| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_nova: Delegate from first available compute host https://review.openstack.org/385221 | 00:04 |
|---|---|---|
| *** sdake_ has joined #openstack-ansible | 00:11 | |
| *** sdake has quit IRC | 00:14 | |
| openstackgerrit | Steve Lewis (stevelle) proposed openstack/openstack-ansible-os_ceilometer: Update ceilometer app.wsgi template per upstream https://review.openstack.org/385697 | 00:14 |
| *** agrebennikov has quit IRC | 00:15 | |
| *** hj-hpe has quit IRC | 00:16 | |
| *** jamesdenton has joined #openstack-ansible | 00:18 | |
| cloudnull | gotta run bbl | 00:20 |
| cloudnull | logan-: still looking into issue btw | 00:20 |
| cloudnull | thanks for the acces. | 00:20 |
| cloudnull | *access. | 00:21 |
| *** aludwar has quit IRC | 00:22 | |
| *** adrian_otto has quit IRC | 00:26 | |
| *** thorst has joined #openstack-ansible | 00:28 | |
| *** aludwar has joined #openstack-ansible | 00:29 | |
| *** javeriak has quit IRC | 00:33 | |
| *** javeriak has joined #openstack-ansible | 00:34 | |
| *** jamesdenton has quit IRC | 00:36 | |
| *** markvoelker has joined #openstack-ansible | 00:40 | |
| *** javeriak has quit IRC | 00:41 | |
| *** markvoelker has quit IRC | 00:45 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_nova: Delegate from first available compute host https://review.openstack.org/385221 | 00:45 |
| *** adrian_otto has joined #openstack-ansible | 00:47 | |
| *** cmart has quit IRC | 00:50 | |
| *** jamesdenton has joined #openstack-ansible | 01:05 | |
| *** fishcried has joined #openstack-ansible | 01:07 | |
| *** rmelero has quit IRC | 01:10 | |
| *** markvoelker has joined #openstack-ansible | 01:12 | |
| *** cmart has joined #openstack-ansible | 01:12 | |
| *** cmart has quit IRC | 01:17 | |
| mhayden | odyssey4me: got it sir | 01:21 |
| *** adrian_otto has quit IRC | 01:30 | |
| *** adrian_otto has joined #openstack-ansible | 01:36 | |
| *** adrian_otto has quit IRC | 01:38 | |
| *** adrian_otto has joined #openstack-ansible | 01:40 | |
| logan- | cloudnull: sure np ill hang on to those nodes till youre done. no rush | 01:45 |
| *** thorst has quit IRC | 01:49 | |
| *** thorst has joined #openstack-ansible | 01:50 | |
| *** Mudpuppy_ has joined #openstack-ansible | 01:52 | |
| *** Mudpuppy has quit IRC | 01:53 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible: Configure Calico specific BIRD settings in OSA https://review.openstack.org/385710 | 01:55 |
| *** thorst has quit IRC | 01:58 | |
| *** adrian_otto has quit IRC | 02:06 | |
| jamesdenton | odyssey4me you still around? | 02:06 |
| *** jcrst has quit IRC | 02:07 | |
| cloudnull | jamesdenton: i think he's sleep | 02:09 |
| cloudnull | something I can help with ? | 02:09 |
| jamesdenton | ha yeah, i just realized that. i didn't pay attention to timestamps :D | 02:09 |
| jamesdenton | naw, i commented on bug and just wanted his feedback. I'm sure I'll get it in the AM :) | 02:10 |
| cloudnull | ok | 02:10 |
| jamesdenton | thanks though! | 02:10 |
| cloudnull | anytime | 02:18 |
| *** LiYuenan has joined #openstack-ansible | 02:23 | |
| openstackgerrit | Merged openstack/openstack-ansible: Change the common proxy cache manage tasks to be stateful https://review.openstack.org/385625 | 02:24 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Prevent overlayfs use in test when kernel < 3.18 or release == trusty https://review.openstack.org/351776 | 02:31 |
| cloudnull | logan-: odyssey4me: after much putzing about I've determined that OverlayFS should just be avoided on trusty. | 02:32 |
| cloudnull | specifically the package libdbus-XXX is a dependency of openssh-server in trusty where it is NOT in xenial | 02:33 |
| cloudnull | i wrote some simple loops to drill deps down to everything potentially required and determined that there's really no way to get around installing that package and it's existance has issues when installed in a contianer using overlayFS | 02:34 |
| *** hughmFLEXin has joined #openstack-ansible | 02:35 | |
| *** hughmFLEXin has quit IRC | 02:39 | |
| *** thorst has joined #openstack-ansible | 02:40 | |
| cloudnull | it seems to manifest in a later kernels but the potential for combinations to break that seem like we should just stick to dir based containers on trusty even if it's running in an AIO | 02:41 |
| *** thorst has quit IRC | 02:42 | |
| *** thorst has joined #openstack-ansible | 02:42 | |
| * neillc heads to the dentist | 02:43 | |
| *** thorst has quit IRC | 02:51 | |
| *** gus has quit IRC | 02:52 | |
| logan- | makes sense cloudnull. probably not too many deployers building greenfield or new containers on n+trusty so i don't think turning it off will affect many. i expect most people's n+trusty will just be a m->n+trusty->n+xenial.. that's how i'm looking at it anyway. | 03:05 |
| logan- | has anyone used any of the various lxc connection plugins? ie https://github.com/Mic92/ansible-lxc or https://github.com/chifflier/ansible-lxc-ssh | 03:06 |
| logan- | curious how well they work and if there's any caveats | 03:07 |
| cloudnull | logan-: last i checked I think we're the only deployment project support newton on trusty | 03:11 |
| cloudnull | *supporting | 03:11 |
| cloudnull | i've looked at chifflier | 03:11 |
| cloudnull | not mic92 | 03:11 |
| cloudnull | in newton the ssh plugin we use is lxc aware | 03:12 |
| cloudnull | https://github.com/openstack/openstack-ansible-plugins/blob/master/connection/ssh.py | 03:12 |
| cloudnull | it looks and feels like normal ssh but its really ssh > host + lxc | 03:13 |
| logan- | oh nice! I didn't realize we're doing that | 03:13 |
| cloudnull | yup. that was the point | 03:13 |
| cloudnull | :) | 03:13 |
| cloudnull | mission accomplished :p | 03:13 |
| logan- | so if i'm understanding correctly i could not have an ansible_host defined for my containers, and as long as the container has a physical_host with ssh host defined I should be able to target tasks at the container | 03:15 |
| logan- | oh and container_name on the container host | 03:16 |
| *** adrian_otto has joined #openstack-ansible | 03:19 | |
| cloudnull | yup | 03:19 |
| *** adrian_otto has quit IRC | 03:19 | |
| cloudnull | ssh is no longer needed into the container. | 03:20 |
| cloudnull | we have it . but its not required from an ansible perspective | 03:20 |
| logan- | awesome | 03:20 |
| cloudnull | there are some wait conditions that still assume ssh | 03:21 |
| cloudnull | but for the work its not needed. | 03:21 |
| *** janki has joined #openstack-ansible | 03:22 | |
| *** adrian_otto has joined #openstack-ansible | 03:25 | |
| *** adrian_otto has quit IRC | 03:26 | |
| *** adrian_otto has joined #openstack-ansible | 03:27 | |
| *** gouthamr has quit IRC | 03:33 | |
| *** adrian_otto has quit IRC | 03:35 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts: Create the netns dir within the base container image https://review.openstack.org/385731 | 03:47 |
| *** thorst has joined #openstack-ansible | 03:49 | |
| *** thorst has quit IRC | 03:56 | |
| logan- | seems like maybe an issue with the plugin detecting the container triggering vars cloudnull http://cdn.pasteraw.com/imog7jq1gt8kcmfojf4b23avrco4s8w | 04:02 |
| logan- | self.vars is an empty dict | 04:03 |
| *** dragonmaster_ has joined #openstack-ansible | 04:11 | |
| *** cmart has joined #openstack-ansible | 04:12 | |
| cloudnull | logan-: http://cdn.pasteraw.com/kple5sbpqxdig2rx0gusx5t6qo3ev6d - the inventory needs a couple more options. | 04:21 |
| cloudnull | you really only need this http://cdn.pasteraw.com/1m489ospt7cya72j2yvrjkijpd392zi | 04:22 |
| cloudnull | in the inventory that is | 04:22 |
| *** hybridpollo has left #openstack-ansible | 04:22 | |
| cloudnull | the plugin does a lookup on the host connects to it and uses lxc after that. | 04:22 |
| cloudnull | note the "Container confirmed" lines | 04:23 |
| cloudnull | and the lxc-attach in the debug output | 04:24 |
| *** appprod0 has quit IRC | 04:35 | |
| *** Mudpuppy_ has quit IRC | 04:43 | |
| *** Mudpuppy has joined #openstack-ansible | 04:44 | |
| *** Mudpuppy has quit IRC | 04:48 | |
| *** gtrxcb has joined #openstack-ansible | 04:52 | |
| *** thorst has joined #openstack-ansible | 04:54 | |
| *** cmart has quit IRC | 04:55 | |
| *** thorst has quit IRC | 05:01 | |
| *** Jack_Iv has joined #openstack-ansible | 05:24 | |
| *** Jack_Iv has quit IRC | 05:25 | |
| *** Jack_Iv has joined #openstack-ansible | 05:25 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Fix container log bind mount regression https://review.openstack.org/385619 | 05:28 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Change the common proxy cache manage tasks to be stateful https://review.openstack.org/385757 | 05:31 |
| *** Jack_Iv has quit IRC | 05:38 | |
| *** Jack_Iv has joined #openstack-ansible | 05:38 | |
| *** javeriak has joined #openstack-ansible | 05:56 | |
| *** thorst has joined #openstack-ansible | 06:00 | |
| *** thorst has quit IRC | 06:06 | |
| *** ZuZuD has quit IRC | 06:09 | |
| *** vishwanathj is now known as vishwanathj_zzz | 06:16 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_heat: Update paste, policy and rootwrap configurations 2016-10-12 https://review.openstack.org/385544 | 06:25 |
| *** rgogunskiy has joined #openstack-ansible | 06:25 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_aodh: Update paste, policy and rootwrap configurations 2016-10-12 https://review.openstack.org/385542 | 06:25 |
| *** dragonmaster_ is now known as allanice001 | 07:00 | |
| openstackgerrit | Merged openstack/openstack-ansible: Configure Calico specific BIRD settings in OSA https://review.openstack.org/385710 | 07:00 |
| *** gus has joined #openstack-ansible | 07:01 | |
| openstackgerrit | Praveen N proposed openstack/openstack-ansible: [docs] Add network config example for test and prod https://review.openstack.org/385272 | 07:02 |
| *** thorst has joined #openstack-ansible | 07:04 | |
| *** admin0 has joined #openstack-ansible | 07:07 | |
| *** thorst has quit IRC | 07:11 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create: Generate Static MAC addresses for containers https://review.openstack.org/385800 | 07:28 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Prevent overlayfs use in test when kernel < 3.18 or release == trusty https://review.openstack.org/351776 | 07:30 |
| *** evilt0ne has quit IRC | 07:30 | |
| *** evilt0ne has joined #openstack-ansible | 07:32 | |
| *** evilt0ne has quit IRC | 07:32 | |
| *** evilt0ne has joined #openstack-ansible | 07:32 | |
| *** drifterza has joined #openstack-ansible | 07:32 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create: Generate Static MAC addresses for containers https://review.openstack.org/385800 | 07:34 |
| *** vdo has left #openstack-ansible | 07:35 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create: Generate Static MAC addresses for containers https://review.openstack.org/385800 | 07:35 |
| *** berendt has quit IRC | 07:36 | |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 07:36 |
| *** evilt0ne has quit IRC | 07:37 | |
| *** evilt0ne has joined #openstack-ansible | 07:38 | |
| *** evilt0ne has quit IRC | 07:38 | |
| *** evilt0ne has joined #openstack-ansible | 07:38 | |
| *** gtrxcb has quit IRC | 07:38 | |
| drifterza | cloudnull, you awake dude ? | 07:38 |
| cloudnull | kinda? | 07:38 |
| cloudnull | :) | 07:38 |
| drifterza | jeeze, pull an all nighter ? | 07:38 |
| cloudnull | on the way to crash in a min | 07:38 |
| cloudnull | restless. | 07:38 |
| drifterza | Ah okay, | 07:39 |
| drifterza | in OSA can we deploy using ipv6 ? | 07:39 |
| evrardjp | hey everyone | 07:42 |
| evrardjp | ipv6 in osa is possible, depending on what you're looking for | 07:42 |
| evrardjp | if you have a ceph cluster, build it from scratch with v6 :D | 07:43 |
| cloudnull | we can use v6 networks, but if you're looking for v6 on the control plane then not yet. | 07:43 |
| drifterza | no only on consumer plane | 07:43 |
| cloudnull | we offer v6 provider networks in the OSIC | 07:44 |
| cloudnull | works great | 07:44 |
| drifterza | do we have a deployment method for that kevin ? | 07:46 |
| *** fops has quit IRC | 07:47 | |
| ioni | yesterday i saw a review with nova novnc and protouri | 07:47 |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 07:47 |
| ioni | i cannot ind it anymore | 07:47 |
| *** fops has joined #openstack-ansible | 07:47 | |
| ioni | i had it on chat but i rebooted it and i don't have the logs enabled | 07:47 |
| ioni | do you happen to know the review id? | 07:48 |
| cloudnull | drifterza: https://twitter.com/cloudnull/status/764475358934544384 | 07:49 |
| cloudnull | its not really anything special . | 07:49 |
| ioni | what's osic? | 07:49 |
| drifterza | would be really cool if we could write plays to pin up the ipv6 part | 07:50 |
| cloudnull | ioni: its a collaboration between intel and rackspace | 07:51 |
| evrardjp | cloudnull: for v6 in the control plane, it's not really different, we had a dual stack control plane in my previous company | 07:51 |
| evrardjp | haproxy needed the rewrite I already shown you, and keepalived was with its own config file, but the rest is kinda ok | 07:51 |
| evrardjp | we implemented weird things here and there, but that's only specific to our install | 07:52 |
| *** fops has quit IRC | 07:52 | |
| evrardjp | but we didn't really move 100% of the traffic to ipv6 to be honest | 07:53 |
| cloudnull | drifterza: from an OSA perspective we simply added the networks into neutron and away it went. | 07:53 |
| cloudnull | the entire config is in that pic. | 07:53 |
| cloudnull | while i love to pick on neutron its fairly robust. | 07:53 |
| drifterza | certainly you need some advertising method for ipv6 | 07:53 |
| cloudnull | for sure. | 07:53 |
| drifterza | oh so it assumes thats in place | 07:54 |
| cloudnull | yes. it does. | 07:54 |
| cloudnull | and we used hardware to do it . | 07:54 |
| drifterza | hmmm | 07:54 |
| cloudnull | its dual stack 10/8 plumbed to a neutron router with a gateway using a public 172/24 | 07:55 |
| evrardjp | cloudnull: you use one provider network right, you didn't go for prefix delegation with one net per tenant, right? | 07:55 |
| drifterza | I thought you needed something like ipv6 radvd | 07:55 |
| jrosser | v6 control plane very interesting for me | 07:56 |
| evrardjp | drifterza: it's included in neutron depending on what you configure/need | 07:56 |
| cloudnull | we did not do prefix delegation | 07:56 |
| cloudnull | that was busted in liberty | 07:56 |
| drifterza | okay Ill read the neutron plays in more depth, we don't have a documented way to do this right | 07:56 |
| evrardjp | I'm pretty sure we hacked it together in liberty :D | 07:56 |
| evrardjp | but yes, experimental tech | 07:57 |
| evrardjp | but you don't need prefix delegation to have an ipv6 cloud for the customers | 07:57 |
| cloudnull | drifterza: https://gist.github.com/cloudnull/4e34276a8e690fb54146d03a73ec44b4 this was the config, its been changed a bit from then to now , but that's the gist (pun intended) :) | 07:57 |
| evrardjp | :D | 07:57 |
| cloudnull | from a neutron setup point of view | 07:57 |
| drifterza | shot! | 07:57 |
| drifterza | would like to get this going in my labs | 07:58 |
| evrardjp | ahah using google | 07:58 |
| cloudnull | jrosser: v6 control plane would be interesting to me too | 07:58 |
| cloudnull | I was shooting to try and make it go in this cycle, maybe the next one. | 07:59 |
| jrosser | how big is the gap | 07:59 |
| jrosser | doing v4 just digs a bigger hole for us here that i'd rather not make worse | 08:00 |
| allanice001 | hey all | 08:00 |
| evrardjp | jrosser: to be IPv6 only wouldn't require much work, it could require hacks, but I think it's doable quite easily, depending on what you really really want | 08:01 |
| cloudnull | jrosser: https://twitter.com/cloudnull/status/717519610761728000 -- i just did get enough time to make it go . :) | 08:02 |
| ioni | is there somebody that uses the last stable/newton and novnc setup? | 08:03 |
| evrardjp | cloudnull: I'd be happy to help you in this battle | 08:03 |
| ioni | i cannot type anything in the console | 08:03 |
| cloudnull | jrosser: my personal plan of attack would be to use the unbound bits that logan- proposed for proper hostname resolution, Move the the IP generation into a proper ip manager (out of the dynamic inventory script), and convert all of the internal services to use DNS. | 08:04 |
| evrardjp | ioni: that's not enough information for us, we don't know how you access it, which os you're trying etc. | 08:04 |
| cloudnull | then we can set everything to use V6 and it "should" be transperent to the deployment itself. | 08:04 |
| evrardjp | cloudnull: that's a good plan | 08:05 |
| evrardjp | we just need to care about rabbit | 08:05 |
| evrardjp | but it's gonna be fine as rabbit works fine with v6 | 08:05 |
| evrardjp | but we have to avoid names there | 08:05 |
| cloudnull | sure. i'm sure there will be exceptions. | 08:05 |
| cloudnull | ioni: is this the iframe console ? | 08:06 |
| ioni | cloudnull, yes | 08:06 |
| cloudnull | can you click on the link to the page with only the console? | 08:06 |
| cloudnull | does that work? | 08:06 |
| evrardjp | when you think about it, http is a problem solved :D | 08:06 |
| cloudnull | ohai btw/ allanice001 | 08:07 |
| evrardjp | so all the openstack services could be quite easily moved to v6, and more so because they are behind a LB which could be v6 quit easily | 08:07 |
| ioni | cloudnull, i'm clicking everywhere, in centos 7 for example any thing i press returns a [ | 08:07 |
| ioni | ubuntu nothing | 08:07 |
| cloudnull | neat. | 08:07 |
| cloudnull | :) | 08:07 |
| evrardjp | but you see the console | 08:07 |
| * cloudnull is a helper | 08:07 | |
| evrardjp | you see the prompt | 08:07 |
| ioni | yes | 08:07 |
| evrardjp | is your locale en_US ? | 08:08 |
| ioni | atkbd serio0: Unknown key pressed (translated set 2, code 0x0 on isa0060/serio0) | 08:08 |
| ioni | centos 7 returns this | 08:08 |
| ioni | :D | 08:08 |
| ioni | it used to work fine before upgrading to N | 08:08 |
| evrardjp | I haven't tried novnc for a while | 08:08 |
| evrardjp | well if it worked and it doesn't anymore, file a bug and give your configs, we could help you better | 08:09 |
| cloudnull | ioni: can you see if you have this commit https://review.openstack.org/#/c/385011/ ? | 08:09 |
| ioni | i have that | 08:09 |
| cloudnull | idk if that's the issue but its novnc and console related. | 08:09 |
| ioni | without that,i wouldn't even had a console | 08:09 |
| cloudnull | ok. | 08:09 |
| cloudnull | ah. so progress . | 08:09 |
| ioni | and it's commited in os_nova so i got the update with latest sha modifications | 08:10 |
| ioni | openstack console url show c13b50e7-a08f-437e-97d7-3341873921cb | 08:11 |
| ioni | opened the link | 08:11 |
| cloudnull | ok. | 08:11 |
| *** thorst has joined #openstack-ansible | 08:11 | |
| cloudnull | ok. i gotta get some shut eye. | 08:11 |
| cloudnull | ioni: if you dont get it figured out in the next few hours. i'd be happy to take a look . | 08:12 |
| cloudnull | let me know. | 08:12 |
| ioni | ok | 08:13 |
| *** fishcried has quit IRC | 08:13 | |
| ioni | everything i found related to nknown key released (translated set 2, code 0x0 on isa0060/serio0). was an issue with bios | 08:14 |
| ioni | and they removed the batery and worked | 08:14 |
| ioni | i was thinking that's a qemu issue but the old machines are still running from an older qemu | 08:14 |
| *** evilt0ne has quit IRC | 08:15 | |
| admin0 | morning | 08:15 |
| *** thorst has quit IRC | 08:16 | |
| *** evilt0ne has joined #openstack-ansible | 08:16 | |
| *** evilt0ne has quit IRC | 08:16 | |
| *** evilt0ne has joined #openstack-ansible | 08:16 | |
| openstackgerrit | Merged openstack/openstack-ansible: Fix container log bind mount regression https://review.openstack.org/385619 | 08:17 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Fix container log bind mount regression https://review.openstack.org/385821 | 08:22 |
| *** LiYuenan has quit IRC | 08:23 | |
| *** haad1 has joined #openstack-ansible | 08:26 | |
| jrosser | evrardjp: i'm under pressure not to build yet more stuff with rfc1918 v4 but thats nice-to-have rather than policy currently | 08:28 |
| cloudnull | jrosser: ++ | 08:29 |
| cloudnull | i support this effort :) | 08:29 |
| cloudnull | ok night all | 08:29 |
| *** evilt0ne has quit IRC | 08:29 | |
| odyssey4me | o/ | 08:30 |
| *** evilt0ne has joined #openstack-ansible | 08:31 | |
| evrardjp | jrosser: well I understand you :) | 08:32 |
| *** jcrst has joined #openstack-ansible | 08:36 | |
| *** karimb has joined #openstack-ansible | 08:36 | |
| odyssey4me | andymccr thanks to mhayden the centralised scripts are now working - it was actually dead simple in the end... so your review of https://review.openstack.org/382314 & https://review.openstack.org/382306 would be appreciated | 08:37 |
| *** karimb has quit IRC | 08:38 | |
| andymccr | ahh so just remove the ' ' | 08:42 |
| drifterza | hey guys, so im including external roles now for monasca. | 08:44 |
| drifterza | have a weird problem though | 08:44 |
| drifterza | http://logs.openstack.org/55/375655/78/check/gate-openstack-ansible-os_monasca-linters-ubuntu-xenial/16a67ea/console.html#_2016-10-13_07_49_03_531755 | 08:44 |
| drifterza | any idea | 08:44 |
| odyssey4me | andymccr that, and switching the order of paramters around in the script itself - https://review.openstack.org/385511 | 08:45 |
| *** asettle has joined #openstack-ansible | 08:46 | |
| odyssey4me | looking for more reviews please: https://review.openstack.org/#/q/starredby:jesse-pretorius+project:%255Eopenstack/openstack-ansible.*+status:open | 08:46 |
| odyssey4me | drifterza see all thelint errors above it | 08:46 |
| odyssey4me | it looks like the storm role could do with some cleaning up | 08:47 |
| drifterza | yeah, but its an external galaxy role | 08:47 |
| odyssey4me | you can tag its inclusion with 'skip_ansible_lint' and it should get ignored | 08:47 |
| *** evilt0ne has quit IRC | 08:47 | |
| odyssey4me | sure - you should push up a patch to fix those anyway - or pick a role which has lint testing | 08:48 |
| odyssey4me | in fact you should encourage the author to properly implement testing, including lint testing | 08:48 |
| drifterza | the sigh | 08:49 |
| *** evilt0ne has joined #openstack-ansible | 08:49 | |
| *** evilt0ne has quit IRC | 08:49 | |
| *** evilt0ne has joined #openstack-ansible | 08:49 | |
| *** jimbaker has quit IRC | 08:50 | |
| odyssey4me | drifterza added a comment to what I think needs to be change | 08:51 |
| odyssey4me | *changed | 08:51 |
| drifterza | thanks dude | 08:51 |
| drifterza | I'm hating this role | 08:51 |
| drifterza | really | 08:51 |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 08:53 |
| *** jimbaker has joined #openstack-ansible | 08:57 | |
| *** jimbaker has quit IRC | 08:57 | |
| *** jimbaker has joined #openstack-ansible | 08:57 | |
| odyssey4me | evrardjp your thoughts? should we go with https://review.openstack.org/#/c/385272/6/etc/network/interfaces.d/openstack_interface.cfg.test.example or adjust it more to be like https://gist.github.com/busterswt/01e706086c6bc6108e66e06fc992d1c0 ? | 08:59 |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 09:00 |
| evrardjp | well that's two different approaches, and I understand them both | 09:00 |
| evrardjp | the last one forces the host management to be vlan tagged | 09:00 |
| evrardjp | which, in most enterprise cases, don't happen (pxe boot etc) | 09:01 |
| *** berendt has joined #openstack-ansible | 09:01 | |
| evrardjp | it also requires change in the switching, but that's most of the time the case | 09:01 |
| *** askb has quit IRC | 09:01 | |
| evrardjp | but here the change has to happen after the host is provisionned (which could be part of the enterprise provisioning) | 09:01 |
| odyssey4me | evrardjp we're looking for something that will work - it can always we iterated on later | 09:02 |
| evrardjp | well it depends on what you call work | 09:02 |
| odyssey4me | but for release we just want something that illustrates how to do it | 09:02 |
| evrardjp | that's my point, it's 2 different scenario :p | 09:02 |
| odyssey4me | there are always many ways to do it | 09:02 |
| evrardjp | you said assigning completely one nic to osa, then I think james denton solution seems more elegant | 09:03 |
| evrardjp | and as you said it's not a problem of losing connectivity, so I'll adapt the review with that in mind | 09:04 |
| *** fops has joined #openstack-ansible | 09:04 | |
| odyssey4me | thanks | 09:04 |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 09:04 |
| *** fops has quit IRC | 09:04 | |
| *** fops has joined #openstack-ansible | 09:04 | |
| odyssey4me | it does look, to me, like james' gist is the most elegant | 09:05 |
| odyssey4me | we just need to adjust it to fit the IP/VLANa llocations | 09:05 |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 09:07 |
| drifterza | apologies for the spam | 09:07 |
| *** adrian_otto has joined #openstack-ansible | 09:08 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_keystone: Use centralised Ansible test scripts https://review.openstack.org/382306 | 09:09 |
| *** evilt0ne has quit IRC | 09:11 | |
| evrardjp | odyssey4me: well I'd not say elegant or not, it's disruptive, but it's clean | 09:11 |
| evrardjp | only one nic | 09:12 |
| odyssey4me | I think at this point we just need to edit the patch to get it done | 09:12 |
| odyssey4me | lemme take a stab at it quickly | 09:12 |
| evrardjp | ok I can edit it if you want | 09:12 |
| *** thorst has joined #openstack-ansible | 09:14 | |
| *** ChanServ has quit IRC | 09:14 | |
| *** jmccrory has quit IRC | 09:17 | |
| *** adreznec has quit IRC | 09:17 | |
| *** kamtamtun has quit IRC | 09:17 | |
| *** dalees has quit IRC | 09:17 | |
| *** raddaoui has quit IRC | 09:17 | |
| *** berendt has quit IRC | 09:17 | |
| *** janki has quit IRC | 09:17 | |
| *** johnmilton has quit IRC | 09:17 | |
| *** grumpycatt has quit IRC | 09:17 | |
| *** arif-ali has quit IRC | 09:17 | |
| *** Jack_Iv has quit IRC | 09:17 | |
| *** sdake_ has quit IRC | 09:17 | |
| *** openstackgerrit has quit IRC | 09:17 | |
| *** Maeca_ has quit IRC | 09:17 | |
| *** jascott1 has quit IRC | 09:17 | |
| *** z- has quit IRC | 09:17 | |
| *** timrc has quit IRC | 09:17 | |
| *** portante has quit IRC | 09:17 | |
| *** jduhamel has quit IRC | 09:17 | |
| *** galstrom_zzz has quit IRC | 09:17 | |
| *** Jolrael has quit IRC | 09:17 | |
| *** andymccr has quit IRC | 09:17 | |
| *** aludwar has quit IRC | 09:17 | |
| *** haasn has quit IRC | 09:17 | |
| *** xar- has quit IRC | 09:17 | |
| *** aetaric_ has quit IRC | 09:17 | |
| *** mgagne_ has quit IRC | 09:17 | |
| *** chris_hultin|AWA has quit IRC | 09:17 | |
| *** NachoDuck has quit IRC | 09:17 | |
| *** vishwanathj_zzz has quit IRC | 09:17 | |
| *** pjm6 has quit IRC | 09:17 | |
| *** mubix_ has quit IRC | 09:17 | |
| *** toan has quit IRC | 09:17 | |
| *** b3rn-n00dl3s has quit IRC | 09:17 | |
| *** calebb has quit IRC | 09:17 | |
| *** nwonknu has quit IRC | 09:17 | |
| *** prometheanfire has quit IRC | 09:17 | |
| *** xek_ has quit IRC | 09:17 | |
| *** dmsimard has quit IRC | 09:17 | |
| *** jwagner has quit IRC | 09:17 | |
| *** cjloader has quit IRC | 09:17 | |
| *** gfa1 has quit IRC | 09:17 | |
| *** d34dh0r53 has quit IRC | 09:17 | |
| *** h1nch has quit IRC | 09:17 | |
| *** fops has quit IRC | 09:17 | |
| *** haad1 has quit IRC | 09:17 | |
| *** irtermit- has quit IRC | 09:17 | |
| *** odyssey4me has quit IRC | 09:17 | |
| *** palendae has quit IRC | 09:17 | |
| *** batkins61 has quit IRC | 09:17 | |
| *** ikalnytskyi has quit IRC | 09:17 | |
| *** ggillies has quit IRC | 09:17 | |
| *** qwang has quit IRC | 09:17 | |
| *** afranc has quit IRC | 09:17 | |
| *** gaudenz_ has quit IRC | 09:17 | |
| *** neith has quit IRC | 09:17 | |
| *** rackertom has quit IRC | 09:17 | |
| *** serverascode has quit IRC | 09:17 | |
| *** maximov_ has quit IRC | 09:17 | |
| *** mgariepy has quit IRC | 09:17 | |
| *** FrankZhang has quit IRC | 09:17 | |
| *** _hanhart has quit IRC | 09:17 | |
| *** dolphm has quit IRC | 09:17 | |
| *** lbragstad has quit IRC | 09:17 | |
| *** mrhillsman has quit IRC | 09:17 | |
| *** adrian_otto has quit IRC | 09:17 | |
| *** asettle has quit IRC | 09:17 | |
| *** jcrst has quit IRC | 09:17 | |
| *** admin0 has quit IRC | 09:17 | |
| *** allanice001 has quit IRC | 09:17 | |
| *** phschwartz has quit IRC | 09:17 | |
| *** basilAB has quit IRC | 09:17 | |
| *** kencjohnston has quit IRC | 09:17 | |
| *** jasondotstar has quit IRC | 09:17 | |
| *** antonym has quit IRC | 09:17 | |
| *** vvro has quit IRC | 09:17 | |
| *** ioni has quit IRC | 09:17 | |
| *** julian1 has quit IRC | 09:17 | |
| *** sulo has quit IRC | 09:17 | |
| *** mhayden has quit IRC | 09:17 | |
| *** kong has quit IRC | 09:17 | |
| *** aleph1 has quit IRC | 09:17 | |
| *** Apsu has quit IRC | 09:17 | |
| *** sigmavirus has quit IRC | 09:17 | |
| *** homerp has quit IRC | 09:17 | |
| *** git-harry has quit IRC | 09:17 | |
| *** automagically has quit IRC | 09:17 | |
| *** mancdaz has quit IRC | 09:17 | |
| *** jroll has quit IRC | 09:17 | |
| *** johnthetubaguy has quit IRC | 09:17 | |
| *** Oku_OS has quit IRC | 09:17 | |
| *** cloudnull has quit IRC | 09:17 | |
| *** jhesketh has quit IRC | 09:17 | |
| *** hughsaunders has quit IRC | 09:17 | |
| *** jhesketh has joined #openstack-ansible | 09:19 | |
| *** rgogunskiy has quit IRC | 09:20 | |
| *** tlbr has quit IRC | 09:20 | |
| *** agireud has quit IRC | 09:20 | |
| *** thorst has quit IRC | 09:20 | |
| evrardjp | woot netsplits! | 09:23 |
| *** askb has joined #openstack-ansible | 09:29 | |
| *** neith has joined #openstack-ansible | 09:35 | |
| *** evrardjp has quit IRC | 09:41 | |
| *** javeriak has quit IRC | 09:45 | |
| * drifterza netsplit | 09:54 | |
| *** grumpycatt has joined #openstack-ansible | 10:06 | |
| *** ioni has joined #openstack-ansible | 10:06 | |
| *** tlbr has joined #openstack-ansible | 10:06 | |
| *** janki has joined #openstack-ansible | 10:06 | |
| *** johnmilton has joined #openstack-ansible | 10:06 | |
| *** arif-ali has joined #openstack-ansible | 10:07 | |
| *** serverascode has joined #openstack-ansible | 10:07 | |
| *** mubix_ has joined #openstack-ansible | 10:07 | |
| *** adrian_otto has joined #openstack-ansible | 10:07 | |
| *** asettle has joined #openstack-ansible | 10:07 | |
| *** admin0 has joined #openstack-ansible | 10:07 | |
| *** phschwartz has joined #openstack-ansible | 10:07 | |
| *** basilAB has joined #openstack-ansible | 10:07 | |
| *** johnthetubaguy has joined #openstack-ansible | 10:07 | |
| *** antonym has joined #openstack-ansible | 10:07 | |
| *** kencjohnston has joined #openstack-ansible | 10:07 | |
| *** vvro has joined #openstack-ansible | 10:07 | |
| *** jasondotstar has joined #openstack-ansible | 10:07 | |
| *** julian1 has joined #openstack-ansible | 10:07 | |
| *** sulo has joined #openstack-ansible | 10:07 | |
| *** mhayden has joined #openstack-ansible | 10:07 | |
| *** sigmavirus has joined #openstack-ansible | 10:07 | |
| *** aleph1 has joined #openstack-ansible | 10:07 | |
| *** homerp has joined #openstack-ansible | 10:07 | |
| *** Apsu has joined #openstack-ansible | 10:07 | |
| *** git-harry has joined #openstack-ansible | 10:07 | |
| *** automagically has joined #openstack-ansible | 10:07 | |
| *** mancdaz has joined #openstack-ansible | 10:07 | |
| *** jroll has joined #openstack-ansible | 10:07 | |
| *** Oku_OS-away has joined #openstack-ansible | 10:07 | |
| *** cloudnull has joined #openstack-ansible | 10:07 | |
| *** evrardjp has joined #openstack-ansible | 10:09 | |
| *** maximov_ has joined #openstack-ansible | 10:09 | |
| *** kong has joined #openstack-ansible | 10:09 | |
| *** raddaoui has joined #openstack-ansible | 10:09 | |
| *** fops has joined #openstack-ansible | 10:09 | |
| *** haad1 has joined #openstack-ansible | 10:09 | |
| *** irtermit- has joined #openstack-ansible | 10:09 | |
| *** palendae has joined #openstack-ansible | 10:09 | |
| *** odyssey4me has joined #openstack-ansible | 10:09 | |
| *** batkins61 has joined #openstack-ansible | 10:09 | |
| *** ikalnytskyi has joined #openstack-ansible | 10:09 | |
| *** ggillies has joined #openstack-ansible | 10:09 | |
| *** qwang has joined #openstack-ansible | 10:09 | |
| *** afranc has joined #openstack-ansible | 10:09 | |
| *** gaudenz_ has joined #openstack-ansible | 10:09 | |
| *** _hanhart has joined #openstack-ansible | 10:09 | |
| *** mgariepy has joined #openstack-ansible | 10:09 | |
| *** FrankZhang has joined #openstack-ansible | 10:09 | |
| *** dolphm has joined #openstack-ansible | 10:09 | |
| *** lbragstad has joined #openstack-ansible | 10:09 | |
| *** mrhillsman has joined #openstack-ansible | 10:09 | |
| *** jmccrory has joined #openstack-ansible | 10:10 | |
| *** adreznec has joined #openstack-ansible | 10:10 | |
| *** kamtamtun has joined #openstack-ansible | 10:10 | |
| *** dalees has joined #openstack-ansible | 10:10 | |
| *** jmccrory has quit IRC | 10:10 | |
| *** adreznec has quit IRC | 10:10 | |
| *** kamtamtun has quit IRC | 10:10 | |
| *** dalees has quit IRC | 10:10 | |
| *** evrardjp has quit IRC | 10:10 | |
| *** maximov_ has quit IRC | 10:10 | |
| *** kong has quit IRC | 10:10 | |
| *** raddaoui has quit IRC | 10:10 | |
| *** fops has quit IRC | 10:10 | |
| *** haad1 has quit IRC | 10:10 | |
| *** irtermit- has quit IRC | 10:10 | |
| *** odyssey4me has quit IRC | 10:10 | |
| *** palendae has quit IRC | 10:10 | |
| *** batkins61 has quit IRC | 10:10 | |
| *** ikalnytskyi has quit IRC | 10:10 | |
| *** ggillies has quit IRC | 10:10 | |
| *** qwang has quit IRC | 10:10 | |
| *** afranc has quit IRC | 10:10 | |
| *** gaudenz_ has quit IRC | 10:10 | |
| *** mgariepy has quit IRC | 10:10 | |
| *** FrankZhang has quit IRC | 10:10 | |
| *** _hanhart has quit IRC | 10:10 | |
| *** lbragstad has quit IRC | 10:10 | |
| *** dolphm has quit IRC | 10:10 | |
| *** mrhillsman has quit IRC | 10:10 | |
| *** serverascode has quit IRC | 10:10 | |
| *** mubix_ has quit IRC | 10:10 | |
| *** adrian_otto has quit IRC | 10:10 | |
| *** asettle has quit IRC | 10:10 | |
| *** admin0 has quit IRC | 10:10 | |
| *** phschwartz has quit IRC | 10:10 | |
| *** basilAB has quit IRC | 10:10 | |
| *** vvro has quit IRC | 10:10 | |
| *** antonym has quit IRC | 10:10 | |
| *** jasondotstar has quit IRC | 10:10 | |
| *** kencjohnston has quit IRC | 10:10 | |
| *** julian1 has quit IRC | 10:10 | |
| *** sulo has quit IRC | 10:10 | |
| *** mhayden has quit IRC | 10:10 | |
| *** aleph1 has quit IRC | 10:10 | |
| *** Apsu has quit IRC | 10:10 | |
| *** automagically has quit IRC | 10:10 | |
| *** sigmavirus has quit IRC | 10:10 | |
| *** homerp has quit IRC | 10:10 | |
| *** mancdaz has quit IRC | 10:10 | |
| *** git-harry has quit IRC | 10:10 | |
| *** jroll has quit IRC | 10:10 | |
| *** Oku_OS-away has quit IRC | 10:10 | |
| *** johnthetubaguy has quit IRC | 10:10 | |
| *** cloudnull has quit IRC | 10:10 | |
| *** arif-ali has quit IRC | 10:10 | |
| *** ioni has quit IRC | 10:10 | |
| *** tlbr has quit IRC | 10:10 | |
| *** janki has quit IRC | 10:10 | |
| *** johnmilton has quit IRC | 10:10 | |
| *** grumpycatt has quit IRC | 10:10 | |
| *** ioni has joined #openstack-ansible | 10:17 | |
| *** mubix_ has joined #openstack-ansible | 10:22 | |
| *** grumpycatt has joined #openstack-ansible | 10:26 | |
| *** serverascode has joined #openstack-ansible | 10:27 | |
| *** asettle has joined #openstack-ansible | 10:27 | |
| *** admin0 has joined #openstack-ansible | 10:27 | |
| *** phschwartz has joined #openstack-ansible | 10:27 | |
| *** basilAB has joined #openstack-ansible | 10:27 | |
| *** johnthetubaguy has joined #openstack-ansible | 10:27 | |
| *** antonym has joined #openstack-ansible | 10:27 | |
| *** kencjohnston has joined #openstack-ansible | 10:27 | |
| *** vvro has joined #openstack-ansible | 10:27 | |
| *** jasondotstar has joined #openstack-ansible | 10:27 | |
| *** julian1 has joined #openstack-ansible | 10:27 | |
| *** sulo has joined #openstack-ansible | 10:27 | |
| *** mhayden has joined #openstack-ansible | 10:27 | |
| *** sigmavirus has joined #openstack-ansible | 10:27 | |
| *** aleph1 has joined #openstack-ansible | 10:27 | |
| *** homerp has joined #openstack-ansible | 10:27 | |
| *** Apsu has joined #openstack-ansible | 10:27 | |
| *** git-harry has joined #openstack-ansible | 10:27 | |
| *** automagically has joined #openstack-ansible | 10:27 | |
| *** mancdaz has joined #openstack-ansible | 10:27 | |
| *** jroll has joined #openstack-ansible | 10:27 | |
| *** Oku_OS-away has joined #openstack-ansible | 10:27 | |
| *** cloudnull has joined #openstack-ansible | 10:27 | |
| *** jmccrory has joined #openstack-ansible | 10:27 | |
| *** adreznec has joined #openstack-ansible | 10:27 | |
| *** kamtamtun has joined #openstack-ansible | 10:27 | |
| *** dalees has joined #openstack-ansible | 10:27 | |
| *** arif-ali has joined #openstack-ansible | 10:27 | |
| *** tlbr has joined #openstack-ansible | 10:27 | |
| *** janki has joined #openstack-ansible | 10:27 | |
| *** johnmilton has joined #openstack-ansible | 10:27 | |
| *** raddaoui has joined #openstack-ansible | 10:27 | |
| *** evrardjp has joined #openstack-ansible | 10:27 | |
| *** maximov_ has joined #openstack-ansible | 10:27 | |
| *** fops has joined #openstack-ansible | 10:27 | |
| *** haad1 has joined #openstack-ansible | 10:27 | |
| *** irtermit- has joined #openstack-ansible | 10:27 | |
| *** palendae has joined #openstack-ansible | 10:27 | |
| *** odyssey4me has joined #openstack-ansible | 10:27 | |
| *** batkins61 has joined #openstack-ansible | 10:27 | |
| *** ikalnytskyi has joined #openstack-ansible | 10:27 | |
| *** ggillies has joined #openstack-ansible | 10:27 | |
| *** qwang has joined #openstack-ansible | 10:27 | |
| *** afranc has joined #openstack-ansible | 10:27 | |
| *** gaudenz_ has joined #openstack-ansible | 10:27 | |
| *** _hanhart has joined #openstack-ansible | 10:27 | |
| *** mgariepy has joined #openstack-ansible | 10:27 | |
| *** FrankZhang has joined #openstack-ansible | 10:27 | |
| *** dolphm has joined #openstack-ansible | 10:27 | |
| *** lbragstad has joined #openstack-ansible | 10:27 | |
| *** mrhillsman has joined #openstack-ansible | 10:27 | |
| *** Jack_Iv has joined #openstack-ansible | 10:28 | |
| *** sdake_ has joined #openstack-ansible | 10:28 | |
| *** openstackgerrit has joined #openstack-ansible | 10:28 | |
| *** Maeca_ has joined #openstack-ansible | 10:28 | |
| *** jascott1 has joined #openstack-ansible | 10:28 | |
| *** z- has joined #openstack-ansible | 10:28 | |
| *** timrc has joined #openstack-ansible | 10:28 | |
| *** portante has joined #openstack-ansible | 10:28 | |
| *** jduhamel has joined #openstack-ansible | 10:28 | |
| *** galstrom_zzz has joined #openstack-ansible | 10:28 | |
| *** Jolrael has joined #openstack-ansible | 10:28 | |
| *** andymccr has joined #openstack-ansible | 10:28 | |
| *** kong has joined #openstack-ansible | 10:28 | |
| *** rackertom has joined #openstack-ansible | 10:28 | |
| *** d34dh0r53 has joined #openstack-ansible | 10:28 | |
| *** haasn has joined #openstack-ansible | 10:28 | |
| *** xar- has joined #openstack-ansible | 10:28 | |
| *** aetaric_ has joined #openstack-ansible | 10:28 | |
| *** mgagne_ has joined #openstack-ansible | 10:28 | |
| *** chris_hultin|AWA has joined #openstack-ansible | 10:28 | |
| *** NachoDuck has joined #openstack-ansible | 10:28 | |
| *** vishwanathj_zzz has joined #openstack-ansible | 10:28 | |
| *** pjm6 has joined #openstack-ansible | 10:28 | |
| *** toan has joined #openstack-ansible | 10:28 | |
| *** b3rn-n00dl3s has joined #openstack-ansible | 10:28 | |
| *** calebb has joined #openstack-ansible | 10:28 | |
| *** nwonknu has joined #openstack-ansible | 10:28 | |
| *** prometheanfire has joined #openstack-ansible | 10:28 | |
| *** xek_ has joined #openstack-ansible | 10:28 | |
| *** dmsimard has joined #openstack-ansible | 10:28 | |
| *** jwagner has joined #openstack-ansible | 10:28 | |
| *** cjloader has joined #openstack-ansible | 10:28 | |
| *** gfa1 has joined #openstack-ansible | 10:28 | |
| *** h1nch has joined #openstack-ansible | 10:28 | |
| odyssey4me | asettle could you please do a final pass through of https://review.openstack.org/385496 ? | 10:28 |
| asettle | odyssey4me: on it already :) | 10:29 |
| odyssey4me | thanks | 10:29 |
| openstackgerrit | Merged openstack/openstack-ansible: [DOCS] Update manual upgrade guide https://review.openstack.org/385582 | 10:29 |
| *** serverascode has quit IRC | 10:29 | |
| *** ChanServ has joined #openstack-ansible | 10:30 | |
| *** tepper.freenode.net sets mode: +o ChanServ | 10:30 | |
| *** d34dh0r53 has quit IRC | 10:30 | |
| pjm6 | good morning all :) | 10:30 |
| odyssey4me | o/ pjm6 | 10:30 |
| *** d34dh0r53 has joined #openstack-ansible | 10:30 | |
| *** rackertom has quit IRC | 10:32 | |
| *** tlbr is now known as Guest76323 | 10:32 | |
| *** ioni is now known as Guest58081 | 10:32 | |
| *** maximov_ has quit IRC | 10:32 | |
| *** karimb has joined #openstack-ansible | 10:32 | |
| *** adrian_otto has joined #openstack-ansible | 10:32 | |
| *** agireud has joined #openstack-ansible | 10:33 | |
| *** sdake_ has quit IRC | 10:35 | |
| *** neith has quit IRC | 10:35 | |
| *** neith has joined #openstack-ansible | 10:35 | |
| *** mubix_ has quit IRC | 10:35 | |
| *** mubix_ has joined #openstack-ansible | 10:35 | |
| *** raddaoui has quit IRC | 10:36 | |
| *** raddaoui has joined #openstack-ansible | 10:36 | |
| *** karimb has quit IRC | 10:37 | |
| *** karimb has joined #openstack-ansible | 10:37 | |
| evrardjp | odyssey4me: that was a hard patchset | 10:38 |
| evrardjp | docs, right :p | 10:38 |
| odyssey4me | docs are hard work | 10:39 |
| odyssey4me | putting all that together took me 6 hours yesterday | 10:39 |
| odyssey4me | the sample configs I mean | 10:39 |
| *** rackertom has joined #openstack-ansible | 10:44 | |
| odyssey4me | andymccr loving how many LOC are deleted in https://review.openstack.org/385852 :) | 10:44 |
| andymccr | odyssey4me: yeah i ahve the fix up too but im on vpn and git review is not being helpful right now :P | 10:44 |
| *** Guest58081 has quit IRC | 10:45 | |
| *** Guest58081 has joined #openstack-ansible | 10:45 | |
| *** Guest58081 is now known as ioni | 10:45 | |
| odyssey4me | andymccr well gert erf der ververpn | 10:45 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_swift: Use dictionary for service mappings https://review.openstack.org/385852 | 10:46 |
| andymccr | meh its ok i added the proxycommand :P | 10:47 |
| *** serverascode has joined #openstack-ansible | 10:51 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [docs] Provide example configurations https://review.openstack.org/385496 | 10:51 |
| odyssey4me | thanks asettle - with regards to 'this environment' it's referring to the environment described on the page itself as a whole | 10:52 |
| odyssey4me | is there a way that I can make the context more clear? | 10:52 |
| odyssey4me | maybe it needs an intro or something? | 10:53 |
| asettle | odyssey4me: yeah I thought so. Hmm let me go have another look. I think it might just be easier instead of 'this environment' you say 'prod' or 'test' environment | 10:53 |
| asettle | Intro would actually be really helpful | 10:53 |
| asettle | Does need some overarching 'this is a thing, and this is what it is' context | 10:53 |
| *** hughsaunders has joined #openstack-ansible | 10:53 | |
| odyssey4me | ok, it does have an intro but I see that it needs some editing | 10:53 |
| odyssey4me | lemme do some quickly | 10:53 |
| odyssey4me | I'll do it fairly free form, then you can correct my horrible arghlish | 10:54 |
| *** maximov_ has joined #openstack-ansible | 10:55 | |
| asettle | odyssey4me: ace :) shoot me a link when you're done | 10:57 |
| asettle | It looks good! Tahnsk for updating this all :) | 10:57 |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 10:59 |
| *** smatzek has joined #openstack-ansible | 11:02 | |
| odyssey4me | evrardjp what is the correct term for a switch that supports bonded interfaces? | 11:03 |
| *** sdake has joined #openstack-ansible | 11:03 | |
| odyssey4me | port aggregation? | 11:03 |
| odyssey4me | link aggregation | 11:04 |
| *** izaakk has quit IRC | 11:06 | |
| drifterza | usually link | 11:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [docs] Provide example configurations https://review.openstack.org/385496 | 11:06 |
| odyssey4me | asettle ^ | 11:06 |
| drifterza | LACP = link agg control prot | 11:06 |
| *** izaakk has joined #openstack-ansible | 11:06 | |
| asettle | odyssey4me: gimme a bit :) just doing other reviews | 11:07 |
| odyssey4me | drifterza yep, I used link aggregation | 11:07 |
| odyssey4me | it's a generic enough term | 11:07 |
| drifterza | yup | 11:07 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_swift: Use dictionary for service mappings https://review.openstack.org/385852 | 11:08 |
| odyssey4me | andymccr I dunno if you noticed my comment in an earlier PS | 11:09 |
| *** sdake_ has joined #openstack-ansible | 11:11 | |
| evrardjp | bond mode 4 is LACP | 11:12 |
| evrardjp | for active-backup a dumb switch is enough | 11:12 |
| evrardjp | lacp is link aggregation | 11:12 |
| evrardjp | well I realise ppl answered here but still answered! | 11:13 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [docs] Provide example configurations https://review.openstack.org/385496 | 11:13 |
| evrardjp | when doing on different switches, it's probably MLAG you're looking for | 11:13 |
| *** sdake has quit IRC | 11:13 | |
| odyssey4me | evrardjp ah yes, I forgot that the config was active-backup and that nothing special is needed there | 11:13 |
| evrardjp | but I doubt it's still what you are looking for :p | 11:13 |
| evrardjp | yup | 11:13 |
| *** markvoelker has quit IRC | 11:14 | |
| odyssey4me | evrardjp correct me if I'm wrong, but IIRC the switch doesn't need to be able to do VLAN's either for you to use VLAN tags | 11:14 |
| odyssey4me | if the switch doesn't do VLAN's it just means that you won't have isolation | 11:14 |
| *** askb has quit IRC | 11:14 | |
| odyssey4me | oh no, I'm wrong | 11:14 |
| odyssey4me | I was thinking about a flat network using multiple IP ranges | 11:15 |
| odyssey4me | nvm | 11:15 |
| evrardjp | well if it doesn't recognize/do something with vlans, it's still a vlan | 11:15 |
| odyssey4me | ok asettle - doing a quick edit | 11:15 |
| evrardjp | the frame doesn't change :p | 11:15 |
| asettle | odyssey4me: okay :) | 11:15 |
| asettle | Sorry, just trying to clear up some docs stuff I'll be there in a second :0 | 11:15 |
| asettle | :) | 11:15 |
| odyssey4me | evrardjp so with a dumb hub/switch using the vlan tags will work? ie servers will be capable of talking to each other? | 11:16 |
| evrardjp | well if the other side of the switch is configured exactly the same yes | 11:16 |
| evrardjp | basically you're giving everything | 11:16 |
| evrardjp | a basic switch won't filter tags | 11:16 |
| odyssey4me | so neither the test or prod environment need anything more than a dumb switch? | 11:16 |
| odyssey4me | an L2 switch | 11:17 |
| evrardjp | well you probably don't want that | 11:17 |
| *** karimb has quit IRC | 11:17 | |
| evrardjp | but yes | 11:17 |
| evrardjp | technically yes | 11:17 |
| odyssey4me | sure, you don't want that - not really, but in terms of setting expectations I want to give the minimum requirement | 11:17 |
| evrardjp | but how do you ensure some native vlans for your storage environment? | 11:17 |
| evrardjp | because sometimes you need that :p | 11:17 |
| odyssey4me | so I'll take the requirement for a VLAN-capable switch out | 11:17 |
| evrardjp | well a plain l2 switch could be enough | 11:18 |
| evrardjp | you could technically remove that | 11:18 |
| evrardjp | although it cost 50 quid, if you can't afford 50 quid maybe worth leaving openstack world | 11:18 |
| evrardjp | :p | 11:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [docs] Provide example configurations https://review.openstack.org/385496 | 11:18 |
| odyssey4me | evrardjp haha, fair point :p | 11:19 |
| odyssey4me | look at you, using english words and all :p | 11:19 |
| evrardjp | and also you have a dumb switch that basically transfers all the frames so thanks for the useless isolation | 11:19 |
| evrardjp | but who am I, right? :p | 11:20 |
| evrardjp | so (optional, highy-recommended) would be my choice | 11:20 |
| odyssey4me | I think anyone building a cloud should know that. If they don't, then they should just pack it up and go home. | 11:21 |
| evrardjp | yup | 11:21 |
| evrardjp | MLAG is different matter 'though | 11:22 |
| evrardjp | and it's quite standardly used in DCs | 11:22 |
| jrosser | would be nice to have some pointers for people at the other end of the spectrum | 11:22 |
| evrardjp | jrosser: ? | 11:22 |
| odyssey4me | yeah, we had some fun getting MLAG working properly between two Arista switches back in a previous life | 11:22 |
| *** thorst has joined #openstack-ansible | 11:22 | |
| jrosser | it's not obvious to me if i can easily have racks of compute which are each their own bit of routed L3 | 11:22 |
| odyssey4me | it was a little weird, but eventually worked very well | 11:23 |
| jrosser | if i can define that grouping in osa, or even if i need to | 11:23 |
| *** johnmilton has quit IRC | 11:23 | |
| jrosser | (from a mgmt plane p.o.v i mean) | 11:23 |
| odyssey4me | jrosser so... it is potentially possible - and something I'd love someone to figure out | 11:23 |
| evrardjp | jrosser: including storage routed? | 11:23 |
| evrardjp | we've done that | 11:24 |
| jrosser | thats easier i think | 11:24 |
| evrardjp | it's a terrible idea but it works | 11:24 |
| evrardjp | we have in osa the additional routes for the networks | 11:24 |
| jrosser | well im starting with a L3 leaf/spine network | 11:24 |
| jrosser | so some of this matters from the start | 11:24 |
| evrardjp | and basically osa is using mgmt network to connect | 11:24 |
| odyssey4me | jrosser I did this as a sketch of how one *might* do it, but I haven't had a chance to try it out: https://gist.github.com/odyssey4me/09d963776f8872f2562e477c5158a3e0 | 11:24 |
| evrardjp | well depending on your requirements, it can change a lot in the topology, and you know it I guess :p | 11:26 |
| *** sdake has joined #openstack-ansible | 11:27 | |
| odyssey4me | jrosser note that I have no idea if that will even work, but I just wanted to capture it at the time after a discussion evrardjp and I had | 11:27 |
| jrosser | odyssey4me: ah there so az<n> corresponds to L3 domain | 11:27 |
| jrosser | which might well be one TOR switch worth of stuff | 11:27 |
| odyssey4me | jrosser yeah, each 'az' (ie availability zone) corresponds to a failure domain - which could be a rack, a rack row, a DC, whatever | 11:28 |
| evrardjp | I don't think for your use case you shouldn't go that far | 11:28 |
| *** sdake_ has quit IRC | 11:28 | |
| evrardjp | jrosser: could you leverage IS-IS to have a l3 controlplane for extending your vlans? | 11:29 |
| odyssey4me | I know we discussed that RabbitMQ, MariaDB, etc would very likely get a performance hit when implementing their traffic via L3 | 11:29 |
| odyssey4me | so what one might have to do is implement a flat network across AZ's for the infra, then do L3 for API's and such | 11:29 |
| evrardjp | well I'm starting to be fancy here, let's discuss that another time :p | 11:29 |
| jrosser | i need to get anything going at all first :) | 11:30 |
| odyssey4me | but yeah - it's all possible, and would likely take some work to figure out how to optimise it, but it *could* be done | 11:30 |
| jrosser | but i have enough stuff to p.o.c it any which way i like | 11:30 |
| evrardjp | odyssey4me: it could be done in a lot of different ways | 11:30 |
| evrardjp | jrosser: ^ | 11:30 |
| evrardjp | the thing is, where do you want to rely on your networking too | 11:30 |
| evrardjp | for me I always rely on putting my l2 matters on switches :p | 11:31 |
| *** thorst has quit IRC | 11:31 | |
| odyssey4me | jrosser with https://github.com/openstack/openstack-ansible-ops/tree/master/multi-node-aio you could make your test hardware go a long way :) | 11:31 |
| odyssey4me | evrardjp andymccr two backports and a simple review in here if you have a moment: https://review.openstack.org/#/q/starredby:jesse-pretorius+project:%255Eopenstack/openstack-ansible.*+status:open | 11:32 |
| jrosser | i have particular requirements pushing me toward calico too but that feels like not first step | 11:33 |
| asettle | odyssey4me: done. I've evened the pages out so they look similar :) sorry it took longer | 11:34 |
| odyssey4me | asettle oh man, that hurts | 11:34 |
| asettle | HAHA sorry odyssey4me | 11:34 |
| odyssey4me | lemme get it done | 11:35 |
| asettle | You'er doing that thing where you do my edits, and tehn you do YOUR edits, adn then I edit your second edits plus your other edits | 11:35 |
| asettle | It's a cool cycle we're in :p | 11:35 |
| odyssey4me | asettle lies, you're not sorry | 11:36 |
| odyssey4me | you take great pleasure in poking fun at my arghlish | 11:36 |
| asettle | Okay, maybe not. But I swear this appendix is going to look SO BEAUTIFUL once you're done :p | 11:36 |
| odyssey4me | yes, no longer the ugly baby it once was | 11:37 |
| odyssey4me | asettle are you OK with the use of the word 'illustrative' instead of 'sample' for the intro? | 11:38 |
| evrardjp | pushing my lucj here: https://review.openstack.org/#/c/351776/ | 11:38 |
| evrardjp | luck* odyssey4me ^ | 11:38 |
| odyssey4me | I feel that it's important to be clear that it's for illustration purposes. | 11:38 |
| asettle | odyssey4me: hmmm so, the only problem I have with 'illustrative' is it is kind of wafty instead of to the point like the rest of the content. I think if you're going to want to use illustration, I would go with any of the following: sample, prescription, explanatory, etc | 11:39 |
| odyssey4me | evrardjp it's for an AIO, we could just reduce it to check for 'trusty' and that's that | 11:39 |
| odyssey4me | asettle ok, so 'sample' or 'example' is a preferred term? | 11:40 |
| asettle | odyssey4me: yep :) let's keep it familiar | 11:40 |
| odyssey4me | I was thinking more along the lines of 'illustrative example' - used only once right at the start | 11:40 |
| odyssey4me | 'sample' on its own still feels a little too prescriptive to me | 11:41 |
| *** retreved has joined #openstack-ansible | 11:41 | |
| odyssey4me | ie it's a small piece of a larger part | 11:41 |
| odyssey4me | ok, I'll go with example | 11:41 |
| asettle | odyssey4me: I'm so glad you decided that on your own :P hahah sorry, wasn't looking at teh chan | 11:43 |
| odyssey4me | it's noisy up here in my head | 11:44 |
| odyssey4me | asettle with regards to 'A production environment' versus 'This production environment'... I'm completely not good with using 'A'. Every production environment has different characteristics... I'm specifically describing *this* environment. | 11:45 |
| asettle | Okay, good point. Change that :) | 11:46 |
| odyssey4me | In fact, I'm starting to think that referring to these as 'test' and 'prod' is a mistake. | 11:46 |
| odyssey4me | We should perhaps rather just refer to 'Example A' and 'Example B'. | 11:46 |
| odyssey4me | A is small and simple - B is highly available | 11:46 |
| asettle | odyssey4me: hmmm but that would cause some issues because we already refer to them as test and prod throughout the guide | 11:50 |
| asettle | Diagrams etc | 11:50 |
| asettle | I don't think we can just change that now. That's a much larger change we CAN make. | 11:50 |
| odyssey4me | yeah, I think that'd be best | 11:51 |
| odyssey4me | I expect that we'll be expanding to more examples in the future. | 11:51 |
| odyssey4me | so for now we go with what we have | 11:51 |
| *** johnmilton has joined #openstack-ansible | 11:52 | |
| *** thorst has joined #openstack-ansible | 11:52 | |
| *** smatzek_ has joined #openstack-ansible | 11:54 | |
| asettle | yep :) | 11:54 |
| asettle | We'll work on a plan later odyssey4me | 11:54 |
| *** smatzek has quit IRC | 11:57 | |
| *** jamesdenton has quit IRC | 11:59 | |
| mhayden | morning | 12:00 |
| mhayden | odyssey4me / andymccr: i've broken enough stuff in OSA so i figured it might be worthwhile to fix something :P | 12:00 |
| andymccr | mhayden: are you OSA's chaos monkey? We're stronger because of it! :P | 12:01 |
| *** kylek3h has quit IRC | 12:01 | |
| mhayden | andymccr: well, i'm more like odyssey4me's chaos monkey | 12:01 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_swift: Use dictionary for service mappings https://review.openstack.org/385852 | 12:02 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [docs] Provide example configurations https://review.openstack.org/385496 | 12:03 |
| odyssey4me | asettle ^ round 15 | 12:03 |
| *** johnmilton has quit IRC | 12:04 | |
| asettle | Okay odyssey4me :) | 12:04 |
| *** johnmilton has joined #openstack-ansible | 12:04 | |
| *** sdake_ has joined #openstack-ansible | 12:05 | |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 12:05 |
| *** sdake has quit IRC | 12:06 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_swift: Use dictionary for service mappings https://review.openstack.org/385852 | 12:12 |
| asettle | odyssey4me: you didn't remove illustrative from line 13? https://review.openstack.org/#/c/385496/15/doc/source/install-guide/app-config-prod.rst | 12:13 |
| *** williamcaban has joined #openstack-ansible | 12:14 | |
| *** markvoelker has joined #openstack-ansible | 12:14 | |
| odyssey4me | asettle you didn't ask me to - I thought that line was OK | 12:15 |
| asettle | odyssey4me: oh I didn't? Bummer. I'm a knob. let's remove all 'illustrative' to be consistent :) | 12:15 |
| asettle | That is legit my only comment. otherwise, looks great dude :) | 12:15 |
| odyssey4me | okie dokey | 12:15 |
| odyssey4me | so change it to 'The configurations here are meant to be examples, not prescriptive.' | 12:16 |
| odyssey4me | ? | 12:17 |
| odyssey4me | hmm, 'an example' | 12:17 |
| odyssey4me | bah, now it ruins the flow of that sentence | 12:18 |
| drifterza | that R proper anglish boet | 12:18 |
| admin0 | "WARNING: The following packages cannot be authenticated!", " ubuntu-cloud-keyring" — why does this come ? | 12:18 |
| odyssey4me | the 'ive' at the end of each word made it all nice and flowing | 12:18 |
| admin0 | TASK [keepalived : Add Ubuntu Cloud Archive keyring] *************************** | 12:18 |
| admin0 | it comes up on others as well | 12:18 |
| admin0 | looks like i am missing something in this new env | 12:18 |
| asettle | odyssey4me: "NOTE: The configurations in this example are not prescriptive" | 12:19 |
| odyssey4me | admin0 in our gates it does that because we're using an apt cache from infra which doesn't authenticate the packages | 12:19 |
| odyssey4me | asettle ah, much betterer | 12:19 |
| *** karimb has joined #openstack-ansible | 12:19 | |
| evrardjp | drifterza: haha boet betekend iets anders in nederlands :p | 12:19 |
| asettle | odyssey4me: here all week | 12:19 |
| drifterza | evrardjp, ne ? | 12:19 |
| odyssey4me | asettle how do you do the markup for note again? | 12:19 |
| asettle | .. note:: | 12:19 |
| asettle | And ensure the start of the note is one line under the 'n' | 12:19 |
| *** markvoelker has quit IRC | 12:20 | |
| evrardjp | what did boet mean in your sentence? | 12:20 |
| drifterza | like brother or bro | 12:20 |
| odyssey4me | asettle to be honest though, I think the sentence has lost its meaning now | 12:20 |
| odyssey4me | I think it better to remove it | 12:20 |
| evrardjp | like broer :) | 12:20 |
| evrardjp | ok | 12:20 |
| drifterza | yup | 12:21 |
| asettle | odyssey4me: was about the say the same thing. I think it's fairly self explanatory that it is. | 12:21 |
| asettle | Considering it's called "example production environment" | 12:21 |
| asettle | :p | 12:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [docs] Provide example configurations https://review.openstack.org/385496 | 12:21 |
| odyssey4me | asettle ^ final round | 12:22 |
| asettle | Love your stuff odyssey4me | 12:22 |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 12:23 |
| odyssey4me | ok, andymccr / mhayden / evrardjp it's your turn: https://review.openstack.org/385496 | 12:26 |
| evrardjp | fine | 12:26 |
| odyssey4me | let's put the puppy to bed now | 12:26 |
| * mhayden turns | 12:26 | |
| evrardjp | done | 12:28 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_swift: Use dictionary for service mappings https://review.openstack.org/385852 | 12:31 |
| drifterza | Jessie, ek skiem you Engels is nogals goed, word jy noe a soutie of wat ? | 12:32 |
| drifterza | ;p | 12:33 |
| odyssey4me | drifterza ek was altyd a soutie :p | 12:33 |
| drifterza | ne ? | 12:33 |
| odyssey4me | ja, my van is Pretorius maar ek het Engels groot geword | 12:33 |
| drifterza | wel kyk my ma's se kant is duits en my pa se kant is nedelands en ek was Engels groot gemaak | 12:34 |
| drifterza | hoe weird. | 12:34 |
| logan- | cloudnull: followup on that ssh thing from last night. sorry i didnt give very good debugging output on what im seeing. this should show things a little better http://cdn.pasteraw.com/n3ulakcdntjo9ge2k4y1ttq34yv6uiz -- the container detection stuff isnt working in the ssh connection module because the dict its looking for the vars in is empty :( but my inventory is almost identical to that test one you put | 12:34 |
| odyssey4me | ja, my Pa se kant is eintlik Afrikaans, maar sy Pa was engels | 12:34 |
| *** kylek3h has joined #openstack-ansible | 12:35 | |
| odyssey4me | logan- did you set the connection plugin to the one we have in our plugins repo? | 12:35 |
| odyssey4me | logan- for Newton 'openstack-ansible' sources /usr/local/bin/openstack-ansible.rc - if you execute 'ansible' you need to do that manually | 12:36 |
| odyssey4me | if we could figure out how to make that happen automatically it'd be nice | 12:36 |
| logan- | yeah it is using the plugins repo one there. that DEBUG line is what i'm using to verify that as it only is added to the osa ssh.py connection plugin | 12:36 |
| asettle | When did this channel turn Afrikaans? | 12:36 |
| logan- | im using an ansible.cfg in this case so not relying on envvars or anything | 12:37 |
| mhayden | asettle: you wanna run the OSA weekly mtg today? i've got a meeting :( | 12:38 |
| odyssey4me | logan- does your ansible.cfg have the locaiton of the conneciton plugin set | 12:38 |
| asettle | mhayden: are you sure you want *me* to do that? | 12:38 |
| asettle | :p | 12:38 |
| logan- | yessir | 12:38 |
| asettle | I mean, I am happy to. | 12:38 |
| asettle | Point me in the direction of all the things. | 12:38 |
| mhayden | asettle: of course -- you're good at it! | 12:38 |
| mhayden | asettle: agenda is here -> https://wiki.openstack.org/wiki/Meetings/openstack-ansible | 12:38 |
| asettle | mhayden: haha okay! Shoot me all the things I need :) | 12:38 |
| asettle | Oh, you already on it. | 12:38 |
| asettle | go mhayden | 12:38 |
| mhayden | asettle: the only two action items are things to be completed at the summit -> http://eavesdrop.openstack.org/meetings/openstack_ansible/2016/openstack_ansible.2016-10-06-16.02.html | 12:39 |
| odyssey4me | logan- FYI I tested the other available lxc connection plugins and learned some things: https://gist.github.com/odyssey4me/ec8f9316f74bca099c9f65f8fefb05a4 | 12:39 |
| asettle | Sweet as :) | 12:39 |
| mhayden | asettle: and you can see the commands i used for the various meeting things -> http://eavesdrop.openstack.org/meetings/openstack_ansible/2016/openstack_ansible.2016-10-06-16.02.log.html | 12:39 |
| * asettle raises hand to mhayden | 12:40 | |
| asettle | Question | 12:40 |
| asettle | Do I have to toot? | 12:40 |
| admin0 | odyssey4me: i am not using the gates, but running the latest newton | 12:40 |
| mhayden | asettle: you may certainly toot | 12:40 |
| asettle | Okay good, I like it. | 12:40 |
| mhayden | but watch out for andymccr -- he's long-winded | 12:40 |
| admin0 | on a set of ubuntu 14.04 machines ( freshly redone ) | 12:40 |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 12:40 |
| asettle | mhayden: I can handle that guy :p | 12:40 |
| odyssey4me | admin0 are you using your own apt mirror? also, is it an AIO? | 12:40 |
| admin0 | so surprized why i am facing this .. never faced this before | 12:40 |
| asettle | "No dinner for you if you keep going, now... " | 12:40 |
| admin0 | its a multi node install | 12:40 |
| * mhayden cannot confirm or deny | 12:40 | |
| admin0 | not an aio | 12:40 |
| odyssey4me | admin0 we introduced the use of UCA for the first time in Newton | 12:40 |
| asettle | On that note, I might go run my errands early then :) | 12:40 |
| * mhayden tips his hat to asettle for the assist! :) | 12:41 | |
| admin0 | how do I fix it :D | 12:41 |
| asettle | mhayden: anytime :) | 12:41 |
| asettle | mhayden: shall I ping people the hour before? | 12:41 |
| admin0 | whats an UCA :D | 12:41 |
| odyssey4me | admin0 ok, so you're using the defaults then I guess? | 12:41 |
| admin0 | universal ca ? | 12:41 |
| odyssey4me | admin0 UCA = Ubuntu Cloud Archive | 12:41 |
| admin0 | oh :D | 12:41 |
| admin0 | :) | 12:41 |
| admin0 | i am using the defaults | 12:41 |
| * admin0 always uses the defaults :D | 12:41 | |
| mhayden | asettle: https://media.giphy.com/media/jt9FFt6f2Q0Te/giphy.gif | 12:41 |
| odyssey4me | UCA has updated packages like libvirt, openvswitch, haproxy, etc | 12:41 |
| logan- | ah cool odyssey4me. i was looking at the other 2 last night and the cloudnull informed me of the osa one so I decided to try it first. will take a look at your results :) | 12:42 |
| admin0 | ok | 12:42 |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 12:42 |
| odyssey4me | logan- yeah, I learned that the stock lxc plugin is localhost only - and that the remote one via SSH was super-slow | 12:42 |
| *** woodard has joined #openstack-ansible | 12:42 | |
| odyssey4me | our one seems to work quite well - although we haven't yet started pulling openssh out of the containers to really give it a go | 12:42 |
| admin0 | odyssey4me: so how do I proceed with this ? how to get past this ? | 12:42 |
| odyssey4me | admin0 is it failing, or just warning? | 12:43 |
| admin0 | failing :( | 12:43 |
| *** woodard has quit IRC | 12:43 | |
| odyssey4me | admin0 and which task - can you pastebin/share your user_variables.yml and a debug log of the failing task? | 12:43 |
| admin0 | https://gist.githubusercontent.com/a1git/c27ffb24ac8e899ceb7b44e2b97479ab/raw/918b29a549dc63e8e979c39a68c57ddf04a5f820/gistfile1.txt | 12:43 |
| *** woodard has joined #openstack-ansible | 12:44 | |
| mhayden | so i did a 13.3.3 to 13.3.4 upgrade yesterday and ended up with duplicated heat service entries | 12:44 |
| mhayden | it didn't hurt anyhting, but i wasn't sure if someone else saw that | 12:44 |
| odyssey4me | admin0 ah, that's in the keepalived role specifically - evrardjp any thoughts? | 12:44 |
| mhayden | i couldn't see anything in the diff that would have caused that | 12:44 |
| *** gouthamr has joined #openstack-ansible | 12:44 | |
| odyssey4me | admin0 it's plausible that the apt cache is stale | 12:45 |
| odyssey4me | try updating the apt cache? | 12:45 |
| admin0 | multiple times | 12:45 |
| admin0 | that was my first thought a well | 12:45 |
| admin0 | as* | 12:45 |
| *** kylek3h has quit IRC | 12:45 | |
| odyssey4me | also validate that you rstandard ubuntu repo has the 'ubuntu-cloud-keyring' packge? | 12:45 |
| admin0 | hmm.. i can try to do the ubuntu-cloud-keyring install by hand | 12:45 |
| asettle | evrardjp: with your comment on here https://review.openstack.org/#/c/385496/ before merge, could you file a bug? | 12:45 |
| asettle | Kthanks :p | 12:45 |
| odyssey4me | admin0 is this trusty or xenial? | 12:46 |
| *** klamath has joined #openstack-ansible | 12:46 | |
| admin0 | trusty | 12:46 |
| *** klamath has quit IRC | 12:46 | |
| admin0 | xenian works fine:) | 12:47 |
| admin0 | this test is on trusty | 12:47 |
| admin0 | xenial* | 12:47 |
| *** klamath has joined #openstack-ansible | 12:47 | |
| odyssey4me | ok, so the ubuntu-cloud-keysring file is in 'universe' - make sure that's configured in your apt sources | 12:48 |
| admin0 | ok | 12:48 |
| admin0 | universe is configured | 12:49 |
| admin0 | E: Failed to fetch http://nl.archive.ubuntu.com/ubuntu/pool/universe/u/ubuntu-cloud-keyring/ubuntu-cloud-keyring_2012.08.14_all.deb 503 Service Unavailable :D | 12:49 |
| odyssey4me | admin0 ah, I think you may be hitting the thing we fixed with https://review.openstack.org/385625 which is on its way to merging into newton right now | 12:50 |
| odyssey4me | ie the package cache is a bit broken - check the repo container and see if you can find what's going on in the apt-cacher-ng log | 12:50 |
| admin0 | i almost never do aio, but only multi nodes, so that description is true for multi-node as well | 12:50 |
| admin0 | please merge now :D | 12:51 |
| admin0 | so that i can git pull and move ahead | 12:51 |
| admin0 | :D | 12:51 |
| odyssey4me | yes evrardjp - none of these drive-by comments :p register a bug already! | 12:51 |
| evrardjp | it's doen | 12:51 |
| evrardjp | done | 12:51 |
| evrardjp | I had my process to not introduce bugs and still follow, I star all I have to do | 12:52 |
| evrardjp | but yes, bug is for long term approach, and other ppl can take it! | 12:52 |
| evrardjp | admin0: got an issue? | 12:52 |
| evrardjp | with your keepalived? | 12:52 |
| admin0 | i hear i need to wait for https://review.openstack.org/385625 to be available | 12:53 |
| evrardjp | oh it was mhayden | 12:53 |
| evrardjp | so issue? | 12:53 |
| admin0 | packages cannot be authenticated issue — on keepalived | 12:53 |
| admin0 | add ubuntu cloud archive keyring | 12:53 |
| admin0 | https://gist.githubusercontent.com/a1git/c27ffb24ac8e899ceb7b44e2b97479ab/raw/918b29a549dc63e8e979c39a68c57ddf04a5f820/gistfile1.txt this one | 12:53 |
| odyssey4me | evrardjp yeah, it's more than likely the package cache in the repo | 12:54 |
| mgariepy | good morning everyone | 12:55 |
| odyssey4me | ie https://bugs.launchpad.net/openstack-ansible/+bug/1632826 & https://bugs.launchpad.net/openstack-ansible/+bug/1632847 | 12:55 |
| openstack | Launchpad bug 1632826 in openstack-ansible "apt/yum cache causes problems when rebuilding" [Undecided,Fix released] - Assigned to Jesse Pretorius (jesse-pretorius) | 12:55 |
| openstack | Launchpad bug 1632847 in openstack-ansible "repo_server package cache causing failures" [High,In progress] - Assigned to Kevin Carter (kevin-carter) | 12:55 |
| evrardjp | odyssey4me: I will still check if keepalived role has the proper keyring | 12:55 |
| evrardjp | auto latest, so yes that could be package cache of repo | 12:57 |
| admin0 | how do i force an apt-get update on everythig ( that the anisble knows of ) ? | 12:58 |
| evrardjp | mhayden: do you mean duplicate entries in the service catalog? | 12:59 |
| odyssey4me | logan- it looks like you have the right bits | 12:59 |
| mhayden | evrardjp: yessir | 12:59 |
| odyssey4me | logan- perhaps try extract those host vars into a yaml file like so? https://github.com/openstack/openstack-ansible-os_swift/blob/master/tests/group_vars/all_containers.yml | 12:59 |
| evrardjp | admin0: ansible -m apt -a "update_cache=yes" all | 12:59 |
| admin0 | thanks :) | 12:59 |
| evrardjp | god damn it we need idempotency in this keystone module! | 13:00 |
| evrardjp | :p | 13:00 |
| evrardjp | mhayden: http https? | 13:00 |
| evrardjp | or does this really look the same? | 13:00 |
| admin0 | we got no clue what you are looking at :D | 13:01 |
| admin0 | [WARNING]: Host file not found: /etc/ansible/hosts :D | 13:01 |
| admin0 | i was hoping to piggyback on the openstack ansible inventory :) | 13:01 |
| odyssey4me | admin0 you need to source /usr/local/bin/openstack-ansible.rc first | 13:01 |
| evrardjp | cd /opt/openstack-ansible/playbooks | 13:02 |
| odyssey4me | and yes, you need to be in that directory | 13:02 |
| admin0 | \o/ :D | 13:02 |
| admin0 | thanks | 13:02 |
| evrardjp | yw | 13:02 |
| *** Mudpuppy has joined #openstack-ansible | 13:02 | |
| openstackgerrit | Donovan Francesco proposed openstack/openstack-ansible-os_monasca: Refactoring Monasca role to align more with OSA. https://review.openstack.org/375655 | 13:03 |
| admin0 | facts removed, all updated, manually installed ubuntu-cloud-keysring — retrying ! | 13:04 |
| Matias | morning | 13:04 |
| admin0 | \o | 13:04 |
| * Matias back to business | 13:04 | |
| Matias | odyssey4me: you are fast, i was going to send a PR for this, but you already fixed it https://github.com/openstack/openstack-ansible/commit/ded68dd775f7187bbe4649d2e31346784c21ac69 | 13:04 |
| evrardjp | hey Matias it's been a while :) | 13:05 |
| logan- | odyssey4me: yep been there too :) http://cdn.pasteraw.com/qyywjcxqwauiukif0za30lcnjyyvv4 | 13:05 |
| admin0 | it runs now ס₪₪₪₪§|(Ξ≥≤≥≤≥≤ΞΞΞΞΞΞΞΞΞΞ> | 13:06 |
| *** cathrichardson has joined #openstack-ansible | 13:06 | |
| *** cathrichardson has quit IRC | 13:06 | |
| *** cathrichardson has joined #openstack-ansible | 13:07 | |
| *** Mudpuppy has quit IRC | 13:07 | |
| logan- | also in case its an issue with that connection plugin working with 2.1.2.0 I also tried 2.1.1.0 | 13:09 |
| logan- | (same result) | 13:09 |
| *** jperry has joined #openstack-ansible | 13:11 | |
| *** Mudpuppy has joined #openstack-ansible | 13:11 | |
| *** jheroux has joined #openstack-ansible | 13:14 | |
| drifterza | http://logs.openstack.org/55/375655/90/check/gate-openstack-ansible-os_monasca-ansible-func-ubuntu-trusty/02e1439/console.html#_2016-10-13_12_53_56_788954 | 13:20 |
| drifterza | Any advise ? I'm using an galaxy role for external dependancies and I can't work out why the dict is not being picked up. | 13:21 |
| drifterza | Anyone that can do a review and help me, would appreciate it. | 13:21 |
| odyssey4me | logan- if you give the container an IP address - perhaps it'll work? that's pretty much the main difference I see here | 13:22 |
| Jack_Iv | Hi guys! Can anybody help me ? https://github.com/ansible/ansible-modules-core/issues/5250 | 13:22 |
| logan- | if i set ansible_host=10.100.100.2 it just tries to connect to ssh on 10.100.100.2 directly | 13:23 |
| odyssey4me | drifterza don't use that role - it requires hash merging enabled, which is a terribad idea | 13:23 |
| odyssey4me | https://github.com/sansible/users_and_groups#ansiblecfg | 13:24 |
| *** drifterza has quit IRC | 13:26 | |
| *** agrebennikov has joined #openstack-ansible | 13:27 | |
| *** pjm6 has quit IRC | 13:27 | |
| *** janki has quit IRC | 13:28 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova: Delegate from first available compute host https://review.openstack.org/385221 | 13:30 |
| evrardjp | I'm avoiding sansible like the plague | 13:31 |
| *** williamcaban has quit IRC | 13:31 | |
| admin0 | is this also a known thing ? RetryError: HTTPConnectionPool(host='internal.stack31.com', port=8181): Max retries exceeded with url: /os-releases/13.3.5/ (Caused by ResponseError('too many 503 error responses',)) | 13:31 |
| *** fguillot has joined #openstack-ansible | 13:31 | |
| evrardjp | I think it's time to kill these 503 | 13:32 |
| admin0 | killall -9 503’s didn’t work :D | 13:32 |
| evrardjp | :p | 13:32 |
| admin0 | of 10 physical servers, fails on pip install -U lxc-python2 | 13:32 |
| evrardjp | admin0: what are you running right now? | 13:32 |
| admin0 | on one | 13:32 |
| admin0 | doh ! | 13:32 |
| *** agrebennikov has quit IRC | 13:33 | |
| *** smatzek_ has quit IRC | 13:33 | |
| admin0 | openstack-ansible setup-hosts | 13:36 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova: Delegate from first available compute host https://review.openstack.org/385991 | 13:37 |
| *** haad1 has quit IRC | 13:37 | |
| admin0 | starting fresh again .. | 13:39 |
| odyssey4me | Jack_Iv perhaps evrardjp can assist, I think he's got a setup to test something like that | 13:40 |
| Jack_Iv | odyssey4me, Thanks! | 13:41 |
| *** haad1 has joined #openstack-ansible | 13:41 | |
| Jack_Iv | evrardjp, If you are not busy, take a look, please https://github.com/ansible/ansible-modules-core/issues/5250 | 13:41 |
| odyssey4me | oh yes, logan- FYI there should be non-voting checks on os_neutron now for the calico environment | 13:42 |
| logan- | nice! | 13:42 |
| logan- | thanks | 13:42 |
| odyssey4me | logan- now we just need a test patch up to see if it works :) | 13:42 |
| logan- | haha yeah i was just looking at https://review.openstack.org/#/q/project:openstack/openstack-ansible-os_neutron to see if anything checked yet | 13:43 |
| logan- | nothing since the 8th | 13:43 |
| evrardjp | Jack_Iv: I can't right now, my openstack playtoys are in rebuilding | 13:43 |
| evrardjp | but I will follow this issue | 13:44 |
| Jack_Iv | evrardjp, Thank you! | 13:44 |
| openstackgerrit | Logan V proposed openstack/openstack-ansible-os_neutron: Testing Calico check https://review.openstack.org/385992 | 13:44 |
| *** markvoelker has joined #openstack-ansible | 13:47 | |
| *** aludwar has joined #openstack-ansible | 13:48 | |
| odyssey4me | andymccr perhaps you'd care to review https://review.openstack.org/378930 now? | 13:53 |
| odyssey4me | seeing as you're aware of how it all works :) | 13:53 |
| *** rgogunskiy has joined #openstack-ansible | 13:56 | |
| ioni | cloudnull, i did not managed to fix the novnc issue. I just tested on a newly installed env and i can replicate it | 13:56 |
| odyssey4me | ioni can you register a bug please with the steps & config to reproduce? | 13:57 |
| ioni | odyssey4me, ok | 13:57 |
| *** adrian_otto has quit IRC | 13:57 | |
| *** smatzek_ has joined #openstack-ansible | 13:57 | |
| odyssey4me | ioni you may need to include something like an apt package list too, just in case | 13:57 |
| odyssey4me | perhaps just have that in a gist for review | 13:57 |
| ioni | odyssey4me, ok, i'll mention the configuration | 13:58 |
| ioni | odyssey4me, i'll also test your patch related to nova-key distribution | 14:00 |
| *** evrardjp has quit IRC | 14:00 | |
| *** evrardjp has joined #openstack-ansible | 14:00 | |
| *** adrian_otto has joined #openstack-ansible | 14:01 | |
| odyssey4me | cloudnull argh, now that the OSIC nodepool max is low we're hitting all the build timeouts again | 14:01 |
| *** jmckind has joined #openstack-ansible | 14:02 | |
| odyssey4me | ioni not my patch - jmccrory's :) I just did the backport | 14:03 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Prevent overlayfs use in test when kernel < 3.18 or release == trusty https://review.openstack.org/351776 | 14:07 |
| *** sdake_ has quit IRC | 14:09 | |
| *** sdake has joined #openstack-ansible | 14:10 | |
| *** rromans has quit IRC | 14:11 | |
| *** rromans has joined #openstack-ansible | 14:11 | |
| *** agrebennikov has joined #openstack-ansible | 14:14 | |
| *** jamesdenton has joined #openstack-ansible | 14:14 | |
| openstackgerrit | Merged openstack/openstack-ansible: Fix container log bind mount regression https://review.openstack.org/385821 | 14:14 |
| *** openstackgerrit has quit IRC | 14:18 | |
| *** openstackgerrit has joined #openstack-ansible | 14:19 | |
| *** Oku_OS-away is now known as Oku_OS | 14:22 | |
| *** spotz_zzz is now known as spotz | 14:22 | |
| cloudnull | evrardjp: regarding https://review.openstack.org/#/c/385800 mind removing -2 even if you disagree? also why would we want to remain on a dynamic hardware address for containers which we know will be longed lived and we know will cause terrible issues (specifically in neutron) should a container be restarted? | 14:23 |
| cloudnull | even if we're generating the mac in the same way being done before? | 14:23 |
| cloudnull | we're just simply not allowing it to regenerate | 14:24 |
| asettle | cloudnull mattt andymccr d34dh0r53 hughsaunders b3rnard0 palendae Sam-I-Am odyssey4me serverascode rromans erikmwilson mancdaz _shaps_ BjoernT claco echiu dstanek jwagner ayoung prometheanfire evrardjp arbrandes mhayden scarlisle luckyinva ntt javeriak automagically spotz vdo jmccrory alextricity25 jasondotstar admin0 michaelgugino ametts v1k0d3n severion bgmccollum darrenc JRobinson__ asettle coli | 14:27 |
| asettle | nmcnamara thorst adreznec eil397 | 14:27 |
| asettle | qwang nishpatwa_ | 14:27 |
| asettle | GUYS DONT FORGET MEETING in an hour and a half :D | 14:27 |
| asettle | I will be your host | 14:27 |
| asettle | Get jiggy with it | 14:27 |
| cloudnull | wooo. | 14:27 |
| hughsaunders | such ping so asettle | 14:27 |
| cloudnull | WOW! | 14:27 |
| adreznec | much early wow | 14:27 |
| odyssey4me | asettle look at you, getting the time right and errithing | 14:27 |
| prometheanfire | asettle: you're turn? | 14:27 |
| asettle | odyssey4me: I know, tell me about it. | 14:27 |
| asettle | prometheanfire: DDAMN STRAIGHT | 14:27 |
| asettle | I'm taking over this joint | 14:27 |
| prometheanfire | k | 14:28 |
| asettle | adreznec: I just like to get you all pumped up, you know? | 14:28 |
| asettle | PREPARE | 14:28 |
| admin0 | i use self-generated HW:address :) tied to the ip bits | 14:28 |
| admin0 | makes troubeshooting easier ) | 14:28 |
| asettle | Just so everyone gets in the appropriate mood for hte kind of meeting I"ll be hosting: https://www.youtube.com/watch?v=Dmt45BF5SOM | 14:28 |
| admin0 | asettle: meetings here: https://www.youtube.com/watch?v=BWPnjp12W5o | 14:29 |
| * prometheanfire was hoping she'd catch that | 14:29 | |
| admin0 | recommended watch :) | 14:29 |
| admin0 | office safe :) | 14:29 |
| *** chris_hultin|AWA is now known as chris_hultin | 14:30 | |
| asettle | And SO is the Lion King | 14:30 |
| asettle | This is a dictatorship and I rule. Everyone watch the Lion King video plz. | 14:31 |
| * asettle takes over in highly dramatic fashion | 14:31 | |
| hughsaunders | asettle: you're planning a coup? | 14:31 |
| asettle | hughsaunders: when *dont* I plan a coup? | 14:31 |
| cloudnull | admin0: ++ | 14:33 |
| admin0 | you watched :) “business.. business ..” | 14:33 |
| * hughsaunders wonders what evrardjp thinks about that | 14:35 | |
| admin0 | all openstack flow/api images are old (not complete) .. you guys ( as in those in rackspace ) are the ones with a billion hours on it .. do you guys have an updated flow diagram ? | 14:35 |
| evrardjp | what? | 14:35 |
| hughsaunders | evrardjp: admin0's video | 14:35 |
| admin0 | evrardjp is from where :D ? | 14:36 |
| evrardjp | A neighbour country | 14:36 |
| prometheanfire | that was good | 14:36 |
| admin0 | either belgium or germany :D | 14:37 |
| admin0 | sorry france | 14:37 |
| admin0 | germany is not metioned there | 14:37 |
| prometheanfire | need to send that to a fiend I have over there | 14:38 |
| evrardjp | omg some parts are so true | 14:39 |
| prometheanfire | The belgian parts? :P | 14:40 |
| evrardjp | he sounds obviously dutch | 14:40 |
| prometheanfire | should send it to my belgian friend too :P | 14:40 |
| odyssey4me | heh admin0 I *just* registered this bug: https://bugs.launchpad.net/openstack-ansible/+bug/1633097 | 14:41 |
| openstack | Launchpad bug 1633097 in openstack-ansible "Replace images with generated sphinx diagrams" [Wishlist,New] | 14:41 |
| evrardjp | well the dutch part compared to the belgian part | 14:41 |
| evrardjp | the belgian part isn't really accurate | 14:41 |
| evrardjp | but the comparison is good | 14:41 |
| admin0 | odyssey4me: thanks | 14:41 |
| odyssey4me | admin0 FYI the newton diagrams are reasonably up to date - we edited them for Newton - what do you think is wrong? | 14:41 |
| admin0 | whats the belgiam part ? | 14:41 |
| evrardjp | just watch it again then :p | 14:42 |
| evrardjp | and the restaurant part... it felt like amsterdam | 14:42 |
| admin0 | not the location.. i mean the terms and how they start or what is involved in meetings | 14:43 |
| admin0 | dutch is pretty accurate .. coffee and done | 14:43 |
| *** adrian_otto has quit IRC | 14:44 | |
| evrardjp | yup | 14:45 |
| cloudnull | odyssey4me: quick lets go recheck everyones jobs to get the nodepool max up :p | 14:47 |
| odyssey4me | lol recheck them all | 14:47 |
| odyssey4me | even the new infracloud is unable to complete an integrated build on time | 14:48 |
| odyssey4me | it gets very, very close | 14:48 |
| odyssey4me | oh good, all 3 jobs in the gate queue are on OSIC cloud1 - that's a good sign | 14:49 |
| *** cmart has joined #openstack-ansible | 14:50 | |
| *** rgogunskiy has quit IRC | 14:51 | |
| odyssey4me | cloudnull I take it that the OSIC 3500/3700 flavors relate to the Intel 3500/3700 SSD's? | 14:52 |
| *** Andrew_jedi has joined #openstack-ansible | 14:53 | |
| *** omiday has quit IRC | 14:56 | |
| *** kylek3h has joined #openstack-ansible | 14:56 | |
| Andrew_jedi | Folks, can i use OSA with CentOS now ? | 14:58 |
| prometheanfire | iirc, yes | 14:59 |
| prometheanfire | there's a centos gate right? | 14:59 |
| prometheanfire | or is that still wip? | 14:59 |
| *** sdake has quit IRC | 15:00 | |
| cloudnull | odyssey4me: yes. | 15:01 |
| cloudnull | we're gating on 3 different hardware profiles. | 15:01 |
| adreznec | prometheanfire: I think there's centos gate (non-voting) for some projects | 15:02 |
| cloudnull | and collecting metrics on those | 15:02 |
| adreznec | No integrated gate though | 15:02 |
| adreznec | So still WIP, I guess | 15:02 |
| prometheanfire | k | 15:02 |
| *** sdake has joined #openstack-ansible | 15:05 | |
| odyssey4me | adreznec centos is in the integrated gate - but experimental only | 15:05 |
| *** adrian_otto has joined #openstack-ansible | 15:05 | |
| adreznec | odyssey4me: Ah ok, must have missed that | 15:05 |
| odyssey4me | Andrew_jedi not just yet - mgariepy has been slowing working through the various issues and is busy working on getting nova to work | 15:05 |
| odyssey4me | Andrew_jedi so we have all the infrastructure working, and voting on centos - and we have keystone, glance, cinder and neutron working on centos | 15:06 |
| odyssey4me | next steps are nova, then swift - then we have a minimum use-case for testing with an integrated build | 15:06 |
| mgariepy | Andrew_jedi, odyssey4me nova is getting further but i still have some issues. | 15:09 |
| *** smatzek_ is now known as smatzek | 15:09 | |
| openstackgerrit | Merged openstack/openstack-ansible: Switch the repo_server role to master https://review.openstack.org/385684 | 15:10 |
| openstackgerrit | Merged openstack/openstack-ansible: [docs] Add network config example for test and prod https://review.openstack.org/385272 | 15:10 |
| *** phalmos has joined #openstack-ansible | 15:11 | |
| mgariepy | Andrew_jedi, i'll do what I can to land it in Ocata | 15:13 |
| *** phalmos_ has joined #openstack-ansible | 15:14 | |
| Andrew_jedi | mgariepy odyssey4me : Thanks! | 15:14 |
| Andrew_jedi | mgariepy: Are you coming to Barcelona ? | 15:15 |
| mgariepy | Andrew_jedi, no, but i'll probably go to boston as it's only 4 hour drive from me. | 15:16 |
| *** phalmos has quit IRC | 15:16 | |
| odyssey4me | Andrew_jedi I'm pretty sure that mgariepy would love some help. :) | 15:16 |
| Andrew_jedi | mgariepy: Gotcha! | 15:16 |
| odyssey4me | We're all helping a bit... but most of the contributors are Ubuntu users - so we implemented CentOS support up to keystone/glance/cinder as a labour of love. | 15:17 |
| odyssey4me | (it took two cycles) | 15:17 |
| Andrew_jedi | odyssey4me: That's what i have in mind. Getting some red hat shops to start trying out OSA instead of Director. | 15:17 |
| odyssey4me | This is the last stretch, and totally doable with 2-3 people across all the integrated roles within this cycle. | 15:18 |
| *** asettle has quit IRC | 15:18 | |
| *** dragonmaster has joined #openstack-ansible | 15:18 | |
| *** dragonmaster is now known as allanice001 | 15:18 | |
| *** asettle has joined #openstack-ansible | 15:18 | |
| *** haad1 has quit IRC | 15:19 | |
| Andrew_jedi | odyssey4me: Got it. Let's see if we can convince some people to help out here. Having OSA support for CentOS will be in the best interest of people sitting with certain red hat licenses. | 15:23 |
| *** afred312 has joined #openstack-ansible | 15:23 | |
| *** omiday has joined #openstack-ansible | 15:23 | |
| odyssey4me | Andrew_jedi Yes, 'Operators' :) | 15:23 |
| odyssey4me | ie not vendors | 15:23 |
| *** weezS has joined #openstack-ansible | 15:24 | |
| Andrew_jedi | Yep, exactly :) | 15:24 |
| odyssey4me | That said, it would be awfully nice if the vendors decided to pitch in and consume the toolset we provide. :) | 15:25 |
| mgariepy | odyssey4me, indeed it would be nice. | 15:25 |
| mgariepy | quick questions: /etc/openstack_deploy/openstack_inventory.json can I destroy this file in order to get it updated with the latest config change i have made ? | 15:26 |
| mgariepy | ^^ (it's on kilo) | 15:26 |
| *** phalmos_ has quit IRC | 15:27 | |
| mgariepy | I removed some nova_nova_conf_overrides in openstack_user_config for containers and it is stuck with the old value. | 15:28 |
| mgariepy | Or should I just edit the json file ? | 15:28 |
| *** Andrew_jedi has quit IRC | 15:29 | |
| odyssey4me | mgariepy if you destroy that file, any existing containers will no longer be in your inventory | 15:29 |
| odyssey4me | mgariepy so rather edit the file | 15:29 |
| odyssey4me | mgariepy FYI In Newton thanks to palendae I believe we've made it clean up after itself. :) | 15:29 |
| mgariepy | haha great. | 15:29 |
| mgariepy | but I have some policital issue before I can upgrade this small cloud ;) | 15:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [docs] Add network config example for test and prod https://review.openstack.org/386049 | 15:31 |
| *** asettle has quit IRC | 15:33 | |
| *** asettle has joined #openstack-ansible | 15:33 | |
| *** fguillot has quit IRC | 15:38 | |
| *** irtermit- is now known as irtermite | 15:39 | |
| *** woodard has quit IRC | 15:43 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Prevent overlayfs use in test when kernel < 3.18 or release == trusty https://review.openstack.org/351776 | 15:43 |
| *** woodard has joined #openstack-ansible | 15:43 | |
| *** appprod0 has joined #openstack-ansible | 15:45 | |
| *** Andrew_jedi has joined #openstack-ansible | 15:46 | |
| ioni | evrardjp, what do you mean "client locales"? client as the one that opens the browser with the console? | 15:48 |
| *** vishwanathj_zzz is now known as vishwanathj | 15:48 | |
| *** admin0 has quit IRC | 15:50 | |
| evrardjp | yes, but also the system locales etc | 15:50 |
| ioni | ok, client is also windows, same issue | 15:50 |
| ioni | nothing changed | 15:50 |
| odyssey4me | any cores around, we need a rapid review of this backport please: https://review.openstack.org/386049 | 15:50 |
| openstackgerrit | Merged openstack/openstack-ansible: Change the common proxy cache manage tasks to be stateful https://review.openstack.org/385757 | 15:51 |
| odyssey4me | jmccrory automagically cloudnull d34dh0r53 stevelle mattt hughsaunders andymccr mhayden evrardjp asettle rromans backport review please: https://review.openstack.org/386049 | 15:54 |
| evrardjp | ok | 15:54 |
| palendae | odyssey4me, mgariepy the scripts/inventory-manage.py file allows you to remove hosts, if that's what you're after | 15:55 |
| palendae | But simply re-running dynamic_inventory,py should update values | 15:55 |
| odyssey4me | palendae in newton, not kilo | 15:55 |
| palendae | Ah | 15:55 |
| * palendae cant' remember when that was addressed | 15:55 | |
| *** javeriak has joined #openstack-ansible | 15:56 | |
| asettle | On that note | 15:56 |
| asettle | cloudnull mattt andymccr d34dh0r53 hughsaunders b3rnard0 palendae Sam-I-Am odyssey4me serverascode rromans erikmwilson mancdaz _shaps_ BjoernT claco echiu dstanek jwagner ayoung prometheanfire evrardjp arbrandes mhayden scarlisle luckyinva ntt javeriak automagically spotz vdo jmccrory alextricity25 jasondotstar admin0 michaelgugino ametts v1k0d3n severion bgmccollum darrenc JRobinson__ asettle coli | 15:57 |
| asettle | nmcnamara thorst adreznec eil397 | 15:57 |
| asettle | qwang nishpatwa_ | 15:57 |
| asettle | Meeting in 5 minutes | 15:57 |
| palendae | asettle, aye aye o7 | 15:57 |
| asettle | HAUL ASS TO MEETING ROOM 4 PLZ | 15:57 |
| asettle | For those playing along at home, that's #openstack-meeting-4 | 15:57 |
| odyssey4me | #openstack-meeting-4 for the lazy | 15:57 |
| * palendae taps ctrl-pgdown | 15:57 | |
| odyssey4me | lol | 15:57 |
| palendae | There, ass hauled | 15:57 |
| *** woodard has quit IRC | 15:57 | |
| asettle | Meeting in openstack-meeting-4 now hoomans :) | 16:01 |
| *** phalmos has joined #openstack-ansible | 16:03 | |
| *** michaelgugino has joined #openstack-ansible | 16:03 | |
| *** hughmFLEXin has joined #openstack-ansible | 16:05 | |
| *** phalmos has quit IRC | 16:08 | |
| *** phalmos has joined #openstack-ansible | 16:14 | |
| *** allanice001 has quit IRC | 16:14 | |
| *** karimb has quit IRC | 16:16 | |
| *** hughmFLEXin has quit IRC | 16:18 | |
| *** Mudpuppy has quit IRC | 16:20 | |
| *** Mudpuppy has joined #openstack-ansible | 16:21 | |
| *** Mudpuppy has quit IRC | 16:25 | |
| *** hughmFLEXin has joined #openstack-ansible | 16:26 | |
| *** Mudpuppy has joined #openstack-ansible | 16:27 | |
| *** TxGirlGeek has joined #openstack-ansible | 16:36 | |
| palendae | Do we support Ansible 2.1 yet? | 16:37 |
| odyssey4me | palendae in newton+ yes | 16:37 |
| palendae | Looks like some of the modules will need at least 2.1 | 16:37 |
| palendae | Cool | 16:37 |
| odyssey4me | we're shipping with 2.1.1 in newton | 16:37 |
| odyssey4me | 2.1.2 is buggy | 16:38 |
| odyssey4me | 2.2 is near RC IIRC - so if we need to do any module rewriting we should probably ensure we support 2.1 and 2.2 | 16:38 |
| jmccrory | does stable/newton still need to be able to run on 1.9? | 16:38 |
| odyssey4me | jmccrory it already can't | 16:38 |
| stevelle | ez meeting, I didn't have to say anything :D | 16:38 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create: Generate Static MAC addresses for containers https://review.openstack.org/385800 | 16:38 |
| jmccrory | yeah, didn't think it did | 16:38 |
| palendae | odyssey4me, Probably, yeah. Was mostly just looking at stuff in the upstream extras/core repos | 16:39 |
| odyssey4me | 2.2 brings some pretty crazy performance gains, so I'd think we want to get on that ASAP - and possibly even make it so in newton | 16:40 |
| palendae | Cool | 16:40 |
| palendae | I'll be out tomorrow and Monday, so module research will probably be slow til at least Tuesday | 16:40 |
| odyssey4me | palendae if you put up an etherpad where you make notes, perhaps jmccrory will add to your notes seeing as he barely sleeps :p | 16:41 |
| palendae | Yep | 16:41 |
| jmccrory | i've looked at os_image a little while ago, it can pretty much a drop in replacement for our glance module | 16:41 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create: Generate Static MAC addresses for containers https://review.openstack.org/385800 | 16:42 |
| *** sdake_ has joined #openstack-ansible | 16:42 | |
| palendae | jmccrory, Nice | 16:42 |
| odyssey4me | jmccrory there was a recent update to our module for magnum - so that may need re-assessment | 16:42 |
| jmccrory | just uploads images, from disk instead of url though so will also be more reliable and fix https://bugs.launchpad.net/openstack-ansible/+bug/1538642 | 16:42 |
| openstack | Launchpad bug 1538642 in openstack-ansible "Cirros Image creation failures" [Undecided,In progress] - Assigned to Hugh Saunders (hughsaunders) | 16:42 |
| cloudnull | evrardjp odyssey4me updated based on some of your feedback | 16:42 |
| cloudnull | it'd be nice to get some more review on that. | 16:42 |
| odyssey4me | cloudnull evrardjp is out for the day | 16:43 |
| *** woodard has joined #openstack-ansible | 16:43 | |
| odyssey4me | logan- so the test scripts make some assumptions | 16:44 |
| openstackgerrit | Merged openstack/openstack-ansible: [docs] Provide example configurations https://review.openstack.org/385496 | 16:44 |
| odyssey4me | one is that the ANSIBLE_INVENTORY is in {toxinidir}/tests/inventory | 16:44 |
| odyssey4me | unless otherwise told | 16:44 |
| odyssey4me | for os_neutron we set https://github.com/openstack/openstack-ansible-os_neutron/blob/master/tox.ini#L26 | 16:44 |
| palendae | jmccrory, https://etherpad.openstack.org/p/osa-ocata-modules | 16:45 |
| odyssey4me | because that's the default inventory for most of the tests | 16:45 |
| palendae | Also added to the Summit etherpad | 16:45 |
| odyssey4me | logan- for the calico test it should get overwritten: https://github.com/openstack/openstack-ansible-os_neutron/blob/master/tox.ini#L154 | 16:45 |
| *** fguillot has joined #openstack-ansible | 16:45 | |
| *** hughmFLEXin has quit IRC | 16:45 | |
| *** hwoarang has quit IRC | 16:45 | |
| *** sdake has quit IRC | 16:46 | |
| *** hughmFLEXin has joined #openstack-ansible | 16:46 | |
| *** hwoarang has joined #openstack-ansible | 16:47 | |
| logan- | interesting odyssey4me. and yet we see a linuxbridge_agent running in the check. maybe we need to move that inventory setting into the setenv of the functional tests themselves | 16:47 |
| logan- | although if the ovs one is working... | 16:47 |
| logan- | ah I do see tox says it is using calico_inventory in the console log | 16:48 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests: Add more diagnostic info to Ansible playbook execution https://review.openstack.org/386082 | 16:48 |
| odyssey4me | logan- ^ that would be useful to figure out what's going on | 16:49 |
| odyssey4me | cloudnull jmccrory reviews please | 16:49 |
| palendae | Debugging hooks are always useful | 16:50 |
| *** woodard has quit IRC | 16:52 | |
| *** woodard has joined #openstack-ansible | 16:53 | |
| odyssey4me | stevelle you're lurking too :) a review would be most kind | 16:54 |
| *** hughmFLEXin has quit IRC | 16:54 | |
| stevelle | much multi-tasking | 16:54 |
| stevelle | good thing we like multi-tasking | 16:55 |
| cloudnull | I think we need to figure someting out here https://review.openstack.org/#/c/385800/ -- evrardjp may be against the approach and that's fine but we need to do something. In Mitaka > Newton (same with liberty > mitaka) there is at least 1 manditory container restart which can, and will, make l3 tenant (using the linuxbridge or OvS agent) networks unusable should a deployment be using l2pop (which many clouds are if they' | 16:55 |
| cloudnull | re long lived). | 16:55 |
| cloudnull | insight would be greatly appreciated. | 16:55 |
| *** hughmFLEXin has joined #openstack-ansible | 16:56 | |
| odyssey4me | cloudnull I think a bug report giving more detail about what the problem is would go a long way to helping reviewers understand why it's needed | 16:57 |
| odyssey4me | either that or the commit message needs to be more whitespace, more sentences, and less brain dump | 16:57 |
| odyssey4me | on initial review I had no idea that this was to assist with upgrades | 16:58 |
| *** weezS has quit IRC | 16:58 | |
| cloudnull | not really upgrades. just issues with rotating macs when a container is restared | 16:59 |
| cloudnull | upgrades cause that | 16:59 |
| mhayden | asettle: thanks for doing the meeeeeeting | 16:59 |
| *** vishwana_ has joined #openstack-ansible | 16:59 | |
| cloudnull | but there are plenty of reasons that a container would need to be restarted. | 16:59 |
| cloudnull | also more data is in the release note. | 16:59 |
| *** fops has quit IRC | 16:59 | |
| cloudnull | and the inline comment | 17:00 |
| *** sdake_ is now known as sdake | 17:01 | |
| palendae | for neutron/rabbitmq, changing the MAC address on a simple restart seems pretty bad to me | 17:01 |
| palendae | Upgrade or no | 17:01 |
| cloudnull | ^ | 17:01 |
| palendae | I don't know enough about the ARP layer to respond to evrardjp's comment though | 17:02 |
| asettle | mhayden: you're most welcome. I think I'm fired though :p | 17:02 |
| *** vishwanathj has quit IRC | 17:03 | |
| asettle | I made bad jokes and did wrong things. | 17:03 |
| asettle | All the above. | 17:03 |
| palendae | asettle, How's that different from any other day? >.> | 17:03 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [docs] Provide example configurations https://review.openstack.org/386091 | 17:03 |
| palendae | cloudnull, iirc only neutron and rabbitmq do the heavy lifting inside containers, right? | 17:03 |
| palendae | Compute is always on a physical host so it matters less there | 17:03 |
| cloudnull | mariadb could be in that group too | 17:04 |
| asettle | palendae: usually I tend to do the write things and make bad jokes (hahahahha ahhh too funny i'll see myself out) | 17:04 |
| cloudnull | though i've not seen issues with rotating macs w/ regard to maria | 17:04 |
| palendae | Gothca | 17:04 |
| palendae | My initial thought was computes, but then I remembered those are pretty much all physical | 17:04 |
| palendae | I suppose someone could do those in containers with the right storage set up, but I've not seen one of those | 17:04 |
| palendae | asettle, Did you start with the wine early today? | 17:04 |
| odyssey4me | another backport review please cores: https://review.openstack.org/386091 | 17:05 |
| *** phalmos has quit IRC | 17:06 | |
| odyssey4me | cloudnull yeah, understood - honestly I didn't realise that LXC did that | 17:06 |
| odyssey4me | jmccrory automagically cloudnull d34dh0r53 stevelle mattt hughsaunders andymccr mhayden evrardjp asettle rromans backport reviews please: https://review.openstack.org/386091 | 17:07 |
| palendae | odyssey4me, LXC is pretty hands off in a lot of cases | 17:08 |
| spotz | I feel left out:( | 17:08 |
| palendae | It could use a few more smarts sometimes | 17:08 |
| odyssey4me | palendae well, it kinda should be | 17:08 |
| palendae | odyssey4me, configurable smarts :p | 17:08 |
| odyssey4me | I I did assume that once a NIC for a container was created, it retained its MAC over restarts | 17:09 |
| odyssey4me | I guess that I shouldn't have assumed that they worked like VM's. :p | 17:09 |
| cloudnull | I knew this was happening however in doing the Liberty > Mitaka upgrade I got to experience the issues first hand at scale under load. | 17:09 |
| cloudnull | I wasn't aware it was a problem | 17:09 |
| cloudnull | until a couple of days ago. | 17:09 |
| palendae | Yeah, sounds like it was only uncovered with load | 17:09 |
| odyssey4me | oh yeah, I can imagine the L3 down time must have been terribad | 17:10 |
| cloudnull | yea. we have roughly 2k vms running and l3 went down | 17:10 |
| odyssey4me | having to rebuild the routing tables due to a new MAC - that's some awful down time right there | 17:10 |
| cloudnull | to recover we had to nuke the vxlan interfaces and restart the linuxbridge agent | 17:10 |
| cloudnull | FDB was all pissed off on the compute side while everything looks happy on the neutron_agent side. | 17:11 |
| cloudnull | it was only after noticing the mac changed that I went down this rabbit hole | 17:11 |
| odyssey4me | cloudnull punch please? https://review.openstack.org/386082 | 17:11 |
| asettle | palendae: lol if only. Still working, can't wine. | 17:12 |
| *** kjw3 has joined #openstack-ansible | 17:12 | |
| odyssey4me | stevelle cloudnull also, punch this backport please? https://review.openstack.org/386091 | 17:12 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [docs] Add network config example for test and prod https://review.openstack.org/386049 | 17:13 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create: Generate Static MAC addresses for containers https://review.openstack.org/385800 | 17:14 |
| *** appprod0 has quit IRC | 17:14 | |
| * palendae piles on the review train | 17:15 | |
| palendae | https://review.openstack.org/#/c/385271/ is a fairly easy one | 17:15 |
| palendae | I think from a new contributor | 17:15 |
| chris_hultin | Got a question: I'm looking at the global overrides section of openstack_user_config.yml, and looking at the "container_interface" and "host_bind_override" values, but my hosts all have different devices for the same network | 17:29 |
| openstackgerrit | Merged openstack/openstack-ansible-tests: Add more diagnostic info to Ansible playbook execution https://review.openstack.org/386082 | 17:31 |
| *** hj-hpe has joined #openstack-ansible | 17:36 | |
| *** Mudpuppy_ has joined #openstack-ansible | 17:37 | |
| *** Mudpuppy has quit IRC | 17:40 | |
| *** appprod0 has joined #openstack-ansible | 17:43 | |
| *** hughmFLEXin has quit IRC | 17:44 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible-os_swift: Move flush handles to the end of the tasks https://review.openstack.org/386109 | 17:47 |
| *** hughmFLEXin has joined #openstack-ansible | 17:49 | |
| *** phalmos has joined #openstack-ansible | 17:50 | |
| *** cathrichardson has quit IRC | 17:53 | |
| *** cfarquhar has joined #openstack-ansible | 17:55 | |
| *** weezS has joined #openstack-ansible | 17:56 | |
| odyssey4me | btw afk for a while | 17:56 |
| *** cathrichardson has joined #openstack-ansible | 17:57 | |
| *** javeriak has quit IRC | 17:59 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_heat: Add os_heat role functional tests for gate https://review.openstack.org/382790 | 18:02 |
| *** shanec has joined #openstack-ansible | 18:03 | |
| *** appprod0 has quit IRC | 18:04 | |
| *** rgogunskiy has joined #openstack-ansible | 18:07 | |
| *** shanec_ has joined #openstack-ansible | 18:07 | |
| jrosser | cloudnull: for some other non openstack stuff we turn ip into mac addr via <const>:<const>:<const>:<lower-3-octets-of-ip> | 18:07 |
| jrosser | that guarantees uniqueness in the l2 domain and makes everything completely deterministic | 18:07 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create: Generate Static MAC addresses for containers https://review.openstack.org/385800 | 18:08 |
| *** rgogunskiy has quit IRC | 18:08 | |
| *** shanec has quit IRC | 18:09 | |
| cloudnull | awesome! | 18:09 |
| cloudnull | jrosser: might you have a script or function that does that which you'd be interested in sharing? | 18:10 |
| cloudnull | i can make one. but it's easier if you already have an implementation . | 18:10 |
| cloudnull | and im lazy :) | 18:10 |
| *** Andrew_jedi has quit IRC | 18:11 | |
| openstackgerrit | Merged openstack/openstack-ansible: Prevent overlayfs use in test when kernel < 3.18 or release == trusty https://review.openstack.org/351776 | 18:12 |
| jrosser | its just a mass of yaml files that then turn into our dhcp/dns config | 18:13 |
| jrosser | very manual | 18:13 |
| cloudnull | ok . | 18:13 |
| stevelle | we should rub some xslt on it | 18:13 |
| cloudnull | I will look at making a contraption | 18:13 |
| *** rgogunskiy has joined #openstack-ansible | 18:13 | |
| *** rgogunskiy has quit IRC | 18:14 | |
| *** fops has joined #openstack-ansible | 18:19 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: Prevent overlayfs use in test when kernel < 3.18 or release == trusty https://review.openstack.org/386128 | 18:19 |
| *** sdake_ has joined #openstack-ansible | 18:19 | |
| *** sdake has quit IRC | 18:20 | |
| *** javeriak has joined #openstack-ansible | 18:22 | |
| openstackgerrit | Merged openstack/openstack-ansible: Remove xrange for run both Python 2 and Python 3 https://review.openstack.org/385271 | 18:23 |
| *** Andrew_jedi has joined #openstack-ansible | 18:23 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create: Generate Static MAC addresses for containers https://review.openstack.org/385800 | 18:23 |
| *** gouthamr has quit IRC | 18:24 | |
| *** vishwana_ has quit IRC | 18:27 | |
| *** vishwanathj has joined #openstack-ansible | 18:28 | |
| *** phalmos has quit IRC | 18:29 | |
| *** phalmos has joined #openstack-ansible | 18:31 | |
| jrosser | cloudnull: if you use that ipv4 <> mac mapping something nicely magical happens when you make an eui-64 ipv6 address :) | 18:33 |
| *** Jack_Iv has quit IRC | 18:35 | |
| *** Jack_Iv has joined #openstack-ansible | 18:35 | |
| cloudnull | oh. good point. OUI + FF FE + NIC == WIN! | 18:35 |
| openstackgerrit | Merged openstack/openstack-ansible-os_heat: Add os_heat role functional tests for gate https://review.openstack.org/382790 | 18:36 |
| cloudnull | no more manual config, no need for DHCP :) | 18:36 |
| cloudnull | tyvm RFC2373 | 18:37 |
| *** Andrew_jedi has quit IRC | 18:37 | |
| palendae | cloudnull, That's weird math for getting Windows | 18:39 |
| cloudnull | ha. | 18:39 |
| *** hughmFLEXin has quit IRC | 18:41 | |
| *** hughmFLEXin has joined #openstack-ansible | 18:44 | |
| *** chris_hultin is now known as chris_hultin|AWA | 18:45 | |
| *** shanec_ has quit IRC | 18:46 | |
| *** Jack_Iv has quit IRC | 18:48 | |
| *** Jack_Iv has joined #openstack-ansible | 18:49 | |
| *** michauds has joined #openstack-ansible | 18:49 | |
| *** galstrom_zzz is now known as galstrom | 18:49 | |
| *** admin0 has joined #openstack-ansible | 18:50 | |
| jroll | it looks like osa-ironic doesn't use upper-constraints when not in dev mode, is that intentional? https://github.com/openstack/openstack-ansible-os_ironic/blob/stable/mitaka/tasks/ironic_install.yml#L51 | 18:53 |
| jroll | (in mitaka) | 18:53 |
| jroll | actually, in master too | 18:54 |
| admin0 | hi .. i have a stuck task: start cluster with wsrep | 18:54 |
| admin0 | latest stable/mitaka | 18:54 |
| jroll | https://github.com/openstack/openstack-ansible-os_ironic/blob/master/tasks/ironic_install.yml#L50 | 18:54 |
| *** gouthamr has joined #openstack-ansible | 18:54 | |
| admin0 | sorry .. i meant newton | 18:55 |
| admin0 | stable/newton | 18:55 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift: Use dictionary for service mappings https://review.openstack.org/385852 | 18:56 |
| odyssey4me | jroll yes, because it's expected that when not in dev mode you're using the integrated build - where we use u-c | 19:00 |
| odyssey4me | if you aren't then you should know what you're doing | 19:00 |
| admin0 | hi odyssey4me .. know a quick fix for start cluster with wsrep ? | 19:00 |
| odyssey4me | jroll I expect that we're going to have to change all that in the next cycle or two as disparate requirements for projects become more of a thing | 19:01 |
| jroll | odyssey4me: ah, ok, I'm not aware of this integrated build thing... can you point me at that? | 19:01 |
| *** sdake has joined #openstack-ansible | 19:01 | |
| odyssey4me | admin0 yes, the docs tell you how to rebootstrap | 19:01 |
| admin0 | its at setup-infra time | 19:01 |
| odyssey4me | jroll https://github.com/openstack/openstack-ansible | 19:01 |
| jroll | odyssey4me: for context, looking at https://bugs.launchpad.net/openstack-ansible/+bug/1632436/ and thinking it's a dependency issue | 19:01 |
| openstack | Launchpad bug 1632436 in openstack-ansible "os_ironic uses wrong keystone_authtoken config keys" [Undecided,In progress] - Assigned to Charles Farquhar (cfarquhar) | 19:01 |
| admin0 | i did followed the docs to do the galera-boostrap command | 19:01 |
| admin0 | but stuck @ the same place | 19:01 |
| admin0 | its greenfield .. so will delete all galera containers and give it another shot | 19:02 |
| odyssey4me | jroll these are the pinned SHA's for the most recent mitaka tag: https://github.com/openstack/openstack-ansible/blob/13.3.4/playbooks/defaults/repo_packages/openstack_services.yml | 19:02 |
| jroll | odyssey4me: oh, the main thing, I'll poke around | 19:02 |
| *** shanec_ has joined #openstack-ansible | 19:02 | |
| jroll | odyssey4me: so libraries aren't pinned via u-c then? | 19:02 |
| odyssey4me | that'll help you figure that out - we use that sha's u-c as a python requirements upper bound | 19:03 |
| jroll | ah ok | 19:03 |
| odyssey4me | jroll whatever's in u-c is the pin | 19:03 |
| odyssey4me | jroll we update those SHA's every two weeks - I'll be doing a release tomorrow and updating the pins | 19:03 |
| jroll | odyssey4me: okay, this makes sense, thanks | 19:04 |
| *** sdake_ has quit IRC | 19:04 | |
| odyssey4me | we only do the services from git sources - we rely on their requirements to determine which libraries to get and use u-c for the upper bound | 19:04 |
| * jroll hrms at this bug then | 19:04 | |
| *** hughmFLEXin has quit IRC | 19:05 | |
| odyssey4me | heh, that kinda makes sense - conductor doesn't use the keystone_authtoken section I don't think | 19:05 |
| odyssey4me | only the API ever will as it's the only thing that needs to talk to users and other services, right? | 19:06 |
| jroll | the API uses it for validating tokens, the conductor uses it for talking to glance and neutron | 19:06 |
| odyssey4me | I wonder if that config section should even be read by conductor | 19:06 |
| odyssey4me | ah | 19:06 |
| odyssey4me | so one is using keystoneauth1 and another is using python-keystoneclient | 19:06 |
| *** adrian_otto has quit IRC | 19:07 | |
| jroll | well, the api uses keystone_middleware in mitaka, and the conductor just uses service clients | 19:07 |
| jroll | e.g. python-neutronclient | 19:07 |
| jroll | I believe, I'd have to look again | 19:07 |
| odyssey4me | that would make sense - and is why it's breaking | 19:07 |
| *** adrian_otto has joined #openstack-ansible | 19:07 | |
| odyssey4me | for the other services - aren't there their own sections with their own auth creds? | 19:08 |
| jroll | idk, we test admin_username etc in CI | 19:08 |
| *** hughmFLEXin has joined #openstack-ansible | 19:08 | |
| jroll | in newton, yes, not in mitaka | 19:08 |
| odyssey4me | it perhaps falls back to using keystone_authtoken if there aren't any | 19:08 |
| jroll | e.g. neutron: https://github.com/openstack/ironic/blob/stable/mitaka/ironic/dhcp/neutron.py#L81 | 19:08 |
| odyssey4me | oh dear, then it's all brokenness | 19:09 |
| odyssey4me | basically you have two incompatible library consumers | 19:09 |
| odyssey4me | so the only way to fix it is to either disable auth, or to find the last available library that supported both the new configs and the old | 19:09 |
| jroll | well, no, we test admin_tenant_name, admin_user, and admin_password in CI | 19:09 |
| jroll | and it works fine in the api as well | 19:10 |
| odyssey4me | then we have the ability to override the upper constraint for that environment | 19:10 |
| odyssey4me | ok, if you can get a pip freeze for a working set, with a working config into that bug log then andymccr and I can take a peek in the morning | 19:11 |
| jroll | although... CI is keystone v2, idk what version mitaka OSA deploys, maybe that's why | 19:11 |
| jroll | sure | 19:11 |
| odyssey4me | aha, yes | 19:11 |
| odyssey4me | the issue is the use of keystone v3 | 19:11 |
| jroll | I'd think the authtoken middleware could use both... maybe not. | 19:11 |
| odyssey4me | alright, that's easy enough then - for the ironic deployment just force the use of keystone v2 | 19:11 |
| jroll | ok, yeah, we didn't fully support v3 in mitaka | 19:11 |
| jroll | yep | 19:11 |
| jroll | want me to make a note there? | 19:12 |
| jroll | (already doing it) | 19:12 |
| odyssey4me | yeah - just note that keystone v3 will not work with ironic for Mitaka | 19:12 |
| ioni | just noticed something | 19:12 |
| odyssey4me | so the service catalog must have v2 endpoints and you must use it | 19:12 |
| ioni | the containers do not retain the system timezone | 19:12 |
| ioni | the new ones | 19:13 |
| ioni | the old ones do | 19:13 |
| jroll | cfarquhar: ^^ fyi, keystone v2 vs v3 | 19:13 |
| *** shanec_ has quit IRC | 19:13 | |
| odyssey4me | stevelle jmccrory anyone around for the second backport review? https://review.openstack.org/#/c/386128/ | 19:13 |
| jroll | odyssey4me: thanks for the help :) | 19:14 |
| *** shanec_ has joined #openstack-ansible | 19:14 | |
| cloudnull | ioni: the system timezone isn't being inherited ? | 19:14 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible: [docs] Provide example configurations https://review.openstack.org/386091 | 19:14 |
| ioni | cloudnull, no | 19:14 |
| cloudnull | hum. | 19:14 |
| ioni | root@ctrl1:~# date | 19:14 |
| ioni | Thu Oct 13 22:14:39 EEST 2016 | 19:14 |
| ioni | root@ctrl1:~# ssh ctrl1_gnocchi_container-700ecc67 | 19:14 |
| ioni | root@ctrl1-gnocchi-container-700ecc67:~# date | 19:15 |
| ioni | Thu Oct 13 19:14:42 UTC 2016 | 19:15 |
| ioni | that's why it was messing with ceilometer and gnocchi | 19:15 |
| ioni | :D | 19:15 |
| odyssey4me | yeah, the TZ is a setting in /etc | 19:15 |
| odyssey4me | so why would it be | 19:15 |
| cloudnull | ioni: can you ``cat /etc/localtime`` | 19:15 |
| ioni | odyssey4me, the old container that i didn't destroy, still have the EEST timezone | 19:15 |
| ioni | cloudnull, container or host? | 19:15 |
| cloudnull | both | 19:16 |
| odyssey4me | we could probably fix that up by changing it in the image | 19:16 |
| odyssey4me | the cache I mean | 19:16 |
| ioni | cloudnull, sure you want cat? | 19:16 |
| cfarquhar | Thanks jroll and odyssey4me | 19:16 |
| ioni | cloudnull, https://paste.xinu.at/u0I94l/ | 19:16 |
| *** jcrst has joined #openstack-ansible | 19:16 | |
| cloudnull | ls -al /etc/localtime to localtime if its a link | 19:16 |
| ioni | not a link | 19:17 |
| ioni | gnocchi container i destroyed today | 19:17 |
| odyssey4me | ok, going afk again for a bit | 19:17 |
| cloudnull | ok. | 19:17 |
| ioni | along with ceilometer | 19:17 |
| cloudnull | my hosts are both /etc/localtime -> /usr/share/zoneinfo/Etc/UTC | 19:17 |
| ioni | sadly i don't setup my controller and compute nodes to utc | 19:18 |
| ioni | i should do that | 19:18 |
| *** javeriak has quit IRC | 19:18 | |
| cloudnull | im always on /usr/share/zoneinfo/Etc/UTC that's interesting though we should change that to ensure they're in-sync | 19:18 |
| ioni | something changed in stable/newton compared with mitaka | 19:19 |
| ioni | it's the first time i destroyed through a container | 19:19 |
| odyssey4me | cloudnull just add it to the files to copy over into the cache in the lxc_hosts role | 19:20 |
| cloudnull | so you're host is on /usr/share/zoneinfo/EET right? | 19:20 |
| cloudnull | odyssey4me: yea that's what I'm thinking we can do | 19:20 |
| ioni | eest | 19:21 |
| ioni | cloudnull, EEST | 19:22 |
| *** klamath_ has joined #openstack-ansible | 19:26 | |
| *** klamath has quit IRC | 19:26 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts: Copy system localtime when creating container cache https://review.openstack.org/386187 | 19:27 |
| cloudnull | ioni: ^ | 19:27 |
| ioni | cloudnull, going to test it now | 19:29 |
| ioni | i guess setup-hosts should fix that | 19:30 |
| cloudnull | it'll fix it in the base container. | 19:30 |
| cloudnull | sadly it wont do much for your existing containers. | 19:30 |
| cloudnull | you can change the /etc/localtime in your existing containers or you can push the file from the hosts into the containers. | 19:31 |
| cloudnull | I'll looking at an upgrade step right now to ensure we get that right too, | 19:31 |
| ioni | hmm | 19:32 |
| ioni | ok, maybe the upgrade step are going to hurt somebody | 19:32 |
| ioni | i'm ok with keeping the same timezone at create | 19:32 |
| *** klamath_ has quit IRC | 19:33 | |
| *** asettle has quit IRC | 19:34 | |
| *** karimb has joined #openstack-ansible | 19:35 | |
| openstackgerrit | Merged openstack/openstack-ansible: [docs] Add network config example for test and prod https://review.openstack.org/386049 | 19:35 |
| *** jcrst has quit IRC | 19:35 | |
| *** TxGirlGeek has quit IRC | 19:45 | |
| *** TxGirlGeek has joined #openstack-ansible | 19:45 | |
| *** Jack_Iv has quit IRC | 19:48 | |
| *** dxiri has joined #openstack-ansible | 19:50 | |
| dxiri | Hi everyone! | 19:50 |
| dxiri | I just finished installing an AIO using openstack-ansible, but I have no idea about where to login into horizon or which credentials to use | 19:51 |
| dxiri | where can I find that? | 19:51 |
| logan- | dxiri: check sudo grep 'auth_admin' /etc/openstack_deploy/user_secrets.yml | 19:52 |
| logan- | that will give you the keystone auth admin password, which is your admin password for horizon | 19:52 |
| *** TxGirlGeek has quit IRC | 19:54 | |
| dxiri | mmm, all is blank in there | 19:54 |
| dxiri | oh no wait...wrong file, | 19:55 |
| dxiri | it does have a bunch of stuff but it's encrypted already | 19:55 |
| ioni | "encrypted" | 19:57 |
| ioni | the long string IS the password | 19:57 |
| ioni | :D | 19:57 |
| dxiri | omg really? lol | 19:57 |
| ioni | a bunch of numbers and letters right? | 19:58 |
| dxiri | right, thought it was some md5-hash or something | 19:58 |
| odyssey4me | http://docs.openstack.org/developer/openstack-ansible/newton/install-guide/verify-operation.html#verifying-the-dashboard-horizon :) | 19:59 |
| *** TxGirlGeek has joined #openstack-ansible | 20:00 | |
| *** weezS has quit IRC | 20:01 | |
| dxiri | thanks! | 20:02 |
| dxiri | root@aio1-utility-container-1a690cbc:/# openstack user list | 20:02 |
| dxiri | An unexpected error prevented the server from fulfilling your request. (HTTP 500) (Request-ID: req-2a19fb1b-f9fb-4be4-90af-0f5bb8874128) | 20:02 |
| dxiri | seems like its not working yet | 20:02 |
| dxiri | and trying horizon gives me "secure connection failed" on firefox | 20:03 |
| stevelle | you do have to 'source openrc' before the CLI would work, not sure if you got that dxiri | 20:06 |
| dxiri | already did :) | 20:06 |
| *** chris_hultin|AWA is now known as chris_hultin | 20:06 | |
| *** weezS has joined #openstack-ansible | 20:06 | |
| dxiri | but I think I know the problem | 20:06 |
| stevelle | yeah, I guess that error msg says yo udid | 20:06 |
| dxiri | I intended to put about 10gigs of ram on the VM | 20:06 |
| dxiri | turns out I put only 1 | 20:07 |
| dxiri | :P | 20:07 |
| stevelle | :D that could be an issue | 20:07 |
| dxiri | so probably a lot of services failed to start | 20:07 |
| dxiri | haha | 20:07 |
| *** askb has joined #openstack-ansible | 20:07 | |
| ioni | dmesg | 20:07 |
| ioni | you should see oom-killer if this is the case | 20:08 |
| stevelle | seems like galera would be pretty well non-functional under 1g, never mind other services | 20:10 |
| odyssey4me | wow, yes - 8 GB RAM min - otherwise it'll fall apart | 20:10 |
| odyssey4me | dxiri what did you deploy - from stable/newton, or from the 14.0.0.0b2 tag? | 20:10 |
| *** jcrst has joined #openstack-ansible | 20:11 | |
| *** johnmilton has quit IRC | 20:12 | |
| dxiri | 14 git checkout stable/mitaka | 20:12 |
| dxiri | 16 git checkout 13.3.4 | 20:12 |
| dxiri | that | 20:12 |
| dxiri | isn't mitaka the stable one? | 20:15 |
| odyssey4me | ah ok - yeah that's stable | 20:15 |
| dxiri | cool | 20:15 |
| odyssey4me | we're issuing the tag RC tag for newton tomorrow | 20:15 |
| odyssey4me | it *should* be the final tag before release next week thu | 20:15 |
| *** javeriak has joined #openstack-ansible | 20:15 | |
| odyssey4me | any testing you can do to help (and register bugs) would be appreciated :) | 20:17 |
| dxiri | this is whole new ground for me :) I'm really confortable with linux but this is completely different | 20:17 |
| dxiri | so many moving parts :) | 20:17 |
| dxiri | but sure! | 20:17 |
| odyssey4me | haha, welcome to openstack | 20:17 |
| dxiri | my goal is to deploy a small 3 node cloud on our lab so we can spin up testing VMs that house some CDN software | 20:18 |
| odyssey4me | well, then this may help you: http://docs.openstack.org/developer/openstack-ansible/install-guide/app-config-test.html | 20:18 |
| *** adrian_otto has quit IRC | 20:18 | |
| *** openstackgerrit has quit IRC | 20:19 | |
| odyssey4me | the newton version of that is almost merged, when it is you'll find new content in http://docs.openstack.org/developer/openstack-ansible/newton/install-guide/app-config-test.html | 20:19 |
| *** openstackgerrit has joined #openstack-ansible | 20:19 | |
| dxiri | uuuuuuuuu dude, this looks extremely similar :) | 20:19 |
| dxiri | to what I want to do | 20:19 |
| dxiri | so, lets say I want to extend that in the near future, can I just add new nodes at will? | 20:20 |
| odyssey4me | check the prod setup for something similar, with a bit more variation | 20:20 |
| *** javeriak has quit IRC | 20:20 | |
| dxiri | say another controller for redundancy, or more block storage | 20:20 |
| odyssey4me | yep - you just add more hosts into the host groups that matter to you | 20:20 |
| dxiri | but the docs advise against re-running the whole thing again, so I guess you just need to run the specific playbooks you are after? | 20:21 |
| odyssey4me | if you don't want some services, you just don't put that group into your openstack_user_config.yml | 20:21 |
| odyssey4me | you absolutely can run again - it's only the AIO that warns against running 'run-playbooks' again because the AIO is a dev tool | 20:21 |
| odyssey4me | for production deployments you don't use run-playbooks - you run the playbooks directly | 20:22 |
| odyssey4me | ie http://docs.openstack.org/developer/openstack-ansible/newton/install-guide/run-playbooks.html | 20:22 |
| dxiri | ah, that makes sense :) | 20:23 |
| odyssey4me | the AIO is a development and CI tool - not meant for production - it does some funny networking stuff | 20:23 |
| dxiri | yep I noticed, but it was the simplest to start with | 20:24 |
| dxiri | so I can get to know how this is laid out | 20:24 |
| dxiri | horizon still doesn't want to come up :( | 20:24 |
| dxiri | openstack user list is working now | 20:24 |
| odyssey4me | if you just rebooted, you may need to wait a bit | 20:25 |
| dxiri | is there such a thing as service horizon restart? | 20:25 |
| odyssey4me | but due to the resource constraints it may also have blown up | 20:25 |
| dxiri | I rebooted about 10-15 mins ago | 20:25 |
| dxiri | 10GB of RAM and 8vcpus | 20:25 |
| odyssey4me | yeah, you can lxc-attach into the horizon container and restart apache | 20:25 |
| dxiri | that's what I am running on | 20:25 |
| odyssey4me | or you can use ansible to reach into the container and do it for you :) | 20:25 |
| dxiri | os-horizon-install.yml | 20:26 |
| dxiri | so just run that playbook again? | 20:26 |
| odyssey4me | jmccrory chris_hultin the os_magnum role tests keep timing out due to the very long build time - is there a way we can cut some of that time down? | 20:26 |
| *** meteorfox has quit IRC | 20:27 | |
| odyssey4me | dxiri to use ansible to do anything you would do via shell you can use ad-hoc ansible commands | 20:27 |
| chris_hultin | odyssey4me: Do you know what part of it is taking the most time? | 20:27 |
| odyssey4me | chris_hultin I honestly haven't dug into the logs at all | 20:27 |
| odyssey4me | just putting it out there in the hopes that someone does :) | 20:27 |
| odyssey4me | dxiri see for example a shell command you may be familiar with: http://docs.openstack.org/developer/openstack-ansible/mitaka/install-guide/ops-galera-recovery.html#multi-node-failure | 20:28 |
| odyssey4me | ansible <target_host> -m shell -a "<shell command to execute>" | 20:29 |
| odyssey4me | dxiri so for instance to restart apache on all horizon hosts: ansible horizon_all -m shell -a "service apache2 restart" | 20:29 |
| *** meteorfox has joined #openstack-ansible | 20:29 | |
| odyssey4me | you'll need to be in the /opt/openstack-ansible/playbooks directory for all ansible command execution | 20:29 |
| *** smatzek has quit IRC | 20:30 | |
| dxiri | I think I have a half baked install | 20:31 |
| dxiri | aio1_horizon_container-4b31ac5b | FAILED | rc=1 >> | 20:31 |
| dxiri | apache2: unrecognized service | 20:31 |
| odyssey4me | oh yeah, sounds like it | 20:31 |
| *** javeriak has joined #openstack-ansible | 20:32 | |
| odyssey4me | trash and rebuild the AIO - this time with enough RAM ;) | 20:32 |
| dxiri | yep, on it now | 20:32 |
| dxiri | I am having fun and boss is cool with this so no pressure :) | 20:32 |
| odyssey4me | FYI the secrets are randonly generated per install - that's why you thought it was 'encrypted' | 20:33 |
| odyssey4me | before the deploy you can change them to something else if you like | 20:33 |
| odyssey4me | but once it's deployed you can't | 20:33 |
| odyssey4me | we have a big read warning note about that :) http://docs.openstack.org/developer/openstack-ansible/newton/install-guide/configure.html#configuring-service-credentials | 20:34 |
| odyssey4me | anytime time for me to go afk again - have fun! | 20:34 |
| *** rmelero has joined #openstack-ansible | 20:35 | |
| spotz | night odyssey4me | 20:35 |
| jmccrory | chris_hultin odyssey4me : looks like it's ~40 minutes to set up everything required, and then the tempest tests taking the rest of the time | 20:38 |
| dxiri | odyssey4me: thanks a lot for all the help! | 20:38 |
| chris_hultin | jmccrory: Problem is, we need every bit of OpenStack, basically. Heat, Nova, Cinder, Identity, Neutron | 20:38 |
| jmccrory | yeah... | 20:38 |
| evrardjp | cloudnull: I'm back for a few moments. If you make it optional, wouldn't it fix my problem with that, and would it solve your problem? | 20:43 |
| cloudnull | yes. but we'd be turning it on by default on the services that matter (neutron, rabbitmq, galera) so it'd be optionally enabled by default. | 20:44 |
| *** sdake has quit IRC | 20:44 | |
| cloudnull | also i'm failing to see what the net negative is ? | 20:44 |
| agrebennikov | hey odyssey4me, not sure you are a right person for this question, but still.... | 20:45 |
| agrebennikov | the dynamic inventory | 20:45 |
| cloudnull | if the mac is always changing which is now known to cause problems and we can stop that using the same general process used to generate the always changing mac what is the apprehension ? | 20:45 |
| agrebennikov | there is a script for dyn inventory as well as group_vars folder | 20:45 |
| agrebennikov | am I supposed to change tha parameters in that folder ever? | 20:46 |
| evrardjp | cloudnull: we have to fix the root cause of the issue | 20:48 |
| cloudnull | which is the changing mac | 20:48 |
| evrardjp | why would that be a problem? fdb tables building is supposed to be fine | 20:48 |
| evrardjp | if I spawn up a new device on a network I am supposed to see it | 20:49 |
| cloudnull | right, and in an AIO that works fine. | 20:49 |
| cloudnull | add 500 hosts and roughly 2000 vms and that goes away | 20:50 |
| evrardjp | so question | 20:50 |
| cloudnull | in the fullness of time it'd work. | 20:50 |
| evrardjp | do we have to implement a workaround for all the cases, where it's not supposed to happen at the first time? | 20:50 |
| cloudnull | but downtime while the fullness of time is taking place is not an option | 20:50 |
| evrardjp | but what is the problem? | 20:51 |
| evrardjp | are the fdb tables full? | 20:51 |
| evrardjp | is the learning too slow? | 20:51 |
| cloudnull | no. they're wrong. | 20:51 |
| evrardjp | is the population failing? | 20:51 |
| evrardjp | should we change the protocol for learning? | 20:51 |
| cloudnull | l2pop only pushes on changes. | 20:51 |
| evrardjp | and it's l2pop only? | 20:51 |
| cloudnull | so you see failed and stale entries | 20:51 |
| evrardjp | well arp stale entries happen all the time | 20:52 |
| evrardjp | they expire | 20:52 |
| cloudnull | they will expire in the fullness of time. | 20:52 |
| cloudnull | the fix then is to delete the vxlan interfaces | 20:52 |
| cloudnull | and restart the agents | 20:52 |
| evrardjp | which is dependent of the implementation we agree | 20:52 |
| cloudnull | then let everything rebuild | 20:52 |
| evrardjp | which is definitely ugly I agree | 20:52 |
| cloudnull | and the root cause is a changed mac address | 20:53 |
| evrardjp | so the problem is the mac learning is too slow or doesn't scale for your use case | 20:53 |
| evrardjp | fixing the mac adress makes everything simple | 20:53 |
| evrardjp | but won't solve any other issue you will have with your arp tables for the rest of your network | 20:53 |
| evrardjp | but anyway, other question | 20:54 |
| openstackgerrit | Merged openstack/openstack-ansible: Prevent overlayfs use in test when kernel < 3.18 or release == trusty https://review.openstack.org/386128 | 20:54 |
| cloudnull | adding a node into the stack will learn everything needed but dealing with those properties everytime we need to restart a container is bad. | 20:54 |
| evrardjp | 2000vms 500 hosts, how much entries are we talking here | 20:54 |
| evrardjp | how much containers? | 20:55 |
| evrardjp | many* | 20:55 |
| evrardjp | it's nothing much | 20:55 |
| cloudnull | just a three node control plane | 20:55 |
| evrardjp | yeah | 20:55 |
| cloudnull | but there's an entry for every compute node, tenant network, router, etc. | 20:55 |
| evrardjp | shouldn't we try to find what's the real cause here? | 20:55 |
| *** smatzek has joined #openstack-ansible | 20:55 | |
| evrardjp | I mean it looks like an interesting scaling issue problem | 20:55 |
| cloudnull | so why do you want to see a random mac on container restart? | 20:57 |
| cloudnull | how is that better? | 20:57 |
| evrardjp | I don't care about the randomness or not -- I care about the base system to work | 20:57 |
| evrardjp | we shouldn't have to hack | 20:57 |
| cloudnull | what's the hack? | 20:57 |
| cloudnull | generally speaking mac addresses dont change. | 20:58 |
| evrardjp | hopefully we are not at static mac addresses, we are still at fixed dynamic mac address | 20:58 |
| cloudnull | IMO the hack is that they currently do | 20:58 |
| evrardjp | right? | 20:58 |
| *** weezS has quit IRC | 20:58 | |
| evrardjp | well it was done for a reason | 20:58 |
| *** sdake has joined #openstack-ansible | 20:59 | |
| cloudnull | w hat was? | 20:59 |
| evrardjp | you restart your container and boom it will force others to refresh their table for it with the new arp that comes with the new interface | 20:59 |
| openstackgerrit | Andrey Grebennikov proposed openstack/openstack-ansible: Added "service_region" to user_variables.yml https://review.openstack.org/386225 | 20:59 |
| evrardjp | but yes we can say it's fixed and not necessary | 20:59 |
| evrardjp | I don't argue negatively about having it "fixed" or not | 21:00 |
| evrardjp | (I hate the term static because it makes me think static arp entry) | 21:01 |
| cloudnull | would 's/static/fixed/g' in the changes make it better ? | 21:01 |
| evrardjp | I argued about the fact that we should fix the problem with mac learning | 21:02 |
| evrardjp | well it scares me less, but it's not the point ;) | 21:02 |
| evrardjp | all the code we have to carry for something that should work... I just don't like it | 21:03 |
| evrardjp | I will remove the vote and leave to others | 21:03 |
| *** weezS has joined #openstack-ansible | 21:03 | |
| cloudnull | mac learning works but i'd rather not take a cloud down while things are being learned. | 21:03 |
| cloudnull | which is compounded by agent intervals and general load. | 21:04 |
| *** jamielennox is now known as jamielennox|away | 21:04 | |
| evrardjp | you had a learning storm | 21:04 |
| *** karimb has quit IRC | 21:04 | |
| evrardjp | for your neutron container I guess then | 21:04 |
| evrardjp | your neutron container restarted and then everything had to learn it! | 21:05 |
| evrardjp | is that your pain? | 21:05 |
| cloudnull | yup. which includes the compute nodes. | 21:05 |
| evrardjp | yup | 21:05 |
| cloudnull | it was around an hour of downtime until we nuked the interfaces and restarted the agents. | 21:06 |
| cloudnull | in that hour some tenant networks came back | 21:06 |
| evrardjp | corrupted fdb tables | 21:06 |
| cloudnull | lots of failed entries. | 21:07 |
| evrardjp | and using l2pop or using standard linux system learning vxlan mechanism? | 21:08 |
| *** retreved has quit IRC | 21:08 | |
| evrardjp | I guess standard | 21:08 |
| *** retreved has joined #openstack-ansible | 21:09 | |
| cloudnull | all pointing to devices no longer on the network. | 21:09 |
| evrardjp | well it doesn't matter | 21:09 |
| cloudnull | we're using l2pop | 21:09 |
| cloudnull | but that's because this env has been up for a long while | 21:09 |
| evrardjp | cloudnull: yup that's what I understood | 21:09 |
| evrardjp | and because it's not static it's a pain to flush tables | 21:09 |
| evrardjp | where is bgp when we need it | 21:10 |
| evrardjp | :p | 21:10 |
| cloudnull | that would've been nice | 21:10 |
| evrardjp | well bgp evpn seems a good thing for your issue but even before going that far | 21:10 |
| evrardjp | I understand why you'd want to prevent this kind of possible errors | 21:11 |
| evrardjp | by having less moving parts for neutron | 21:11 |
| evrardjp | still if I destroy a neutron instance and recreate a new one, it shouldn't take hours to converge | 21:11 |
| evrardjp | :p | 21:11 |
| cloudnull | that's the thing. dstroy 1 / add 1 no big deal. | 21:12 |
| cloudnull | except for routers. | 21:12 |
| cloudnull | which are scheduled using the chance schduler | 21:12 |
| cloudnull | distributing various routers throughout the stack | 21:12 |
| cloudnull | that'll over within 5 min | 21:13 |
| cloudnull | unless l3ha is enabled | 21:13 |
| evrardjp | let's not go that far | 21:13 |
| evrardjp | we have a problem right before that | 21:13 |
| cloudnull | our issue was load and scale. | 21:14 |
| cloudnull | we cycled the containers all over 30 min period because we we're trying to roll the restart out without impacting uptime | 21:15 |
| evrardjp | yes indeed, but let's not change the architecture right now, because it's not that big | 21:15 |
| cloudnull | it didn't help | 21:15 |
| evrardjp | sure | 21:15 |
| evrardjp | why doesn't lxc for example drop the current mac address and reuse it | 21:15 |
| evrardjp | let's start with that | 21:16 |
| cloudnull | it does with that change :) | 21:16 |
| evrardjp | yes because you force it yourself | 21:16 |
| cloudnull | i mean i have code to handle it but setting a fixed mac is part of lxc | 21:16 |
| cloudnull | we have a dynamic mac | 21:16 |
| evrardjp | I understand that we use the xx.xx.xx part | 21:17 |
| evrardjp | just making sure it's the right thing too | 21:17 |
| evrardjp | why one first instantiation of the container, wouldn't the mac address be kept in the config? | 21:17 |
| openstackgerrit | Frédéric Guillot proposed openstack/openstack-ansible-os_almanach: Add .gitreview file https://review.openstack.org/386233 | 21:18 |
| openstackgerrit | Charles Farquhar proposed openstack/openstack-ansible-os_ironic: Configure ironic to use keystone v2 https://review.openstack.org/386234 | 21:18 |
| cloudnull | lxc doesn't manage the config nor is there a daemon that keeps track of such things | 21:18 |
| *** chris_hultin is now known as chris_hultin|AWA | 21:19 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible: Create minor playbook to sync TZ data https://review.openstack.org/386235 | 21:22 |
| evrardjp | I hate bringing the proxy arp | 21:22 |
| evrardjp | on the table | 21:22 |
| cloudnull | http://www.funtoo.org/Linux_Containers | 21:22 |
| cloudnull | "It is a very good idea to assign a static MAC address to your container using lxc.network.hwaddr. If you don't, LXC will auto-generate a new random MAC every time your container starts, which may confuse network equipment that expects MAC addresses to remain constant." | 21:23 |
| openstackgerrit | Charles Farquhar proposed openstack/openstack-ansible-os_ironic: Configure ironic to use keystone v2 https://review.openstack.org/386234 | 21:24 |
| cloudnull | also "LXD will setup a fixed MAC address for you (as a volatile key). You can either change that with “lxc config edit” or you can just add the macvlan nic device to the container directly and then set the hwaddr key there." -- https://www.stgraber.org/2016/03/19/lxd-2-0-your-first-lxd-container-312/ | 21:24 |
| cloudnull | so in LXD they're system will do this for you. | 21:24 |
| evrardjp | I understand that network equipment don't like mac flapping | 21:25 |
| evrardjp | :p | 21:25 |
| evrardjp | I lost a switch once with that, but whatever, they are meant for that! | 21:26 |
| evrardjp | and it makes sense that lxd handles that properly | 21:26 |
| evrardjp | so it's a good step forward | 21:26 |
| jrosser | i have to deal with .1x mac auth | 21:26 |
| evrardjp | oh | 21:26 |
| jrosser | all this talk of random mac is making me feel quite unwell :) | 21:27 |
| cloudnull | death to random mac | 21:27 |
| evrardjp | well it's not supposed to be random all the time | 21:27 |
| evrardjp | it's for the life of the container | 21:27 |
| evrardjp | we are just fixing it for its multiple lifes in case of reboots | 21:28 |
| evrardjp | that's the question right? | 21:28 |
| evrardjp | like I said above, I don't see a problem itself of having it fixed | 21:28 |
| cloudnull | yes. when the container is restarted it currently gets a new mac. and that makes cloud sad. | 21:29 |
| *** fguillot has quit IRC | 21:29 | |
| evrardjp | I just want to make sure the learning is fine, and resilient | 21:29 |
| *** Andrew_jedi has joined #openstack-ansible | 21:29 | |
| evrardjp | cloudnull: I understand the problem | 21:29 |
| cloudnull | and the internet seems to agree that defining a fixed mac makes cloud not sad. :) | 21:30 |
| logan- | i don't get why rabbit/galera keeps coming up | 21:30 |
| evrardjp | all of that we agreed | 21:30 |
| *** michauds has quit IRC | 21:30 | |
| logan- | just curious why would they be sensitive to mac address changes | 21:31 |
| logan- | network node makes sense, thats a lot of cam table entries to update, just not seeing the translation to rabbit/galera | 21:32 |
| cloudnull | galera is an assumption due to it being part of a Synchronous cluster. | 21:33 |
| cloudnull | rabbitmq same thing however we know it has a lot more network sensitivity | 21:34 |
| *** jheroux has quit IRC | 21:34 | |
| *** Jeffrey4l__ has quit IRC | 21:35 | |
| *** Jeffrey4l_ has joined #openstack-ansible | 21:35 | |
| evrardjp | logan-: I don't know but I'm not surprised. Let's see this erlang conversation "I can't contact my container" "Please try to contact your container" " But I'm trying to contact it" "Do it again" | 21:35 |
| evrardjp | terrible joke | 21:35 |
| cloudnull | generally frustrations in erlang | 21:35 |
| spotz | accurate though evrardjp | 21:35 |
| evrardjp | cloudnull: what if we template the mac address in the container-interface.ini.j2, and we use a conditional? | 21:36 |
| evrardjp | if dynamic: 00:16:3e:xx:xx:xx , and if not, boom, generated mac address? | 21:36 |
| odyssey4me | agrebennikov generally the ideal is that all config happens in user space - ie in /etc/openstack_deploy/ - never in the code tree itself | 21:37 |
| cloudnull | like i said before, im ok with that. but what's the negative of just setting it? | 21:37 |
| evrardjp | well that's not better | 21:37 |
| evrardjp | what I just proposed is terrible | 21:37 |
| evrardjp | ignore this | 21:37 |
| cloudnull | letting it be dynamic has known downsides but is there an upside? | 21:38 |
| evrardjp | god I wish the world would be v6 with static mac addresses and no privacy extensions + good firewalling | 21:38 |
| *** shasha_t_ is now known as shasha_tavil | 21:39 | |
| evrardjp | cloudnull: I guess it's due to our non-ephemeral nature of the containers | 21:39 |
| evrardjp | if it was ephemeral you'd probably prefer new macs/ips and grat arp | 21:39 |
| odyssey4me | our containers are not ephemeral though | 21:40 |
| coolj | since we are treating containers like hosts it makes sense to have persistent hardware addresses. 2c. | 21:40 |
| evrardjp | the question is what's the risk now for current existing things, and how do we upgrade | 21:40 |
| odyssey4me | they are long lived | 21:40 |
| odyssey4me | that's the point cloudnull brought up right at the start | 21:40 |
| evrardjp | the transition seems ok and the workaround seems not bad | 21:40 |
| odyssey4me | why should our containers not be treated like VM's in a VMWare world? | 21:41 |
| odyssey4me | and if we choose to rebuild a container with a new MAC, so be it | 21:41 |
| evrardjp | well maybe we should use lxd then | 21:41 |
| neillc | morning | 21:41 |
| *** javeriak has quit IRC | 21:41 | |
| odyssey4me | I am with logan- on this though - let's be selective about where we do this. We treat some of our containers as ephemeral and some as not. | 21:41 |
| evrardjp | odyssey4me: oh I think we all agree | 21:42 |
| odyssey4me | This will be a concept we need to start applying to upgrades soon to cut some of the upgrade nonsense that happens. | 21:42 |
| evrardjp | odyssey4me: I'd prefer see in the implementation: having it optional with the minimum of code to maintain | 21:43 |
| evrardjp | that was my point | 21:43 |
| evrardjp | the option could then simply be a var overriden in the group_vars | 21:43 |
| odyssey4me | coolj FYI, treating them like hosts may just go away ;) but that doesn't change the fact that some are longer lived and that MAC consistency is important for easier transitions, even if we replace a container's content | 21:44 |
| evrardjp | depending on which group | 21:44 |
| odyssey4me | evrardjp so yes, then switch to a -1 vote and allow the review process to achieve that | 21:44 |
| odyssey4me | and if the initial implementation merges and you don't like it, improve on it | 21:44 |
| evrardjp | done | 21:44 |
| odyssey4me | this is very early in the cycle, there is a lot of time to try this out and see how it goes - then improve | 21:45 |
| evrardjp | but that's what I said in my first comment | 21:45 |
| odyssey4me | this is why we have a 6 month dev cycle where nothing is released | 21:45 |
| agrebennikov | odyssey4me, so literally if I need to allow keystone region variable to be changed - I put the variable to the group_vars and explain how to specify it in the user_variables, right? | 21:45 |
| odyssey4me | agrebennikov no need - just add your var and value into user_variables.yml | 21:45 |
| odyssey4me | there is already a var for it in group vars, so all you do is override the var & value in user_vars | 21:46 |
| evrardjp | cloudnull: I still think l2pop was a terrible choice for the mac population in the tunnelz | 21:46 |
| evrardjp | but not your fault :) | 21:46 |
| odyssey4me | agrebennikov that's why this is in the user_vars file: https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/user_variables.yml#L16-L19 | 21:47 |
| *** weezS has quit IRC | 21:48 | |
| *** javeriak has joined #openstack-ansible | 21:48 | |
| agrebennikov | odyssey4me, I see now, thanks) | 21:50 |
| odyssey4me | agrebennikov so if you add 'service_region: RegionTwo' in user_variables.yml then that will override the value in group_vars | 21:51 |
| agrebennikov | odyssey4me, yep, I got it | 21:51 |
| agrebennikov | odyssey4me, we are pretty new to ansible, were in the puppet world before, so sometimes it is difficult to realize all these nuances | 21:51 |
| agrebennikov | especially with vars overrides | 21:52 |
| odyssey4me | agrebennikov we do use variable precedence in a slightly odd way due to our desire to try and separate code from config | 21:54 |
| odyssey4me | we're getting better at simplifying things | 21:54 |
| *** jamielennox|away is now known as jamielennox | 21:55 | |
| *** spotz is now known as spotz_zzz | 21:59 | |
| *** jperry has quit IRC | 22:00 | |
| agrebennikov | odyssey4me, did you guys actually try to deploy osa into different regions with shared keystone? | 22:03 |
| odyssey4me | agrebennikov sure, it has been done - typically with individual inventories per region | 22:03 |
| agrebennikov | like when I deploy the first cloud entirely, and then change the region and update inventory leaving same keystone nodes | 22:03 |
| odyssey4me | ie each region is its own deployment | 22:04 |
| odyssey4me | you can then setup a global cluster if you want, or a central keystone - it all depends on the architecture you're looking for | 22:04 |
| agrebennikov | no, my question is if I need same keystone nodes in the second region | 22:04 |
| odyssey4me | ok, so you want to use the same keystone back-end for multiple regions? | 22:04 |
| *** jmckind has quit IRC | 22:04 | |
| odyssey4me | same keystone endpoints | 22:04 |
| agrebennikov | not just the backend | 22:05 |
| cloudnull | ok im off, cheers all. | 22:05 |
| agrebennikov | odyssey4me, nope)) have same keystone services, but have another set of endpoints with another region name | 22:05 |
| odyssey4me | so dyou you want to setup the db and service independently - then have multiple compute/object storage regions use that set? | 22:05 |
| agrebennikov | odyssey4me, pretty much like that, but I can set up keystone together with the first region | 22:06 |
| agrebennikov | but it doesn't really matter | 22:06 |
| odyssey4me | there are many ways to skin that cat | 22:06 |
| odyssey4me | ok, so setup region one with everything | 22:06 |
| coolj | later cloudnull | 22:06 |
| agrebennikov | so literally the second region will have to utilize existing keystone | 22:07 |
| *** jamesdenton has quit IRC | 22:07 | |
| agrebennikov | does it work fine? | 22:07 |
| agrebennikov | from the OSA perspective I mean | 22:07 |
| odyssey4me | then for region 2, don't allocate any hosts to the identity_hosts host group - and for that setup override all the galera/keystone settings to use the first one | 22:07 |
| odyssey4me | oh sure - it will be a lot of variable overriding and a bit of a pain to get setup - but it'll work | 22:07 |
| agrebennikov | odyssey4me, ah, I see | 22:07 |
| agrebennikov | well, will try really soon)) | 22:08 |
| logan- | agrebennikov: i've been looking at this multi region stuff some lately. what odyssey4me is what i'm looking at doing for a 1st phase. full osa stack in region 1, separate osa deployment/inventory in region 2 with no keystone or horizon. | 22:08 |
| logan- | the more resilient way to do it is put keystone on a separate galera cluster that spans both regions | 22:08 |
| agrebennikov | logan-, agree | 22:09 |
| logan- | https://openstack.nimeyo.com/58101/openstack-operators-milti-site-keystone-%26-galera https://etherpad.opnfv.org/p/multisite_identity_management have some great info about that | 22:09 |
| odyssey4me | yes - the far better way is to create a global galera cluster and use the capabilities it has to do wan sync | 22:09 |
| odyssey4me | then use that for your keystone endpoints in each region | 22:09 |
| *** thorst has quit IRC | 22:09 | |
| odyssey4me | that way each region has 2 or more galera nodes, but the data is synchronised across all regions | 22:10 |
| odyssey4me | one data back-end, regional endpoints | 22:10 |
| logan- | yep-- i think for phase one we're going to have keystone in one region only, and then once we work thru the logistics of managing the multi-region galera cluster and the config tuning differences from the regular galera setup, we'll migrate to keystone per region | 22:10 |
| odyssey4me | but if you do that, be sure to use fernet tokens :) | 22:10 |
| agrebennikov | odyssey4me, I completely disagree with this approach | 22:10 |
| agrebennikov | I mean replicate keystone galera | 22:10 |
| *** Andrew_jedi has quit IRC | 22:11 | |
| odyssey4me | of course another way is to use an LDAP back-end, with local DB's which are only used for shadow accounts and services | 22:11 |
| agrebennikov | we debated with keystone folks on the last summit regarding exactly this point | 22:11 |
| agrebennikov | right | 22:11 |
| agrebennikov | something like that | 22:11 |
| *** TxGirlGeek has quit IRC | 22:11 | |
| odyssey4me | all-in-all there are many options, and all of them are possible - it really depends on your environment and usage profile | 22:12 |
| agrebennikov | but right now I have a bit another usecase - it is literally centralized keystone, not 2 different ones with shared DB | 22:12 |
| odyssey4me | sure, that is often how people start | 22:13 |
| openstackgerrit | Merged openstack/openstack-ansible: [docs] Provide example configurations https://review.openstack.org/386091 | 22:13 |
| agrebennikov | that's how real customer manages 20 clouds of 3000 nodes overall right now :( | 22:13 |
| odyssey4me | another deployment model often discussed is centralised keystone and horizon, them regionalised compute which use the centralised keystone | 22:13 |
| agrebennikov | this ^^ is the model I'm talking about right nwo | 22:14 |
| agrebennikov | now | 22:14 |
| agrebennikov | I personally don't like it | 22:14 |
| agrebennikov | I prefer having fully independent clouds with same ldap backend for users, same roles IDs and tenants created automatically based on the ldap groups | 22:15 |
| logan- | http://cdn.pasteraw.com/s5nd9xtvwzdxo0lmhsxsqgtndmqm2ss is a POC I did to set up something similar to what you're saying agrebennikov. you would set up your regionOne as normal. and this config would be part of your regionTwo deploy | 22:15 |
| agrebennikov | that's how you get your fernet token valid in both clouds | 22:15 |
| odyssey4me | haha, if only we could be handed decisions we liked all the time - wouldn't life be grand? :) | 22:16 |
| openstackgerrit | Bjoern Teipel proposed openstack/openstack-ansible-os_cinder: Add remote procedure call tunables to cinder https://review.openstack.org/360749 | 22:16 |
| logan- | so as odyssey4me was saying, no identity_hosts defined in regionTwo, and this would null out your horizon deploy in regionTwo also. | 22:17 |
| agrebennikov | odyssey4me, well, I'm pretty lucky - I managed to get my customer share my point of view | 22:17 |
| odyssey4me | yep, although you wouldn't need to modify the env.d/horizon file in newton - you could just use the more granular host groups and opt to not deploy the dashboard_hosts | 22:18 |
| agrebennikov | logan-, does master_region_keystone_address exist in the playbooks? | 22:18 |
| logan- | no it is a variable i created to reflect the keystone endpoint in regionOne | 22:18 |
| agrebennikov | logan-, or it is something specific? | 22:18 |
| agrebennikov | hm | 22:18 |
| odyssey4me | agrebennikov nope - that could be 'foo_baz: blah' - it was used so as not to repeat the value in the variables lower down | 22:18 |
| agrebennikov | so is it just used somewhere in "when" statement? | 22:18 |
| odyssey4me | (in that same paste) | 22:19 |
| logan- | it is only used relative to that config. keystone_service_adminuri and keystone_service_internaluri reference it | 22:19 |
| *** smatzek has quit IRC | 22:19 | |
| agrebennikov | ah)) | 22:19 |
| agrebennikov | crap | 22:19 |
| agrebennikov | didn't notice | 22:19 |
| odyssey4me | :) | 22:19 |
| agrebennikov | wy always overcomplicate it? :) | 22:19 |
| agrebennikov | why | 22:19 |
| coolj | dry | 22:20 |
| *** agireud has quit IRC | 22:21 | |
| *** batkins61 has quit IRC | 22:21 | |
| odyssey4me | alright, the last newton patch needed for the next RC has merged - lemme put the RC request in, then I'm out for the night | 22:21 |
| *** woodard_ has joined #openstack-ansible | 22:22 | |
| logan- | the other big catch with doing this is you will need to either copy all of the service keystone auth passwords from user_secrets across your regions, or you will need to create override vars for all of these https://github.com/openstack/openstack-ansible-os_nova/blob/master/defaults/main.yml#L168 --ie maybe set nova_service_user_name: "nova_{{ nova_service_region }}" etc.. | 22:22 |
| *** woodard has quit IRC | 22:25 | |
| *** galstrom is now known as galstrom_zzz | 22:25 | |
| *** woodard_ has quit IRC | 22:26 | |
| *** kylek3h has quit IRC | 22:27 | |
| dxiri | odyssey4me: had to start from scratch :P | 22:27 |
| dxiri | lets see if this time it works out :) | 22:27 |
| odyssey4me | dxiri the best way to learn is to fail, and fail again | 22:28 |
| dxiri | odyssey4me: totally agree :) | 22:28 |
| dxiri | as long as you don't fail on the same place everytime | 22:28 |
| odyssey4me | andymccr and everyone, Newton RC3 has been proposed: https://review.openstack.org/386258 | 22:30 |
| odyssey4me | I expect that we'll see the tag created tomorrow, giving us all next week to find any critical bugs. | 22:30 |
| odyssey4me | well, this tag will become the release if we don't before Thu | 22:31 |
| odyssey4me | alright, that's me out for the night - chat tomorrow all! | 22:32 |
| dxiri | lets say I want to try this rc3, how can I upgrade after when its stable? | 22:33 |
| dxiri | just wondering if I should be doing the deployment with that instead of mitaka | 22:33 |
| coolj | night odyssey4me | 22:34 |
| *** cathrichardson has quit IRC | 22:37 | |
| *** asettle has joined #openstack-ansible | 22:37 | |
| *** cmart has quit IRC | 22:40 | |
| jmccrory | dxiri: for upgrading from rc3, you'd just need to checkout a newer newton release tag, update roles, and re-run the playbooks http://docs.openstack.org/developer/openstack-ansible/newton/upgrade-guide/minor-upgrade.html | 22:41 |
| *** asettle has quit IRC | 22:42 | |
| *** phalmos has quit IRC | 22:44 | |
| *** admin0 has quit IRC | 22:45 | |
| dxiri | jmccrory: nice! thank you! | 22:45 |
| *** evilt0ne has joined #openstack-ansible | 22:56 | |
| *** evilt0ne has quit IRC | 22:57 | |
| *** evilt0ne has joined #openstack-ansible | 22:58 | |
| *** weezS has joined #openstack-ansible | 22:58 | |
| *** vishwanathj has quit IRC | 22:59 | |
| *** agireud has joined #openstack-ansible | 22:59 | |
| *** weezS has quit IRC | 23:00 | |
| *** hughmFLE_ has joined #openstack-ansible | 23:04 | |
| *** hughmFLEXin has quit IRC | 23:05 | |
| *** hughmFLE_ has quit IRC | 23:08 | |
| *** jamielennox is now known as jamielennox|away | 23:20 | |
| *** v1k0d3n has joined #openstack-ansible | 23:22 | |
| *** agrebennikov has quit IRC | 23:28 | |
| *** cfarquhar has quit IRC | 23:32 | |
| *** jamielennox|away is now known as jamielennox | 23:35 | |
| *** galstrom_zzz is now known as galstrom | 23:37 | |
| *** rmelero has quit IRC | 23:41 | |
| *** galstrom is now known as galstrom_zzz | 23:43 | |
| *** javeriak has quit IRC | 23:45 | |
| *** shanec_ has quit IRC | 23:48 | |
| *** sdake_ has joined #openstack-ansible | 23:51 | |
| *** retreved has quit IRC | 23:52 | |
| *** sdake has quit IRC | 23:52 | |
| *** retreved has joined #openstack-ansible | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
