Friday, 2020-09-11

« Thursday, 2020-09-10 Index Saturday, 2020-09-12 »
*** zzzeek has joined #openstack-lbaas00:02
*** rcernin has quit IRC00:28
*** rcernin has joined #openstack-lbaas00:31
johnsomCores: the pyroute2 revert patch appears to be good. It would be super groovy if we could get that merged tonight to remove that hack. The UC change has merged.  https://review.opendev.org/75111900:39
openstackgerritlikui proposed openstack/octavia master: Stop install unnecessary packages when run tox -edocs  https://review.opendev.org/75114501:59
*** zzzeek has quit IRC02:27
*** zzzeek has joined #openstack-lbaas02:33
*** rcernin has quit IRC03:53
*** rcernin has joined #openstack-lbaas04:15
*** lemko8 has joined #openstack-lbaas04:25
*** lemko has quit IRC04:28
*** lemko8 is now known as lemko04:28
*** hongbin has quit IRC04:46
*** lemko7 has joined #openstack-lbaas04:53
*** lemko has quit IRC04:54
*** lemko7 is now known as lemko04:54
*** gcheresh has joined #openstack-lbaas06:03
*** zzzeek has quit IRC06:23
*** zzzeek has joined #openstack-lbaas06:24
*** rcernin has quit IRC06:28
*** rcernin has joined #openstack-lbaas06:45
*** rcernin has quit IRC06:45
*** rcernin has joined #openstack-lbaas06:46
*** zzzeek has quit IRC06:51
*** zzzeek has joined #openstack-lbaas06:53
johnsomyep, that isn't the case.... those are required for docs07:09
*** AlexStaf has joined #openstack-lbaas07:13
openstackgerritwu.shiming proposed openstack/python-octaviaclient master: Remove install unnecessary packages  https://review.opendev.org/75117707:51
*** gcheresh has quit IRC07:52
openstackgerritwu.shiming proposed openstack/python-octaviaclient master: Remove Babel  https://review.opendev.org/75118108:01
*** AlexStaf has quit IRC08:53
*** vishalmanchanda has joined #openstack-lbaas09:00
*** rcernin has quit IRC09:10
*** AlexStaf has joined #openstack-lbaas09:20
openstackgerritMerged openstack/octavia master: Revert pyroute2 workaround  https://review.opendev.org/75111909:52
*** AlexStaf has quit IRC10:11
*** emccormick has quit IRC10:11
*** stevenglasford has quit IRC10:11
*** mnaser has quit IRC10:11
*** aannuusshhkkaa has quit IRC10:11
*** fyx has quit IRC10:11
*** jrosser has quit IRC10:11
*** AlexStaf has joined #openstack-lbaas10:11
*** emccormick has joined #openstack-lbaas10:11
*** stevenglasford has joined #openstack-lbaas10:11
*** jrosser has joined #openstack-lbaas10:11
*** fyx has joined #openstack-lbaas10:16
*** mnaser has joined #openstack-lbaas10:16
*** aannuusshhkkaa has joined #openstack-lbaas10:16
*** aannuusshhkkaa has quit IRC10:17
*** aannuusshhkkaa has joined #openstack-lbaas10:17
*** mnaser has quit IRC10:17
*** mnaser has joined #openstack-lbaas10:17
*** fyx has quit IRC10:17
*** fyx has joined #openstack-lbaas10:17
*** tkajinam has quit IRC10:22
*** AlexStaf has quit IRC10:55
*** rcernin has joined #openstack-lbaas11:06
*** rcernin has quit IRC11:20
*** zzzeek has quit IRC12:50
*** zzzeek has joined #openstack-lbaas12:51
*** zzzeek has quit IRC13:04
*** njohnston has joined #openstack-lbaas13:06
*** zzzeek has joined #openstack-lbaas13:07
*** njohnston has quit IRC13:14
*** njohnston has joined #openstack-lbaas13:19
*** TrevorV has joined #openstack-lbaas13:27
*** AlexStaf has joined #openstack-lbaas13:32
*** AlexStaf has quit IRC13:39
*** dmsimard has quit IRC13:44
*** dmsimard has joined #openstack-lbaas13:45
*** ccamposr has joined #openstack-lbaas14:01
*** ccamposr__ has quit IRC14:03
*** gmann is now known as gmann_afk14:05
*** AlexStaf has joined #openstack-lbaas14:08
*** AlexStaf has quit IRC14:15
*** gcheresh has joined #openstack-lbaas14:56
*** spatel has joined #openstack-lbaas15:28
*** spatel has quit IRC15:32
*** zzzeek has quit IRC15:32
*** zzzeek has joined #openstack-lbaas15:34
*** spatel has joined #openstack-lbaas15:43
*** spatel has quit IRC15:48
*** spatel has joined #openstack-lbaas15:48
*** zzzeek has quit IRC15:55
*** zzzeek has joined #openstack-lbaas15:56
*** ccamposr has quit IRC15:57
*** zzzeek has quit IRC16:03
*** zzzeek has joined #openstack-lbaas16:05
*** zzzeek has quit IRC16:26
*** zzzeek has joined #openstack-lbaas16:26
*** gcheresh has quit IRC16:34
*** spatel has quit IRC16:46
*** ccamposr has joined #openstack-lbaas16:55
*** tobberydberg has quit IRC16:57
openstackgerritMichael Johnson proposed openstack/octavia-tempest-plugin master: Add no-voting aarch64 scenario job  https://review.opendev.org/74762917:00
openstackgerritMichael Johnson proposed openstack/octavia master: Add aarch64/amd64 support to amphora and devstack  https://review.opendev.org/73809617:04
openstackgerritMichael Johnson proposed openstack/octavia master: Drop diskimage-builder from root requirements.txt  https://review.opendev.org/74196017:10
rm_workjohnsom: will test that patch shortly. Got sidetracked by another issue where the AAP port/IP isn't passing traffic (I think)17:38
rm_workBut I am pretty sure that's an environmental issue17:39
johnsomThat would be a new one17:39
rm_workTrying to figure out what neutron actually does to make those route... Like, does it run iptables commands on the HV? Or is it connected into a router somewhere that I can check routes on17:40
rm_workI assume it depends on the driver in use17:40
johnsomYeah, definitely will be different depending on the drivers in use17:40
rm_workYeah anyway, sidetracked. But will test shortly17:41
*** gmann_afk is now known as gmann17:42
openstackgerritMerged openstack/python-octaviaclient stable/victoria: Update .gitreview for stable/victoria  https://review.opendev.org/75107917:52
openstackgerritMerged openstack/python-octaviaclient stable/victoria: Update TOX_CONSTRAINTS_FILE for stable/victoria  https://review.opendev.org/75108017:52
rm_workif i change the LVS config to use the vrrp_ip and reload it, and then try to connect with the vrrp_ip, then it works. but if it's using the ha_ip it doesn't (just times out)17:52
rm_workso it seems like the ha_ip (AAP address) is not getting routing set up correctly17:52
rm_workha, yeah... looking at a WORKING vs BROKEN example: http://paste.openstack.org/show/797792/17:59
rm_workI expect it to not be listening on the random port I chose17:59
rm_workso I expect a timeout for both17:59
rm_workbut for the broken one, the ha_ip (AAP port) gets `no route to host` instead18:00
rm_worksomehow the routes aren't getting set up18:00
rm_workif I do a failover, the routes get set up properly on the new one18:00
rm_workthose routes are coming from the vrrp stuff right?18:00
rm_workso it's like... the keepalived garps didn't go out properly or something?18:00
rm_workwhich is an issue on the amps?18:01
johnsomUgh, not following this stream. Where were you nc'ing from?18:06
rm_workyeah anyway, yes, keepalived-octavia service just failed to start18:06
rm_work`Entering FAULT STATE` on both master and backup amp18:06
rm_workummmmm18:14
rm_workthere's a check_script for haproxy18:15
rm_workwhich shouldn't exist on these amps because they're UDP only18:15
rm_workobviously it's going to fail18:15
johnsomI thought we fixed that....18:15
rm_workknown bug?18:16
rm_workit seems to resolve itself after a LB failover18:16
rm_worknew amps won't have it18:16
rm_workbut on initial create they seem to18:16
rm_workah18:18
rm_workso it wouldn't get added if a UDP listener exists18:18
rm_workbut... on initial create of the LB before any listeners...  wouldn't it have this?18:18
rm_workand then what would  ever delete it?18:18
rm_workthen failover would set it up initially with the full config and so it wouldn't ever be created, which is why it works on failover18:19
*** vishalmanchanda has quit IRC18:20
openstackgerritAdam Harwell proposed openstack/octavia master: Remove haproxy_check_script for UDP-only LBs  https://review.opendev.org/75137918:37
rm_workjohnsom: ^^  this IMO18:37
rm_workwhat was the original bug report? do you remember?18:37
johnsomNo, I looked for the patch but didn't find one18:40
rm_worki'm beating my head against storyboard trying to search18:42
*** gcheresh has joined #openstack-lbaas18:42
johnsomI may be confusing it with this method I added during failover testing.18:42
rm_workwhatever, this is fine18:46
rm_workdon't need a story18:46
rm_workmaybe need a relnote18:46
johnsomSo what is the scenario, lb had a TCP listener, then removed and UDP added?18:46
*** armax has quit IRC18:57
*** armax has joined #openstack-lbaas18:57
*** armax has quit IRC18:57
*** tobberydberg has joined #openstack-lbaas18:58
openstackgerritAdam Harwell proposed openstack/octavia master: Use routed network filter if it exists  https://review.opendev.org/70615319:03
rm_workjohnsom: no actually19:03
rm_workjohnsom: LB was created with no listeners (like, duh) then a UDP listener was added later19:03
rm_workthis is actually the default UDP workflow19:03
rm_workmakes me wonder if no one actually uses this19:03
johnsomHow do any of our UDP tests pass?19:03
rm_workI seriously don't know19:03
rm_workfollowing the logic though, it makes sense19:04
rm_workdo our UDP tests include LBs that also have TCP listeners?19:06
rm_workare we trying to "reuse"19:06
johnsomProbably19:06
openstackgerritAdam Harwell proposed openstack/octavia master: [DNM] Create yamas driver for statistics  https://review.opendev.org/74365319:08
johnsomWell, this bug was introduced in the failover patch, so yeah, probably not many have deployed it yet19:09
johnsomrm_work I will add a story and rel note so we can backport this.19:11
rm_workk19:11
openstackgerritArkady Shtempler proposed openstack/octavia-tempest-plugin master: Adding failover test. Sending HTTP traffic while MASTER Amphore is rebooted. BACKUP Amphore should serve the traffic.  https://review.opendev.org/75138919:11
*** gcheresh has quit IRC19:16
openstackgerritMichael Johnson proposed openstack/octavia master: Remove haproxy_check_script for UDP-only LBs  https://review.opendev.org/75137919:16
openstackgerritMerged openstack/octavia master: Change tls_ciphers init in Pool API classes  https://review.opendev.org/73040719:29
*** armax has joined #openstack-lbaas19:47
rm_workah yeah if the LB EVER came up with a TCP listener in tempest, it would have gARP'd once at least and so even if it wasn't continuing to gARP, the routes would exist when it was UDP only20:18
rm_workneed to check to make sure but my guess is the UDP LB tests piggyback on TCP LB tests20:19
rm_workif we even have a UDP traffic test?20:19
johnsomYes we do20:21
johnsomYou probably haven't seen the fruits of my recent labors yet: https://1011f633dbb64bb3f6e0-d72dd3f5e858ba8ddc3e5e0b32c41186.ssl.cf2.rackcdn.com/751119/1/check/octavia-v2-dsvm-scenario/7693bad/testr_results.html20:22
johnsomThere are now 194 scenario tests.20:23
johnsomUp from 25...20:24
rm_worknice20:25
*** TrevorV has quit IRC20:37
*** servagem has quit IRC21:14
*** rcernin has joined #openstack-lbaas22:02
*** rcernin has quit IRC22:03
*** rcernin has joined #openstack-lbaas22:03
*** njohnston has quit IRC22:41
« Thursday, 2020-09-10 Index Saturday, 2020-09-12 »

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!