Tuesday, 2021-01-26

« Monday, 2021-01-25 Index Wednesday, 2021-01-27 »
*** k_mouza has quit IRC00:01
*** brinzhang has quit IRC00:05
*** luksky has quit IRC00:08
openstackgerritJames Kirsch proposed openstack/kolla-ansible master: Add support for LetsEncrypt-managed certs  https://review.opendev.org/c/openstack/kolla-ansible/+/74134000:09
*** brinzhang has joined #openstack-kolla00:17
*** LinPeiWen has joined #openstack-kolla00:36
*** LinPeiWen has quit IRC00:41
*** LinPeiWen has joined #openstack-kolla00:46
*** ivan_lin has joined #openstack-kolla00:48
openstackgerritZhouHeng proposed openstack/kolla-ansible master: Replace db-sock with db-nb-sock and db-sb-sock  https://review.opendev.org/c/openstack/kolla-ansible/+/77159400:49
*** k_mouza has joined #openstack-kolla01:32
*** k_mouza has quit IRC01:58
*** zzzeek has quit IRC02:19
*** zzzeek has joined #openstack-kolla02:22
*** zzzeek has quit IRC02:49
*** zzzeek has joined #openstack-kolla02:51
*** k_mouza has joined #openstack-kolla03:58
*** k_mouza has quit IRC04:03
*** brinzhang has quit IRC05:20
*** brinzhang has joined #openstack-kolla05:21
*** cah_link has joined #openstack-kolla06:07
*** spotz has quit IRC06:54
*** ivan_lin has quit IRC06:56
*** antonym has quit IRC06:56
*** Reepicheep has quit IRC06:57
*** ivan_lin has joined #openstack-kolla06:57
*** antonym has joined #openstack-kolla06:57
*** Reepicheep has joined #openstack-kolla06:57
yoctozeptoMr_Freezeex: sounds reasonable06:59
yoctozeptomorning koalas06:59
osmanlicilegimorning07:06
*** xinliang has joined #openstack-kolla07:12
openstackgerritPiotr Parczewski proposed openstack/kolla-ansible master: [docs] Unify project's naming convention  https://review.opendev.org/c/openstack/kolla-ansible/+/77128207:24
*** luksky has joined #openstack-kolla07:50
*** luksky has quit IRC07:53
*** rpittau|afk_ is now known as rpittau07:57
kevkomorning08:14
kevkoyoctozepto: i've checked your change of my change for mariadb ... why did you remove grouping in ansible/site.yml ?08:14
kevko:q!08:14
yoctozeptokevko: quitting vim, are we? :D08:16
kevko:D08:16
kevkoyes08:16
kevkoyoctozepto: i've checked what you've changed .. ok rename of variables make sense ..but why you remove dynamic grouping in ansible/site.yml ? of course without it will not work ...08:17
kevkoit will not work without it , of couse :D08:17
*** andrewbonney has joined #openstack-kolla08:19
yoctozeptokevko: it will, I broke something else :-)08:20
yoctozeptoyeah, I killed the gathering conditional08:23
yoctozeptohmm08:23
*** bengates has joined #openstack-kolla08:24
kevko:D08:24
*** bengates has quit IRC08:26
*** bengates has joined #openstack-kolla08:26
*** maharg101 has joined #openstack-kolla08:28
kevkoi'm afraid that it can be written different ..08:29
kevkocan't08:29
*** bengates has quit IRC08:30
kevkoyou have to also ensure that if you running with limit to one host for example .. you always have to gather info about shards ...08:30
kevkoto be able generate configs ...etc ..08:30
*** bengates has joined #openstack-kolla08:30
yoctozeptowe always gather all facts08:33
kevkoyoctozepto: ok, can u upload that killed gathering conditional as u said above ?08:46
hrwosmanlicilegi: can you go to https://review.opendev.org/c/openstack/kolla/+/772382 and help with cleanup?08:47
yoctozeptokevko: you mean the fixed one? :P08:52
yoctozeptoyeah, I'll look into it08:52
hrwtrivial fixes for morning coffee: https://review.opendev.org/c/openstack/kolla/+/771846 https://review.opendev.org/c/openstack/kolla/+/77238208:52
openstackgerritMark Goddard proposed openstack/kolla stable/stein: DNM: Test Stein builds  https://review.opendev.org/c/openstack/kolla/+/77247808:58
*** bengates has quit IRC08:58
*** bengates has joined #openstack-kolla08:59
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: WIP: move to Debian 'bullseye'  https://review.opendev.org/c/openstack/kolla/+/77247909:03
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible master: Refactor mariadb to support shards  https://review.opendev.org/c/openstack/kolla-ansible/+/76695209:03
yoctozeptokevko: ^ this should go green09:04
hrwThat bullseye patch is from my tests. Decided that it is too good to be wasted.09:05
yoctozeptohrw: ack09:07
hrwHard Freeze of Bullseye is on 12th March. So Xavier cycle can get it probably09:08
yoctozeptomgoddard: could you confirm whether the approach we have with kevko regarding shards in mariadb would make sense with our approach to cells?09:08
hrwwe moved to Buster before it was released09:08
yoctozeptohrw: that would be cool; I asked zigo about making debian an official openstack distro but he was not too eager to make it happen09:09
mgoddardyoctozepto: I don't know enought to say yet09:09
mgoddardwhat does it mean to be an official distro?09:09
yoctozeptomgoddard: honestly, I did not expect a ready yes/no :D just please review sooner than later so that we go in the right direction for things we want09:10
yoctozeptomgoddard: TC approves it and includes in the tables; and QA (devstack) has jobs on debian09:10
yoctozeptothat is a very simplistic answer09:10
yoctozeptobut it boils down to it09:10
*** jbadiapa has joined #openstack-kolla09:10
yoctozeptoand then we could officially support py3.9 because it is supported by debian09:11
yoctozepto:D09:11
hrwyoctozepto: Debian Openstack team would have to invest time in CI doing git based builds. As now they work on current release only on RC/final tarballs09:12
hrwyoctozepto: note that debian/binary target still has victoria09:12
yoctozeptohrw: yeah, I figured; but for the purposes of upstream the packaging is irrelevant09:12
yoctozeptodevstack wallaby uses uca victoria repos09:13
yoctozeptobecause that's how life goes ;d09:13
hrwdevstack.09:13
yoctozeptowe recently dropped opensuse09:13
yoctozeptoyeah, this is what QA means here09:13
*** kevko has quit IRC09:13
hrwwhen I want to curse and I am in a group where I should not I use devstack09:13
yoctozeptoit's not far from the truth in the team as well09:14
yoctozeptothere are too few cores to make it work legitly09:15
yoctozeptoand it accrued a lot of hacks during its lifetime09:15
yoctozeptoso basically it's one big hack09:15
yoctozeptothat works for most09:15
yoctozeptobut not always09:16
yoctozeptoand not best09:16
hrwI had to use devstack few years ago due something. Learnt quickly that proper way of running devstack is adding "2>/dev/null >/dev/null", run and going to the kitchen to sacrify chicken to gods09:16
hrwand then retry and hope that this time it will happen09:17
yoctozeptopretty much that's the way it works09:18
yoctozeptowell, that's what you get with hacking on bash09:18
hrw;d09:20
*** murphyslawbbs has quit IRC09:26
*** kevko has joined #openstack-kolla09:27
kevkoyoctozepto: i think it is very usable with cells ..09:29
kevkoyoctozepto: you just need to define nova_cell_database_shard_id per cell's group vars09:30
yoctozeptokevko: yeah, I feel similarly, but it's not my view on cells so I leave it to those whose it is :D09:30
kevkoi've better not add cells to my patches ... but i've tested locally09:31
kevkoyoctozepto: regarding offictial openstack distro ... what's needed to do that ?09:38
kevkoyoctozepto: i've asked him to switch  "https://openstack.debian.net/ buster-victoria main"09:38
kevkois this only thing ?09:38
kevko(he answered that it is usefull and silly to use https ..)09:39
kevkoand also said there is osbpo.debian.net ...09:40
kevkowhat is requirements ... ?09:40
yoctozeptokevko: I did not ask about the url, I think it was hrw09:42
kevkook, so what is needed to have debian as official openstack distro ?09:42
kevkohrw: ^^ ?09:42
yoctozeptokevko: it would consist of telling TC about it and committing to supporting debian upstream which basically boils down to creating a devstack job for it and making sure it is green most of the time09:43
yoctozeptosimilarly to how it was with opensuse09:43
yoctozeptoand is with centos and ubuntu09:43
yoctozeptoit is not that hard as the base image is already available (as used by kolla and a few others)09:44
yoctozeptoI am not sure how devstack fares on debian though; it should be pretty easy to get it working because it works already on ubuntu09:44
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible master: Refactor mariadb to support shards  https://review.opendev.org/c/openstack/kolla-ansible/+/76695209:47
kevkoyoctozepto: I think i can do it ..but have to ask my employer if i have a time :D09:47
kevkobut I am pro09:47
*** e0ne has joined #openstack-kolla09:49
*** bengates has quit IRC09:49
*** bengates has joined #openstack-kolla09:50
openstackgerritMichal Arbet proposed openstack/kolla-ansible master: Refactor mariadb to support shards  https://review.opendev.org/c/openstack/kolla-ansible/+/76695209:55
kevkoyoctozepto: i've changed commit message and change default shard_id to 1 (as i am doing multiplication in proxysql role..and it is not good to do it with 0 :)) ..09:56
*** k_mouza has joined #openstack-kolla09:59
yoctozeptokevko: why? it makes total sense to have it starting with 0 :D10:01
kevkoyoctozepto: well, as i said ..if you check proxysql role you will see10:01
kevkoproxysql has to have groups defined per galera cluster10:02
kevkoexample   shard1 : writer_group = N , backup_writer = N + 1, reader_group = N + 2 , offline_group = N + 310:03
kevkothis has to be defined per cluster shard ..10:03
*** k_mouza has quit IRC10:03
yoctozeptokevko: and it works with 010:03
kevkoagree10:04
kevkoBUT :) ..10:04
yoctozeptoyou don't like zeros?10:04
yoctozeptoI believe it is actually safer10:04
yoctozeptoto default to 0 and teach people to use 110:04
yoctozeptoetc.10:04
kevkoi also didn't want to solve spacing of 4 between shards ... so i decided to have shard starting with 1, galera hostgroup writer 10, 11,12,13 20,21,22,2310:04
yoctozeptoalso do note the variable names have changed10:04
yoctozeptothis is to not confuse the client side database_10:04
yoctozeptowith server side mariadb_10:04
kevkocheck my message above ..10:05
yoctozeptokevko: yeah, but it works with 010:05
yoctozepto:D10:05
kevkow8 a minute10:06
yoctozeptoa minute has passed10:07
kevko:D10:10
kevkoyoctozepto: http://paste.openstack.org/show/801976/10:10
*** e0ne has quit IRC10:11
yoctozeptokevko: yes, precisely10:11
yoctozeptoit makes sense to me10:11
*** e0ne has joined #openstack-kolla10:11
yoctozeptoespecially to default to 0 if user is *not* using shards10:11
yoctozeptoalso, note my comment about var names10:12
yoctozeptoI hope you understand the database_ / mariadb_ prefix difference10:12
*** brinzhang has quit IRC10:14
kevkohmm, ok10:15
kevkook, i will revert to your patchset ...and wait if it will be green ..10:16
openstackgerritMark Goddard proposed openstack/kolla stable/stein: CI: Use Docker proxy for pulls + Ruby fixes + get_pip  https://review.opendev.org/c/openstack/kolla/+/77249010:19
openstackgerritMichal Arbet proposed openstack/kolla-ansible master: Refactor mariadb to support shards  https://review.opendev.org/c/openstack/kolla-ansible/+/76695210:19
*** ricolin has quit IRC10:28
openstackgerritPierre Riteau proposed openstack/kayobe stable/train: Fix pygments style  https://review.opendev.org/c/openstack/kayobe/+/77249310:28
*** bengates has quit IRC10:40
*** bengates has joined #openstack-kolla10:41
openstackgerritMark Goddard proposed openstack/kolla stable/train: CI: revert to public package mirrors after build  https://review.opendev.org/c/openstack/kolla/+/77009010:44
*** e0ne has quit IRC10:50
openstackgerritMerged openstack/kolla master: base: drop Linaro OBS key for CentOS and Ubuntu  https://review.opendev.org/c/openstack/kolla/+/77238210:51
openstackgerritMerged openstack/kolla stable/ussuri: nova-compute: 'pmdk-tools' is x86-64 only in Ubuntu 'bionic'  https://review.opendev.org/c/openstack/kolla/+/77184610:51
*** also_stingrayza has quit IRC10:54
*** stingrayza has joined #openstack-kolla10:56
*** xinliang has quit IRC10:58
*** brinzhang has joined #openstack-kolla11:04
*** k_mouza has joined #openstack-kolla11:05
openstackgerritMerged openstack/kolla master: Build Masakari for Ubuntu binary  https://review.opendev.org/c/openstack/kolla/+/76825511:09
*** rpittau is now known as rpittau|bbl11:09
*** bengates has quit IRC11:18
yoctozeptomgoddard: psst, you reviewed some old version11:18
mgoddarddammit11:19
yoctozeptoxD11:19
*** bengates has joined #openstack-kolla11:19
mgoddardsitting in my browser for too long11:19
hrwok, looks like our pip fix for train is coming.11:19
*** bengates_ has joined #openstack-kolla11:23
yoctozeptomgoddard: fwiw, I addressed most of your concers (if not all) - how's that for being proactive?11:23
openstackgerritMerged openstack/kayobe stable/train: Fix pygments style  https://review.opendev.org/c/openstack/kayobe/+/77249311:24
yoctozepto:D11:24
yoctozeptoyeah, I think all11:24
openstackgerritMark Goddard proposed openstack/kolla-ansible stable/stein: CI: restrict pip to <21  https://review.opendev.org/c/openstack/kolla-ansible/+/77250111:25
*** bengates has quit IRC11:27
openstackgerritMark Goddard proposed openstack/kolla stable/stein: CI: Use Docker proxy for pulls + Ruby fixes + get_pip  https://review.opendev.org/c/openstack/kolla/+/77249011:29
hrwmgoddard: we still maintain stein?11:34
mgoddardhrw: we are not obliged to, but that doesn't mean we can't11:34
hrwsure11:35
mgoddardnice to at least keep it building11:35
hrw+211:36
mgoddardwell let's see it pass first11:50
*** bengates_ has quit IRC11:51
*** bengates has joined #openstack-kolla11:52
openstackgerritMerged openstack/kolla master: Remove two missed unicode flags  https://review.opendev.org/c/openstack/kolla/+/77235211:54
openstackgerritMerged openstack/kolla stable/train: Improve pip install process for offline deployment.  https://review.opendev.org/c/openstack/kolla/+/75805711:55
hrwyay!12:00
openstackgerritMarcin Juszkiewicz proposed openstack/kolla stable/train: nova-compute: provide needed deps for debian/aarch64  https://review.opendev.org/c/openstack/kolla/+/77199412:00
openstackgerritMarcin Juszkiewicz proposed openstack/kolla stable/train: nova-compute: 'pmdk-tools' is x86-64 only in Ubuntu 'bionic'  https://review.opendev.org/c/openstack/kolla/+/77192512:00
hrwrebased12:00
*** jbadiapa has quit IRC12:06
*** kevko has quit IRC12:07
mgoddardphew, proxysql chain reviewed12:08
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: base: print release version in case of running unsupported one  https://review.opendev.org/c/openstack/kolla/+/77250712:10
hrwsmall tweak for those moments when user makes an error12:11
openstackgerritMichal Nasiadka proposed openstack/kolla master: Add octavia-driver-agent  https://review.opendev.org/c/openstack/kolla/+/77119112:22
*** ricolin has joined #openstack-kolla12:34
*** LinPeiWen has quit IRC12:42
*** kevko has joined #openstack-kolla12:55
openstackgerritPierre Riteau proposed openstack/kayobe stable/train: CI: add dummy1 port for seed upgrade and VM jobs  https://review.opendev.org/c/openstack/kayobe/+/76418412:58
*** rpittau|bbl is now known as rpittau12:59
*** Marianito has joined #openstack-kolla13:00
MarianitoHi guys!13:01
*** jbadiapa has joined #openstack-kolla13:03
*** Marianito has quit IRC13:17
*** dougsz has joined #openstack-kolla13:18
kevkohi13:19
yoctozeptohi kevko again13:44
yoctozeptois there a chance that we can get away from creating a user per shard?13:44
yoctozeptoI know you said it is related to how sharding is configured13:44
yoctozeptobut perhaps there are alternatives we would like to know/consider with mgoddard13:45
kevkowell, when I was implementing maxscale ..i had every shard directly available via unix socket on some docker volume which was mounted to kolla_toolbox and users/databases was created against unix socket ..no special user ...13:46
kevkocheck my old patch ...it was dirty ... this is more clear i think ..13:46
kevkoanother solution could be to have rules in proxysql ...13:47
kevkorule for 'CREATE DATABASE neutron' ... etc13:47
kevkolike , rule for create user ..rule for create database ..13:48
kevkoit means it will be matched OR not against set of rules ... and i didn't like also ..13:48
kevkotried lot of combinations ..and this was most clear i think on first glance ..13:49
kevkoyoctozepto: not to mention that they really are different database clusters ..so i think it is not dirty to have different root users ..13:50
kevkohaving special user is most simplest and most straightforward13:51
mgoddardkevko: do mariadb queries go haproxy -> proxysql -> mariadb, or just proxysql -> mariadb?13:53
kevkojust proxysql-> mariadb13:53
yoctozeptokevko: yeah, it's not that wrong, don't take us wrong13:54
yoctozeptowe are just evaluating our options13:54
mgoddardok13:55
kevkothat's ok, as I said .. I tried similar combinations ..if you have some idea .. just ask ..maybe i already tried and can say something regarding it :)13:55
kevkoI've just learned argumenting, then it's easier to get something into upstream, even though it's always a tiring fight :D13:56
yoctozeptokevko: oh! I do hope you don't find working with us a "tiring fight" :-)14:03
yoctozeptoit is all about getting the top quality solutions14:03
yoctozeptoand we know that you can deliver those :-)14:04
kevkohh, thanks14:04
kevkoi'm trying to ..14:05
kevkojust curious, have you already checked all pieces ? to have some better picture ?14:06
yoctozeptoyes, me and mgoddard did14:06
yoctozeptoand we are satisfied with the general direction14:07
mgoddardyoctozepto speaks for mgoddard? :)14:08
yoctozeptomgoddard: I read your reviews14:09
yoctozeptotell me I was wrong :-)14:09
mgoddardhonestly I'm still mulling over the big picture14:09
yoctozeptooh, I see, sorry then; remove the plural!14:10
yoctozeptos/we are/I am14:10
yoctozepto:-)14:10
mgoddardI definitely feel that for features like this a spec process would help14:11
hrwhttps://review.opendev.org/c/openstack/kolla/+/771925 anyone? mgoddard yoctozepto?14:13
kevkoyoctozepto, mgoddard: btw, somewhere I've seen SSL blueprint for db connections ... with proxysql merged it will be very easy to implement .. just btw14:15
openstackgerritMark Goddard proposed openstack/kolla-ansible stable/stein: CI: restrict pip to <21, pin alembic to avoid ARA failure  https://review.opendev.org/c/openstack/kolla-ansible/+/77250114:15
*** TrevorV has joined #openstack-kolla14:16
mgoddardkevko: I expect proxysql will make it more complicated to implement14:16
mgoddardhaproxy already supports tls14:16
*** hjensas is now known as hjensas|afk14:16
kevkoreally ? connections between haproxy and mariadb can be enrypted now ? :) I mean haproxy  <-> mariadb14:18
kevkomgoddard: https://blueprints.launchpad.net/kolla-ansible/+spec/mariadb-ssl-support14:18
openstackgerritMark Goddard proposed openstack/kolla-ansible stable/stein: CI: restrict pip to <21, pin alembic to avoid ARA failure  https://review.opendev.org/c/openstack/kolla-ansible/+/77250114:18
dmsimardmgoddard: apologies for the ara breakage, 0.x hasn't been maintained for a while :(14:19
mgoddardkevko: well the main concern is the frontend connection - clients to haproxy/proxysql14:19
dmsimardif I can help upgrading let me know, hopefully it shouldn't be too complicated considering https://review.opendev.org/c/openstack/kolla-ansible/+/72090214:20
dmsimardthere's probably the generate html command to change as well but it should otherwise just work14:20
kevkomgoddard: well, I meant above blueprint14:20
mgoddardkevko: perhaps proxysql could help with the backend14:20
mgoddardkevko: the blueprint is ambiguous, but I would assume frontend would come first14:21
mgoddarddmsimard: no problem, it happens14:21
*** stand has quit IRC14:24
*** brinzhang has quit IRC14:25
*** brinzhang has joined #openstack-kolla14:25
*** stand has joined #openstack-kolla14:26
mgoddarddmsimard: should we just take that patch?14:29
dmsimardmgoddard: not that particular one but I can send something14:30
dmsimard2 mins14:30
mgoddardgreat14:30
openstackgerritDavid Moreau Simard proposed openstack/kolla-ansible master: Upgrade ara to the latest version in CI jobs  https://review.opendev.org/c/openstack/kolla-ansible/+/77253214:33
dmsimardmgoddard: ^ maybe14:33
hrw+2 and let zuul show does it work14:36
hrwtripleo-build-containers-centos-7 suxx14:40
frickleris "Ubuntu 16.04" still correct here https://docs.openstack.org/kayobe/victoria/installation.html ? or did you just forget to update?14:42
*** stand has quit IRC14:43
frickleralso https://docs.openstack.org/kayobe/victoria/support-matrix.html doesn't list ubuntu at all14:43
hrwwhat is ubuntu?14:43
*** dave-mccowan has joined #openstack-kolla14:44
*** dave-mccowan has quit IRC14:46
*** brinzhang has quit IRC14:48
*** brinzhang has joined #openstack-kolla14:51
*** brinzhang has quit IRC14:51
*** brinzhang has joined #openstack-kolla14:52
yoctozeptofrickler: kayobians are adding ubuntu support in wallaby15:03
yoctozeptoand it will be 20.04 of course15:05
frickleryoctozepto: that's sad to learn, /me is looking for something to replace the current automation starting from our stein deployments15:06
priteaufrickler: Ubuntu is supported as a control host, but not yet as seed / overcloud hosts (but work is in progress)15:09
*** gary_perkins has quit IRC15:28
*** strigazi has quit IRC15:45
*** strigazi has joined #openstack-kolla15:47
*** ysirndjuro has joined #openstack-kolla16:05
*** kevko has quit IRC16:06
mgoddardfrickler: I expect we'll be open to backporting ubuntu support, depending on how large it turns out to be. I'm not sure we'd go as far as stein upstream, but if you were open to keeping a fork perhaps we could work something out?16:18
yoctozeptoyes, also: you can start easy with kolla-ansible and then move to kayobe16:25
yoctozeptodepends on the level of automation you need from the day zero :-)16:26
*** e0ne has joined #openstack-kolla16:27
priteau+1 yoctozepto16:31
*** kevko has joined #openstack-kolla16:32
kevkostill here ?16:32
kevkosorry16:32
kevkoyoctozepto, mgoddard: ^^16:33
mgoddardkevko: yes16:33
kevkoi replied to your comment about users ... just let me know if i explain good :)16:34
kevkonow i need to move for 30 minutes ..but I look forward to reading when I return :P16:35
*** gary_perkins has joined #openstack-kolla16:36
openstackgerritMark Goddard proposed openstack/kolla-ansible master: Persist nova libvirt secrets in a Docker volume  https://review.opendev.org/c/openstack/kolla-ansible/+/65085316:46
hrwhttps://review.opendev.org/c/openstack/kolla/+/771994 anyone? zuul passed17:17
hrwtrain aarch64 fix17:17
hrwand https://review.opendev.org/c/openstack/kolla/+/77192517:17
*** gfidente is now known as gfidente|afk17:22
*** rpittau is now known as rpittau|afk17:27
*** bengates has quit IRC17:27
*** dougsz has quit IRC17:29
*** zul has quit IRC17:52
*** maharg101 has quit IRC17:55
kevkoyoctozepto, mgoddard still here ? or already out ?17:58
*** TrevorV has quit IRC18:00
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible master: Refactor mariadb to support shards  https://review.opendev.org/c/openstack/kolla-ansible/+/76695218:15
openstackgerritMerged openstack/kolla stable/train: Bump prometheus-openstack-exporter version to 1.3.0  https://review.opendev.org/c/openstack/kolla/+/77007618:19
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible master: Refactor mariadb to support shards  https://review.opendev.org/c/openstack/kolla-ansible/+/76695218:27
yoctozeptokevko: I guess I was here18:29
yoctozeptoand am18:29
kevkoi was just checking your change ..18:30
kevkoi'm little bit afraid if now it is everything passing as before ..because little bit lost now :)18:30
kevkoyou didn't rebase all comits, did you ?18:31
yoctozeptokevko: I did not, they will require some love18:36
yoctozeptoI am trying to get this in bit by bit18:37
kevkook, thank you :)18:37
yoctozeptonow all mgoddard's concerns should have been addressed except for the user part but I find it the cleanest so far after reading the docs18:37
yoctozeptokevko: you are very welcome18:37
yoctozeptothe idea and contribution is great :-)18:37
kevkothank you , I am blushing18:38
kevkoyoctozepto: one thing that bothers me ..18:38
yoctozeptokevko: no need to be bothered about blushing18:38
yoctozeptoit's pretty common18:38
kevkoyoctozepto: :D18:39
kevkoyoctozepto: https://review.opendev.org/c/openstack/kolla/+/76938518:39
kevkoyoctozepto: if you read docs for proxysql ..you found that proxysql has configration as cisco :D18:39
kevkodisk -> memory -> runtime18:39
kevkoand config is used only for bootstrap18:39
yoctozeptoyeah, pretty much, except for syntax that is18:39
yoctozeptoyeah, yeah, I figured18:40
yoctozeptoyou are doing magic for it18:40
yoctozeptoinside the script18:41
yoctozeptoit feels really hacky18:41
yoctozeptobut perhaps sane with the iac approach18:41
kevkobecause of this I had to choose some nice way to do it like haproxy for example ..18:41
kevkoi went by a way of yaml ..18:41
yoctozeptoyaml is the way to go18:42
yoctozeptounless one can use toml18:42
yoctozeptothen use toml18:42
yoctozeptobut not here18:42
yoctozeptonot now18:42
yoctozepto:D18:42
kevko(didn't want to use libconfig library ..if i remember well)18:42
yoctozeptoyes, that's right18:42
yoctozeptojust a general remark18:42
yoctozeptofor the real life18:42
yoctozeptotoml > yaml > json18:42
yoctozeptopick the best you are allowed to18:43
yoctozeptooh, that sounded as if kolla is not real18:43
yoctozeptoperhaps that's true18:44
kevko:D18:44
yoctozeptoit's just the illusion of containers18:44
yoctozeptowell, anyhow, you get the point18:44
kevkoi wanted to be just sure that it is ok18:44
kevkobecause i had a felling that noone checks ..18:44
yoctozeptoyeah, the yaml sounds fine18:44
kevkoand also had a feeling that especially this will be problem18:44
yoctozeptoI am not 100% about the logic18:45
yoctozeptobut it's better than leaving it to proxysql I guess?18:45
kevkohonestly, i didn't want to have like it is proposed ..18:45
yoctozeptoI feel you18:45
kevkobut again ..it was most simple and clean solution18:45
yoctozeptowell, sad we had to give up on the maxscale18:46
yoctozeptobut life is life18:46
yoctozeptothanks for adapting sooo quickly :D18:46
kevkofirstly i was trying to inject percona scripts ... write own script to "download" users from backends ..18:46
kevkofor reloading configuration i wanted to use proxysql and flag --reload flag ..but it was not working always ..18:47
*** k_mouza has quit IRC18:48
kevkoso, i decided to remove internal DB, generate proxysql's libconfig from several yaml sub-configs and run bootstrap internal db again18:48
kevkoyoctozepto: no problem18:48
kevkoyoctozepto: we have this delivered in production on client's side ..and it is working very nice ..(i mean proxysql thing)18:49
kevkoyoctozepto: but most things were deployed by hand ... kolla's proxysql image and keepalived image built and configuration was created from hand (as ansible will do it)18:50
kevkobut, image is working :)18:50
kevkobut i'm nervous when I am sending tutorial to change something manually, will have better sleep if it will be merged and I will cherry-pick to ussuri :P18:51
yoctozeptohaha, totally relatable18:51
kevkothese changes are big and I don't have ambition to do it every release and solve conflicts again and again18:52
kevkoand store in house18:52
kevkoanother story is logging :) ..but i am communicating with author of python-fluent-logger and he promised that he will rewrite logger to msgpack >1.018:53
yoctozeptoyeah, that logging is a sad story18:53
yoctozeptoso close, yet so far18:53
kevkoeven if there is pull request to bump msgpack :D18:54
kevkoi asked him if it is in good shape (because tests are passing) ..but he said that he has to edit his unit tests ..18:55
kevkoi think i should resend to some kolla mail list18:55
yoctozeptowhat do you mean18:55
* yoctozepto bbl18:55
kevkocomminication about that logger with author18:56
kevkojust to record that there should be a progress in short (hoping) period of time18:57
*** e0ne has quit IRC18:59
*** ianw_pto is now known as ianw18:59
*** kevko has quit IRC19:12
*** kevko has joined #openstack-kolla19:12
*** andrewbonney has quit IRC19:16
*** kevko has quit IRC19:25
yoctozeptono stress19:27
*** e0ne has joined #openstack-kolla19:30
*** e0ne has quit IRC19:32
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible master: Rename role haproxy to loadbalancer  https://review.opendev.org/c/openstack/kolla-ansible/+/77061819:33
*** kevko has joined #openstack-kolla19:39
openstackgerritJames Kirsch proposed openstack/kolla-ansible master: Add support for LetsEncrypt-managed certs  https://review.opendev.org/c/openstack/kolla-ansible/+/74134019:50
*** maharg101 has joined #openstack-kolla19:51
*** maharg101 has quit IRC19:56
ozzzoI'm experimenting with changing neutron defaults. I edited /etc/kolla/config/neutron.conf and set "quota_port = 5" under [quotas] and then ran "kolla-ansible -i multinode deploy --tags neutron" but that didn't do anything to my neutron containers. What tags do I need to specify, to make it reconfigure neutron?20:02
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible master: Add proxysql support for database  https://review.opendev.org/c/openstack/kolla-ansible/+/77021520:05
yoctozeptoozzzo: based on your message you have done everything correctly20:06
yoctozeptowhat did the kolla-ansible do?20:06
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible master: Edit services roles to support database sharding  https://review.opendev.org/c/openstack/kolla-ansible/+/77021620:07
ozzzooic what happened; I modified kolla-ansible for the adjutant install. I need to put the default one back to work on neutron20:07
openstackgerritRadosław Piliszek proposed openstack/kolla-ansible master: [CI] Enable proxysql test  https://review.opendev.org/c/openstack/kolla-ansible/+/77062120:07
*** cah_link has quit IRC20:09
yoctozeptokevko: rebased the whole stack20:10
kevkoyoctozepto: i'm curious what it will do :)20:10
kevkoyoctozepto: if some service providing crontab ..what is best practise in k-a ? have it in services's container ..or in cron container20:11
openstackgerritMerged openstack/kolla stable/victoria: Bump up rally and rally-openstack  https://review.opendev.org/c/openstack/kolla/+/77223520:11
openstackgerritMerged openstack/kolla stable/ussuri: Bump up rally and rally-openstack  https://review.opendev.org/c/openstack/kolla/+/77239620:11
openstackgerritMerged openstack/kayobe stable/train: CI: add dummy1 port for seed upgrade and VM jobs  https://review.opendev.org/c/openstack/kayobe/+/76418420:11
yoctozeptokevko: per container20:12
kevkook20:12
yoctozeptokevko: I hope it passes20:12
yoctozeptoif not, I will try to fix it20:12
yoctozeptoas that would be me who broke it20:12
yoctozeptothehe20:12
kevkook :D20:12
* kevko will be checking whole nigt 20:13
kevko:D20:13
yoctozeptooh geez20:13
yoctozeptoi am very sorry20:13
yoctozeptook, leaving now20:14
yoctozeptohave fun and good night :D20:14
*** stackedsax has quit IRC20:14
kevkobye bye20:14
kevkogoog night20:14
*** r3ap3r has quit IRC20:14
*** jopdorp_ has quit IRC20:15
*** r3ap3r has joined #openstack-kolla20:16
*** stackedsax has joined #openstack-kolla20:16
*** murphyslawbbs has joined #openstack-kolla20:28
yoctozeptobed time delayed20:31
yoctozeptohttps://blog.qualys.com/vulnerabilities-research/2021/01/26/cve-2021-3156-heap-based-buffer-overflow-in-sudo-baron-samedit20:31
openstackgerritPierre Riteau proposed openstack/kayobe stable/train: Fix overcloud host image resolv workaround on CentOS 8.3  https://review.opendev.org/c/openstack/kayobe/+/77124620:35
openstackgerritMerged openstack/kolla stable/train: CI: revert to public package mirrors after build  https://review.opendev.org/c/openstack/kolla/+/77009020:35
openstackgerritMerged openstack/kolla stable/train: nova-compute: provide needed deps for debian/aarch64  https://review.opendev.org/c/openstack/kolla/+/77199420:35
openstackgerritMerged openstack/kolla stable/train: nova-compute: 'pmdk-tools' is x86-64 only in Ubuntu 'bionic'  https://review.opendev.org/c/openstack/kolla/+/77192520:36
yoctozeptooh, the one time I love that I have selinux and the exploit is unable to run20:36
yoctozeptoI mean, unable to get "real root"20:36
r3ap3ryoctozepto: doesn't kolla-ansible disable SElinux though? ^20:44
r3ap3rsorry, I know it is late for you...20:44
*** brinzhang_ has joined #openstack-kolla20:44
yoctozeptor3ap3r: our users never have any local access to openstack hosts20:45
yoctozeptor3ap3r: k-a indeed disables it20:46
r3ap3rDoesn't mean "somebody" doesn't. ;-)20:46
yoctozeptor3ap3r: well, then I failed at securing anyway so I don't mind :-)20:46
r3ap3rhaha20:47
yoctozeptoyou get the point20:47
*** brinzhang has quit IRC20:47
r3ap3rYeah, I gotcha. Always a way. I was just poking fun at you. Have a good night. :-)20:47
openstackgerritMerged openstack/kolla stable/ussuri: base: install setuptools to fix build on aarch64  https://review.opendev.org/c/openstack/kolla/+/77182520:48
openstackgerritMerged openstack/kolla master: openstack-base: move some packages from child images  https://review.opendev.org/c/openstack/kolla/+/77118120:48
openstackgerritMerged openstack/kolla master: Fix Mistral source images to respect upper-constraints  https://review.opendev.org/c/openstack/kolla/+/77220520:48
yoctozeptor3ap3r: good night :-)20:48
*** k_mouza has joined #openstack-kolla21:06
*** hjensas|afk is now known as hjensas21:09
*** k_mouza has quit IRC21:10
*** k_mouza has joined #openstack-kolla21:25
ozzzohow can I display quota usage on CLI?21:28
*** k_mouza has quit IRC21:30
*** jbadiapa has quit IRC22:06
openstackgerritPierre Riteau proposed openstack/kayobe master: Adapt to changes from the Bifrost Victoria release  https://review.opendev.org/c/openstack/kayobe/+/77260722:20
*** k_mouza has joined #openstack-kolla22:26
openstackgerritPierre Riteau proposed openstack/kayobe master: Adapt to changes from the Bifrost Victoria release  https://review.opendev.org/c/openstack/kayobe/+/77260722:29
*** k_mouza has quit IRC22:31
openstackgerritPierre Riteau proposed openstack/kayobe master: Adapt to changes from the Bifrost Victoria release  https://review.opendev.org/c/openstack/kayobe/+/77260722:36
openstackgerritPierre Riteau proposed openstack/kayobe master: [DNM] Work in progress of multiple host image support  https://review.opendev.org/c/openstack/kayobe/+/77260922:48
*** kevko has quit IRC22:58
*** kevko has joined #openstack-kolla22:58
*** kevko has quit IRC23:18
*** k_mouza has joined #openstack-kolla23:26
*** k_mouza has quit IRC23:30
*** jonaspaulo has quit IRC23:31
*** admin0 has quit IRC23:39
openstackgerritMerged openstack/kayobe stable/train: Fix overcloud host image resolv workaround on CentOS 8.3  https://review.opendev.org/c/openstack/kayobe/+/77124623:58
*** admin0 has joined #openstack-kolla23:59
« Monday, 2021-01-25 Index Wednesday, 2021-01-27 »

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