damiandabrowski[m] | hello, I'm using ubuntu focal(keepalived 2.0.19) and victoria(neutron 17.2.1) | 07:20 |
---|---|---|
damiandabrowski[m] | when router's l3 agent switches to master, keepalived should send gratuitous ARP after 1min due to `garp_master_delay 60` but for some reason it does it immediately(when interface is down) | 07:20 |
damiandabrowski[m] | Have You ever seen this behavior? | 07:20 |
slaweq | damiandabrowski: I'm not sure now but I think it sends garps immediately and repeat after some time, no? | 07:40 |
lajoskatona | slaweq, damiandabrowski[m]: we set it really to 60: https://opendev.org/openstack/neutron/src/branch/master/neutron/agent/linux/keepalived.py#L40 | 07:45 |
damiandabrowski[m] | slaweq: if I understand it correctly, keepalived does not mention anything about sending GARPs twice(immediately and after specified delay): https://github.com/acassen/keepalived/blob/master/ChangeLog#L8334 | 07:51 |
damiandabrowski[m] | but it's not that important in my case because I don't see any ARPs on tcpdump after 60s | 07:51 |
damiandabrowski[m] | (i started looking into this because my router stores outdated ARPs for floating IPs in its cache) | 07:51 |
slaweq | hmm, I don't remember it exactly as I've been playing with it pretty long time ago | 07:52 |
slaweq | maybe something changed in the keepalived, idk | 07:52 |
opendevreview | Slawek Kaplonski proposed openstack/neutron stable/rocky: Don't setup bridge controller if it is already set https://review.opendev.org/c/openstack/neutron/+/816471 | 07:59 |
opendevreview | Slawek Kaplonski proposed openstack/neutron stable/queens: Don't setup bridge controller if it is already set https://review.opendev.org/c/openstack/neutron/+/816472 | 07:59 |
damiandabrowski[m] | thanks, i'll have a deeper look at this next week, eventually i may end up with filling keepalived's bug report | 07:59 |
lajoskatona | damiandabrowski[m]: by github it seems that it sends 1 garp immediately: https://github.com/acassen/keepalived/blob/v2.0.19/keepalived/vrrp/vrrp.c#L1649 | 08:05 |
damiandabrowski[m] | ouh, thanks for clarification. So I have 2 problems to solve:... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/zBrpXbUgSuqxgIbmxTejYuwB) | 08:15 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN] Execute OVN migration transactions independently https://review.opendev.org/c/openstack/neutron/+/814181 | 08:17 |
lajoskatona | ralonsoh, slaweq, liuyulong: regarding https://review.opendev.org/c/openstack/neutron/+/797120 and https://review.opendev.org/c/openstack/neutron-lib/+/807224 is it ok have it topic for next weeks drivers meeting? | 08:20 |
ralonsoh | lajoskatona, we should yes | 08:20 |
ralonsoh | if not, we are going to be reviewing in circles on those two patches | 08:20 |
ralonsoh | damiandabrowski[m], lajoskatona https://keepalived-devel.narkive.com/axSrQ5Zd/garp-master-delay-could-someone-clarify | 08:21 |
ralonsoh | from this link | 08:21 |
ralonsoh | "When host changes to MASTER role, he garps immediately, | 08:21 |
ralonsoh | and after garp_master_delay seconds garps again. | 08:21 |
ralonsoh | " | 08:21 |
frickler | slaweq: lajoskatona: ralonsoh: seems the rally jobs are failing when devstack stopped creating the keystone admin endpoint, does one of you have time to look into that? | 08:22 |
ralonsoh | frickler, I'll check it | 08:23 |
frickler | ralonsoh: thx, https://review.opendev.org/c/openstack/devstack/+/777345 is the devstack change that pulled it off | 08:23 |
frickler | ralonsoh: if you can't solve, you can set that flag for the affected job, similar to what I did in the dependency for openstacksdk | 08:24 |
ralonsoh | frickler, I think that will be the easiest thing | 08:25 |
ralonsoh | I think I'll do it for rally | 08:25 |
lajoskatona | ralonsoh, damiandabrowski[m]: good this means at least I can still read some C/C++ code :-) | 08:30 |
ralonsoh | hehehe | 08:30 |
lajoskatona | damiandabrowski[m]: +1 perhaps just make it documented how it works and why :-) | 08:31 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: Create Keystone admin endpoint for Rally job https://review.opendev.org/c/openstack/neutron/+/816792 | 08:31 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN][Placement] Add support for minimum bandwidth QoS rules https://review.opendev.org/c/openstack/neutron/+/786441 | 08:34 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN][Placement] Drive binding by placement allocation https://review.opendev.org/c/openstack/neutron/+/786478 | 08:34 |
opendevreview | Lajos Katona proposed openstack/neutron master: WIP: remove admin role reference from rally tasks https://review.opendev.org/c/openstack/neutron/+/816794 | 08:53 |
lajoskatona | ralonsoh, frickler: an alternative aproach ^ , if I understand well it should work as well..... | 08:54 |
ralonsoh | lajoskatona, if this patch works, maybe we could remove this role then | 08:55 |
ralonsoh | I'll wait for the CI | 08:55 |
ralonsoh | if passes, I'll +2 your patch | 08:55 |
damiandabrowski[m] | lajoskatona: thanks for Your input! | 09:00 |
lajoskatona | damiandabrowski[m]: yw | 09:07 |
lajoskatona | ralonsoh: not sure, but we will see :-) | 09:07 |
opendevreview | liuyulong proposed openstack/neutron master: Add table for pps limitaion https://review.opendev.org/c/openstack/neutron/+/804213 | 09:33 |
opendevreview | liuyulong proposed openstack/neutron master: Meter flows and ovsdb action for ovs bridge https://review.opendev.org/c/openstack/neutron/+/816800 | 09:33 |
opendevreview | liuyulong proposed openstack/neutron master: Add QoS pps limit rule to support list https://review.opendev.org/c/openstack/neutron/+/816801 | 09:33 |
opendevreview | liuyulong proposed openstack/neutron master: Support pps limitation for openvswitch agent https://review.opendev.org/c/openstack/neutron/+/816802 | 09:33 |
lajoskatona | ralonsoh, frickler: removing "roles:.... " is not enough, but as I see it's failing for others as well, see: https://zuul.openstack.org/builds?job_name=cinder-rally-task | 10:07 |
ralonsoh | so maybe just enabling the flag again | 10:08 |
ralonsoh | that's the easiest way | 10:08 |
ralonsoh | it is passing in https://zuul.opendev.org/t/openstack/status#816792 | 10:08 |
ralonsoh | I think this is because of NovaServers | 10:11 |
ralonsoh | that tries to retrieve the users | 10:11 |
ralonsoh | that needs admin role | 10:11 |
frickler | lajoskatona: ralonsoh: ok, so 816792 should be the fix for now. I'm testing rally jobs directly now with https://review.opendev.org/c/openstack/rally-openstack/+/816814 | 10:14 |
lajoskatona | ralonsoh, frickler: thanks | 10:30 |
opendevreview | Slawek Kaplonski proposed openstack/neutron master: Add functional and fullstack jobs with FIPS enabled https://review.opendev.org/c/openstack/neutron/+/814009 | 10:55 |
opendevreview | Slawek Kaplonski proposed openstack/neutron master: Add functional and fullstack jobs with FIPS enabled https://review.opendev.org/c/openstack/neutron/+/814009 | 11:05 |
opendevreview | Slawek Kaplonski proposed openstack/ovn-octavia-provider master: Use f-strings instead of regular strings to make pep8 happy https://review.opendev.org/c/openstack/ovn-octavia-provider/+/816829 | 12:10 |
opendevreview | Slawek Kaplonski proposed openstack/ovn-octavia-provider master: Fix lower-constanints and pep8 jobs https://review.opendev.org/c/openstack/ovn-octavia-provider/+/816829 | 12:27 |
opendevreview | Lajos Katona proposed openstack/neutron master: WIP: Change default domain to Default for rally https://review.opendev.org/c/openstack/neutron/+/816794 | 12:31 |
*** jlibosva is now known as Guest5001 | 12:38 | |
opendevreview | Przemyslaw Szczerbik proposed openstack/neutron-tempest-plugin master: QoS min pps API tests https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/806689 | 12:46 |
opendevreview | Przemyslaw Szczerbik proposed openstack/neutron-tempest-plugin master: QoS min pps API tests https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/806689 | 12:51 |
opendevreview | Merged openstack/neutron stable/xena: Check a namespace existence by checking only its own directory https://review.opendev.org/c/openstack/neutron/+/816512 | 13:16 |
opendevreview | Slawek Kaplonski proposed openstack/ovn-octavia-provider master: Fix lower-constanints and pep8 jobs https://review.opendev.org/c/openstack/ovn-octavia-provider/+/816829 | 13:44 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: Create Keystone admin endpoint for Rally job https://review.opendev.org/c/openstack/neutron/+/816792 | 13:56 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [WIP] Improve DHCP RPC handler https://review.opendev.org/c/openstack/neutron/+/816850 | 14:03 |
opendevreview | Lajos Katona proposed openstack/neutron master: Document the effects of admin_state_up https://review.opendev.org/c/openstack/neutron/+/816725 | 14:52 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [WIP][OVN] Check new added segments in OVN mech driver https://review.opendev.org/c/openstack/neutron/+/816856 | 15:42 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: [WIP][OVN] Check new added segments in OVN mech driver https://review.opendev.org/c/openstack/neutron/+/816856 | 15:59 |
seba | I'm writing tests for a neutron ml2 mechanism driver. while debugging the tests I'd like to see all the warn/info/debug loglines of the driver, but I only see the error lines. is there an easy way I can configure logging from within the tests? maybe a place where I could stealâ„¢ code for this? | 16:05 |
opendevreview | Slawek Kaplonski proposed openstack/ovn-octavia-provider master: Fix lower-constanints and pep8 jobs https://review.opendev.org/c/openstack/ovn-octavia-provider/+/816829 | 16:07 |
seba | okay, cfg.CONF.set_override('debug', True); neutron.common.config.setup_logging() seems to do the trick | 16:13 |
opendevreview | Slawek Kaplonski proposed openstack/ovn-octavia-provider master: Check gateway IP while looking for LR plugged to LS https://review.opendev.org/c/openstack/ovn-octavia-provider/+/816868 | 17:52 |
opendevreview | Merged openstack/networking-ovn stable/train: [ovn] Stop monitoring the SB MAC_Binding table to reduce mem footprint https://review.opendev.org/c/openstack/networking-ovn/+/814878 | 19:21 |
opendevreview | Merged openstack/networking-ovn stable/train: Fix OVN migration workload creation order https://review.opendev.org/c/openstack/networking-ovn/+/815622 | 19:22 |
*** blmt is now known as Guest5036 | 20:04 | |
*** yoctozepto8 is now known as yoctozepto | 21:28 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!