Monday, 2020-11-23

« Sunday, 2020-11-22 Index Tuesday, 2020-11-24 »
*** kevko has joined #openstack-kolla00:00
*** evrardjp has quit IRC00:05
*** evrardjp has joined #openstack-kolla00:15
*** devfaz_ has joined #openstack-kolla00:27
*** zzzeek has quit IRC00:28
*** devfaz has quit IRC00:29
*** zzzeek has joined #openstack-kolla00:32
-openstackstatus- NOTICE: Our Gerrit upgrade maintenance has concluded successfully; please see the maintenance wrap-up announcement for additional details: http://lists.opendev.org/pipermail/service-announce/2020-November/000014.html00:54
*** kevko has quit IRC01:07
*** ivan_lin has quit IRC01:39
*** LinPeiWen46 has quit IRC01:40
*** zzzeek has quit IRC02:53
*** zzzeek has joined #openstack-kolla02:54
*** ricolin has quit IRC03:07
*** sorin-mihai has quit IRC03:08
*** LinPeiWen97 has joined #openstack-kolla03:11
*** dave-mccowan has quit IRC03:17
*** skramaja has joined #openstack-kolla04:06
*** evrardjp has quit IRC05:33
*** evrardjp has joined #openstack-kolla05:33
*** ricolin has joined #openstack-kolla05:42
*** cah_link has joined #openstack-kolla06:49
*** abdysn has joined #openstack-kolla06:56
*** rpittau|afk is now known as rpittau07:29
*** openstackgerrit has joined #openstack-kolla07:46
openstackgerritwu.chunyang proposed openstack/kolla-ansible master: add  prepare task  for octavia upgrade  https://review.opendev.org/c/openstack/kolla-ansible/+/74600507:46
openstackgerritwu.chunyang proposed openstack/kolla-ansible master: add  prepare task  for octavia upgrade  https://review.opendev.org/c/openstack/kolla-ansible/+/74600507:50
*** kevko has joined #openstack-kolla07:59
kevkoyoctozepto: hi, regarding my review about maxscale image, what do you think by this ? "So you now only need to add the Ubu/Deb keys to the comments as well ? "   << I removed keys for deb/ubuntu and remained only one key :/ ...08:01
yoctozeptokevko: hi, yes08:03
yoctozeptokevko: but all should be mentioned in the comments above array08:03
*** bengates has joined #openstack-kolla08:07
kevkoyoctozepto: ah, ok, i see it now ..08:08
kevkouploading08:08
openstackgerritMichal Arbet proposed openstack/kolla master: Add maxscale image  https://review.opendev.org/c/openstack/kolla/+/76363408:08
kevkoyoctozepto: one more thing, I have working maxscale in DEV now, I've alredy done some work in kolla-ansible ...but I don't know how to deal with keepalived ... or let's say I don't know what do you prefer guys.. Now keepalived is installed in role haproxy, and only haproxy is using keepalived now ...08:10
kevkoIf I consider that maxscale can be running on other set of machines .. I have to add some VIP and start new keepalived_maxscale container (in case it will be runnning on same host where keepalived running now) ... OR rework check in keepalived to check also maxscale ...and start maxscale containers always next to haproxy through inventory ... For me all ways are "half-ways" ... what do you mean ?08:13
kevkomaybe the best way is to create new VIP, add check to keepalived image for maxscale ..and start container with different name .. This has also benefits - DB traffic is always tageting different VIP - let's say database_internal_vip , and I am able to separate these things to separate hosts ...08:16
yoctozeptokevko: using another vip address sounds good to me08:16
openstackgerritLin PeiWen proposed openstack/kolla-ansible master: Delete the default parameters of the gluster NFS driver  https://review.opendev.org/c/openstack/kolla-ansible/+/76374608:16
kevkook thanks08:20
yoctozeptokevko: so, essentially, it would make most sense to refactor the role to spawn keepalived08:20
yoctozeptoI wonder if it is better to use a single daemon with multiple configs08:21
yoctozeptoor multiple daemons08:21
kevkowell, one containr08:22
kevkoone container with multiple configs sounds better ...08:22
kevkobut you can have maxscale on another set of machines while haproxy on another ..so ..you have to have two containers with two configs :/08:23
yoctozeptokevko: you can have different configs on different machines08:24
yoctozeptojust need to consider more than one keepalived ring08:24
kevkoyes i know08:24
*** luksky has joined #openstack-kolla08:24
kevkook, so I have to expect that I can schedule more than one container per host .. and inject config for every container ..08:38
yoctozeptokevko: I'd say keep to at most one, but be able to merge config08:40
yoctozeptoso that one process is handling more than one vip address08:40
yoctozeptoI just don't remember the keepalived recommendation on that08:40
yoctozeptoperhaps they do recommend to use separate ones08:40
*** dougsz has joined #openstack-kolla08:49
kevkoyoctozepto: understand ! ok , going to work on that :) , btw, maxscale image in kolla is now OK please ?08:55
*** mgoddard has joined #openstack-kolla08:58
yoctozeptokevko: commented; only alphabetic sort; we still want to see it in action  in k-a before merging the image itself09:00
kevkounderstand09:01
*** dardelean has joined #openstack-kolla09:04
*** e0ne has joined #openstack-kolla09:19
*** dougsz has quit IRC09:22
*** soumplis has joined #openstack-kolla09:26
hrwmorning09:28
hrwkevko: maxscale image failed to build on aarch6409:29
kevkohi09:29
kevkobut it is non-voting, isn't it ? :)09:30
soumplisHi all, does anyone has experience with the merge_config ansible module that it is heavily used in kolla-ansible ? I try to debug a case where a ceph.conf file is created empty.09:30
hrwkevko: hehehe09:30
kevko:D09:31
hrwkevko: if it does not build then it should be in UNBUILDABLE from start09:31
* hrw -> gerrit docs to check how to ignore 09:31
kevkook, didn't know it .. will add to unbuildable ..09:31
kevkosorry, i don't have machines with aarch64 :(09:32
kevkobut, maxscale is supporting some distros ...09:32
*** abdysn has quit IRC09:32
kevkobut haven't test it09:32
hrwkevko: you add something, you look at CI09:33
kevkoof course, sorry for it, i have currently several terminals open and working on several changes ..09:34
hrwsorry to shout on you ;D09:34
hrwtrying to findout why qemu does not install as well09:35
hrwuf. more backports09:36
*** wuchunyang has joined #openstack-kolla09:36
mgoddardmorning09:38
*** dougsz has joined #openstack-kolla09:38
kevkomorning09:38
mgoddardsoumplis: if the file is created empty, it probably means that you have not provided ceph.conf in the correct location09:40
hrwI have a patch which should make aarch64 nova-{libvirt,compose} buildable again09:41
openstackgerritMark Goddard proposed openstack/kolla master: WIP: CI: revert to public package mirrors after build  https://review.opendev.org/c/openstack/kolla/+/76192809:42
soumplismgoddard: I have checked with ansible and the node_custom_config has the correct value. I suppose that the value of this variable (changed from the default) is ignored when the module is running09:43
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: base: fix QEMU dependencies for Debian/aarch64  https://review.opendev.org/c/openstack/kolla/+/76375709:45
mgoddardsoumplis: it should not be ignored09:48
mgoddardsoumplis: suggest adding a debug task that shows the expected path of the file09:48
hrwplease review ^^09:49
*** ricolin has quit IRC09:53
*** k_mouza has joined #openstack-kolla09:59
*** gfidente has joined #openstack-kolla09:59
hrwit built!10:12
hrwso RP bumped ;D10:14
*** wuchunyang has quit IRC10:17
openstackgerritRiccardo Pittau proposed openstack/kayobe master: Fix ansible-lint star expansion  https://review.opendev.org/c/openstack/kayobe/+/76376210:22
mgoddardWallaby priorities now decided: http://lists.openstack.org/pipermail/openstack-discuss/2020-November/018972.html10:35
mgoddardWhiteboard updated to match10:35
mgoddardFeel free to pick one up10:35
* yoctozepto owning running CI jobs locally10:43
*** abdysn has joined #openstack-kolla10:44
openstackgerritMark Goddard proposed openstack/kolla master: DNM: testing mirror revert  https://review.opendev.org/c/openstack/kolla/+/76335110:49
*** muhaha has joined #openstack-kolla10:49
openstackgerritMerged openstack/kayobe-config master: Update TOX_CONSTRAINTS_FILE  https://review.opendev.org/c/openstack/kayobe-config/+/76333810:59
openstackgerritMerged openstack/kayobe-config-dev master: Update TOX_CONSTRAINTS_FILE  https://review.opendev.org/c/openstack/kayobe-config-dev/+/76333910:59
*** wuchunyang has joined #openstack-kolla11:08
openstackgerritlikui proposed openstack/kolla-ansible master: Add enable_manila_backend_lvm support  https://review.opendev.org/c/openstack/kolla-ansible/+/74278211:17
openstackgerritMark Goddard proposed openstack/kolla-ansible master: docs: Add magnum guide  https://review.opendev.org/c/openstack/kolla-ansible/+/76377011:20
hrwmgoddard: I am planning to mark bifrost unbuildable on aarch64. Just to get rid of that @$@#%@^@11:30
mgoddardhrw: sure11:30
hrwINFO:kolla.common.utils.bifrost-base:    "msg": "No package matching 'grub-efi-amd64-signed' is available"11:31
hrwI do not plan to fix it11:31
mnasiadkamorning11:33
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: bifrost: mark as UNBUILDABLE for aarch64  https://review.opendev.org/c/openstack/kolla/+/76377111:34
hrwnot reporting issue upstream even11:34
mnasiadkageez, even gerrit links got changed11:35
*** e0ne has quit IRC11:36
mgoddardmnasiadka: yeah, backwards compatible though11:38
mnasiadkamgoddard: I still have a problem finding anything :)11:39
openstackgerritwu.chunyang proposed openstack/kolla-ansible master: add  upgrade tasks  for octavia  https://review.opendev.org/c/openstack/kolla-ansible/+/74600511:41
hrwworks same for me11:42
hrwUI change does not bother me as I used that UI in other projects already11:42
hrwand enabled numbers in lists to make it easier to look for thsoe11:43
openstackgerritMark Goddard proposed openstack/kayobe master: DNM: debug seed-upgrade job  https://review.opendev.org/c/openstack/kayobe/+/76377311:48
openstackgerritLin PeiWen proposed openstack/kolla-ansible master: Delete the default parameters of the gluster NFS driver  https://review.opendev.org/c/openstack/kolla-ansible/+/76374611:50
kevkowhere can I see failed tests in new gerrit ?11:57
kevkooh, i am blind :D11:57
mnasiadkano worries, I'm blind as well11:57
kevkono, I am :D , i can't see it :D11:58
hrwkevko: in comments11:58
hrwuserscript I use needs adaptations ;(11:58
kevkovrr, don't like new look, it's not better than before :(11:58
openstackgerritMichal Arbet proposed openstack/kolla master: Add maxscale image  https://review.opendev.org/c/openstack/kolla/+/76363411:59
kevkohrw: thank you12:00
*** jonaspaulo has joined #openstack-kolla12:01
openstackgerritwu.chunyang proposed openstack/kolla-ansible master: fix a ansible warnning  https://review.opendev.org/c/openstack/kolla-ansible/+/76377512:03
*** wuchunyang has quit IRC12:05
openstackgerritPierre Riteau proposed openstack/kolla-ansible master: Fix an Ansible warning  https://review.opendev.org/c/openstack/kolla-ansible/+/76377512:11
*** jonaspaulo has quit IRC12:14
*** jonaspaulo has joined #openstack-kolla12:15
*** Luzi has joined #openstack-kolla12:25
openstackgerritMarcin Juszkiewicz proposed openstack/kolla master: bifrost: mark as UNBUILDABLE for aarch64  https://review.opendev.org/c/openstack/kolla/+/76377112:34
hrwrebased12:34
*** jonaspaulo has quit IRC13:00
*** happyhemant has joined #openstack-kolla13:02
*** soumplis has quit IRC13:03
*** kevko_ has joined #openstack-kolla13:17
*** klippo has quit IRC13:17
*** e0ne has joined #openstack-kolla13:18
*** kevko has quit IRC13:20
*** kevko_ has quit IRC13:24
*** Luzi has quit IRC13:29
*** kevko has joined #openstack-kolla13:30
*** cah_link has quit IRC13:31
*** cah_link has joined #openstack-kolla13:32
*** kevko_ has joined #openstack-kolla13:41
*** kevko has quit IRC13:44
*** wuchunyang has joined #openstack-kolla13:59
yoctozeptohrw: yeah, I'm using the same script14:02
*** dave-mccowan has joined #openstack-kolla14:20
hrwyoctozepto: I know ;D14:24
hrwreview ^^ ;D14:32
kevko_hmm, guys , maybe I am asking stupid question ...but why in ansible/site.yml is grouping by enable to all hosts ?14:34
kevko_i know it is for &something and matching hosts ..14:34
kevko_but it makes sense more if it is applicated to specific hosts14:35
*** dave-mccowan has quit IRC14:35
kevko_as example, group_by enable mariadb to mariadb hosts14:35
kevko_not all :/14:36
*** rpittau is now known as rpittau|brb14:36
*** muhaha has quit IRC14:56
*** rpittau|brb is now known as rpittau15:01
mgoddardkevko_: if you did it for specific groups you would need a play for each group. Then you would lose the performance benefit of doing it at all15:12
openstackgerritMerged openstack/kolla master: base: fix QEMU dependencies for Debian/aarch64  https://review.opendev.org/c/openstack/kolla/+/76375715:13
kevko_mgoddard: ok, I'm implementing the maxscale thing .. maxscale shoold have keepalived ...15:14
kevko_but keepalived is not separate role ... so i can copy whole keepalived folder from haproxy ..change things ..and vuala ..15:14
kevko_but .. i have two quite similar set of files in two roles ..15:15
kevko_so, i can refactor keepalived and do separate role ..and make group by maxscale_enabled haproxy_enabled ..and apply role to hosts ..15:16
kevko_which is better way ?15:17
*** skramaja has quit IRC15:50
*** abdysn has quit IRC15:51
*** wuchunyang has quit IRC16:03
*** mgoddard has quit IRC16:12
*** jbadiapa has joined #openstack-kolla16:13
*** mgoddard has joined #openstack-kolla16:15
*** dardelean has quit IRC16:23
*** sorin-mihai has joined #openstack-kolla16:28
*** kevko_ has quit IRC16:44
*** kevko has joined #openstack-kolla16:44
*** bengates has quit IRC16:45
*** bengates has joined #openstack-kolla16:47
*** bengates has quit IRC16:51
openstackgerritMark Goddard proposed openstack/kolla master: DNM: testing mirror revert  https://review.opendev.org/c/openstack/kolla/+/76335117:00
*** rpittau is now known as rpittau|afk17:03
headphoneJameskevko: FWIW, Refactoring out keepalive makes sense to me.17:03
mgoddardkevko: unsure about keepalived. It's currenty quite tied to haproxy, and we have to coordinate restarts with haproxy17:04
mgoddardwould maxscale need a separate VIP?17:04
kevkomgoddard: no, it doesn't need ..but when i am doing something i like it to do in good way ..so if i will have maxscale ..i want to have check in keepalived to switch VIP17:05
kevkomgoddard: also , i would like to have option to schedule maxscale to other hosts then controllers ...17:06
kevkomgoddard: and don't want to add check to keepalived's haproxy check ..17:06
mgoddardso in your design maxscale has its own VIP and separate keepalived?17:07
kevkoin my working design i have maxscale on nodes where haproxy currently is ..17:07
kevkobut it is still DEV ..so now i am writing patchset to master ..and wanted to have it really HA17:07
mgoddardwhat is really HA?17:08
kevkoreally HA is that if I am running two servers on VIP ..i have to expect that one node will fail ..so VIP will switch to another host where backup maxscale is running ..17:09
kevkoas it is for haproxy of course17:09
kevkoi can also go through VIP -> haproxy -> maxscale ...but it's ugly solution ..17:10
mgoddardit would be possible to let maxscale listen on the same VIP as haproxy. That would still be HA17:11
kevkoso if i want to run for example two instances of maxscale ..i need to have option to switch to another working node ..and this is achievable when : I always will schedule maxscale to haproxy hosts ..and add check to keepalived haproxy check ..which is ugly17:11
kevkoyes, sure, but then i have to add check to keepalived's check ... which will be composed from  haproxy-OK && maxscale-OK17:12
mgoddardI'm not necessarily in favour of sharing the VIP with HAproxy, just trying to understand the trade offs17:12
kevkoand, i am limited to run on haproxy hosts .. :/17:13
mgoddardan extra VIP (& FQDN) comes with complexity17:13
kevkook, step by step17:13
kevkoExample : Two haproxy along with keepalived on ctrl0 and ctrl1, keepalived has haproxy-socket docker volume mounted and checking liveness of haproxy through socket ...17:15
kevkoOK, if I want to listen on same VIP, it is OK, but I am limited to schedule maxscale to this group of hosts ...17:16
kevkomoreover , if i want to ensure that VIP will switch after maxscale fail .. I have to add check into /check_alive.sh of keeplived container ..so one check(consist of two checks) for one vip ..17:18
mgoddardyeah, got it17:19
kevkoso, what i wanted to do is something like this :17:20
kevkorefactor out keepalived, do some common tasks .. and include tasks per haproxy, maxscale which will inject configuration inside keepalived17:21
*** cah_link has quit IRC17:22
kevkobest practise is one check per VIP17:23
kevkoI can do it the way it works by default even now ... only users which will want to switch to maxscale will have little bit more complexity ..17:25
kevkomoreover .. i think there is a precondition for the future to use keepalived as a role ... not only haproxy/keepalived17:26
kevkofor potentional other services ..17:26
*** jbadiapa has quit IRC17:33
*** e0ne has quit IRC17:33
*** dougsz has quit IRC17:35
mgoddardkevko: so how many keepalived services would there be?17:38
kevkowell, if you schedule haproxy and maxscale to same group of servers ..nothing is changed ..only one block of keepalived.conf more -> that means, one more check, one more VIP17:39
kevkoif you schedule maxscale to another group of servers ..you will have keepalived also on that group ... to fullfill HA ..17:40
kevkoi don't think it is some overkill or something :/17:41
*** kevko_ has joined #openstack-kolla17:49
*** kevko has quit IRC17:51
*** mgoddard has quit IRC17:54
*** k_mouza has quit IRC18:13
*** muhaha has joined #openstack-kolla18:32
*** kevko_ has quit IRC18:37
*** gfidente is now known as gfidente|afk18:51
openstackgerritPiotr Parczewski proposed openstack/kayobe master: Fix Python setup when venv is not used  https://review.opendev.org/c/openstack/kayobe/+/76365918:52
*** zzzeek has quit IRC18:55
*** zzzeek has joined #openstack-kolla18:57
*** muhaha has quit IRC18:59
openstackgerritMark Goddard proposed openstack/kayobe stable/victoria: DNM: debug seed-upgrade job  https://review.opendev.org/c/openstack/kayobe/+/76373419:19
openstackgerritMark Goddard proposed openstack/kayobe stable/victoria: DNM: debug seed-upgrade job  https://review.opendev.org/c/openstack/kayobe/+/76373419:20
openstackgerritMark Goddard proposed openstack/kayobe master: DNM: debug seed-upgrade job  https://review.opendev.org/c/openstack/kayobe/+/76377319:21
*** mgoddard has joined #openstack-kolla19:26
*** zzzeek has quit IRC19:27
*** zzzeek has joined #openstack-kolla19:29
*** hjensas has quit IRC19:31
*** hjensas has joined #openstack-kolla19:31
*** kevko_ has joined #openstack-kolla19:33
*** dougsz has joined #openstack-kolla19:37
*** mgoddard has quit IRC19:38
*** dougsz has quit IRC19:52
*** zzzeek has quit IRC19:53
*** zzzeek has joined #openstack-kolla19:55
-openstackstatus- NOTICE: The Gerrit service on review.opendev.org is being restarted quickly to troubleshoot an unexpected error condition, downtime should be less than 5 minutes20:00
*** openstackgerrit has quit IRC20:02
*** k_mouza has joined #openstack-kolla20:14
*** k_mouza has quit IRC20:19
*** zzzeek has quit IRC20:30
*** zzzeek has joined #openstack-kolla20:31
*** sorin-mihai has quit IRC20:34
*** sorin-mihai has joined #openstack-kolla20:47
*** sorin-mihai has quit IRC20:53
*** sorin-mihai has joined #openstack-kolla20:53
*** kevko_ has quit IRC21:12
*** happyhemant has quit IRC21:21
*** gfidente|afk has quit IRC21:23
*** sorin-mihai has quit IRC21:37
*** sorin-mihai has joined #openstack-kolla21:38
*** sorin-mihai has quit IRC22:01
*** sorin-mihai has joined #openstack-kolla22:02
*** luksky has quit IRC22:05
*** sorin-mihai has quit IRC22:08
*** sorin-mihai has joined #openstack-kolla22:10
*** sorin-mihai has quit IRC22:12
*** luksky has joined #openstack-kolla22:18
*** waxfire has quit IRC23:05
*** waxfire has joined #openstack-kolla23:05
*** luksky has quit IRC23:10
« Sunday, 2020-11-22 Index Tuesday, 2020-11-24 »

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