Wednesday, 2020-06-24

openstackgerritStephanie Djajadi proposed openstack/octavia master: WIP: Refactoring amphora stats driver interface
*** armax has joined #openstack-lbaas00:29
openstackgerritMichael Johnson proposed openstack/octavia master: Add support for nftables
*** ivve has quit IRC00:35
*** rcernin has quit IRC01:12
*** rcernin has joined #openstack-lbaas01:17
*** yamamoto has joined #openstack-lbaas01:40
openstackgerritMichael Johnson proposed openstack/octavia master: Add support for nftables
*** yamamoto has quit IRC02:04
*** shtepanie has quit IRC02:20
*** rcernin has quit IRC02:34
*** rcernin has joined #openstack-lbaas02:36
*** yamamoto has joined #openstack-lbaas02:37
*** psachin has joined #openstack-lbaas03:43
*** rcernin has quit IRC03:46
*** rcernin has joined #openstack-lbaas03:55
*** gthiemonge has quit IRC04:04
*** rcernin has quit IRC04:04
*** gthiemonge has joined #openstack-lbaas04:04
*** rcernin has joined #openstack-lbaas04:05
lxkonghi team, i have a feature request from our customer. Currently, when adding members (Nova VMs) to an LB, it means the VMs need to be open to its subnet. The request is, if Octavia can provider a SG that could be referenced as source to define the SGs of the members, i.e. the customer prefer to only allow the amphorae to access the VMs.04:20
lxkongIs that something you've already considered in the past? May I get some suggestion?04:21
johnsomYeah, we have talked about it and tried a few things. I think we keep coming down to needing changes to neutron to make ot04:22
johnsomIt work04:22
johnsomPrivate networks and backend client authentication are currently options, but don’t always meet the need.04:24
johnsomI think rm_work was the last to give it a try04:25
*** gcheresh has joined #openstack-lbaas04:37
*** aannuusshhkkaa has quit IRC04:45
*** riuzen has joined #openstack-lbaas05:34
*** riuzen has quit IRC05:34
*** gthiemonge has quit IRC05:49
*** gthiemonge has joined #openstack-lbaas05:49
*** jamesdenton has quit IRC05:52
*** jamesdenton has joined #openstack-lbaas06:00
openstackgerritGregory Thiemonge proposed openstack/octavia master: WIP Fix invalid DOWN status when updating a UDP pool
rm_worklxkong yes it's close, there's a spec in neutron for the feature we need, and a patch in Octavia that's waiting for it06:48
rm_worklxkong: unfortunately I haven't been able to work on the neutron spec... If you could take that over, it'd speed things up06:49
rm_workThat's the best way you could help06:49
*** rpittau|afk is now known as rpittau06:57
*** born2bake has joined #openstack-lbaas06:58
*** ccamposr has joined #openstack-lbaas07:20
*** maciejjozefczyk has joined #openstack-lbaas07:29
*** vishalmanchanda has joined #openstack-lbaas07:33
*** rcernin has quit IRC07:39
*** luksky has joined #openstack-lbaas07:40
*** dmellado has joined #openstack-lbaas07:43
openstackgerritCarlos Goncalves proposed openstack/octavia master: Remove amp_ssH_access_allowed option
cgoncalves"ssH" oops!08:26
*** TMM has quit IRC08:44
*** TMM has joined #openstack-lbaas08:44
*** AlexStaf has joined #openstack-lbaas08:51
*** AlexStaf has quit IRC08:52
*** AlexStaf has joined #openstack-lbaas08:52
*** rcernin has joined #openstack-lbaas09:12
*** rcernin has quit IRC09:17
fricklercgoncalves: seems sdk job is still failing after your patch
lxkongrm_work: may i have the link for neutron spec?09:52
lxkongjohnsom: also thank you09:52
openstackgerritCarlos Goncalves proposed openstack/octavia master: Fix missing params in amphora base and noop driver
cgoncalvesfrickler, err, sorry about that. can you check if ^ fixes the issue by updating the depends-on?10:02
cgoncalveslxkong, I think it is
*** tkajinam has quit IRC10:03
cgoncalveslxkong, also: , and
openstackgerritCarlos Goncalves proposed openstack/octavia master: Fix missing params in amphora base and noop driver
lxkongcgoncalves: ta10:05
*** rcernin has joined #openstack-lbaas10:10
*** rpittau is now known as rpittau|bbl10:11
*** rcernin has quit IRC10:15
*** luksky has quit IRC10:16
fricklercgoncalves: updated, thx10:19
*** luksky has joined #openstack-lbaas10:28
openstackgerritMerged openstack/python-octaviaclient master: Show operating_status in listing LB output
*** psachin has quit IRC10:57
*** rcernin has joined #openstack-lbaas11:27
*** rcernin has quit IRC11:34
*** rcernin has joined #openstack-lbaas11:39
*** gcheresh has quit IRC11:42
*** servagem has joined #openstack-lbaas11:58
*** rcernin has quit IRC12:08
*** TMM has quit IRC12:08
*** TMM has joined #openstack-lbaas12:08
*** rpittau|bbl is now known as rpittau12:13
*** ataraday_ has quit IRC12:14
*** ccamposr__ has joined #openstack-lbaas12:31
*** ccamposr has quit IRC12:33
*** rcernin has joined #openstack-lbaas12:35
*** gcheresh has joined #openstack-lbaas12:51
*** rcernin has quit IRC13:29
*** TrevorV has joined #openstack-lbaas13:54
*** luksky has quit IRC14:05
*** also_stingrayza is now known as stingrayza14:08
openstackgerritMichael Johnson proposed openstack/octavia-tempest-plugin master: Add an optional tenant flow log offload test
openstackgerritMichael Johnson proposed openstack/octavia-tempest-plugin master: Add an optional amphora admin log offloading test
rm_workgonna miss the meeting today T_T have fun14:11
*** luksky has joined #openstack-lbaas14:19
*** gcheresh has quit IRC14:37
*** also_stingrayza has joined #openstack-lbaas15:06
*** stingrayza has quit IRC15:10
*** gcheresh has joined #openstack-lbaas15:19
*** gcheresh has quit IRC15:30
*** yamamoto has quit IRC15:34
*** ataraday_ has joined #openstack-lbaas15:50
openstackgerritMichael Johnson proposed openstack/octavia-tempest-plugin master: Fix availability zone API tests
*** aannuusshhkkaa has joined #openstack-lbaas15:58
*** shtepanie has joined #openstack-lbaas15:59
johnsom#startmeeting Octavia16:00
openstackMeeting started Wed Jun 24 16:00:25 2020 UTC and is due to finish in 60 minutes.  The chair is johnsom. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.16:00
*** openstack changes topic to " (Meeting topic: Octavia)"16:00
openstackThe meeting name has been set to 'octavia'16:00
johnsomHi everyone (anyone?)16:00
johnsomPretty light agenda today16:01
johnsom#topic Announcements16:01
*** openstack changes topic to "Announcements (Meeting topic: Octavia)"16:01
johnsomI don't think I have any real announcements this week.16:02
johnsomTime flies, we are now in the MS2 phase of Victoria16:02
johnsomThere was an e-mail to the openstack-discuss list asking for input on multi-architecture Octavia.16:03
johnsomPlease add your input16:04
cgoncalvesjust a note about the MS2, that we set to "Aspire to make AmphoraV2 the default driver at MS2 checkpoint" from the PTG16:04
*** yamamoto has joined #openstack-lbaas16:04
johnsomRight, that will be the end of July, so coming up quickly16:05
johnsomWe had some criteria for that on the PTG etherpad:16:06
johnsomline 8416:06
cgoncalvesI agree with your reply email. I think Octavia can support multiple amphora architecture types and that can be done via Octavia flavors16:06
johnsomYeah, I don't think this is a bad thing. I just don't have any cycles to help that effort.16:07
johnsomAny other announcements this week?16:07
cgoncalvesmaybe the multi-arch SIG in openstack can help16:07
johnsomYep. I don't think it would be a lot of work to add architecture to the flavors. Just some time.16:08
johnsom#topic Brief progress reports / bugs needing review16:09
*** openstack changes topic to "Brief progress reports / bugs needing review (Meeting topic: Octavia)"16:09
johnsomI have been busy with reviews and minor patches to fix / address a few issues.16:09
johnsomI am also working on adding support for nftables to Octavia.16:09
johnsomIf you are not familiar, nftables is proposed to replace iptables. Debian buster has already made the switch.16:10
*** yamamoto has quit IRC16:10
johnsomIt's just a slightly different syntax. Bionic supports it, but does not use it by default.16:10
gthiemongeI'm working on the UDP health-monitor issues in tempest. I may have an idea, but I need to reproduce the bug with more debug messages to confirm it would work... Reproduction rate is really low, we hit that issue 3 times in the CI in the last 7 days16:11
cgoncalvesI worked on fixing the grenade job in multiple branches, patches are up for review. these are gate blockers.16:11
johnsomAh, are they good now? do you have a list?16:11
cgoncalvesI also fixed some oopsies in the failover refactor patch, thanks to the SDK for finding and reporting them!16:12
johnsomYeah, I noticed the timeout_dict wasn't there, but I saw it had a default and figured it didn't matter. I should have flagged it.16:13
cgoncalvesI took a few minutes to post a patch that adds amphora image tag capability to octavia flavors. needs is in line with the multi-arch email discussed earlier16:14
johnsomOn the grenade jobs, I assume we need to review from the oldest to the newest but I see the queens patch failed the gates. Do these just need rechecks?16:14
cgoncalvesorder is:16:15
cgoncalvesI will follow-up then with forward-portings to newer stable branches16:15
johnsomOk, thanks16:16
cgoncalveswe can't fix grenade in rocky and queens until devstack gets fixed after the new uwsgi release16:16
johnsomAh, that is still open?16:16
cgoncalveslooks like so16:17
cgoncalvesthis was caught in stable/rocky
johnsomUgh, ok16:18
johnsomAny other updates?16:18
johnsomI think shtepanie had a stats driver patch proposed as well.16:19
shtepanieyup! continued to work on refactoring the driver with aannuusshhkkaa16:20
shtepanieand rm_work16:20
johnsomYeah, great! I think you are looking for early stage review feedback on the approach right?16:20
johnsom#topic Open Discussion16:21
*** openstack changes topic to "Open Discussion (Meeting topic: Octavia)"16:21
johnsomAny other topics for today?16:21
ataraday_I've got question about experimental jobs with amphorav2. Is there any guide how to create such jobs or does someone looked into it?16:22
johnsomYes, let me see what I can find16:23
cgoncalvesshould be just adding an "experimental" section in zuul.d/projects.yaml like "gate", "periodic" and "check", no?16:25
johnsomYeah, I think so. I thought that was discussed in the project guide, but I'm not finding it right away16:26
ataraday_I'm not really familiar how jobs created, so will check zuul.d/projects.yaml16:28
johnsomYeah, it would be a section at the bottom of projects.yaml. Where we have "gate" and "periodic" you would add a pipeline section for "experimental".16:29
cgoncalves^ just an example from neutron16:29
johnsomYeah, I was just going to look there.16:29
ataraday_cgoncalves, Thanks!16:29
johnsomThere is some mention of it here: but not much16:30
johnsomI think to trigger it, you comment on the patch with "experimental"16:32
cgoncalvesthought it was "check-experimental"16:32
cgoncalves"check experimental" actually16:33
johnsomYeah, ok. Ugh, wish that was easy to find in the infra docs16:33
johnsomYeah, I think "check experimental"16:33
ataraday_got it16:33
johnsomAny other topics today?16:35
johnsomOk, thanks everyone. Have a great week!16:37
*** openstack changes topic to "Discussions for OpenStack Octavia | Priority bug review list:"16:37
openstackMeeting ended Wed Jun 24 16:37:37 2020 UTC.  Information about MeetBot at . (v 0.1.4)16:37
openstackMinutes (text):
*** rpittau is now known as rpittau|afk16:41
*** ataraday_ has quit IRC17:31
openstackgerritAnushka Singh proposed openstack/octavia master: WIP: Refactoring amphora stats driver interface
*** ccamposr__ has quit IRC17:50
*** ccamposr has joined #openstack-lbaas17:51
aannuusshhkkaaHey johnsom, did you get a chance to review our changes yet?18:00
johnsom Not since we last chatted about it. Is there a specific section you would like me to look at?18:01
aannuusshhkkaaYeap.. we removed the driver interface for the health updater. Did you review that already?18:02
johnsomNo I have not. If you would like me to take a look at that again I can prioritize another review pass for this morning. I just have a patch I'm wrapping up to finish first.18:03
mlozaHello, I have amphoras that has status error probably cause after upgrading to train and doing failover to re-issue the certs. I can confirm that they orphan. What's the safest way to clean them in the DB?18:05
johnsomThese are amphora that are not associated with a load balancer?18:06
mlozajohnsom: yes18:06
johnsomCan you paste a "openstack loadbalancer amphora show" for one of them?18:06
aannuusshhkkaajohnsom: Yeah it's not urgent or anything. We just wanted to know if it looked okay and were hoping to do the release note this afternoon if everything worked out fine.18:07
johnsommloza Yeah, ok, they did get hung up due to a certificate issue. We have tools to help with this coming in Victoria, but for now it would be a DB update. You would set status = DELETED, cert_busy = False, then delete the nova instance by hand. This assumes you know that the load balancer has other amps that are not in ERROR. I.e. that load balancer should have a BACKUP and MASTER amphora that are not in error.18:11
mlozajohnsom: Got it. Thanks18:14
*** vishalmanchanda has quit IRC18:17
*** gcheresh has joined #openstack-lbaas18:31
*** irclogbot_1 has quit IRC18:34
*** irclogbot_0 has joined #openstack-lbaas18:38
mlozajohnsom: cert_busy column has type int18:51
johnsomOh, yeah, that is sqlalchemy strangeness. False is 018:51
*** yamamoto has joined #openstack-lbaas18:53
*** yamamoto has quit IRC18:58
*** gcheresh has quit IRC18:59
openstackgerritMerged openstack/octavia master: Fix missing params in amphora base and noop driver
*** ccamposr has quit IRC20:18
*** ccamposr has joined #openstack-lbaas20:19
*** gcheresh has joined #openstack-lbaas20:30
*** gcheresh has quit IRC20:54
openstackgerritMichael Johnson proposed openstack/octavia master: Add support for nftables
*** gthiemonge has quit IRC21:04
*** gthiemonge has joined #openstack-lbaas21:04
openstackgerritMichael Johnson proposed openstack/octavia master: Add support for nftables
dawzonjohnsom Is there any work that still needs to be done for the minimum version patch?
*** maciejjozefczyk has quit IRC21:08
johnsomdawzon Hi there. Frankly I lost track of that patch. It looks like I didn't have any open comments. I will try to have another look later today.21:08
dawzonjohnsom Thanks, just wanted to make sure everything was done before I purge my devstack VM :)21:11
johnsomdawzon Ha, yeah, you are good. Thanks again for all of the great work!21:12
*** servagem has quit IRC21:59
johnsomaannuusshhkkaa shtepanie rm_work I have done another review on the stats patch.22:07
*** luksky has quit IRC22:08
aannuusshhkkaajohnsom, thanks! will look into those..22:10
openstackgerritMichael Johnson proposed openstack/octavia master: Add support for nftables
*** TrevorV has quit IRC22:41
*** rcernin has joined #openstack-lbaas22:42
openstackgerritMichael Johnson proposed openstack/octavia-tempest-plugin master: Tags are strings attached to an entity with the purpose of classification into groups. Tags support API tests were added.
*** tkajinam has joined #openstack-lbaas22:51
aannuusshhkkaajohnsom: not sure if gerrit lets you know if i reply to your comment on the review.. does it?22:52
johnsomaannuusshhkkaa It does, but sometime it can get lost in the volume of messages I get, so if you have a question feel free to ping me here.22:54
aannuusshhkkaaLine 97:22:55
aannuusshhkkaaSo this is still a single driver call, are you going to change that in this patch or a different one?22:55
aannuusshhkkaaI still think, under this new model, this could become a singleton.22:55
aannuusshhkkaa-- What do you mean by "this could become a singleton"? Do you mean we should entirely do away with the driver call and directly incorporate the code in the handler?22:55
johnsomBy a singleton I mean that it could load the drivers once, the first time, and then reuse that loaded set of drivers for future calls.22:56
johnsomaannuusshhkkaa Something similar to this:
johnsomBasically we should only need to load them once, then future calls will just be map calls on the loaded drivers.23:00
aannuusshhkkaaYeah that makes sense. Thank you!23:02
rm_workI think we will make it do the "multiple driver loading" thing in a later patch23:03
rm_workI think we will do the singleton thing as part of that change23:06
rm_workjust need to decide how we'll do that... i guess it could be in this patch23:07
rm_worklooking at the options you linked from earlier still and trying to determine which is the correct approach for fanout23:07
*** gthiemonge has quit IRC23:36
*** gthiemonge has joined #openstack-lbaas23:36

Generated by 2.17.2 by Marius Gedminas - find it at!