Friday, 2020-01-10

*** sep has quit IRC00:14
*** DanyC has quit IRC00:31
*** zbr|rover has quit IRC01:11
*** zbr has joined #openstack-ansible01:21
*** zbr has quit IRC01:24
*** zbr has joined #openstack-ansible01:27
*** zbr has quit IRC02:16
*** zbr_ has joined #openstack-ansible02:16
*** zbr_ has quit IRC02:17
*** gyee has quit IRC02:23
*** zbr has joined #openstack-ansible02:29
*** macz has joined #openstack-ansible03:24
maczhey all, im running into an issue on a pike deployment, nothing can reach the galera servers from the VIP, I get the following error: ERROR 2013 (HY000): Lost connection to MySQL server at 'reading initial communication packet', system error: 0 "Internal error/check (Not system error)"03:26
poopcat@macz, in the galera container, is there a service called 'xinetd' running?03:34
maczhi @poopcat yup its up, i see the mysqlchk config in the /etc/xinetd.d/ dir too03:35
poopcatk... Do you see a list where IPs are? on the same line, add 0.0.0.0/0 ans restart xinetd service03:36
poopcatIf that resolves it, you can add the following override03:37
poopcat'galera_monitoring_allowed_source: 0.0.0.0/0'03:38
poopcator you can scope it to the IPs you need, you can do that as well03:40
maczthanks, i tried that but i'm still getting the issue04:00
macz@poopcat does anything else need a restart? galera or haproxy?04:03
*** macz_ has joined #openstack-ansible04:07
*** macz has quit IRC04:08
*** macz_ has quit IRC04:14
*** macz has joined #openstack-ansible04:15
maczi ended up removing check port 9200 from the galera haproxy configs and it works. ill have to look into this some more, im not getting the issue on a dev env with the same configs04:16
*** hwoarang has quit IRC04:24
*** hwoarang has joined #openstack-ansible04:26
*** goldyfruit has quit IRC05:30
*** goldyfruit has joined #openstack-ansible05:30
*** evrardjp has quit IRC05:33
*** evrardjp has joined #openstack-ansible05:34
*** macz has quit IRC05:46
*** macz has joined #openstack-ansible06:04
*** macz has quit IRC06:14
*** gregwork has quit IRC06:42
*** cshen has joined #openstack-ansible06:55
*** jawad_axd has joined #openstack-ansible07:12
*** DanyC has joined #openstack-ansible07:37
*** rpittau|afk is now known as rpittau07:44
*** DanyC has quit IRC07:47
*** pcaruana has joined #openstack-ansible07:53
*** cshen has quit IRC07:56
*** cshen has joined #openstack-ansible08:01
*** tosky has joined #openstack-ansible08:12
jawad_axdHi all, morning question is, In stein centos7 linux bridge environment, firewall v1 is deprecated while firewall v2 is not supported with linux bridge, so which option do we have in specified environment for Firewall ?08:30
*** jbadiapa has joined #openstack-ansible08:43
*** pcaruana has quit IRC08:46
*** cshen has quit IRC08:49
*** DanyC has joined #openstack-ansible08:53
*** pcaruana has joined #openstack-ansible08:53
*** DanyC has quit IRC08:53
*** cshen has joined #openstack-ansible08:59
*** DanyC has joined #openstack-ansible08:59
*** sep has joined #openstack-ansible09:00
*** cshen has quit IRC09:03
*** zbr is now known as zbr|rover09:19
*** cshen has joined #openstack-ansible10:04
*** aedc has joined #openstack-ansible10:58
*** aedc has quit IRC11:03
*** rpittau is now known as rpittau|bbl11:21
*** ansmith_ has joined #openstack-ansible11:29
*** ansmith has quit IRC11:32
noonedeadpunkjawad_axd: sorry, not really an expert... Maybe jamesdenton can help you to answer this question...11:47
*** sshnaidm is now known as sshnaidm|off11:52
*** cshen has quit IRC12:20
*** pcaruana has quit IRC12:34
*** DanyC has quit IRC12:34
*** pcaruana has joined #openstack-ansible12:38
*** mgariepy has joined #openstack-ansible12:41
*** DanyC has joined #openstack-ansible12:43
*** DanyC has quit IRC12:44
*** DanyC has joined #openstack-ansible12:53
*** rpittau|bbl is now known as rpittau12:55
*** goldyfruit has quit IRC13:05
*** goldyfruit has joined #openstack-ansible13:06
*** openstackgerrit has joined #openstack-ansible13:09
openstackgerritDmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Add image for manila test  https://review.opendev.org/67465313:09
mgariepys13:09
*** trident has quit IRC13:13
*** trident has joined #openstack-ansible13:15
*** aedc has joined #openstack-ansible13:35
*** ansmith has joined #openstack-ansible13:40
*** ansmith_ has quit IRC13:40
*** aedc has quit IRC13:51
*** cshen has joined #openstack-ansible13:53
openstackgerritDmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Update ansible to 2.9.1  https://review.opendev.org/68100314:15
openstackgerritDmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Update ansible to 2.9.2  https://review.opendev.org/68100314:15
mnasernoonedeadpunk: so wrt to manila, i was thinking of a way we can test it in a simpler way for a more popular likely use case14:19
mnasernoonedeadpunk: im thinking instead of using the hosted image one, why dont we use the ceph job, deploy cephfs and use the cephfs+nfs/ganesha driver by default? -- https://docs.openstack.org/manila/latest/admin/cephfs_driver.html14:20
mnaserspecifically the "CephFS NFS shares" here, not the native shares14:20
noonedeadpunkI kinda adjusted to use ceph https://review.opendev.org/#/c/674653/14:20
noonedeadpunkBut I thought that we need to use their image anyway?14:21
mnasernoonedeadpunk: if you're using ceph native, i dont think its needed anymroe.  i think you only need the image if you're using a specific driver14:22
* mnaser looks for it14:22
noonedeadpunkOh, I think that makes sense14:22
noonedeadpunkAnyway I'm trying now to deploy aio, and weirdly fail while deploying ceph...14:23
*** theintern_ has joined #openstack-ansible14:23
mnasernoonedeadpunk: interesting read here -- https://object-storage-ca-ymq-1.vexxhost.net/swift/v1/6e4619c416ff4bd19e1c087f27a43eea/www-assets-prod/presentation-media/What-the-heck-DHSS.pdf14:23
openstackgerritDmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: Drop unnecessary sections  https://review.opendev.org/67593414:23
mnaserso we can decide to use dhss or not14:24
tbarronnoonedeadpunk: mnaser that service image was used for the "generic" driver for nova VMs, but also in some scenario tests for VMs that mount shares.14:24
noonedeadpunkI think ceph native do not use it14:24
noonedeadpunkhttps://docs.openstack.org/manila/latest/contributor/development-environment-devstack.html#sample-local-conf-files-that-get-you-started14:24
mnasertbarron: ah i see, how's your luck with running nested virt in CI?14:24
tbarronnoonedeadpunk: mnaser: you probably dont' need it for ceph or ceph-nfs unless you are running those scenario tests (which14:25
tbarronI'd like to redo with contqiner based mounts anyways to be lighter weight)14:25
mnasertbarron: a k8s driver? that'd be neat.14:25
tbarronnoonedeadpunk: mnaser: and cephfs back ends (with or without nfs) are dhss=false14:25
tbarronsome day we may do a dhss=true variant but not yet14:26
cjloadermnaser: I know theintern_ is trying to get manila/cephfs with ganesha working14:26
mnaserok cool, so this simplifies our stuff a lot for someone whos trying to get a ceph based deployment14:26
cjloadermight want to pick his brain too14:26
mnasertbarron: would the scenario jobs be compatible with these backends too so we can just run them?14:26
noonedeadpunktbarron: so, what tempest scenraios can be passed with cephfs native? I'm was thinking about smth basic, like test_share_basic_ops14:27
theintern_well, more that I've gotten it working. What would y'all need?14:27
theintern_also, howdy. Long time no see14:27
mnasertheintern_: oh cool, can you share the actual configs so we can update the CI so we can test that its working? afaik there's no proper CI for it right now14:27
theintern_I've got a doc that I can send to you that details how I got it installed on Stien14:27
tbarronnoonedeadpunk: i'd recommend just getting the api tempest tests working first, though called "api" they'll do CRUD of share/snapshot resources14:28
noonedeadpunkthat would be really helpfull14:28
tbarronnoonedeadpunk: just in the spirt of walk then run14:28
mnasernoonedeadpunk: i think it would be easier for us to test cephfs with nfs because thats more likely of a popular use case cause its easier to mount nfs than use ceph14:28
theintern_https://gist.github.com/Alfano93/6eccf28bdbd6a95ac594956740bc2e2e14:28
noonedeadpunktbarron: oh, I see. Just usually API tests do just api mocks, which do not represent in most cases really correctly configured service14:29
theintern_most providers are using CephFS with NFS because it isolates the ceph networks from VM users14:29
tbarronmnaser: if ou are interested in k8s and cephfs-with-nfs then there's a manila-csi-plugin referenced here:14:29
mnasertbarron: we already know it works fwiw, as theintern_ is using it there, we're trying to actaully get an integration test going14:29
*** cshen has quit IRC14:29
theintern_it also makes it so we don't have to provide VMs with the admin key14:29
mnasertheintern_: ++ i agree, that's why i think going to cephfs+nfs is a much better option out of the box generally14:30
tbarronhttps://github.com/shiftstack/manila-kube14:30
theintern_with K8s, the game changes a bit14:30
mnasertbarron: yeah i was thinking actually more of manila using k8s for dhss=true with an openstack deployment :p14:30
tbarronand here: https://github.com/tombarron/pocketmanilakube14:30
cjloadermnaser: sorry I saw you discussing it, was trying to help point to a user of it14:31
mnasercjloader: yeah no those docs are ultra useful from theintern_ ! if we can get those integrated into ci, EXTREMELY useful14:31
tbarronmnaser: well I'm intereseted in that too but so far that's slideware14:31
mnasertbarron: agreed :)14:31
theintern_I usually lurk in here, but I can be more active if I can be of use to y'all14:31
mnaseri think that doc is ultra useful for noonedeadpunk :>14:32
noonedeadpunk++14:32
cjloaderI saw a perfect opportunity theintern_14:34
cjloaderlol14:34
tbarrontheintern_: that's a cool doc; we use pacemaker (in tripleo context) for ganesha HA but I am *delighted* to see this lighter weight/simpler (in good sense) approach14:35
cjloaderFWIW: I used these docs to install Manila with Train mnaser noonedeadpunk14:36
*** jawad_axd has quit IRC14:36
cjloadermanila playbooks may need some work14:36
*** jawad_axd has joined #openstack-ansible14:37
cjloadertheintern_ helped me wwith it14:37
*** jawad_axd has quit IRC14:37
cjloaderwith*14:37
theintern_tbarron lol, I looked at pacemaker and honestly it scared me14:39
theintern_that's why keepalived got used14:39
mnasertheintern_: the cool thing is we already run keepalived inside OSA14:39
mnaserso in this case we can leverage that for the ganesha nfs!14:39
theintern_yup, for haproxy. That's where I copied my initial config from14:40
*** jawad_axd has joined #openstack-ansible14:41
tbarrontheintern_: we were already using pacemaker all over the place in tripleo or I liekly would have avoided it14:42
tbarronin upstream ganesha/ceph/rook jeff layton has WIP to make ganesha run active-active, etc. and have its HA more or less self-managed like the ceph daemons14:44
theintern_yea, I think in N ceph it's a first class process14:45
theintern_actually monitored by the status collecting14:45
*** jawad_axd has quit IRC14:45
theintern_the only way I can think of making ganesha active-active in its current state is to do some HAProxy magic14:46
theintern_but I haven't gone too far down that rabbit hole, so not too sure if it's possible14:47
tbarronnoonedeadpunk: so back to scenario tests.  They launch a VM to mount shares.  Cirros image NFS client doesn't work right.  But you might be able to use14:47
tbarrona lighter weight VM image than the one we have.14:47
tbarronScenario tests in manila are a pain in OpenStack CI, always running out of resoruces, etc. and failing for extraneours reasons.14:48
tbarronextraneous14:48
mnaser:<14:48
noonedeadpunkYeah, I've actually faced with that when was trying to achieve the same some time ago... Now have some time to return back to it...14:49
noonedeadpunkSo then we'll probably do at least api tests, if you say they're testing share resources14:50
tbarronnoonedeadpunk: here are the tests we're running in upstream manila for cephfs with  nfs gateway, including some scenario tests:14:50
tbarronhttps://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_781/697957/4/check/manila-tempest-minimal-dsvm-cephfs-nfs/7817237/logs/testr_results.html.gz14:50
tbarronusing this tempest.conf (which defines the manila capabilities to be tested against):14:51
tbarronhttps://zuul.opendev.org/t/openstack/build/7817237f809e4adb8f38207c4e8d84ea/log/logs/tempest_conf.txt.gz14:51
noonedeadpunkoh, this is super helpful!14:51
tbarronnoonedeadpunk: I pulled these from the manila-tempest-minimal-dsvm-cephfs-nfs job  logs in this review: https://review.opendev.org/#/c/697957/14:52
tbarronreview selection was pretty much arbitrary, it's just a review that triggers that job14:52
noonedeadpunkI was kinda deploying devstack with tempest, but I think it wasn't configured enough (maybe I've missed to define smth)....14:53
tbarronnoonedeadpunk: the local.conf used in upstream manila ci is unnecessarily complicated, the ones cited earlier are probably a better base in general14:53
tbarronupstream manila has weird legacy jobs with pre and post hook setup stuff and conventions about LONDON, PARIS back ends  etc. that are  obscure14:54
tbarronnone of that stuff is essential if you are working in another green(er) field context14:55
noonedeadpunkok, I see. Thanks for your help! I need to investigate things, as there was lot's of information here. Won't mind if I ping you in case of some questions?14:55
tbarronnoonedeadpunk: feel free to ping, and maybe also do it in #openstack-manila if you don't get action here right away.  There are others there who would be glad to help as well.14:56
noonedeadpunktbarron: sure! thanks again14:57
theintern_if y'all could keep me in the loop as well, I'll try and help as best I can15:02
*** macz has joined #openstack-ansible15:02
noonedeadpunktheintern_: much appreciated :)15:04
cjloaderI think between tbarron and theintern_ we should be able to setup manila testing in OSA15:07
cjloader(knowledge wise)15:07
*** theintern_ has quit IRC15:08
*** DanyC has quit IRC15:09
*** theintern_ has joined #openstack-ansible15:19
*** zul has joined #openstack-ansible15:21
*** DanyC has joined #openstack-ansible15:42
*** DanyC has quit IRC15:48
openstackgerritMerged openstack/openstack-ansible-memcached_server master: Document access memcached via haproxy  https://review.opendev.org/70057015:55
openstackgerritMerged openstack/openstack-ansible-os_octavia master: tox: Keeping going with docs  https://review.opendev.org/69067416:08
*** jawad_axd has joined #openstack-ansible16:08
*** jawad_axd has quit IRC16:13
openstackgerritMerged openstack/openstack-ansible-os_congress master: tox: Keeping going with docs  https://review.opendev.org/69064316:20
openstackgerritMerged openstack/openstack-ansible-os_tempest master: tox: Keeping going with docs  https://review.opendev.org/69068616:21
openstackgerritMerged openstack/openstack-ansible-os_placement master: tox: Keeping going with docs  https://review.opendev.org/69064616:22
openstackgerritMerged openstack/openstack-ansible-os_trove master: tox: Keeping going with docs  https://review.opendev.org/69063916:22
openstackgerritMerged openstack/openstack-ansible-rabbitmq_server master: tox: Keeping going with docs  https://review.opendev.org/69065116:22
openstackgerritMerged openstack/openstack-ansible-repo_server master: tox: Keeping going with docs  https://review.opendev.org/69066316:23
openstackgerritMerged openstack/openstack-ansible-os_ceilometer master: tox: Keeping going with docs  https://review.opendev.org/69062716:23
*** DanyC has joined #openstack-ansible16:29
*** jawad_axd has joined #openstack-ansible16:44
openstackgerritMerged openstack/openstack-ansible-os_glance master: tox: Keeping going with docs  https://review.opendev.org/69064516:47
*** jawad_axd has quit IRC16:49
openstackgerritMerged openstack/openstack-ansible-os_neutron master: tox: Keeping going with docs  https://review.opendev.org/69068416:51
*** gyee has joined #openstack-ansible16:56
*** rpittau is now known as rpittau|afk17:10
openstackgerritDmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Update ansible to 2.8.7  https://review.opendev.org/70201317:16
*** jawad_axd has joined #openstack-ansible17:20
openstackgerritMerged openstack/openstack-ansible-os_zun master: tox: Keeping going with docs  https://review.opendev.org/69063817:20
*** jawad_axd has quit IRC17:25
*** evrardjp has quit IRC17:33
*** evrardjp has joined #openstack-ansible17:34
*** errr_ has quit IRC17:38
openstackgerritDmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: Add manila CI  https://review.opendev.org/67593418:02
*** pcaruana has quit IRC18:09
*** theintern_ has quit IRC18:09
*** DanyC has quit IRC18:12
*** theintern_ has joined #openstack-ansible18:34
openstackgerritMerged openstack/openstack-ansible master: Deploy heat with magnum  https://review.opendev.org/70123318:41
*** aedc has joined #openstack-ansible18:51
*** pcaruana has joined #openstack-ansible18:57
*** citil has joined #openstack-ansible19:18
citilI'm following https://docs.openstack.org/project-deploy-guide/openstack-ansible/train/overview.html and using a setup similar to Production Environment. setup-openstack fails while deploying the first compute node. Infra nodes seem to work fine. Failure in nova-status upgrade check to validate a healthy configuration task. The upgrade check fails19:21
citilfor "Placement API endpoint not found".Do I need to configure anything special for placement that's missing in the docs? Please point me in the right direction.19:21
*** theintern_ has quit IRC19:33
jrossercitil: check that you have a host for the placement service defined - thats missing from some of the example configs and should be a bit like this https://review.opendev.org/#/c/701548/19:34
citiljrosser: thanks much. I'll add it now and try again. Appreciate it.19:47
*** macz has quit IRC19:51
*** jawad_axd has joined #openstack-ansible20:15
*** jawad_axd has quit IRC20:20
*** jawad_axd has joined #openstack-ansible20:36
*** jawad_axd has quit IRC20:41
*** jawad_axd has joined #openstack-ansible20:57
*** jawad_axd has quit IRC21:01
*** hwoarang has quit IRC21:16
*** hwoarang has joined #openstack-ansible21:16
*** DanyC has joined #openstack-ansible21:39
*** gokhani has quit IRC21:43
*** ahosam has joined #openstack-ansible21:57
openstackgerritMerged openstack/openstack-ansible master: [doc] Add placement definition to config sample  https://review.opendev.org/70154822:01
openstackgerritMerged openstack/openstack-ansible-ops master: Updates to Grafana role  https://review.opendev.org/70135022:04
*** ivve has quit IRC22:26
openstackgerritMerged openstack/openstack-ansible-os_horizon stable/rocky: Fix panels enable/disable for distro install.  https://review.opendev.org/69280422:29
*** KeithMnemonic has quit IRC22:46
*** DanyC has quit IRC22:46
*** pcaruana has quit IRC23:35
*** jbadiapa has quit IRC23:35

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!