| opendevreview | Tianhao Chen proposed openstack/octavia master: Fix IPv6 session persistence failed https://review.opendev.org/c/openstack/octavia/+/888530 | 02:36 |
|---|---|---|
| opendevreview | Tianhao Chen proposed openstack/octavia master: Fix IPv6 session persistence failed https://review.opendev.org/c/openstack/octavia/+/888530 | 03:07 |
| opendevreview | Tianhao Chen proposed openstack/octavia master: Fix IPv6 session persistence failed https://review.opendev.org/c/openstack/octavia/+/888530 | 06:31 |
| opendevreview | Gregory Thiemonge proposed openstack/octavia stable/2023.1: Fix amphorav1 member deletion bug https://review.opendev.org/c/openstack/octavia/+/888669 | 09:26 |
| opendevreview | Merged openstack/octavia-tempest-plugin master: Remove xena and wallaby jobs from the gates https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/888134 | 11:00 |
| opendevreview | Gregory Thiemonge proposed openstack/octavia stable/2023.1: Fix amphorav1 member deletion bug https://review.opendev.org/c/openstack/octavia/+/888669 | 12:53 |
| danfai | Hi, I got a question regarding the amphora reboot capability. In our setup the automatic assignment of the interface to the amphora netns does not work, due to /sbin/ip link having the altname as a 3rd line. With this the awk to concatinate the lines for the interface and thus the detection of the amphora-netns.service does not work. Do I miss a kernel parameter for removing | 14:14 |
| danfai | the altname or is this maybe unsupported, because the output of ip changed? | 14:14 |
| danfai | (I am building a diskimage based on AlmaLinux/Rhel, the ip link output is used in amphora-netns.systemd.j2 ) | 14:17 |
| gthiemonge | danfai: hi, you're talking about https://opendev.org/openstack/octavia/src/branch/master/octavia/amphorae/backends/agent/api_server/templates/amphora-netns.systemd.j2#L29 right? | 14:23 |
| gthiemonge | I haven't tested amphora reboot for a long time, rebooting doesn't work with TLS terminated listeners, because the certificates/keys are stored in a tmpfs, so they are not preserved across reboots | 14:24 |
| gthiemonge | maybe we should get rid of this line, it's really ugly | 14:28 |
| danfai | gthiemonge: yes, this is the line I was talking about. | 14:54 |
| danfai | I am currently fixing it by adding a `grep -e mtu -e link`, but I guess a smarter solution, maybe based on pyroute might make sense to add the interfaces to the other netns | 14:55 |
| gthiemonge | danfai: can you report it on https://bugs.launchpad.net/octavia ? | 15:04 |
| danfai | will do | 15:06 |
| gthiemonge | ok, I'm reproducing it here, after rebooting, the vip/member interfaces are in the "default" namespace | 15:07 |
| opendevreview | Elod Illes proposed openstack/octavia stable/victoria: [stable-only] Cap virtualenv/setuptools https://review.opendev.org/c/openstack/octavia/+/888707 | 15:09 |
| danfai | are there plans to support rebooting for terminated TLS as well? | 15:10 |
| johnsom | Not at the moment. The keys are stored in an encrypted ramfs volume in the amphora, on reboot that volume is lost. | 15:35 |
| opendevreview | Tom Weininger proposed openstack/octavia master: Add support for HTTP Strict Transport Security https://review.opendev.org/c/openstack/octavia/+/880806 | 16:16 |
| opendevreview | Tom Weininger proposed openstack/octavia master: Add support for HTTP Strict Transport Security https://review.opendev.org/c/openstack/octavia/+/880806 | 16:18 |
| opendevreview | Tom Weininger proposed openstack/octavia master: Add support for HTTP Strict Transport Security https://review.opendev.org/c/openstack/octavia/+/880806 | 16:22 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!