| *** cjloader_ has joined #openstack-ansible | 00:08 | |
| *** cjloader_ has quit IRC | 00:08 | |
| *** cjloader has quit IRC | 00:12 | |
| *** thorst has joined #openstack-ansible | 00:16 | |
| *** thorst has quit IRC | 00:21 | |
| *** vnogin has quit IRC | 00:21 | |
| *** dayou has joined #openstack-ansible | 00:37 | |
| *** TxGirlGeek has quit IRC | 00:38 | |
| *** chyka has quit IRC | 00:38 | |
| *** chyka has joined #openstack-ansible | 00:38 | |
| *** chyka has quit IRC | 00:42 | |
| *** cjloader has joined #openstack-ansible | 00:50 | |
| *** thorst has joined #openstack-ansible | 00:50 | |
| *** esberglu has quit IRC | 00:52 | |
| *** thorst has quit IRC | 00:52 | |
| *** cjloader has quit IRC | 00:54 | |
| *** esberglu has joined #openstack-ansible | 01:02 | |
| *** esberglu has quit IRC | 01:07 | |
| *** dayou has quit IRC | 01:07 | |
| *** dave-mcc_ has joined #openstack-ansible | 01:09 | |
| *** phalmos has quit IRC | 01:12 | |
| *** vnogin has joined #openstack-ansible | 01:23 | |
| *** vnogin has quit IRC | 01:29 | |
| *** thorst has joined #openstack-ansible | 01:29 | |
| *** dayou has joined #openstack-ansible | 01:33 | |
| *** thorst has quit IRC | 01:33 | |
| *** gouthamr has quit IRC | 01:58 | |
| *** gouthamr has joined #openstack-ansible | 02:04 | |
| *** thorst has joined #openstack-ansible | 02:08 | |
| *** thorst has quit IRC | 02:09 | |
| *** chyka has joined #openstack-ansible | 02:09 | |
| *** chyka has quit IRC | 02:14 | |
| *** vnogin has joined #openstack-ansible | 02:25 | |
| *** gouthamr has quit IRC | 02:27 | |
| *** vnogin has quit IRC | 02:29 | |
| openstackgerrit | Merged openstack/openstack-ansible-galera_client master: Fix SSL handling for galera clusters https://review.openstack.org/524059 | 02:42 |
|---|---|---|
| *** thorst has joined #openstack-ansible | 02:43 | |
| *** thorst has quit IRC | 02:48 | |
| *** thorst has joined #openstack-ansible | 02:55 | |
| *** thorst has quit IRC | 02:55 | |
| *** cjloader has joined #openstack-ansible | 03:09 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_glance master: Fix systemd init template for program override https://review.openstack.org/524778 | 03:10 |
| *** dave-mcc_ has quit IRC | 03:11 | |
| *** udesale has joined #openstack-ansible | 03:13 | |
| *** cjloader has quit IRC | 03:14 | |
| savvas | cloudnull: reboot did the trick on that weird mtu issue. Was a fresh install, guess some network settings may not have been applied, rebooting the infra nodes one by one did the trick | 03:22 |
| *** thorst has joined #openstack-ansible | 03:23 | |
| *** thorst has quit IRC | 03:23 | |
| *** cjloader has joined #openstack-ansible | 03:24 | |
| *** vnogin has joined #openstack-ansible | 03:26 | |
| *** cjloader has quit IRC | 03:26 | |
| *** vnogin has quit IRC | 03:31 | |
| *** cjloader has joined #openstack-ansible | 03:44 | |
| *** askb has quit IRC | 03:47 | |
| *** cjloader has quit IRC | 03:49 | |
| *** thorst has joined #openstack-ansible | 03:59 | |
| *** germs has quit IRC | 04:01 | |
| *** cmart has joined #openstack-ansible | 04:02 | |
| *** germs has joined #openstack-ansible | 04:02 | |
| *** thorst has quit IRC | 04:04 | |
| *** cmart has quit IRC | 04:08 | |
| *** cmart has joined #openstack-ansible | 04:08 | |
| *** thorst has joined #openstack-ansible | 04:32 | |
| *** thorst has quit IRC | 04:37 | |
| *** cjloader has joined #openstack-ansible | 04:40 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Remove send_arp_for_ha option https://review.openstack.org/523930 | 04:43 |
| *** cjloader has quit IRC | 04:44 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_glance master: Fix systemd init template for program override https://review.openstack.org/524778 | 04:56 |
| *** savvas_ has joined #openstack-ansible | 05:04 | |
| *** savvas_ has quit IRC | 05:04 | |
| *** savvas has quit IRC | 05:04 | |
| *** savvas has joined #openstack-ansible | 05:05 | |
| *** thorst has joined #openstack-ansible | 05:06 | |
| *** thorst has quit IRC | 05:11 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts stable/pike: Add ExecStartPre command to lxc-dnsmasq service unit https://review.openstack.org/524785 | 05:22 |
| *** germs has quit IRC | 05:23 | |
| *** chyka has joined #openstack-ansible | 05:27 | |
| *** vnogin has joined #openstack-ansible | 05:27 | |
| *** vnogin has quit IRC | 05:31 | |
| *** chyka has quit IRC | 05:32 | |
| *** thorst has joined #openstack-ansible | 05:42 | |
| *** thorst has quit IRC | 05:47 | |
| *** cjloader has joined #openstack-ansible | 05:56 | |
| *** cjloader has quit IRC | 06:01 | |
| *** lihi has quit IRC | 06:13 | |
| *** lihi has joined #openstack-ansible | 06:15 | |
| *** thorst has joined #openstack-ansible | 06:16 | |
| *** thorst has quit IRC | 06:21 | |
| *** cjloader has joined #openstack-ansible | 06:40 | |
| *** cjloader has quit IRC | 06:44 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Do not implement reload for novnc/spice service https://review.openstack.org/522731 | 06:54 |
| *** thorst has joined #openstack-ansible | 06:55 | |
| *** thorst has quit IRC | 06:59 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: Disable offloading in test by default https://review.openstack.org/523541 | 07:00 |
| *** vnogin has joined #openstack-ansible | 07:28 | |
| *** pcaruana has joined #openstack-ansible | 07:32 | |
| *** thorst has joined #openstack-ansible | 07:33 | |
| *** vnogin has quit IRC | 07:33 | |
| *** thorst has quit IRC | 07:37 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-specs master: Add spec for centralized nginx https://review.openstack.org/524812 | 08:08 |
| *** thorst has joined #openstack-ansible | 08:09 | |
| *** gouthamr has joined #openstack-ansible | 08:11 | |
| *** thorst has quit IRC | 08:13 | |
| *** vnogin has joined #openstack-ansible | 08:29 | |
| *** vnogin has quit IRC | 08:33 | |
| *** cjloader has joined #openstack-ansible | 08:42 | |
| *** cjloader has quit IRC | 08:46 | |
| *** thorst has joined #openstack-ansible | 08:47 | |
| *** thorst has quit IRC | 08:52 | |
| *** chyka has joined #openstack-ansible | 08:58 | |
| *** Jeffrey4l has quit IRC | 09:07 | |
| *** Jeffrey4l has joined #openstack-ansible | 09:08 | |
| *** gouthamr has quit IRC | 09:25 | |
| *** thorst has joined #openstack-ansible | 09:26 | |
| *** thorst has quit IRC | 09:30 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: scripts: scripts-library.sh: Use pgrep -f to find the dstat process https://review.openstack.org/523904 | 09:32 |
| *** cjloader has joined #openstack-ansible | 09:36 | |
| *** Smeared_Beard has quit IRC | 09:36 | |
| *** sxc731 has joined #openstack-ansible | 09:37 | |
| *** cjloader has quit IRC | 09:41 | |
| *** vnogin has joined #openstack-ansible | 09:44 | |
| *** vnogin has quit IRC | 09:45 | |
| *** vnogin has joined #openstack-ansible | 09:46 | |
| *** vnogin has quit IRC | 09:48 | |
| *** m|y|k has joined #openstack-ansible | 09:55 | |
| *** m|y|k has quit IRC | 09:55 | |
| *** udesale__ has joined #openstack-ansible | 09:55 | |
| *** udesale has quit IRC | 09:56 | |
| *** thorst has joined #openstack-ansible | 09:57 | |
| *** thorst has quit IRC | 10:01 | |
| wlmbasson | odyssey4me: hi, I successfully removed designate as you suggested on one of my AIOs (centos7/pike 16.0.2) and the dataprocessing cluster template creation from horizon is working | 10:19 |
| wlmbasson | odyssey4me: I repeated the same process on another AIO (centos7/pike 16.0.1) but I can't re-create the Horizon container | 10:20 |
| *** hamza21 has joined #openstack-ansible | 10:20 | |
| wlmbasson | it fails at 'TASK [pip_install : Install EPEL and yum priorities plugin]' with error msg: http://mirror.centos.org/centos7/virt/x86_64/kvm-common/repodata/repomd.xml: [Errno -1] Error importing repomd.xml for qemu-ev: Damaged repomd.xml | 10:22 |
| wlmbasson | I destroyed the container and then re-created it with setup-everything limited to horizon. Also tried running setup everything without limiting to horizon, but same issue. Any ideas? | 10:23 |
| *** chyka has quit IRC | 10:24 | |
| *** chyka has joined #openstack-ansible | 10:25 | |
| wlmbasson | Also CentOS AIO takes more than 10 hours on metal! And it's a big machine with a decent internet connection. I remember Ubuntu builds only taking a couple of hours. Why is this? | 10:26 |
| *** thorst has joined #openstack-ansible | 10:28 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/pike: Disable offloading in test by default https://review.openstack.org/524860 | 10:29 |
| *** chyka has quit IRC | 10:30 | |
| *** thorst has quit IRC | 10:32 | |
| wlmbasson | odyssey4me: I lied, I didn't do it exactly the same. Didn't use --limit lxc_hosts and os-horizon-install.yml play, but rather setup-everything.yml. Thought it would have the same effect. Testing now... | 10:39 |
| *** cjloader has joined #openstack-ansible | 10:40 | |
| openstackgerrit | Merged openstack/openstack-ansible-ceph_client stable/newton: Fix update APT cache condition https://review.openstack.org/524632 | 10:42 |
| evrardjp | wlmbasson: which branch? | 10:45 |
| xdfil | mhayden: You run ipa-client on rhel OSA hosts? any issues? | 10:45 |
| *** cjloader has quit IRC | 10:45 | |
| xdfil | hello evrardjp :) | 10:45 |
| evrardjp | hello xdfil | 10:45 |
| evrardjp | how are you? | 10:45 |
| xdfil | pretty good just pulling an all nighter | 10:45 |
| xdfil | need to have an openshift deployment ready for monday | 10:46 |
| wlmbasson | evrardjp: which branch takes long to build? | 10:46 |
| xdfil | have to add more compute nodes to openstack first though hehe | 10:46 |
| wlmbasson | 16.0.0/1/2 | 10:46 |
| evrardjp | wlmbasson: yes, what you are doing? Are you using centos and stable/pike ? | 10:47 |
| evrardjp | xdfil: isn't our docs working for you? (on the how to add a compute node) | 10:47 |
| evrardjp | if not we should probably fix it asap | 10:47 |
| xdfil | hehe yeah I'm good on that | 10:48 |
| xdfil | I was just wondering about freeipa | 10:48 |
| wlmbasson | CentOS7. I've had that experience on all the 16.0.x tags as far as I remember | 10:48 |
| xdfil | i've never installed in on any openstack hosts before | 10:48 |
| xdfil | it will probably be fine but it does change ssh settings a bit and I am a tiny bit worried it would break tunnled migrations | 10:49 |
| xdfil | or something else I haven't thought of | 10:49 |
| xdfil | I figure if anyone has experience with it it would be mhayden | 10:50 |
| wlmbasson | evrarjdp: definitely 16.0.0.1/2 and main branch stable/pike. Takes no less than 8 hours. The machine has a 200Mbit connection and 48 cores. I re-ran setup-everything on an existing 16.0.1 AIO build last night on a 64 core machine with a 10Mbit connection and it took 10 hours before it failed (it was only stuck in the failing step for 5 minutes or so) | 10:51 |
| wlmbasson | evrarjdp: and that was still part of the setup-infrastructure step | 10:52 |
| wlmbasson | odyssey4me: hmmm...same issue. Not sure if I changed tags somewhere along the way. Too many builds going on...I'm going to try and upgrade to the latest version. Also, I'm using a proxy on this build, but my config hasn't changed. | 10:54 |
| evrardjp | wlmbasson: that sounds terribad | 10:55 |
| evrardjp | do you know which step it failed? | 10:55 |
| wlmbasson | evrarjdp: | 10:55 |
| evrardjp | centos is very slow in our gates, and we haven't managed to find out why. So if you find out why you are winning. | 10:55 |
| wlmbasson | https://www.irccloud.com/pastebin/x98kajbO/ | 10:55 |
| evrardjp | I can offer you a beer | 10:55 |
| evrardjp | but maybe you should have a look at this patch too | 10:56 |
| evrardjp | just in case: https://review.openstack.org/#/c/524860/ | 10:56 |
| wlmbasson | hahaha, thanks man | 10:56 |
| evrardjp | http://mirror.centos.org/centos7/virt/x86_64/kvm-common/repodata/repomd.xml is a 404 for me | 10:57 |
| evrardjp | are you sure your mirrors are fine? | 10:57 |
| evrardjp | and updated? | 10:57 |
| evrardjp | and your proxy doesn't have a bad cache? | 10:57 |
| evrardjp | ^ is my biggest concern | 10:57 |
| xdfil | wlmbasson: is your deploy host and your target host seperated by WAN or something? those are strange connection speeds :) | 10:58 |
| wlmbasson | I hate proxies... | 10:59 |
| wlmbasson | xdfil: those are two different AIOs | 10:59 |
| wlmbasson | evrarjdp: yes...proxy seems to have a bad cache | 10:59 |
| wlmbasson | tested on another connection with 404 but still able to access the file via proxy | 10:59 |
| xdfil | wlmbasson: oh ok I get it | 11:00 |
| *** thorst has joined #openstack-ansible | 11:00 | |
| evrardjp | wlmbasson: that doesn't help | 11:00 |
| evrardjp | :p | 11:00 |
| evrardjp | "let me deliver you a broken file you don't need!" | 11:00 |
| wlmbasson | "aaah...thanks proxy :)" | 11:00 |
| evrardjp | but still, 10h. | 11:01 |
| evrardjp | he was delivering it to you by 1bytes pieces every 5 seconds? | 11:01 |
| evrardjp | :p | 11:01 |
| wlmbasson | yea, same effect on a non-proxy 200Mbit connection | 11:01 |
| evrardjp | oh | 11:01 |
| evrardjp | that's a different thing then | 11:01 |
| evrardjp | that's what we need to fix | 11:01 |
| evrardjp | like I said, maybe this patch would help you | 11:01 |
| evrardjp | could you check with an ethtool? | 11:02 |
| evrardjp | see if you have sg enabled or not | 11:02 |
| odyssey4me | morning all | 11:03 |
| wlmbasson | builds on digital ocean also take an inordinate amount of time. Think gun1x waited 6 hours for one. | 11:03 |
| evrardjp | I have to go though. Sorry I can't help your right now, but I am very interested by fixing your thing. You might be the key to fasten our gates! | 11:03 |
| evrardjp | odyssey4me: morning | 11:03 |
| evrardjp | how are you? | 11:03 |
| evrardjp | wlmbasson: that is very weird. I should ask a digital ocean server to test. | 11:03 |
| wlmbasson | evrardjp: sg wasn't disabled on those AIO builds | 11:03 |
| odyssey4me | not too bad, care to give https://review.openstack.org/524785 a nudge - it'll help do the right thing for upgrades | 11:04 |
| evrardjp | wlmbasson: could you try to re-run and disable it? | 11:04 |
| wlmbasson | unless the AIO disabled it automatically | 11:04 |
| evrardjp | odyssey4me: sounds good. | 11:05 |
| *** thorst has quit IRC | 11:05 | |
| evrardjp | odyssey4me: I am very confused about tempest right now. | 11:05 |
| xdfil | odyssey4me too?! Hi! I should stay up until 3am more often :) | 11:05 |
| odyssey4me | haha, happy saturday morning :p | 11:05 |
| wlmbasson | evrarjdp: Will test with sg disabled on DO | 11:05 |
| evrardjp | odyssey4me: https://review.openstack.org/#/c/524179/ for example was verified on first patchset. A few rechecks later, boom, failed to delete resources on time with tempest, PERMANENTLY | 11:06 |
| evrardjp | I mean it happens all the time in gates. | 11:06 |
| evrardjp | (those last 4 builds) | 11:07 |
| evrardjp | that's frustrating | 11:07 |
| evrardjp | :) | 11:07 |
| odyssey4me | it's quite possibly a bug upstream? | 11:07 |
| evrardjp | I am wondering if we can increase tempest timeouts | 11:07 |
| odyssey4me | oh, this is a stable branch so it should be a fixed SHA/version | 11:07 |
| evrardjp | didn't see a patch | 11:07 |
| odyssey4me | oh bother, we don't have logs either | 11:09 |
| odyssey4me | ok, it's time we get that all sorted out | 11:10 |
| wlmbasson | evrarjdp: to tun sg off for AIO builds is it sufficient to place this in /usr/sbin/ifup-local? https://www.irccloud.com/pastebin/NUboYPBR/ | 11:11 |
| evrardjp | odyssey4me: if we need to sort out thing, the most urgent for tempest is master | 11:12 |
| evrardjp | I can reproduce this on my machine | 11:12 |
| evrardjp | this is no AZ available | 11:12 |
| evrardjp | you want access to one of my nodes? I can't figure this out. | 11:12 |
| odyssey4me | "The requested availability zone is not available" is the error I'm seeing come back from tempest in https://review.openstack.org/#/c/516448/11 | 11:14 |
| evrardjp | yeah | 11:15 |
| evrardjp | exactly | 11:15 |
| xdfil | https://www.irccloud.com/pastebin/29EEexC0/ | 11:15 |
| odyssey4me | ok, that means that either one of the nova services or one of the neutron services are busted | 11:16 |
| evrardjp | the AZ seems ok | 11:16 |
| xdfil | is there a reason openstack-ansible is using a much older version of ansible? | 11:16 |
| evrardjp | they are not | 11:16 |
| evrardjp | let me show you something real quick | 11:16 |
| odyssey4me | sure, if you have one running pm me the ip once you've got my pubkey on it | 11:16 |
| odyssey4me | xdfil looks like you're on newton? if you upgrade to ocata you'll get ansible 2.2, then pike 2.3 | 11:17 |
| xdfil | oh ok so the bootstrap does it | 11:17 |
| xdfil | thats what i wanted to confirm | 11:17 |
| odyssey4me | and for queens we're hoping to get to ansible 2.4 | 11:17 |
| xdfil | thanks | 11:17 |
| odyssey4me | yes, we have to pin the version of ansible per series because ansible breaks things with each rev | 11:17 |
| evrardjp | http://paste.openstack.org/show/628008/ | 11:17 |
| xdfil | I thought maybe I broke it at some point :) | 11:17 |
| evrardjp | odyssey4me: ^ | 11:17 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts stable/pike: Add ExecStartPre command to lxc-dnsmasq service unit https://review.openstack.org/524785 | 11:17 |
| evrardjp | xdfil: I have something pending for 2.4 | 11:18 |
| evrardjp | I just have a few issues to fix first. | 11:18 |
| xdfil | Thats cool | 11:18 |
| *** michelv has joined #openstack-ansible | 11:18 | |
| xdfil | I'm not in a rush hehe | 11:18 |
| odyssey4me | evrardjp I suspect that the issue could be that the node isn't registered in a cell? | 11:19 |
| xdfil | This is a production deployment I haven't touched in a while i wanted to make sure I ran the playbooks with the right verion of ansible | 11:19 |
| odyssey4me | otherwise you'll have to tail the nova/neutron logs and trace why the scheduler thinks that there's no az available | 11:19 |
| xdfil | hopefully I can upgrade it to O next week | 11:19 |
| xdfil | I spent all week this week fixing Xen pool issues | 11:20 |
| odyssey4me | xdfil you can always verify by checking at the top of the bootstrap-ansible script | 11:20 |
| odyssey4me | https://github.com/openstack/openstack-ansible/blob/stable/newton/scripts/bootstrap-ansible.sh#L25 | 11:20 |
| evrardjp | odyssey4me: could you check this: http://paste.openstack.org/show/628009/ ? | 11:20 |
| evrardjp | odyssey4me: oh good thought! | 11:20 |
| xdfil | odyssey4me: thanks for pointing out how lazy I am lol | 11:21 |
| evrardjp | I was more concerned about the first error in the link above | 11:21 |
| odyssey4me | ah, I rcognise that error | 11:21 |
| evrardjp | Yeah I remember at some point we changed flags and all | 11:21 |
| odyssey4me | that nova-compute error came up previously when the python-libvirt wheel used was not compiled with the same version of libvirt-bin as the one being used on the host | 11:22 |
| evrardjp | it's an update of libvirt than changes, so nova needed changes in code or config, right? | 11:22 |
| evrardjp | mmm | 11:22 |
| odyssey4me | check the nova venv for the libvirt wheel version | 11:22 |
| odyssey4me | and check the version of the package used and where it came from | 11:22 |
| odyssey4me | check whether the uca repo is configured on the host, and on the repo container | 11:23 |
| evrardjp | ok. | 11:24 |
| evrardjp | Will check. | 11:24 |
| evrardjp | http://paste.openstack.org/show/628010/ | 11:24 |
| evrardjp | drumroll | 11:24 |
| evrardjp | odyssey4me: 1000 points for you! | 11:25 |
| evrardjp | I thought an upgrade of libvirt, but that was simpler. | 11:25 |
| gun1x | wlmbasson: it was a centos image on digitalocean, yes... took half a day to deploy. however with ubuntu on a private cloud we got, it took one hour for aio. also on my home PC it took me 1 hour 20 mins to deploy 5 nodes | 11:25 |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: test-set-nodepool-vars: Fix mirror variable on CentOS7 https://review.openstack.org/524597 | 11:26 |
| evrardjp | odyssey4me: ofc the uca isn't on the repo container. | 11:26 |
| odyssey4me | aha, ok so this is likely caused by the recent uca repo config changes | 11:26 |
| evrardjp | because the merge of the remove on the pip install role | 11:26 |
| odyssey4me | yup | 11:26 |
| evrardjp | so I need to go forward with this | 11:27 |
| odyssey4me | so now we need to figure out why it's not there, especially if the openstack_hosts role ran before lxc_hosts | 11:27 |
| evrardjp | yeah a gerrit log might help here. | 11:27 |
| odyssey4me | does the container have the repo config file? | 11:27 |
| wlmbasson | gun1x: I'll be so glad if disabling sg fixes this... | 11:28 |
| odyssey4me | gun1x wlmbasson I know that we did pick up that using dnf instead of yum improved the speed quite a lot due to the way the ansible module handles updates of the cache and packages | 11:29 |
| odyssey4me | apparently yum always refreshes all caches for any yum module actions, whereas dnf is more surgical and around 10x faster | 11:29 |
| odyssey4me | but for pike I thought we had switched to make it all work using dnf for that reason exactly | 11:30 |
| xdfil | any reason for why we use chrony vs ntpd other than because thats what the cool kids use? | 11:33 |
| gun1x | odyssey4me: is dnf available on centos7 ? | 11:33 |
| evrardjp | odyssey4me: http://logs.openstack.org/48/516448/11/check/openstack-ansible-deploy-ceph-ubuntu-xenial/04e768b/job-output.txt.gz#_2017-12-01_14_31_50_877281 does that seem right to you? | 11:33 |
| evrardjp | I'd say /etc/apt/sources.list.d/ is missing? | 11:33 |
| xdfil | gun1x: not in base or epel | 11:34 |
| evrardjp | alternatively I can apply my openstack_hosts to it too. | 11:34 |
| odyssey4me | I know the issue. | 11:34 |
| evrardjp | ? | 11:34 |
| *** thorst has joined #openstack-ansible | 11:35 | |
| odyssey4me | we don't have the UCA repo in the copy_from_host list: https://github.com/openstack/openstack-ansible-lxc_hosts/blob/master/vars/ubuntu-16.04.yml#L56-L62 | 11:35 |
| evrardjp | that's exactly what I said :) | 11:36 |
| odyssey4me | we'd have to set the filename in https://github.com/openstack/openstack-ansible-openstack_hosts/blob/1016c0446b1da5310db42ec2fff3250a3c39be55/vars/ubuntu-16.04.yml#L94 so that we always know the filename | 11:37 |
| evrardjp | yeah | 11:37 |
| evrardjp | I agree there | 11:37 |
| odyssey4me | then we'd have to use the same filename var in group_vars | 11:37 |
| odyssey4me | and then use that filename in the copy_from_hosts list | 11:37 |
| evrardjp | ubuntu_cloud_archive_canonical_com_ubuntu.list is basically what you always get by default, but I don't like that | 11:37 |
| evrardjp | I just kept what was there before | 11:37 |
| evrardjp | yeah. | 11:38 |
| evrardjp | we could also, alternatively apply openstack_hosts after container creation, but I know you don't like it :p | 11:38 |
| odyssey4me | xdfil apparently chrony handles time differences between the host and the time source better - I think it doesn't panic if the difference is too high, but instead sets it regardless by default | 11:38 |
| odyssey4me | evrardjp yeah, in my view the repo filename should always be set and should not be based on what the repo URL is | 11:39 |
| *** dayou has quit IRC | 11:39 | |
| evrardjp | odyssey4me: I agree with you there | 11:39 |
| evrardjp | let me fix that quick, with no variable at all. | 11:39 |
| xdfil | odyssey4me: yeah last time i looked into it i believe the recommendation was chrony for laptops/workstations and ntp for servers (always on) | 11:39 |
| evrardjp | static | 11:39 |
| evrardjp | if someone wants a different file name, he can change the list. | 11:39 |
| odyssey4me | we'll have to do something like https://github.com/openstack/openstack-ansible-galera_client/blob/master/tasks/galera_client_install_apt.yml#L58-L85 to transition from unnamed to a named file | 11:40 |
| evrardjp | I will still keep the same name for consistency | 11:40 |
| xdfil | i just got annoyed because I already configured ntpd for my CDMA clocks and setup-hosts uninstalled it lol | 11:40 |
| odyssey4me | xdfil I defer to mhayden on that one. | 11:40 |
| *** thorst has quit IRC | 11:40 | |
| xdfil | the differences between the two services are not worth debating in my opinion | 11:41 |
| openstackgerrit | Merged openstack/openstack-ansible-os_horizon master: Remove enable_{firewall,vpn} options https://review.openstack.org/523916 | 11:41 |
| odyssey4me | evrardjp so personally I think we should just use the filename 'uca' | 11:41 |
| evrardjp | we can probably do differently: we can remove default name, we can remove the one with the variable name, and add the static with a new name | 11:41 |
| evrardjp | like uca | 11:41 |
| odyssey4me | whatever we do, it has to work for using an external repo instead of uca too | 11:41 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/pike: test-log-collect.sh: Collect package and repository info on Ubuntu https://review.openstack.org/523348 | 11:42 |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: Gather more logs when running tests https://review.openstack.org/523647 | 11:42 |
| evrardjp | if the user defined its repo, there is no issue there. | 11:43 |
| odyssey4me | evrardjp we have an issue in RPC where an external repo is used, when we upgrade the old repo URL stays in the file... so if we could resolve that at the same time then we'd be golden | 11:43 |
| odyssey4me | that's why this task is so important: https://github.com/openstack/openstack-ansible-galera_client/blob/master/tasks/galera_client_install_apt.yml#L79-L85 | 11:43 |
| odyssey4me | this set of tasks gets rid of any repo configured with the appropriate URL in it: https://github.com/openstack/openstack-ansible-galera_client/blob/master/tasks/galera_client_install_apt.yml#L58-L77 | 11:44 |
| evrardjp | what this does is adding tasks to remove automatically old things, instead of relying on our upgrades to add the necessary variables | 11:45 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/pike: Gather more logs when running tests https://review.openstack.org/524863 | 11:46 |
| evrardjp | I kinda like the automation better | 11:46 |
| evrardjp | instead of carrying variables | 11:46 |
| evrardjp | but I kinda recognize the fact that it should be a var | 11:46 |
| odyssey4me | yeah, we (RPC) need it as a var because we override it to use our frozen repo | 11:47 |
| evrardjp | because that's how you should do it: explicitly remove things/adding things. I see the usage of the removal of revoked keys for example | 11:47 |
| evrardjp | I mean the var should have something like: old repo state removed, new repo state added | 11:47 |
| evrardjp | which is tedious and error prone | 11:47 |
| evrardjp | but better for auditability | 11:48 |
| odyssey4me | yeah, except that we don't have a way of knowing the old state - except to do that's been done for galera | 11:48 |
| evrardjp | here we do | 11:48 |
| evrardjp | because we haven't changed variable names | 11:48 |
| odyssey4me | of course if we switch to have config in etcd then we could version the config and reach into the past to compare old and new states... but that's an evolution that's not for now | 11:48 |
| evrardjp | so we now the filename is either default, or variable name | 11:48 |
| evrardjp | yeah | 11:49 |
| evrardjp | definitely not | 11:49 |
| odyssey4me | sure, go for it - once you have a patch or two up I'll review | 11:49 |
| evrardjp | when ubuntu will have an etcd3 ppa :) | 11:49 |
| evrardjp | I will work on that with the appropriate ppl | 11:49 |
| evrardjp | lihi have one, so we can go ahead with that | 11:50 |
| evrardjp | yeah let me fix that up | 11:50 |
| odyssey4me | yeah, once I have the python builds simplified I'd like to dig into using etcd better - I don't think we need etcd3, but the built-in distributed lock capability would be handy | 11:53 |
| odyssey4me | I think we have *a lot* of possible optimisations we can do with it... although before that I think I'd like to work on using the deployment node more as a conduit for things (generating certs, distributing keys/certs, etc) | 11:54 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/pike: Do not implement reload for novnc/spice service https://review.openstack.org/522735 | 11:55 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/ocata: Do not implement reload for novnc/spice service https://review.openstack.org/522736 | 11:55 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/newton: Do not implement reload for novnc/spice service https://review.openstack.org/522737 | 11:55 |
| evrardjp | no but openstack is basically recommending etcd3 | 11:56 |
| evrardjp | even if it's not clear | 11:56 |
| evrardjp | odyssey4me: that's the more ansibley way to do it indeed | 11:57 |
| odyssey4me | evrardjp sure, but the things we would be doing for the first phase would not need etcd3 I don't think - only for that second phase and hopefully by then RDO/SuSE/UCA have packages etcd3 | 11:58 |
| evrardjp | rdo suse already have them, and we have a ppa for ubuntu, but I'd like to bring that into UCA | 11:59 |
| evrardjp | it makes sense for me that etcd3 is in uca | 11:59 |
| odyssey4me | yeah, of course - sounds like a good project for R | 12:00 |
| odyssey4me | you got that patch ready to go yet? | 12:01 |
| odyssey4me | oh, and by the way - well done for breaking master all on your own this time ;) | 12:02 |
| odyssey4me | mhayden would be proud | 12:02 |
| evrardjp | Well I didn't want to merge this one yet, but I didn't -W on time. | 12:03 |
| evrardjp | But I had a comment there to wait. | 12:03 |
| evrardjp | :p | 12:03 |
| *** dayou has joined #openstack-ansible | 12:04 | |
| evrardjp | ok I checked, these 3 tasks would be fine with someone would have a known url and things like that. It sounds more complex than just telling how to remove it. | 12:04 |
| odyssey4me | comments are no good - if you want to prevent a merge right now, -w is the only way to signal that | 12:04 |
| evrardjp | yeah I know I generally do it, but I have weaknesses too. | 12:04 |
| odyssey4me | of course, but this is how we learn ;) | 12:04 |
| odyssey4me | not sure I understand what you meant by "these 3 tasks would be fine with someone would have a known url and things like that. It sounds more complex than just telling how to remove it" | 12:05 |
| *** thorst has joined #openstack-ansible | 12:05 | |
| xdfil | odyssey4me: I remember you commenting about this at PTG haha | 12:06 |
| xdfil | https://www.irccloud.com/pastebin/vKtvMpdh/ | 12:06 |
| odyssey4me | haha, yeah | 12:07 |
| odyssey4me | it got fixed | 12:07 |
| xdfil | north-america is the best continent though honestly | 12:07 |
| odyssey4me | so the americans keep telling me ;) | 12:07 |
| odyssey4me | see, all fixed now: https://github.com/openstack/ansible-hardening/blob/c54fc86bfd0139d04959cf405b1f34449550f841/defaults/main.yml#L265-L269 | 12:08 |
| *** thorst has quit IRC | 12:10 | |
| xdfil | thats good. I customize it anyway though | 12:12 |
| xdfil | got local stratum 1 sources | 12:12 |
| xdfil | I'm a time snob | 12:12 |
| odyssey4me | alright - I got weekend things to do and evrardjp's taking too long to get a patch up to fix what he broke :p | 12:14 |
| odyssey4me | I'll check back later, or maybe tomorrow. | 12:15 |
| xdfil | cya | 12:15 |
| odyssey4me | have a great w/end y'all! | 12:15 |
| *** michelv has quit IRC | 12:18 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_glance master: Remove jinja2 delimiters from when https://review.openstack.org/524662 | 12:19 |
| *** sxc731 has quit IRC | 12:23 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-openstack_hosts master: Force the uca filename https://review.openstack.org/524865 | 12:38 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-openstack_hosts master: Remove the curly braces in when https://review.openstack.org/524866 | 12:41 |
| evrardjp | odyssey4me: good weekend! | 12:41 |
| evrardjp | odyssey4me: I needed my second breakfast. | 12:42 |
| evrardjp | it's the week-end too for me! | 12:42 |
| evrardjp | :p | 12:42 |
| *** thorst has joined #openstack-ansible | 12:44 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance stable/pike: Remove jinja2 delimiters from when https://review.openstack.org/524867 | 12:45 |
| *** thorst has quit IRC | 12:48 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Adapt to openstack_hosts/pip_install changes https://review.openstack.org/515405 | 12:52 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-openstack_hosts master: Force the uca filename https://review.openstack.org/524865 | 13:03 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-openstack_hosts master: Force the uca filename https://review.openstack.org/524865 | 13:03 |
| wlmbasson | evrarjdp: I keep getting these error whenever I build and AIO (and always have on CentOS). It's not a blocker, because re-running the play solves the issue. It's just annoying. Is there a way to solve this? https://www.irccloud.com/pastebin/F2wmrTgN/ | 13:06 |
| *** pcaruana|afk| has joined #openstack-ansible | 13:06 | |
| wlmbasson | ^evrardjp | 13:07 |
| evrardjp | mmm I am refactoring this whole thing in Queens, but I don't change that part | 13:07 |
| evrardjp | what would be the cause? | 13:07 |
| *** mattt has quit IRC | 13:08 | |
| evrardjp | could you connect manually on the first issue, see what's the mirror that causes you pain? | 13:08 |
| evrardjp | Cannot find a valid baseurl for repo: base/7/x86_64 | 13:09 |
| evrardjp | that sounds bad | 13:09 |
| wlmbasson | This has happened on every CentOS install I've done in every environment. | 13:10 |
| wlmbasson | evrardjp: not sure how to connect manually to test? | 13:11 |
| evrardjp | I mean it's on your host right | 13:11 |
| evrardjp | so ssh into it and use the same command | 13:11 |
| evrardjp | or simply a yum update should do the trick | 13:12 |
| *** michelv has joined #openstack-ansible | 13:12 | |
| evrardjp | wait | 13:12 |
| evrardjp | maybe we SHOULD do a yum update ourselves | 13:12 |
| evrardjp | I think it's worth filing a bug | 13:12 |
| evrardjp | this way we can fill mhayden in too | 13:12 |
| evrardjp | so you say on every single install, you have that issue when using yum-config-manager | 13:13 |
| evrardjp | Could you export your repos configuration somewhere? | 13:14 |
| *** chyka has joined #openstack-ansible | 13:14 | |
| *** sboyron has joined #openstack-ansible | 13:15 | |
| *** chyka has quit IRC | 13:15 | |
| evrardjp | I mean for this bug to be complete, we'd need: the paste showing the issue (the one above), the branch you're using (Pike), the configuration of your repos before running OSA (showing it's working fine), and after the bug triggered in OSA | 13:15 |
| evrardjp | wlmbasson: ^ | 13:15 |
| evrardjp | if you provide us all of that, that can help us. | 13:15 |
| *** thorst has joined #openstack-ansible | 13:16 | |
| *** chyka has joined #openstack-ansible | 13:19 | |
| *** sboyron has quit IRC | 13:19 | |
| *** sboyron has joined #openstack-ansible | 13:20 | |
| *** thorst has quit IRC | 13:21 | |
| *** sboyron has quit IRC | 13:21 | |
| *** sboyron has joined #openstack-ansible | 13:23 | |
| *** chyka has quit IRC | 13:23 | |
| *** hamza21 has quit IRC | 13:37 | |
| wlmbasson | evrardjp: thanks I'll provide that | 13:37 |
| *** coffee_cat is now known as ildikov | 13:38 | |
| *** sxc731 has joined #openstack-ansible | 13:39 | |
| *** cjloader has joined #openstack-ansible | 13:40 | |
| *** cjloader has quit IRC | 13:44 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Adapt to openstack_hosts/pip_install changes https://review.openstack.org/515405 | 13:50 |
| *** thorst has joined #openstack-ansible | 13:51 | |
| *** thorst has quit IRC | 13:52 | |
| *** thorst has joined #openstack-ansible | 13:52 | |
| *** ianychoi has joined #openstack-ansible | 13:55 | |
| xdfil | wlmbasson, evrardjp: I like to think of myself as being more of a rhel engineer than a deb engineer. But I've never run OSA on CentOS. So I'm curious where yum-config-manager is fitting into the picture | 14:03 |
| xdfil | Does OSA use it? I haven't used that command in a long time I typically edit the repo files or the yum conf files | 14:03 |
| xdfil | is the repo container using it? | 14:04 |
| wlmbasson | xdfil: not 100% sure. I just run the default quickstart playbooks on a fresh CentOS7 install | 14:05 |
| xdfil | I remember mhayden talking about challenges in modifying the repo files at PTG | 14:05 |
| xdfil | oh so its not something you are trying to do custom | 14:05 |
| xdfil | can you show me the repo file on the system where it's failing? | 14:07 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Change opendaylight default port https://review.openstack.org/524873 | 14:12 |
| *** thorst has quit IRC | 14:16 | |
| *** michelv has quit IRC | 14:18 | |
| xdfil | wlmbasson: did you try manually uncommenting "baseurl" and commenting "mirrorlist" in the repo file containing the /etc/yum.repos.d/CentOS-Base.repo? | 14:20 |
| xdfil | also just try manually running "yum install epel-release" | 14:20 |
| *** indistylo has joined #openstack-ansible | 14:21 | |
| xdfil | or yum update manually | 14:21 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_neutron master: Fix swift port conflict https://review.openstack.org/524874 | 14:21 |
| xdfil | I sort of suspect the ansible is not expanding the $releasever variables in the repo file | 14:21 |
| xdfil | if that is the case try putting static values in there | 14:22 |
| *** chhavi has joined #openstack-ansible | 14:23 | |
| xdfil | when I say yum update manually I mean that just to test if yum works outside of the plays | 14:23 |
| wlmbasson | xdfil: now that you mention it I have actually commented out 'baseurl'... can't remeber if it worked. Will test and give update. But I'm afk rest of the day. | 14:23 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-repo_server master: Fix swift port conflict https://review.openstack.org/524875 | 14:24 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Change opendaylight default port https://review.openstack.org/524873 | 14:24 |
| wlmbasson | xdfil: yum update does work outside of the plays as far as I remember, but it's probably because the play already ran once. It always works the second time I run the play. | 14:25 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-repo_server master: Fix swift port conflict https://review.openstack.org/524875 | 14:25 |
| xdfil | wlmbasson: that's interesting | 14:25 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_neutron master: Fix swift port conflict https://review.openstack.org/524874 | 14:25 |
| wlmbasson | xdfil: and that's without commenting out baseurl and there is no proxy involved | 14:26 |
| xdfil | by default baseurl is commented in a centos install | 14:26 |
| xdfil | so if its uncommented that means OSA did it | 14:27 |
| wlmbasson | xdfil: I mean without me uncommenting it manually | 14:27 |
| xdfil | ahhh ok | 14:27 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Build everything on metal https://review.openstack.org/504224 | 14:27 |
| xdfil | I've run into this before trying to wirte config management scripts | 14:28 |
| xdfil | you go into it thinking you can just parse the repo file | 14:28 |
| wlmbasson | I'll send repo lists on clean install and after first error. | 14:28 |
| xdfil | but then you realize yum knows how to evaluate the variables and your script doesnt | 14:28 |
| evrardjp | odyssey4me: cloudnull mhayden or anyone really available, that would be nice to have a review of https://review.openstack.org/#/c/524875/ | 14:28 |
| evrardjp | it's simple | 14:28 |
| xdfil | so So i wouldnt be suprized if OSA is encountering a similar situation | 14:29 |
| xdfil | its an easy oversight when writing a large script | 14:30 |
| xdfil | ansible or whatever it may be | 14:30 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Adapt to openstack_hosts/pip_install changes https://review.openstack.org/515405 | 14:31 |
| *** indistylo has quit IRC | 14:32 | |
| *** cjloader has joined #openstack-ansible | 14:40 | |
| *** cjloader has quit IRC | 14:44 | |
| *** thorst has joined #openstack-ansible | 14:47 | |
| xdfil | evrardjp: is that a new port for the repo container? | 14:48 |
| xdfil | it was 8181 previously right? | 14:48 |
| *** michelv has joined #openstack-ansible | 14:49 | |
| *** thorst has quit IRC | 14:52 | |
| *** dave-mccowan has joined #openstack-ansible | 14:55 | |
| *** armaan has joined #openstack-ansible | 14:55 | |
| *** dave-mcc_ has joined #openstack-ansible | 14:57 | |
| xdfil | so i'm assuming someone changed the port but chose 8080 :) | 14:58 |
| *** thorst has joined #openstack-ansible | 14:58 | |
| *** dave-mccowan has quit IRC | 15:00 | |
| *** armaan has quit IRC | 15:01 | |
| *** thorst has quit IRC | 15:03 | |
| *** armaan has joined #openstack-ansible | 15:03 | |
| wlmbasson | evrardjp: with sg disabled using /usr/sbin/ifup-local '/sbin/ethtool -K $1 sg off' deployment time 3 hours on DigitalOcean!!!! | 15:04 |
| wlmbasson | evrardjp: that's more than 2x as fast then any of my previous deployments. Unless something changed in 16.0.4, it was sg that gave the boost. | 15:05 |
| wlmbasson | *disabling sg | 15:06 |
| *** indistylo has joined #openstack-ansible | 15:14 | |
| *** sxc731 has quit IRC | 15:17 | |
| *** vnogin has joined #openstack-ansible | 15:20 | |
| *** michelv has quit IRC | 15:20 | |
| *** cjloader has joined #openstack-ansible | 15:28 | |
| *** yifei has joined #openstack-ansible | 15:28 | |
| *** yifei has quit IRC | 15:30 | |
| *** thorst has joined #openstack-ansible | 15:32 | |
| *** cjloader has quit IRC | 15:32 | |
| *** vnogin has quit IRC | 15:32 | |
| *** thorst has quit IRC | 15:37 | |
| evrardjp | wlmbasson: ok | 15:42 |
| evrardjp | good to know | 15:42 |
| *** SerenaFeng has joined #openstack-ansible | 15:43 | |
| *** SerenaFeng has quit IRC | 15:44 | |
| *** SerenaFeng has joined #openstack-ansible | 15:45 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-tests master: Only use ethtool if present https://review.openstack.org/524881 | 15:46 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_neutron master: Fix swift port conflict https://review.openstack.org/524874 | 15:47 |
| *** hamza21 has joined #openstack-ansible | 16:00 | |
| *** SerenaFeng has quit IRC | 16:02 | |
| *** thorst has joined #openstack-ansible | 16:02 | |
| *** thorst has quit IRC | 16:06 | |
| *** indistylo has quit IRC | 16:08 | |
| *** germs has joined #openstack-ansible | 16:13 | |
| *** germs has quit IRC | 16:15 | |
| *** sboyron has quit IRC | 16:17 | |
| *** savvas_ has joined #openstack-ansible | 16:24 | |
| *** savvas_ has quit IRC | 16:26 | |
| *** savvas has quit IRC | 16:26 | |
| *** savvas has joined #openstack-ansible | 16:27 | |
| evrardjp | xdfil: you'll have to check into the haproxy config to have that info, or into the repo server. I don't know by heart | 16:32 |
| xdfil | I'm pretty sure it was | 16:36 |
| evrardjp | we have 8080 recently for pypi server, else I think it's 8181 | 16:37 |
| evrardjp | but I can be wrong | 16:37 |
| xdfil | about a month ago I was playing with ODL | 16:37 |
| evrardjp | and? | 16:37 |
| xdfil | the port conflict kept me up all night | 16:37 |
| evrardjp | you got the issue? | 16:37 |
| evrardjp | ahah ok | 16:37 |
| xdfil | yeah lol | 16:37 |
| evrardjp | if you patched it, nobody would have go through that trouble :p | 16:37 |
| xdfil | and i'm 98% positive repo and ODL wanted 8181 | 16:38 |
| evrardjp | odl was 8080 | 16:38 |
| evrardjp | swift was 8080 too | 16:38 |
| xdfil | I reported it to the ODL guys and to you too in IRC if i remember correctly | 16:38 |
| evrardjp | 8180 seems not taken | 16:38 |
| evrardjp | my bad | 16:38 |
| xdfil | I think 8080 is a legacy ODL port not sure though | 16:38 |
| xdfil | but I know for sure it currently needs 8181 for mgmt by default now | 16:39 |
| xdfil | so by my logic, if someone knows ODL and starts using OSA and they need to troubleshoot they will instinctivley try to reach 8181 | 16:39 |
| xdfil | first anyway | 16:39 |
| xdfil | and then they will complain that OSA uses 8180 | 16:40 |
| xdfil | right ? :) | 16:40 |
| *** thorst has joined #openstack-ansible | 16:40 | |
| *** thorst has quit IRC | 16:45 | |
| xdfil | evrardjp: Yeah I just verified the ODL WebGUI also uses port 8181 too | 16:46 |
| xdfil | wait let me double verify | 16:46 |
| *** pcaruana|afk| has quit IRC | 16:47 | |
| evrardjp | yeah so 8181 is already taken by us | 16:48 |
| evrardjp | same as 8080 | 16:48 |
| evrardjp | so I thought that moving odl to 8180 is not a bad idea | 16:48 |
| *** germs has joined #openstack-ansible | 16:48 | |
| xdfil | http://docs.opendaylight.org/en/stable-nitrogen/user-guide/service-function-chaining.html | 16:48 |
| evrardjp | and a part of the repo should be in 8280 | 16:48 |
| evrardjp | while the other part of the repo can continue to be 8181 | 16:48 |
| xdfil | yeah I suppose an alternate port is not a big deal | 16:49 |
| evrardjp | maybe I should have used 8183 for the other part of the rpeo | 16:49 |
| evrardjp | yeah it's a problem right now because it conflicts | 16:49 |
| *** armaan has quit IRC | 16:50 | |
| *** germs1 has joined #openstack-ansible | 16:50 | |
| *** armaan has joined #openstack-ansible | 16:50 | |
| xdfil | evrardjp: I just imagine a lot of questions comming up on IRC as people are following ODL docs to access the restAPI and GUI for ODL while using OSA and not realizing they need to use a different port because of OSA | 16:51 |
| xdfil | as an open source project, support resources are limited | 16:52 |
| xdfil | whereas the repos server port is something that most users are blissfully unaware of | 16:53 |
| xdfil | they will never try to hit it in a web browser | 16:53 |
| evrardjp | yeah but repo isn't the problem | 16:53 |
| evrardjp | swift is | 16:53 |
| evrardjp | there is no overlap with repo | 16:53 |
| xdfil | swift is 8080 right? | 16:53 |
| xdfil | the repos is 8181 | 16:53 |
| xdfil | i thought | 16:53 |
| xdfil | sorry if i'm wrong | 16:53 |
| evrardjp | rn: | 16:54 |
| *** germs has quit IRC | 16:54 | |
| evrardjp | 8080 is swift + repo pypi + odl , 8181 is repo | 16:54 |
| evrardjp | I move to: 8080 swift, 8180 odl, 8181 repo, 8280 repo pypi | 16:54 |
| xdfil | yeah I'm sure it will be fine that way too | 16:55 |
| xdfil | :) | 16:55 |
| xdfil | I just think ODL has a little bit more twitter followers than OSA haha | 16:56 |
| xdfil | for now! OSA will rule the world soon though | 16:57 |
| evrardjp | haha | 16:57 |
| evrardjp | it's just a port... if ppl using ODL don't understand what a port is, that's a different story :p | 16:58 |
| xdfil | thats a good point I considered also | 16:58 |
| evrardjp | :) | 16:58 |
| xdfil | but... | 16:58 |
| xdfil | Consider that other applications and devices might be configured to use 8181 for ODL | 16:59 |
| xdfil | lots of things an potentially talk to ODL | 16:59 |
| xdfil | only OSA talks to the repos container | 16:59 |
| *** chyka has joined #openstack-ansible | 17:01 | |
| xdfil | if we changed the repo contnainer port, it has a good chance of going unnoticed. setup-inrastructure would change the port on the container and on HAproxy before the other playbooks would need to access it... i think... | 17:02 |
| evrardjp | xdfil: it won't go unnoticed. | 17:03 |
| xdfil | idk sometimes i need to manually delete 00aptchache proxy~~~ | 17:03 |
| evrardjp | haha :) | 17:04 |
| *** vnogin has joined #openstack-ansible | 17:04 | |
| xdfil | this is all theoretically manageable via host-vars right? | 17:05 |
| *** chyka has quit IRC | 17:06 | |
| evrardjp | not really | 17:07 |
| xdfil | evrardjp: yeah I was wondering if https://review.openstack.org/#/c/524875/2/defaults/main.yml could refrence the variable | 17:07 |
| evrardjp | in the future it should be | 17:07 |
| xdfil | forgive me I've been awake for over 24 hours now | 17:07 |
| evrardjp | lol | 17:07 |
| evrardjp | go to bed then :) | 17:07 |
| evrardjp | for work? | 17:08 |
| xdfil | i can sleep tomorrow | 17:08 |
| evrardjp | your body will complain :p | 17:08 |
| xdfil | This is the only time I can get work done lol | 17:08 |
| evrardjp | tomorrow :p | 17:08 |
| xdfil | I can shut up in IRC though lol | 17:08 |
| xdfil | until tomorrow only though | 17:08 |
| xdfil | then I will start talking again | 17:08 |
| *** thorst has joined #openstack-ansible | 17:10 | |
| *** chhavi has quit IRC | 17:10 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-repo_server master: Fix swift port conflict https://review.openstack.org/524875 | 17:10 |
| evrardjp | lol | 17:11 |
| evrardjp | I am not asking that | 17:11 |
| evrardjp | I just don't think we should mess up with port we are using for generations. I see the upgrade and ops problems in the future. | 17:12 |
| evrardjp | ODL inclusion is recent | 17:12 |
| evrardjp | pypi server is a few days old | 17:12 |
| evrardjp | these need to change | 17:12 |
| xdfil | True, OSA+ODL user-base is very small | 17:14 |
| xdfil | for now... haha | 17:16 |
| *** thorst has quit IRC | 17:18 | |
| xdfil | I'd love to hear from someone working on the OVS ml2 driver what they think will happen | 17:19 |
| xdfil | the ODL driver seems to have solved a few problems the OVS driver has | 17:19 |
| xdfil | but I wouldnt put any faith in my assesment | 17:20 |
| xdfil | I could be way out of date | 17:20 |
| xdfil | my knowledge is all based on youtube videos that are over a year old | 17:21 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Build everything on metal https://review.openstack.org/504224 | 17:27 |
| evrardjp | XCI is one use case, so ODL developers will use this | 17:28 |
| evrardjp | not all, but some | 17:28 |
| *** cmart has joined #openstack-ansible | 17:31 | |
| *** cmart has quit IRC | 17:32 | |
| xdfil | XCI? | 17:32 |
| xdfil | 91 | 17:33 |
| xdfil | I found something on OPNFV site, reading it now | 17:34 |
| xdfil | oh dev lifecycle stuff. not my forte | 17:35 |
| xdfil | which you are aware of based on my lack of merge requests :) | 17:36 |
| xdfil | I'm working on it... | 17:37 |
| *** sxc731 has joined #openstack-ansible | 17:39 | |
| *** thorst has joined #openstack-ansible | 17:48 | |
| *** cjloader has joined #openstack-ansible | 17:50 | |
| *** thorst has quit IRC | 17:53 | |
| savvas | Hi guys, does anyone know if there is a role inside Openstack Anisble to update the endpoint urls after deployment? Or would you have to update them manually, in let's say a case where you change the external vip ip | 17:57 |
| *** thorst has joined #openstack-ansible | 18:02 | |
| *** thorst has quit IRC | 18:07 | |
| *** cmart has joined #openstack-ansible | 18:07 | |
| *** cmart has quit IRC | 18:08 | |
| *** armaan has quit IRC | 18:16 | |
| *** armaan has joined #openstack-ansible | 18:16 | |
| *** armaan has quit IRC | 18:17 | |
| *** vnogin has quit IRC | 18:17 | |
| *** armaan has joined #openstack-ansible | 18:17 | |
| *** sxc731 has quit IRC | 18:25 | |
| evrardjp | savvas: you should probably update them manually if you haven't deployed with the proper endpoints | 18:28 |
| evrardjp | please note our keystone ansible module does it, if you're ready to run all the playbooks | 18:28 |
| evrardjp | which you probably should, to restart the services etc. | 18:28 |
| *** ianychoi_ has joined #openstack-ansible | 18:31 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Build everything on metal https://review.openstack.org/504224 | 18:32 |
| *** ianychoi has quit IRC | 18:34 | |
| *** thorst has joined #openstack-ansible | 18:39 | |
| *** thorst has quit IRC | 18:44 | |
| *** armaan has quit IRC | 18:44 | |
| *** armaan has joined #openstack-ansible | 18:45 | |
| evrardjp | xdfil: thanks for the review! | 18:47 |
| *** dave-mcc_ has quit IRC | 19:00 | |
| *** thorst has joined #openstack-ansible | 19:14 | |
| *** thorst has quit IRC | 19:19 | |
| *** hamza21 has quit IRC | 19:19 | |
| *** chyka has joined #openstack-ansible | 19:25 | |
| *** chyka has joined #openstack-ansible | 19:29 | |
| *** storshoo has quit IRC | 19:32 | |
| *** dave-mccowan has joined #openstack-ansible | 19:33 | |
| *** dave-mccowan has quit IRC | 19:35 | |
| *** storshoo has joined #openstack-ansible | 19:39 | |
| *** thorst has joined #openstack-ansible | 19:54 | |
| *** sboyron has joined #openstack-ansible | 19:55 | |
| *** thorst has quit IRC | 19:59 | |
| *** cjloader has quit IRC | 20:02 | |
| *** armaan has quit IRC | 20:07 | |
| *** armaan has joined #openstack-ansible | 20:08 | |
| *** armaan has quit IRC | 20:12 | |
| *** chyka has quit IRC | 20:13 | |
| *** chyka has joined #openstack-ansible | 20:14 | |
| *** chyka has quit IRC | 20:18 | |
| *** vnogin has joined #openstack-ansible | 20:22 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/pike: Do not force using the remote "origin" when testing upgrades https://review.openstack.org/523033 | 20:24 |
| *** vnogin has quit IRC | 20:25 | |
| *** sboyron has quit IRC | 20:26 | |
| *** thorst has joined #openstack-ansible | 20:29 | |
| *** thorst has quit IRC | 20:34 | |
| *** vnogin has joined #openstack-ansible | 20:36 | |
| *** indistylo has joined #openstack-ansible | 20:43 | |
| *** vnogin has quit IRC | 20:44 | |
| openstackgerrit | Merged openstack/openstack-ansible-repo_server master: Fix swift port conflict https://review.openstack.org/524875 | 20:46 |
| *** thorst has joined #openstack-ansible | 20:50 | |
| *** thorst has quit IRC | 20:54 | |
| *** chyka has joined #openstack-ansible | 21:01 | |
| *** chyka has quit IRC | 21:06 | |
| *** thorst has joined #openstack-ansible | 21:30 | |
| *** thorst has quit IRC | 21:35 | |
| *** gouthamr has joined #openstack-ansible | 21:44 | |
| *** indistylo has quit IRC | 21:56 | |
| *** thorst has joined #openstack-ansible | 22:03 | |
| *** thorst has quit IRC | 22:08 | |
| *** thorst has joined #openstack-ansible | 22:35 | |
| *** thorst has quit IRC | 22:40 | |
| *** goldenfri has quit IRC | 22:41 | |
| cloudnull | savvas: just update the keystone DB. the change is immediate. | 22:42 |
| cloudnull | if all you have to do is update the endpoints. | 22:42 |
| cloudnull | evrardjp: you around? | 22:43 |
| cloudnull | we still having issues with tempest ? | 22:43 |
| cloudnull | this is a full breakdown of a tempest run http://paste.openstack.org/show/628001/ | 22:44 |
| cloudnull | shows it's failing basic ops and nova az | 22:44 |
| cloudnull | wlmbasson: check out https://github.com/openstack/openstack-ansible/commit/c2402a0723eff832a78fb82a966a608fa45b150f | 22:45 |
| cloudnull | that's a good way of passing ethtool options to the interfaces. | 22:45 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_swift master: Add basic functional testing for swift3 https://review.openstack.org/490248 | 22:58 |
| cloudnull | if anyone has a moment to review https://review.openstack.org/#/c/490689/ it'd be greatly appreciated | 22:59 |
| cloudnull | the changes are being tested here https://review.openstack.org/#/c/477017/ and while that review is failing it's showing functional with the nspawn container type | 23:00 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: [WIP] Add nspawn container driver https://review.openstack.org/477017 | 23:01 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add nspawn container driver https://review.openstack.org/477017 | 23:06 |
| *** cjloader has joined #openstack-ansible | 23:08 | |
| *** cjloader has quit IRC | 23:12 | |
| *** thorst has joined #openstack-ansible | 23:12 | |
| *** thorst has quit IRC | 23:17 | |
| *** thorst has joined #openstack-ansible | 23:45 | |
| *** thorst has quit IRC | 23:51 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!