Wednesday, 2020-08-12

*** armax has joined #openstack-lbaas00:01
*** armax has quit IRC00:18
*** yamamoto has joined #openstack-lbaas00:29
rm_workjohnsom: https://i.imgur.com/RsVbQcI.png00:42
*** shtepanie has joined #openstack-lbaas00:42
rm_workinteresting note -- we saw it was slower than other LBs, possibly locationally -- so i did a failover at the 17:30 line00:43
rm_workso that's what happens to metrics during a failover event, and also yes, on a new HV it sped up significantly00:43
rm_workred line is total_connections, yellow line is active_connections00:44
*** yamamoto has quit IRC01:03
johnsomrm_work: yeah, you should see some short delays during failover, we can tune that sub second if needed. There should not be any errors, but there may be a patch needed to make it zero I have in the back of my head.01:24
rm_workyeah the interesting thing to me was the active_connections going up to like double01:25
johnsomSo host over subscribed maybe?01:25
johnsomThis is tcp right?01:26
johnsomLike http with keepalive I suspect01:26
rm_workyes01:26
johnsomStill not stressing the LB really01:27
rm_worknot clear on why this is the max we can seem to get01:28
rm_workit's beyond my skillset01:28
rm_workeven just VM to VM the max i can get is like 4000-6000 RPS01:29
rm_workwithout a LB involved01:29
johnsomHa, yeah, that can go deep to figure out01:29
rm_workLB really doesn't influence it at all01:29
rm_workVM -> VM == 6000 top RPS01:29
rm_workVM -> LB -> 2 VMs == 6000 top RPS01:29
johnsomYeah, something is not right in the underlying network01:30
johnsomIn my experience the LB overhead is a small percent, so at those low levels it is probably hard to measure01:30
johnsomSo, you can say you have zero overhead and full available bandwidth! Lol01:32
*** yamamoto has joined #openstack-lbaas01:33
rm_workheh01:35
johnsom100% of available bandwidth! ???? Grin01:36
*** rcernin has quit IRC03:10
*** rcernin has joined #openstack-lbaas03:17
*** ramishra has joined #openstack-lbaas03:21
*** psachin has joined #openstack-lbaas03:41
*** shtepanie has quit IRC03:52
*** dmsimard2 has joined #openstack-lbaas04:12
*** dmsimard has quit IRC04:13
*** dmsimard2 is now known as dmsimard04:13
*** logan- has joined #openstack-lbaas04:40
*** gcheresh_ has joined #openstack-lbaas04:52
*** tow has quit IRC05:33
*** spatel has joined #openstack-lbaas06:02
*** tobberydberg has quit IRC06:05
*** spatel has quit IRC06:08
*** tobberydberg has joined #openstack-lbaas06:11
*** gcheresh_ has quit IRC06:49
*** maciejjozefczyk has joined #openstack-lbaas07:04
*** rcernin has quit IRC07:07
*** gcheresh_ has joined #openstack-lbaas07:08
*** mvorwerk has joined #openstack-lbaas07:12
*** sapd1 has joined #openstack-lbaas07:33
*** ataraday_ has joined #openstack-lbaas07:33
*** ccamposr has joined #openstack-lbaas07:55
*** ccamposr__ has quit IRC07:57
*** gthiemon1e has joined #openstack-lbaas08:02
*** gthiemonge has quit IRC08:02
*** gthiemon1e is now known as gthiemonge08:10
*** born2bake has joined #openstack-lbaas08:14
*** tkajinam has quit IRC08:19
*** kobis1 has joined #openstack-lbaas08:32
*** kobis1 has quit IRC08:32
*** sapd1 has quit IRC08:42
*** spatel has joined #openstack-lbaas08:44
*** spatel has quit IRC08:48
cgoncalvesopenstackgerrit is down :/10:11
cgoncalvesah, it was announced and to be "offline for a short period". it's been 12 hours now10:12
cgoncalveshmm, openstackgerrit is online actually but did not join here the channel10:20
*** yamamoto has quit IRC10:41
*** yamamoto has joined #openstack-lbaas10:50
*** vishalmanchanda has joined #openstack-lbaas10:57
*** mvorwerk has quit IRC11:32
*** mvorwerk has joined #openstack-lbaas11:40
*** yamamoto has quit IRC11:41
*** yamamoto has joined #openstack-lbaas11:41
*** yamamoto has quit IRC11:42
*** yamamoto has joined #openstack-lbaas11:57
*** yamamoto has quit IRC12:28
*** ianychoi has quit IRC12:40
*** dulek has quit IRC13:01
*** yamamoto has joined #openstack-lbaas13:07
*** dulek has joined #openstack-lbaas13:12
*** yamamoto has quit IRC13:18
*** TrevorV has joined #openstack-lbaas13:23
*** sapd1 has joined #openstack-lbaas13:35
*** hongbin has joined #openstack-lbaas13:40
*** ianychoi has joined #openstack-lbaas13:44
johnsomYeah, I noticed it never came back yesterday, but gave opendev the benefit of not complaining until today. Did you mention it over there?13:54
cgoncalvesyes, #opendev. gave some examples of changes14:01
*** yamamoto has joined #openstack-lbaas14:07
*** yamamoto has quit IRC14:14
*** hongbin has quit IRC14:14
*** armax has joined #openstack-lbaas14:24
*** hongbin has joined #openstack-lbaas14:39
*** mvorwerk has quit IRC15:46
*** mvorwerk has joined #openstack-lbaas15:47
*** shtepanie has joined #openstack-lbaas15:51
*** mvorwerk has quit IRC15:54
johnsom#startmeeting Octavia16:00
openstackMeeting started Wed Aug 12 16:00:27 2020 UTC and is due to finish in 60 minutes.  The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot.16:00
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
rm_worko/16:00
gthiemongehi16:00
aannuusshhkkaahello!16:00
johnsomHello everyone! Welcome to another week in Octavia land.16:01
cgoncalveshi16:01
johnsom#topic Announcements16:01
*** openstack changes topic to "Announcements (Meeting topic: Octavia)"16:01
johnsomThe Open Infrastructure Summit is virtual and free October 19-23, to register you can go to this link:16:02
*** ZhuXiaoYu has joined #openstack-lbaas16:02
johnsom#link https://openinfrasummit2020.eventbrite.com16:02
johnsomI know we have one presentation proposed, plus the usual project updates, etc.16:02
johnsomThe Project Teams Gathering is also virtual and free October 26-30:16:02
johnsom#link https://october2020ptg.eventbrite.com/16:03
johnsomSo if you are interested in either event, they are free so sign up today!16:03
johnsomAlso of note, gerritbot is broken, so patches posted are not being announced in the channel. opendev is on it.16:03
johnsomThe last item I have is I am taking a long weekend, so will not be around Friday through Monday.16:04
johnsomAny other announcements this week?16:04
rm_worknice16:04
cgoncalvesone other note related to infra, there are some IPv6 issues somewhere on the internetzz where CI jobs can't reach pypi so they fail16:05
cgoncalvesplease refrain from rechecking patches until this is resolved16:05
johnsomAh, I heard something about that but didn't see the details. Joy.16:06
johnsom#topic Brief progress reports / bugs needing review16:07
*** openstack changes topic to "Brief progress reports / bugs needing review (Meeting topic: Octavia)"16:07
ataraday_Give a round of reviews, and finally fixed https://review.opendev.org/#/c/685337/ default ciphers change16:08
johnsomI have been heads down working on the tempest tests. I think the bulk of the work is now complete. I just have some cleanup to do and another patch to pull out bogus configuration settings and add a release note.16:09
cgoncalvesA recent DIB change broke our CentOS amphora images. I've posted patches to add non-voting amphora image build jobs to DIB CI. hopefully DIB folks will consider those job results when reviewing changes.16:09
ataraday_And want to hightlight fix for octavia-v2-dsvm-scenario-amphora-v2 job #link https://review.opendev.org/#/c/744701/16:09
cgoncalvesContinuing on the "boot faster, damn it!" saga, I noticed both Ubuntu and CentOS amps have Grub timeout set to 5 seconds. Posted a patch to drop it to 0.16:09
cgoncalvesI proposed a new DIB element "octavia-lib". it will allow us to have unreleased octavia-lib changes installed in amp images. Thanks for the reviews so far, Ann!16:10
rm_workah nice16:10
rm_workto all of that16:10
gthiemongeNeed some reviews on #link https://review.opendev.org/#/c/735126/ (we discussed that change 2 weeks ago and I've already received CR+2)16:10
johnsomNice, sounds like great progress!16:11
cgoncalvesalso proposed a patch to fix TLS versions and ciphers unsets (+ add unset support to the CLI)16:12
johnsomYeah, bummer that got overlooked, but thanks for catching it.16:13
johnsom#topic Open Discussion16:13
*** openstack changes topic to "Open Discussion (Meeting topic: Octavia)"16:13
johnsomAny topics for today?16:13
*** yamamoto has joined #openstack-lbaas16:13
ZhuXiaoYuhttps://review.opendev.org/#/c/723864/ need review :)16:14
cgoncalvesplease help review patches. ready for review patch list is getting longer and longer16:15
cgoncalves#link https://etherpad.opendev.org/p/octavia-priority-reviews16:15
johnsomAh, thank you for bringing that up again. Yes, please take some time to review that specification for Active/Active16:15
ZhuXiaoYuby the way, the ECMP spec in Neutron is approved16:16
johnsomExcellent!16:16
cgoncalvesnice!16:16
cgoncalvesZhuXiaoYu, do you have a link to the ECMP spec?16:17
cgoncalvesthis one? https://review.opendev.org/#/c/729532/16:17
johnsomSo it really is down to us reviewing the spec. I have already reviewed it once, but it looks like there haven't been comments by others yet.16:17
ZhuXiaoYuhttps://review.opendev.org/#/c/729532/ here16:18
cgoncalvesthanks16:18
johnsom#link https://review.opendev.org/#/c/729532/16:18
johnsomAny other topics this week?16:20
johnsomOk, thank you everyone! Have a great week.16:21
johnsom#endmeeting16:21
*** openstack changes topic to "Discussions for OpenStack Octavia | Priority bug review list: https://etherpad.openstack.org/p/octavia-priority-reviews"16:21
openstackMeeting ended Wed Aug 12 16:21:38 2020 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)16:21
openstackMinutes:        http://eavesdrop.openstack.org/meetings/octavia/2020/octavia.2020-08-12-16.00.html16:21
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/octavia/2020/octavia.2020-08-12-16.00.txt16:21
openstackLog:            http://eavesdrop.openstack.org/meetings/octavia/2020/octavia.2020-08-12-16.00.log.html16:21
rm_workpart of the problem on that is i just don't have the background to understand what any of that stuff means16:21
gthiemongethank you johnsom!16:21
rm_worki get the concept of ECMP but it gets pretty complex16:22
*** yamamoto has quit IRC16:23
johnsomIt's basically a load balancer. Typically no L7.16:23
rm_worki mean those active-active specs16:23
rm_workbut yeah, I can look16:23
johnsomPlease do, thanks!16:24
*** ataraday_ has quit IRC16:26
*** maciejjozefczyk has quit IRC16:28
*** sapd1 has quit IRC16:31
*** psachin has quit IRC16:39
*** ccamposr has quit IRC16:40
*** hongbin has quit IRC16:43
*** hongbin has joined #openstack-lbaas17:00
cgoncalvesall: go for CI rechecks17:00
johnsomThanks cgoncalves17:03
*** ZhuXiaoYu has quit IRC17:36
*** maciejjozefczyk has joined #openstack-lbaas17:52
*** yamamoto has joined #openstack-lbaas18:23
*** yamamoto has quit IRC18:32
*** servagem has quit IRC18:39
*** servagem has joined #openstack-lbaas19:13
*** shtepanie has quit IRC19:21
*** hongbin has quit IRC19:26
*** maciejjozefczyk has quit IRC19:39
*** hongbin has joined #openstack-lbaas19:44
*** numans has quit IRC19:51
*** TrevorV has quit IRC20:24
*** yamamoto has joined #openstack-lbaas20:32
*** yamamoto has quit IRC20:39
*** gcheresh_ has quit IRC21:14
*** ccamposr has joined #openstack-lbaas21:23
*** yamamoto has joined #openstack-lbaas21:49
*** yamamoto has quit IRC21:54
*** vishalmanchanda has quit IRC21:55
*** ramishra has quit IRC22:37
*** rcernin has joined #openstack-lbaas22:43
*** rcernin has quit IRC22:43
*** rcernin has joined #openstack-lbaas22:43
*** born2bake has quit IRC22:51
*** haleyb has quit IRC22:55
*** haleyb has joined #openstack-lbaas22:55
*** tkajinam has joined #openstack-lbaas22:57
*** spatel has joined #openstack-lbaas23:00
*** spatel has quit IRC23:00

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