reliably reproduce this behavior by deleting the router out of the project and creating a new one | 04:20 |
| stupidnic | first boot... astara can't update... terminate it... second time around astara has no issues and the config is updated | 04:20 |
| elo | can you open up a bug on this since you can reproduce it reliably. I'll see if I can replicate. Is this main or liberty branch? | 04:21 |
| stupidnic | stable/liberty | 04:21 |
| stupidnic | where do I open a bug? I know you have told me this before | 04:22 |
| elo | https://bugs.launchpad.net/astara | 04:23 |
| elo | this isn interesting issue that I'm curious. | 04:25 |
| stupidnic | Let me nail down the particulars to make sure I have it all down and I will create the bug shortly there after | 04:26 |
| elo | eth0 should be the mgmt interface for the router, eth1 would be the external network, and eth2 is the tenant network. | 04:26 |
| elo | any additional networks create after the router is created by the tenant will be hot plugged into the instance as ethX | 04:27 |
| stupidnic | Yep. It's only the initial creation of the router that has this issue. | 04:29 |
| stupidnic | I'll open up the bug now | 04:29 |
| elo | I'll be on/off all weekend... and need to build some additional environments for testing | 04:37 |
| elo | I'll see if I can replicate and figure out the issue | 04:38 |
| stupidnic | No problem. Thanks for being around. | 04:40 |
| stupidnic | elo: https://bugs.launchpad.net/astara/+bug/1535032 | 05:00 |
| openstack | Launchpad bug 1535032 in Astara "Initial router configuration fails" [Undecided,New] | 05:00 |
| *** stanchan has joined #openstack-astara | 05:53 | |
| *** stanchan has quit IRC | 08:42 | |
| *** openstackgerrit has quit IRC | 09:20 | |
| *** openstackgerrit has joined #openstack-astara | 09:21 | |
| *** leonstack has joined #openstack-astara | 19:41 | |
| *** leonstack1 has quit IRC | 19:42 | |
| stupidnic | So after digging at the metadata problem (the metadata_proxy really lacks proper logging by default), I have determined that the metadata proxy is trying to contact the rug server at some IPv6 address. | 20:08 |
| stupidnic | It looks like this address is actually hard coded in the application | 20:08 |
| stupidnic | The address isn't the address of the RUG server... I am not sure why this is happening | 20:15 |
| elo | there were some hardwood IP address that were fixed recently. Let me see if there is a bug for this | 20:33 |
| elo | this was for the metadata proxy, correct? | 20:34 |
| stupidnic | elo: yes | 20:35 |
| stupidnic | I am looking at the mainline repo and I don't see where it is corrected | 20:35 |
| elo | brb. biobreak. | 20:36 |
| stupidnic | This could be an issue with how we are configuring the management network I guess. | 20:41 |
| stupidnic | I am using the same subnet (fdca:3ba5:a17a:acda::/64) | 20:42 |
| *** davidlenwell has quit IRC | 22:25 | |
| *** justinlund1 has joined #openstack-astara | 22:47 | |
| *** justinlund has quit IRC | 22:48 | |
| *** cleverdevil has quit IRC | 22:53 | |
| *** owlbot has quit IRC | 22:53 | |
| *** davidlenwell has joined #openstack-astara | 22:57 | |
| *** cleverdevil has joined #openstack-astara | 22:57 | |
| *** owlbot has joined #openstack-astara | 22:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!