*** vnogin has quit IRC | 00:00 | |
*** aredinger has joined #openstack-ansible | 00:03 | |
*** hamzy has joined #openstack-ansible | 00:05 | |
ThiagoCMC | And here is the controller bare metal syslog (while OSA is trying to start glance/cinder nspawn): systemd-nspawn[21746]: Failed to resolve interface mv-storage: No such device | 00:06 |
---|---|---|
ThiagoCMC | The bridge name looks wrong: mv-storage? I have "br-storage"... | 00:06 |
*** aredinger has quit IRC | 00:07 | |
*** jbadiapa has quit IRC | 00:08 | |
nsmeds | ThiagoCMC: not sure if I'm helpful here, but: mv = macvlan and br = bridge | 00:09 |
nsmeds | so its possible some .nspawn config needs to be changed? | 00:09 |
ThiagoCMC | Hmmm... | 00:09 |
ThiagoCMC | I was planning to just rename "br-storage" to "mv-storage" lol | 00:10 |
ThiagoCMC | I think that I found something here: https://cloudnull.io/2017/06/networkd-for-nspawn-with-openstack-ansible/ | 00:11 |
*** jbadiapa has joined #openstack-ansible | 00:11 | |
nsmeds | yep cloudnull's blog has some good posts related to nspawn | 00:11 |
*** priteau has joined #openstack-ansible | 00:12 | |
nsmeds | on the server, compare the actual interfaces to /etc/systemd/nspawn/ files | 00:12 |
nsmeds | can either be BRIDGE= or MACVLAN=, and results in different interfaces inside the containers | 00:12 |
nsmeds | err, Bridge= and MACVLAN= | 00:13 |
ThiagoCMC | That sounds very helpful, thanks! | 00:13 |
*** priteau has quit IRC | 00:16 | |
*** nsmeds has quit IRC | 00:38 | |
*** fatdragon has quit IRC | 00:38 | |
*** cshen has joined #openstack-ansible | 00:40 | |
*** cshen has quit IRC | 00:46 | |
cloudnull | ThiagoCMC yo | 00:48 |
cloudnull | hows it going? | 00:49 |
ThiagoCMC | Trying to create the mv-storage for nspawn! =P | 00:49 |
ThiagoCMC | The macvlan thing | 00:49 |
cloudnull | ha | 00:49 |
ThiagoCMC | lol | 00:49 |
cloudnull | so with the nspawn I generally use a radically simpler approach | 00:49 |
* cloudnull gets the config to show | 00:49 | |
cloudnull | https://gist.github.com/cloudnull/eab22ba500843343470a7fd1e3f55d1c#file-openstack_user_config-yml-L15-L74 | 00:50 |
cloudnull | so everything is built off the named interface. | 00:51 |
ThiagoCMC | My glance and cinder nspawn containers was failing to start because they have access to the storage area (NFS share, iSCSI net) via this mv-storage interface. First time ever trying this! Fun times. :-D | 00:51 |
ThiagoCMC | Let me check that! | 00:51 |
cloudnull | in this case bond0 | 00:51 |
cloudnull | so my host setup is just that. I make bond0 and that's it | 00:51 |
cloudnull | all of the mv-* interfaces will be automatically created for me | 00:51 |
cloudnull | and built off of what's defined in the "container_bridge" option | 00:52 |
ThiagoCMC | :-O | 00:52 |
ThiagoCMC | This looks awesome! | 00:52 |
cloudnull | its a lot less to setup before you get started. | 00:52 |
ThiagoCMC | Wait, I don't need to manually create the systemd networkd files for mv-* ? | 00:52 |
cloudnull | https://review.openstack.org/#/q/project:%255Eopenstack/openstack-ansible-nspawn.*+branch:stable/rocky+branch:stable/rocky+status:open,n,z | 00:54 |
cloudnull | ThiagoCMC nope | 00:54 |
ThiagoCMC | You just configure all mv-* on top of bond0!? No need for bond0.10, bond.11 (vlans for each mv-*)? | 00:54 |
ThiagoCMC | cloudnull, I have to try this now! LOL | 00:54 |
cloudnull | you can setup the network in any way you want. hell, you can use eth0 for all the system cars | 00:54 |
cloudnull | **cares | 00:54 |
cloudnull | I'd probably still use VLANs to isolate traffic | 00:55 |
cloudnull | but no, its not required. | 00:55 |
ThiagoCMC | Daammnn! Now we're talking! | 00:55 |
cloudnull | the one requirement is that you have the interface name you use for "container_bridge" consistent throughout the environment. IE: bond0, eth0, p1p0, etc | 00:56 |
ThiagoCMC | I was ask you about that now! :-) | 00:57 |
ThiagoCMC | "container_bridge: "bond0"", but, "bond0" isn't a bridge! Right? | 00:57 |
ThiagoCMC | Just your regular Linux bond0 interface...? | 00:57 |
cloudnull | correct its not a bridge. | 00:57 |
cloudnull | yes. | 00:57 |
cloudnull | when you have a heterogeneous environment it may be useful to define specific VLAN tagged interfaces, bonds, or use a bridge to get a consistent name across your hosts. but again, none of that is required. | 00:58 |
cloudnull | In my case, I have a mess of servers a single bond, and I'm done. | 00:59 |
ThiagoCMC | LOL | 01:00 |
cloudnull | that review link i shared before, there are two nspawn related patches that need to be pulled into rocky to make it go | 01:00 |
ThiagoCMC | Sounds cool! =) | 01:00 |
cloudnull | it is cool, IMHO :) | 01:00 |
ThiagoCMC | Have you ever tried nspawn / networkd with OpenvSwitch (i.e., the bond0 being a OVS interface)? | 01:01 |
cloudnull | nope | 01:02 |
* cloudnull does not use OvS | 01:02 | |
*** aredinger has joined #openstack-ansible | 01:03 | |
ThiagoCMC | ^_^ | 01:03 |
ThiagoCMC | Going SmartNICs ? lol | 01:03 |
cloudnull | couple screenshots of nspawn: containers -- https://i.snag.gy/Mg7Lmy.jpg -+- links on the controller -- https://i.snag.gy/hDUH2C.jpg?nocache=1540343122999 | 01:06 |
cloudnull | most of my deployments are all linuxbridge | 01:06 |
*** aredinger has quit IRC | 01:07 | |
ThiagoCMC | Niiice! Actually, my OSA "mv-nspawn0@nspawn0" and "mv-mgmt@veth-mgmt2", first time that I'm seeing those! :-D | 01:08 |
ThiagoCMC | I'm trying to figure out why mv-storage isn't there but, I believe that I have to try yours: https://gist.github.com/cloudnull/eab22ba500843343470a7fd1e3f55d1c#file-openstack_user_config-yml-L15-L74 | 01:08 |
cloudnull | yea get that type of configuration a go. and let me know what you think. | 01:10 |
*** medberry has joined #openstack-ansible | 01:11 | |
cloudnull | for prod I'd still recommend VLANs for traffic isolation but with this type of setup the "container_bridge" would be something like "bond0.10" for the storage traffic and "bond0.20" for the management traffic | 01:12 |
ThiagoCMC | That's what I already have but, my "bond0.20" is being used by "br-storage" Linux Bridge (OSA Recommendation for LXC)... | 01:13 |
cloudnull | yes. | 01:13 |
ThiagoCMC | So, I don't need that "br-storage" bridge anymore? Since it will be directly on top of bond0.20 ??! | 01:13 |
ThiagoCMC | :-O | 01:13 |
cloudnull | yup | 01:14 |
ThiagoCMC | Daammnn!!! :-D | 01:14 |
cloudnull | no more bridges, with nspawn that is | 01:14 |
cloudnull | unless you want them :) | 01:14 |
ThiagoCMC | No wonder why you don't use OVS! LOL | 01:14 |
ThiagoCMC | This is so simple. | 01:14 |
* cloudnull has a hate relationship with OvS | 01:14 | |
ThiagoCMC | LMFAO! | 01:15 |
ThiagoCMC | Thank you so much for your today's help! | 01:15 |
ThiagoCMC | I learned a LOT! | 01:15 |
ThiagoCMC | :-D | 01:15 |
cloudnull | happy to help :) | 01:18 |
ThiagoCMC | Thanks! | 01:19 |
ThiagoCMC | Cool! I manually created the mv-storage, by following: https://cloudnull.io/2017/06/networkd-for-nspawn-with-openstack-ansible/ - glance and cinder nspawn container started! | 01:19 |
cloudnull | nice! | 01:20 |
*** mmercer has quit IRC | 01:21 | |
*** cshen has joined #openstack-ansible | 01:40 | |
*** cshen has quit IRC | 01:45 | |
*** fatdragon has joined #openstack-ansible | 01:54 | |
*** aredinger has joined #openstack-ansible | 02:03 | |
*** fatdragon has quit IRC | 02:04 | |
*** aredinger has quit IRC | 02:08 | |
*** priteau has joined #openstack-ansible | 02:13 | |
*** priteau has quit IRC | 02:18 | |
*** fatdragon has joined #openstack-ansible | 02:38 | |
*** fatdragon has quit IRC | 02:45 | |
*** aredinger has joined #openstack-ansible | 03:03 | |
*** aredinger has quit IRC | 03:07 | |
*** fatdragon has joined #openstack-ansible | 03:14 | |
*** priteau has joined #openstack-ansible | 03:14 | |
*** priteau has quit IRC | 03:18 | |
*** Leo_m has quit IRC | 03:19 | |
*** fatdragon has quit IRC | 03:23 | |
*** cshen has joined #openstack-ansible | 03:41 | |
*** cshen has quit IRC | 03:45 | |
*** udesale has joined #openstack-ansible | 03:50 | |
*** udesale has quit IRC | 03:50 | |
*** udesale has joined #openstack-ansible | 03:51 | |
*** fatdragon has joined #openstack-ansible | 03:57 | |
*** aredinger has joined #openstack-ansible | 04:03 | |
*** fatdragon has quit IRC | 04:06 | |
*** aredinger has quit IRC | 04:07 | |
*** priteau has joined #openstack-ansible | 04:15 | |
*** priteau has quit IRC | 04:19 | |
*** spsurya has joined #openstack-ansible | 04:24 | |
*** fatdragon has joined #openstack-ansible | 04:39 | |
*** hamzaachi has joined #openstack-ansible | 04:42 | |
*** chhagarw has joined #openstack-ansible | 04:43 | |
*** fatdragon has quit IRC | 04:45 | |
*** dave-mccowan has quit IRC | 04:57 | |
*** aredinger has joined #openstack-ansible | 05:03 | |
*** chkumar|off is now known as chandankumar | 05:03 | |
*** aredinger has quit IRC | 05:07 | |
*** hamzaachi has quit IRC | 05:14 | |
*** hamzaachi has joined #openstack-ansible | 05:15 | |
*** pvradu has joined #openstack-ansible | 05:16 | |
*** fatdragon has joined #openstack-ansible | 05:18 | |
*** faizy_ has joined #openstack-ansible | 05:21 | |
*** fatdragon has quit IRC | 05:23 | |
*** faizy98 has quit IRC | 05:25 | |
*** priteau has joined #openstack-ansible | 05:26 | |
*** priteau has quit IRC | 05:31 | |
*** cshen has joined #openstack-ansible | 05:41 | |
*** ThiagoCMC has quit IRC | 05:43 | |
*** cshen has quit IRC | 05:46 | |
*** fatdragon has joined #openstack-ansible | 05:54 | |
*** fatdragon has quit IRC | 05:59 | |
*** aredinger has joined #openstack-ansible | 06:03 | |
*** aredinger has quit IRC | 06:07 | |
*** cshen has joined #openstack-ansible | 06:17 | |
*** pvradu has quit IRC | 06:18 | |
*** cshen has quit IRC | 06:25 | |
*** raginbajin has quit IRC | 06:32 | |
*** fatdragon has joined #openstack-ansible | 06:33 | |
*** raginbajin has joined #openstack-ansible | 06:34 | |
*** bertys has joined #openstack-ansible | 06:36 | |
*** pvradu has joined #openstack-ansible | 06:38 | |
*** fatdragon has quit IRC | 06:38 | |
*** hamzaachi has quit IRC | 06:40 | |
bertys | cores - may I draw your attention to https://review.openstack.org/#/c/549956/? thanks in advance for your reviews/feedback. | 06:40 |
*** aedc has quit IRC | 06:41 | |
*** aedc has joined #openstack-ansible | 06:41 | |
*** pvradu has quit IRC | 06:42 | |
*** pvradu has joined #openstack-ansible | 06:43 | |
*** aedc has quit IRC | 06:46 | |
*** aedc has joined #openstack-ansible | 06:47 | |
*** cshen has joined #openstack-ansible | 06:47 | |
*** aludwar has quit IRC | 06:50 | |
*** aedc has quit IRC | 06:52 | |
*** aludwar has joined #openstack-ansible | 06:53 | |
*** faizy_ has quit IRC | 06:55 | |
*** aredinger has joined #openstack-ansible | 07:03 | |
*** ivve has joined #openstack-ansible | 07:03 | |
*** fatdragon has joined #openstack-ansible | 07:05 | |
*** pcaruana has joined #openstack-ansible | 07:05 | |
*** aredinger has quit IRC | 07:07 | |
*** fatdragon has quit IRC | 07:10 | |
*** threestrands has quit IRC | 07:13 | |
openstackgerrit | Merged openstack/ansible-role-systemd_networkd master: Set container_name as host_var in test inventory https://review.openstack.org/612343 | 07:17 |
*** thuydang has joined #openstack-ansible | 07:18 | |
*** faizy98 has joined #openstack-ansible | 07:19 | |
openstackgerrit | Merged openstack/ansible-role-systemd_service master: Set container_name as host_var in test inventory https://review.openstack.org/612344 | 07:19 |
*** thuydang has quit IRC | 07:22 | |
*** hamzaachi has joined #openstack-ansible | 07:30 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cloudkitty stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612898 | 07:32 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_watcher stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612899 | 07:33 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_zun stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612900 | 07:34 |
openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/rocky: Create virtualenv without downloading https://review.openstack.org/612640 | 07:34 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_masakari stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612901 | 07:34 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_zaqar stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612902 | 07:35 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_panko stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612903 | 07:35 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_searchlight stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612904 | 07:36 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_blazar stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612906 | 07:36 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_monasca-agent stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612907 | 07:36 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tempest stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612908 | 07:38 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612909 | 07:38 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_magnum stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612910 | 07:38 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_octavia stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612911 | 07:39 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612912 | 07:39 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612913 | 07:40 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ironic stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612914 | 07:40 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612915 | 07:40 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_aodh stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612917 | 07:41 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_barbican stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612918 | 07:41 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ceilometer stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612919 | 07:42 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_designate stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612920 | 07:42 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_horizon stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612921 | 07:42 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tacker stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612922 | 07:43 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612923 | 07:43 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_sahara stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612924 | 07:43 |
*** fatdragon has joined #openstack-ansible | 07:44 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_keystone stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612925 | 07:44 |
openstackgerrit | Merged openstack/openstack-ansible-tests master: test-log-collect: Disable repositories when collecting installed pkgs on SUSE https://review.openstack.org/612661 | 07:44 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_trove stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612926 | 07:44 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_congress stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612927 | 07:45 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_rally stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612928 | 07:45 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_heat stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612929 | 07:45 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_gnocchi stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612930 | 07:46 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-rabbitmq_server stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612931 | 07:46 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-galera_server stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612932 | 07:47 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_server stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612933 | 07:47 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612934 | 07:47 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_container_create stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612935 | 07:48 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-nspawn_container_create stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612936 | 07:48 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612937 | 07:48 |
*** fatdragon has quit IRC | 07:48 | |
*** shyamb has joined #openstack-ansible | 07:50 | |
*** tosky has joined #openstack-ansible | 07:52 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Upgrade ceph to mimic release https://review.openstack.org/610165 | 07:53 |
openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron stable/rocky: V2 driver is needed to work with ODL 9.0 https://review.openstack.org/612939 | 07:53 |
*** shyamb has quit IRC | 07:58 | |
*** shyamb has joined #openstack-ansible | 07:59 | |
*** aredinger has joined #openstack-ansible | 08:01 | |
*** ianychoi_ is now known as ianychoi | 08:03 | |
*** aredinger has quit IRC | 08:06 | |
*** hw_wutianwei_ has quit IRC | 08:06 | |
*** pvradu_ has joined #openstack-ansible | 08:08 | |
*** shyamb has quit IRC | 08:09 | |
*** pvradu has quit IRC | 08:11 | |
*** FuzzyFerric has joined #openstack-ansible | 08:22 | |
*** fatdragon has joined #openstack-ansible | 08:23 | |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: SUSE: Add support for openSUSE Leap 15 https://review.openstack.org/604366 | 08:28 |
*** fatdragon has quit IRC | 08:28 | |
*** thuydang has joined #openstack-ansible | 08:30 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance master: Install missing libxml2 distro package https://review.openstack.org/612704 | 08:30 |
*** suggestable has joined #openstack-ansible | 08:31 | |
odyssey4me | jrosser: great catch on that one, I wonder where else we could save some time by using the right package for the install part | 08:31 |
jrosser | probably quite a few - debtree is ace | 08:31 |
openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible master: Integrate Blazar with OpenStack Ansible https://review.openstack.org/549956 | 08:34 |
*** shyamb has joined #openstack-ansible | 08:43 | |
openstackgerrit | Merged openstack/ansible-role-python_venv_build master: Mark build task changed only when wheels are built https://review.openstack.org/611770 | 08:45 |
openstackgerrit | Merged openstack/openstack-ansible-openstack_openrc stable/rocky: Add default for openrc_os_domain_name and update example https://review.openstack.org/612764 | 08:47 |
*** priteau has joined #openstack-ansible | 08:48 | |
odyssey4me | hmm, that's nice - it seems that ansible 2.5.10 is causing breakage all over the place | 08:51 |
*** faizy98 has quit IRC | 08:52 | |
openstackgerrit | Merged openstack/ansible-config_template master: Set container_name as host_var in test inventory https://review.openstack.org/612312 | 08:52 |
openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server stable/ocata: rabbitmq_upgrade_check: Do not complain for upgrades on distro packages https://review.openstack.org/607490 | 08:55 |
*** bertys has quit IRC | 09:00 | |
evrardjp | I am looking at the keystone fetch issue | 09:00 |
jrosser | it's that across everything isnt it | 09:00 |
openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server stable/pike: rabbitmq_upgrade_check: Do not complain for upgrades on distro packages https://review.openstack.org/607489 | 09:00 |
odyssey4me | everything that uses keystone it seems | 09:03 |
*** Emine has joined #openstack-ansible | 09:13 | |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_nova master: zuul: Switch to distro job template https://review.openstack.org/612957 | 09:16 |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_neutron master: zuul: Switch to distro job template https://review.openstack.org/612958 | 09:19 |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_nova master: zuul: Switch the Xenial lxd job to Bionic https://review.openstack.org/612959 | 09:20 |
evrardjp | so currently the state for rocky is indeed brokenness. I checked the diff of 2.5.8 and 2.5.10 and nothing changed on fetch module. Fetch module itself uses different things, including slurp module. slurp hasn't changed. So right now I am guessing connection changes. | 09:26 |
evrardjp | I can not tell as I don't have a failing environment yet. My environment is re-creating with a 2.5.8 and said patches, and I will auto bissect the ansible commits. | 09:26 |
*** thuydang has quit IRC | 09:27 | |
evrardjp | it might take a while (vs fixing directly) | 09:27 |
evrardjp | but it's easier to do on the side :D | 09:27 |
evrardjp | if you want I can rush this too | 09:27 |
jrosser | odyssey4me: looks like some good opportunities to prune -dev packages http://paste.openstack.org/show/732951/ | 09:27 |
evrardjp | python-dev is required :p | 09:28 |
odyssey4me | jrosser: the ones on the repo container may need to stay - but yeah all the others, for sure | 09:28 |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: zuul: Make SUSE and Ubuntu distro jobs non-voting https://review.openstack.org/612961 | 09:28 |
evrardjp | generally they are required for building wheels and pip packages | 09:28 |
*** thuydang has joined #openstack-ansible | 09:28 | |
evrardjp | so they probably need to stay on build machines | 09:29 |
odyssey4me | jrosser evrardjp yep, but they should be moved into a *_dev_distro_packages var, then only used for building the wheels | 09:29 |
evrardjp | except python-dev which should be everywhere | 09:29 |
odyssey4me | jrosser evrardjp much like https://review.openstack.org/#/c/612704/ | 09:29 |
odyssey4me | I think python-dev may only need to be wherever the wheels are built. | 09:31 |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: zuul: Make SUSE and Ubuntu distro jobs non-voting https://review.openstack.org/612961 | 09:31 |
*** shyamb has quit IRC | 09:37 | |
*** shyamb has joined #openstack-ansible | 09:37 | |
evrardjp | ok so for rocky/keystone failure, a downgrade to 2.5.8 didn't work. So we have something else going on. | 09:38 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: MNAIO: Implement branch var for pinning Ansible & Plugins https://review.openstack.org/612963 | 09:39 |
openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Added support for installing tempest from distro https://review.openstack.org/591424 | 09:39 |
odyssey4me | evrardjp jrosser FYI the reason I pushed up that series of backports is because I saw other rocky patches failing after the ansible 2.5.10 upgrade merged... only for role tests though | 09:43 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build stable/rocky: Mark build task changed only when wheels are built https://review.openstack.org/612966 | 09:44 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Only install the appropriate distro packages https://review.openstack.org/611508 | 09:45 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/rocky: test-log-collect: Disable repositories when collecting installed pkgs on SUSE https://review.openstack.org/612967 | 09:45 |
evrardjp | odyssey4me: I think there was a race condition that made us believe the 2.5.10 could have been the cause | 09:46 |
evrardjp | as the rollback to 2.5.8 should have deployed just fine, but wasn't on my env | 09:47 |
evrardjp | except if pebkac ofc | 09:47 |
openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: WIP - Enable python-tempestconf support https://review.openstack.org/612968 | 09:47 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Update repo_build SHA for virtualenv creation fix https://review.openstack.org/612971 | 09:48 |
hwoarang | jrosser: isn't https://review.openstack.org/#/c/611610/ already fixed by https://review.openstack.org/#/c/611843/ ? | 10:01 |
openstackgerrit | Merged openstack/openstack-ansible-os_heat stable/rocky: Replace Chinese punctuation with English punctuation https://review.openstack.org/609449 | 10:03 |
openstackgerrit | Merged openstack/openstack-ansible-repo_server stable/rocky: Replace Chinese punctuation with English punctuation https://review.openstack.org/610062 | 10:03 |
odyssey4me | evrardjp: http://logs.openstack.org/04/612704/2/check/openstack-ansible-functional-ubuntu-bionic/a1e1de8/job-output.txt.gz#_2018-10-24_09_40_28_627386 | 10:04 |
odyssey4me | that's from a patch to master | 10:04 |
evrardjp | same issue indeed | 10:05 |
evrardjp | worried now | 10:06 |
odyssey4me | https://github.com/openstack/keystone/commit/ea6755a82588ab4881f41280e981364439b4b4a2#diff-e80f17310109447772dca82b45ef35a5 | 10:07 |
openstackgerrit | Merged openstack/openstack-ansible-os_neutron stable/rocky: Replace Chinese punctuation with English punctuation https://review.openstack.org/610063 | 10:07 |
odyssey4me | something is fishy | 10:07 |
odyssey4me | rocky still has it | 10:08 |
evrardjp | when I will have an env ready with 2.5.8 I can tell you more | 10:13 |
evrardjp | I am waiting for my env to be ready as pie! | 10:13 |
*** shyamb has quit IRC | 10:14 | |
*** shyam89 has joined #openstack-ansible | 10:14 | |
*** maharg101 has joined #openstack-ansible | 10:14 | |
*** shyam89 has quit IRC | 10:18 | |
maharg101 | I have a question about multiple horizon instances with different customization modules. It looks like openstack-ansible-os_horizon just supports a single customization module. Has anyone here had experience of setting up multiple horizon instances with different customisations ? | 10:20 |
evrardjp | ok I got a machine running! that will change things | 10:20 |
evrardjp | yeah it's only that file that's missing | 10:21 |
evrardjp | the other fetch instruction work | 10:21 |
jrosser | hwoarang: yes you are right https://review.openstack.org/#/c/611610/ got missed in reviews and the later patch got merged | 10:22 |
evrardjp | found it | 10:22 |
evrardjp | so keystone removed the api-paste 2 hours ago on master | 10:23 |
evrardjp | in setup.cfg | 10:23 |
evrardjp | so that's it :) | 10:23 |
odyssey4me | ok, that means that rocky's not fetching the paste file from the rocky branch - which is not good | 10:23 |
evrardjp | we are using that package setup.cfg content in rocky branch. | 10:23 |
*** strobelight has joined #openstack-ansible | 10:25 | |
benkohl | gettting this with osa 18.0.0: https://gist.github.com/benkuly/69a8735deb574988c6060c056c651a00 | 10:27 |
benkohl | the logs in the repo container's /var/log/repo say :8181/simple/pkg-resources/ returns a 404. | 10:27 |
odyssey4me | evrardjp https://github.com/openstack/openstack-ansible-tests/blob/stable/rocky/test-vars.yml#L21 | 10:29 |
*** strobelight has quit IRC | 10:30 | |
odyssey4me | the changes for that env var to work properly have not merged into rocky - only master | 10:31 |
odyssey4me | so we need to change that default | 10:31 |
evrardjp | yup that would be important to merge very quick | 10:32 |
odyssey4me | currently for almost all the rocky roles, tox.ini does not allow that env var to be passed into the tox venv | 10:32 |
*** strobelight has joined #openstack-ansible | 10:32 | |
evrardjp | when this will be in, the patches of jimmy can go in and we can return to happy life | 10:32 |
evrardjp | so summary: os_keystone master needs cleanup to remove references of api-paste, all roles for rocky and below needs adaption of tox.ini, and we need to merge jmccrory 's patches. | 10:35 |
*** pvradu_ has quit IRC | 10:38 | |
*** pvradu has joined #openstack-ansible | 10:38 | |
*** maddtux_ has joined #openstack-ansible | 10:45 | |
*** shyamb has joined #openstack-ansible | 10:45 | |
jrosser | benkohl: this might be related to your issue https://review.openstack.org/#/c/612640/ | 10:45 |
openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-tests stable/rocky: Do not use master branch for role tests https://review.openstack.org/612983 | 10:45 |
evrardjp | odyssey4me: jrosser : https://review.openstack.org/#/c/612983/ please? | 10:46 |
*** cshen has quit IRC | 10:48 | |
openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-os_neutron stable/rocky: V2 driver is needed to work with ODL 9.0 https://review.openstack.org/612939 | 11:01 |
*** udesale has quit IRC | 11:01 | |
*** chhagarw has quit IRC | 11:06 | |
*** hamzaachi has quit IRC | 11:10 | |
*** faizy98 has joined #openstack-ansible | 11:11 | |
*** cshen has joined #openstack-ansible | 11:13 | |
*** cshen has quit IRC | 11:18 | |
openstackgerrit | Merged openstack/openstack-ansible stable/rocky: Restore OpenSUSE voting jobs https://review.openstack.org/612391 | 11:26 |
*** cshen has joined #openstack-ansible | 11:28 | |
*** shyamb has quit IRC | 11:30 | |
*** cshen has quit IRC | 11:34 | |
*** fatdragon has joined #openstack-ansible | 11:36 | |
openstackgerrit | Arx Cruz proposed openstack/openstack-ansible-os_tempest master: WIP - Enable python-tempestconf support https://review.openstack.org/612968 | 11:41 |
*** fatdragon has quit IRC | 11:41 | |
*** cshen has joined #openstack-ansible | 11:44 | |
noonedeadpunk | folks, probably we would like to install some text editor inside each container? As in bionic it seems there is no less, vi, vim, nano inside containers | 11:44 |
noonedeadpunk | what do you think about that? | 11:44 |
noonedeadpunk | or, probably, it's a topic for the next meeting | 11:45 |
jrosser | there are vars for adding your own extra things to the containers if you would like them | 11:45 |
jrosser | here for example https://github.com/openstack/openstack-ansible-lxc_hosts/blob/master/defaults/main.yml#L182 | 11:46 |
noonedeadpunk | I mean, that in xenial there was vi by default, but nothing present in bionic. | 11:46 |
noonedeadpunk | yep, thanks, i'll use it) | 11:47 |
noonedeadpunk | *less is present, but it's not about editor | 11:47 |
noonedeadpunk | jrosser: should I lcear lxc cache after adding some package there? | 11:48 |
*** ansmith has quit IRC | 11:48 | |
jrosser | you will need to make it rebuild the base image yes, plus it wont put any extra packages into the containers | 11:49 |
jrosser | *containers that are already running | 11:49 |
noonedeadpunk | ok, got it, thanks | 11:50 |
jrosser | noonedeadpunk: there are other places you can specify extra packages too https://review.openstack.org/#/q/topic:distro_extra+(status:open+OR+status:merged) | 11:51 |
*** shyamb has joined #openstack-ansible | 11:52 | |
noonedeadpunk | great patch, thank you! | 11:53 |
jrosser | i never quite got the openstack_hosts one to pass tests - must get back to that, unless you want to take a look? | 11:53 |
*** hamzaachi has joined #openstack-ansible | 11:55 | |
noonedeadpunk | jrosser: if you wish, I may try to look into it once have some time. Now I'm upgrading Q->R and xenial -> bionic, but I'm almost done with that | 11:56 |
*** shyamb has quit IRC | 11:58 | |
*** pvradu_ has joined #openstack-ansible | 12:04 | |
*** pvradu has quit IRC | 12:08 | |
*** fatdragon has joined #openstack-ansible | 12:10 | |
*** fatdragon has quit IRC | 12:14 | |
*** udesale has joined #openstack-ansible | 12:17 | |
*** udesale has quit IRC | 12:25 | |
*** smoke87 has joined #openstack-ansible | 12:28 | |
*** rgogunskiy has joined #openstack-ansible | 12:32 | |
*** maddtux_ has quit IRC | 12:33 | |
*** rgogunskiy has quit IRC | 12:34 | |
noonedeadpunk | are repos on zuul upgraded via periodic tasks? | 12:37 |
nicolasbock | Yes noonedeadpunk (IRC) | 12:38 |
noonedeadpunk | thanks, than I should ait for a while before re-checking then | 12:41 |
noonedeadpunk | s/ait/wait | 12:41 |
*** fatdragon has joined #openstack-ansible | 12:42 | |
*** fatdragon has quit IRC | 12:47 | |
*** hamzaachi has quit IRC | 12:50 | |
jrosser | argh http://logs.openstack.org/83/612983/1/check/openstack-ansible-functional-centos-7/c594bf1/job-output.txt.gz#_2018-10-24_12_44_13_579055 | 12:51 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_keystone stable/rocky: Set container_name as host_var in test inventory https://review.openstack.org/612925 | 12:52 |
*** ansmith has joined #openstack-ansible | 12:57 | |
*** FuzzyFerric has quit IRC | 12:59 | |
*** dave-mccowan has joined #openstack-ansible | 13:04 | |
*** hamzaachi has joined #openstack-ansible | 13:05 | |
openstackgerrit | Merged openstack/openstack-ansible stable/queens: Fix missing virtualenv package https://review.openstack.org/606910 | 13:12 |
*** fatdragon has joined #openstack-ansible | 13:14 | |
*** fatdragon has quit IRC | 13:19 | |
*** strattao has joined #openstack-ansible | 13:20 | |
*** udesale has joined #openstack-ansible | 13:22 | |
*** chhagarw has joined #openstack-ansible | 13:31 | |
*** vakuznet has joined #openstack-ansible | 13:35 | |
arxcruz | is os_keystone with problems ? | 13:44 |
noonedeadpunk | arxcruz: it seems that we need https://review.openstack.org/#/c/612673/ to be implemented in zuul via periodic tasks | 13:46 |
*** smoke87 has quit IRC | 13:46 | |
arxcruz | noonedeadpunk: i'm getting this http://logs.openstack.org/68/612968/1/check/openstack-ansible-functional-opensuse-423/c7aa515/job-output.txt.gz#_2018-10-24_11_30_01_394873 | 13:47 |
arxcruz | not sure if is related | 13:47 |
noonedeadpunk | yeah, probably not related. | 13:48 |
openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_keystone master: Remove paste references in Keystone https://review.openstack.org/613030 | 13:48 |
*** strattao has quit IRC | 13:49 | |
evrardjp | arxcruz: I just forgot to push a patch this morning that I wrote. | 13:49 |
arxcruz | evrardjp: are you going to summit? | 13:50 |
chandankumar | odyssey4me: http://logs.openstack.org/24/591424/42/check/openstack-ansible-functional-centos-7/3ba2141/job-output.txt.gz#_2018-10-24_11_21_54_615105 | 13:51 |
chandankumar | failed: [openstack1] (item=keystone-paste.ini) => {"changed": false, "item": "keystone-paste.ini", "msg": "file not found: /openstack/venvs/keystone-testing/etc/keystone/keystone-paste.ini"} | 13:51 |
evrardjp | arxcruz: I am | 13:53 |
arxcruz | evrardjp: no beer for you :P | 13:54 |
evrardjp | chandankumar: see patch above | 13:54 |
evrardjp | arxcruz: Don't worry I am Belgian :D | 13:54 |
arxcruz | dammit! | 13:54 |
arxcruz | lol | 13:54 |
arxcruz | evrardjp: i need to visit you | 13:54 |
evrardjp | arxcruz: and also I am fixing it, I didn't break it! :D | 13:54 |
evrardjp | hahahah | 13:55 |
*** fatdragon has joined #openstack-ansible | 13:55 | |
arxcruz | need to increase my number of belgian beers | 13:55 |
arxcruz | evrardjp: here in berlin there's a belgian pub | 13:55 |
evrardjp | There is a pub like maybe a km from my place that has 2018 beers | 13:55 |
evrardjp | just saying ;) | 13:55 |
*** vnogin has joined #openstack-ansible | 13:56 | |
evrardjp | arxcruz: I am fine everywhere where there is beer (as long as its no 1664). | 13:56 |
evrardjp | it is not 1664* | 13:56 |
arxcruz | hehehe | 13:56 |
evrardjp | 'it's no' is confusing | 13:56 |
arxcruz | evrardjp: we'll drink some beers on summit :D | 13:56 |
evrardjp | for sure! | 13:56 |
*** strattao has joined #openstack-ansible | 13:58 | |
* noonedeadpunk wants to drink some beer as well | 13:58 | |
*** fatdragon has quit IRC | 14:00 | |
*** cshen has quit IRC | 14:03 | |
chandankumar | evrardjp: what is the suse package name for subunit-filters | 14:03 |
chandankumar | ? | 14:03 |
chandankumar | evrardjp: regartding this failure http://logs.openstack.org/24/591424/42/check/openstack-ansible-functional-distro_install-opensuse-423/7ce063c/job-output.txt.gz#_2018-10-24_11_40_50_057647 | 14:05 |
arxcruz | evrardjp: so, when the patch get merged for keystone, let me know so i can recheck :) | 14:05 |
chandankumar | evrardjp: this gate is timing pout http://logs.openstack.org/24/591424/42/check/openstack-ansible-functional-distro_install-ubuntu-bionic/a02986a/job-output.txt.gz#_2018-10-24_11_48_14_138334 | 14:05 |
arxcruz | or if you point me the review i can monitor | 14:06 |
*** ThiagoCMC has joined #openstack-ansible | 14:06 | |
*** udesale has quit IRC | 14:07 | |
evrardjp | chandankumar: let me check | 14:08 |
evrardjp | arxcruz: for sure | 14:08 |
evrardjp | chandankumar: oh that's interesting | 14:08 |
evrardjp | chandankumar: that's generally when systemd manages to start the service and everything is fine, but in fact it's not fine | 14:09 |
evrardjp | we can check that later | 14:09 |
evrardjp | let me first check the package name for subunit | 14:09 |
evrardjp | chandankumar: 423? | 14:09 |
ThiagoCMC | cloudnull, good morning! Quick question, with LXC, I can easily attach to each container, but with nspawn, there is the `machinectl login ...` and it asks for login/pass! Where is it configured? Or, is there another way to attach to nspawn containers without user/pass? | 14:10 |
evrardjp | ThiagoCMC: nsenter | 14:10 |
chandankumar | evrardjp: https://review.openstack.org/591424 it is a massive change | 14:10 |
evrardjp | find the root pid and use that | 14:10 |
evrardjp | chandankumar: we've been there, don't worry we'll get you there :) | 14:11 |
evrardjp | (sorry for the current state of master) | 14:11 |
ThiagoCMC | evrardjp, nester? | 14:12 |
evrardjp | nsenter | 14:12 |
evrardjp | allows you to enter the namspace | 14:12 |
chandankumar | evrardjp: no problem | 14:12 |
ThiagoCMC | Oh, I see | 14:13 |
ThiagoCMC | =) | 14:13 |
evrardjp | chandankumar: I am not sure if that would help you but the only thing I see for 42.3 is python-python-subunit on a search | 14:13 |
chandankumar | evrardjp: that got added yesterday | 14:13 |
evrardjp | ThiagoCMC: I am using machinectl to find info about pids then enter with nsenter | 14:13 |
evrardjp | ThiagoCMC: But I can't remember everything | 14:13 |
evrardjp | my old brain :p | 14:13 |
evrardjp | chandankumar: what do you mean? | 14:13 |
ThiagoCMC | Sounds good to start! Thanks! | 14:14 |
odyssey4me | ThiagoCMC: I think we've added a convenience script called something like 'nspawn-attach' | 14:14 |
*** pvradu_ has quit IRC | 14:14 | |
evrardjp | chandankumar: oh you mean python-python-subunit ? | 14:14 |
chandankumar | evrardjp: https://review.openstack.org/#/c/591424/42/vars/suse.yml@22 | 14:14 |
evrardjp | yeah that sounds like it | 14:14 |
evrardjp | so what's missing? | 14:14 |
*** pvradu has joined #openstack-ansible | 14:14 | |
chandankumar | subunit-filters | 14:14 |
evrardjp | I am not sure it's packaged | 14:15 |
chandankumar | the same error we hit for centos then we added subunit-filters now it is working | 14:15 |
ThiagoCMC | nspawn-attach worked! Thanks! | 14:15 |
evrardjp | cool | 14:16 |
*** shyamb has joined #openstack-ansible | 14:16 | |
* chandankumar will wait for sensei odyssey4me to come and have a look :-) | 14:16 | |
evrardjp | I am not sure odyssey4me will be helpful on packaging new things for an old suse :p | 14:17 |
evrardjp | hahaha | 14:17 |
*** aredinger has joined #openstack-ansible | 14:18 | |
evrardjp | chandankumar: what's the command you are looking for? | 14:20 |
evrardjp | checking if I can find a package that could map | 14:20 |
hwoarang | maybe there is some value to have this package in the OBS cloud repo | 14:20 |
*** maharg101 has quit IRC | 14:21 | |
hwoarang | the rules are more relaxed there | 14:21 |
hwoarang | i think... | 14:21 |
chandankumar | evrardjp: https://github.com/openstack/openstack-ansible-os_tempest/blob/master/tasks/tempest_run.yml#L52 | 14:22 |
evrardjp | hwoarang: I am checking on obs | 14:22 |
evrardjp | I see other packages matching subunit | 14:22 |
evrardjp | hwoarang: but I don't think the exact match is there | 14:22 |
evrardjp | someone with the srpm can probably do something on obs indeed | 14:22 |
suggestable | jamesdenton: ping. Are you online yet? I have some news! | 14:22 |
jamesdenton | i am! but will be OOO soon. Whatcha got? | 14:22 |
chandankumar | evrardjp: https://review.openstack.org/#/c/591424/42/tasks/tempest_run.yml@61 | 14:22 |
suggestable | jamesdenton: I got it working! | 14:23 |
suggestable | L3HA works! | 14:23 |
jamesdenton | hey, cool! | 14:23 |
jamesdenton | what did you do? | 14:23 |
suggestable | Basically, I took https://bugs.launchpad.net/openstack-ansible/+bug/1782195 and added a bit extra... | 14:24 |
openstack | Launchpad bug 1782195 in openstack-ansible "neutron l3 agent fails to spawn HA router" [High,In progress] | 14:24 |
*** cshen has joined #openstack-ansible | 14:24 | |
suggestable | jamesdenton: The key seems to be that the neutron-server containers have the HA interface set in neutron.conf to eth11, which doesn't exist on the neutron-server containers. Switching it to eth1 makes it work. | 14:25 |
evrardjp | chandankumar: in the meantime I wouldn't mind if we add two conditionals : running said step when not distro installs and not suse. With an extra TODO | 14:25 |
jamesdenton | interesting. did you see the errors in the l3 agent log? | 14:25 |
suggestable | jamesdenton: I had to adjust the path for the neutron venv in the symlink, of course, as I'm running 18.0.0, but it just worked! | 14:26 |
suggestable | jamesdenton: I saw the exact same error in the l3 agent log, but no stacktrace. | 14:26 |
jamesdenton | hmmm, ok. well.. good job! i'm deploying now, so ifthere's something to fix i'll try and nail it down | 14:27 |
suggestable | OK. Hopefully what I've just said will help somebody to make a patch for this, so that others deploying with L3HA won't face this problem. That bug's been open since July... | 14:28 |
chandankumar | evrardjp: http://logs.openstack.org/24/591424/42/check/openstack-ansible-functional-distro_install-opensuse-423/7ce063c/logs/ara-report/result/f8efe2c6-9157-4d13-a4eb-d13f83f66577/ | 14:28 |
chandankumar | evrardjp: i got the issue | 14:29 |
chandankumar | evrardjp: python-junitxml will fix the issue | 14:29 |
evrardjp | let me check the package real quick then | 14:29 |
jamesdenton | suggestable do you have the actual config option you modified? | 14:30 |
suggestable | jamesdenton: Sure. Let me get it for you. | 14:30 |
suggestable | jamesdenton: In addition to what was mentioned in the patch (fixing the keepalived state change stuff), I edited this line in neutron.conf on the neutron-server container: | 14:32 |
suggestable | jamesdenton: ha_network_physical_name = eth1 | 14:32 |
chandankumar | evrardjp: https://pkgs.org/download/python-junitxml | 14:32 |
jamesdenton | suggestable what is l3_ha_network_type set to? anything? | 14:33 |
suggestable | jamesdenton: By default, it is set to eth11. I set it to eth1. | 14:33 |
jamesdenton | right, sorry. it's a different one: l3_ha_network_type | 14:34 |
suggestable | jamesdenton: Sorry. ha_network_type = vlan | 14:34 |
jamesdenton | ahh ok | 14:34 |
suggestable | As you know, we're running VXLAN, but eth1 (br-mgmt) is a flat network, so VLAN seemed appropriate there. | 14:35 |
openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Added support for installing tempest from distro https://review.openstack.org/591424 | 14:35 |
*** fatdragon has joined #openstack-ansible | 14:35 | |
jamesdenton | l3_ha_network_physical_name should be the provider label and not necessarily a physical interface name. | 14:38 |
suggestable | jamesdenton: Within the container, there's only eth0 (br-lxc0) and eth1 (br-mgmt). The container doesn't know any names other than eth0 and eth1. | 14:39 |
jamesdenton | right - neutron-server container isn't used for any network traffic, though. | 14:40 |
*** fatdragon has quit IRC | 14:41 | |
suggestable | jamesdenton: OK, but neutron-server itself runs inside the container, no? | 14:41 |
evrardjp | chandankumar: I was on obs to check | 14:41 |
chandankumar | evrardjp: what is the full url of obs? | 14:41 |
evrardjp | https://build.opensuse.org/package/view_file/openSUSE:Leap:42.3/python-junitxml/python-junitxml.spec?expand=1 | 14:41 |
*** shyamb has quit IRC | 14:42 | |
evrardjp | seems it would do the trick for you | 14:42 |
chandankumar | evrardjp: thanks :-) | 14:42 |
evrardjp | Sorry I was sidetracked | 14:42 |
evrardjp | :) | 14:42 |
jamesdenton | it runs inside a container, yes. those config options are used when implementing the ha network on the l3-agent node(s). if l3_ha_network_type is defined as vlan, then l3_ha_network_physical_name can be used as an override to specify the desired provider interface label. if l3_ha_network_type is left undefined, then the default tenant network type is used (so vxlan) | 14:42 |
suggestable | jamesdenton: There is no l3_ha_network_type specified in neutron.conf in the container, only ha_network_type, which is set to vlan by default. | 14:43 |
jamesdenton | it may be ha_network_type | 14:44 |
jamesdenton | k | 14:44 |
*** shyamb has joined #openstack-ansible | 14:44 | |
*** spatel has joined #openstack-ansible | 14:47 | |
yetiszaf | jamesdenton: I fear I need to bother you with a question as well | 14:50 |
yetiszaf | Is there anything I can do if the l2 population in a dvr-ovs-vlan-environment is inconsistent apart from restarting the agents or rebooting? | 14:51 |
*** thuydang has quit IRC | 14:53 | |
spotz | About if needed | 14:53 |
*** chhagarw has quit IRC | 14:57 | |
suggestable | Anyone got any advice for how to migrate from an old Icehouse POC-turned-Prod environment to our shiny new everything-done-properly Rocky environment? Manually exporting/importing images of the ~300 instances will cost us a lot of disk space (and lose the benefits of QCOW). | 14:57 |
jamesdenton | yetiszaf i have to step away for a few hours | 14:59 |
jamesdenton | yetiszaf l don't care much for l2pop and just end up disabling in favor of multicast/broadcast | 14:59 |
jamesdenton | or better yet, vlan | 14:59 |
*** shyamb has quit IRC | 15:00 | |
yetiszaf | jamesdenton: Ok, Thanks. | 15:01 |
*** Leo_m has joined #openstack-ansible | 15:01 | |
spotz | suggestable: could you potentially script and export, import with delete before moving on? Might be able to do it without losing anything using the APIs | 15:06 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-repo_server master: Use distro packages only after they are installed https://review.openstack.org/613069 | 15:11 |
*** fatdragon has joined #openstack-ansible | 15:13 | |
suggestable | spotz: "import with delete"? What's that? | 15:14 |
spotz | suggestable: so write a script that does an export, imports into the new system then deletes on the filesystem:) | 15:15 |
*** gyee has joined #openstack-ansible | 15:15 | |
spotz | You were worried about disk space but if you do it as you go no disk issues:) | 15:15 |
suggestable | spotz: Won't that impact my ability to migrate instances or shutdown/restart them if they've lost their base image? | 15:17 |
*** fatdragon has quit IRC | 15:18 | |
suggestable | spotz: Ideally, I'd like to migrate the base images, then migrate the QCOW delta for each instance, along with all its metadata (adjusted for the new environment) so that the instances don't really notice much has changed other than their IPs. | 15:18 |
spotz | suggestable: I'm just saying delete the image you've exported from where it's exported too, not from the cluster itself. Basically you'd export from cluster a on a machine1 running the script. import into cluster b and then clean up machine1 before moving to the next image | 15:18 |
spotz | Nothing has changed on cluster a | 15:19 |
suggestable | spotz: OK, that's easy. What results from it, though, is cluster b ends up with images for every instance that was migrated, and the instances have lost their association with the base image they were built from. | 15:19 |
*** Emine has quit IRC | 15:22 | |
spotz | suggestable: could just clean up afterwards with a database update. Other options are to bring everything over on the filesystem to the same location providing that's where glance is keeping the images and do a dump and load with any fixes. Not sure of DB schema compatability though | 15:22 |
spotz | This might be a better question for the ops channel in case someone has done it already | 15:22 |
suggestable | spotz: OK. Thank you :-) | 15:23 |
suggestable | I'll join #openstack-ops. | 15:24 |
spotz | #openstack-operators :) | 15:24 |
spotz | Nice folks over there too:) | 15:24 |
suggestable | That would be why there was nobody in -ops... :-( | 15:24 |
spotz | may not be as active so be patient getting a response | 15:24 |
suggestable | No worries. Thank you! :-) | 15:25 |
*** vnogin has quit IRC | 15:26 | |
spotz | You're welcome:) | 15:28 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Minimise distro packages installed into the lxc image https://review.openstack.org/613078 | 15:31 |
*** fatdragon has joined #openstack-ansible | 15:32 | |
*** ThiagoCMC has quit IRC | 15:40 | |
*** fghaas has quit IRC | 15:42 | |
*** pcaruana has quit IRC | 15:49 | |
mbuil | I am having some issues with the rabbitmq role because the amount of forks increases over the limit of the cgroup when the service is started. I saw that the role includes a config which limits the amount of threads to 128 and fixes my problem: https://github.com/openstack/openstack-ansible-rabbitmq_server/blob/master/templates/rabbitmq-env.j2#L4. Nevertheless, the first time the service is started is in the rabbitmq_install.yml and | 15:51 |
mbuil | at that point the config generated by OSA is not there yet. That config is generated in the rabbitmq_post_install.yml (and the service is restarted) | 15:51 |
openstackgerrit | Merged openstack/openstack-ansible-ops master: MNAIO: Implement branch var for pinning Ansible & Plugins https://review.openstack.org/612963 | 15:51 |
mbuil | Does anyone understand why the service is started in the install.yml playbook? Just checking before I create a patch | 15:51 |
*** fghaas has joined #openstack-ansible | 15:57 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Minimise distro packages installed into the lxc image https://review.openstack.org/613078 | 15:57 |
*** dave-mccowan has quit IRC | 16:02 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-memcached_server master: Add netcat as a required package https://review.openstack.org/613099 | 16:11 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Minimise distro packages installed into the lxc image https://review.openstack.org/613078 | 16:14 |
*** suggestable has quit IRC | 16:15 | |
*** vnogin has joined #openstack-ansible | 16:16 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_keystone master: Add cron as a required package https://review.openstack.org/613102 | 16:17 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Minimise distro packages installed into the lxc image https://review.openstack.org/613078 | 16:18 |
cloudnull | mbuil I thought that we disabled the service start on install. however its possible Im mistaken there. | 16:19 |
*** nsmeds_ has joined #openstack-ansible | 16:19 | |
cloudnull | ThiagoCMC evrardjp there's also `nspawn-attach $CONTAINERNAME` or `machinectl shell $CONTAINERNAME`` | 16:20 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_glance master: Add cron as a required package https://review.openstack.org/613103 | 16:22 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Minimise distro packages installed into the lxc image https://review.openstack.org/613078 | 16:23 |
*** vnogin has quit IRC | 16:24 | |
*** pvradu has quit IRC | 16:26 | |
*** pvradu has joined #openstack-ansible | 16:26 | |
*** aedc has joined #openstack-ansible | 16:27 | |
openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-rabbitmq_server master: Start rabbitmq service when config is in place https://review.openstack.org/613104 | 16:29 |
*** pvradu has quit IRC | 16:31 | |
nsmeds_ | also, can specify a shell with machinectl, which I find handy | 16:33 |
nsmeds_ | `machinectl shell $CONTAINERNAME /bin/bash` for example | 16:33 |
*** vnogin has joined #openstack-ansible | 16:33 | |
antonym | anyone familiar with the cinder service reload vs restart in os-cinder-install? it appears that it was set to restart to avoid a cpu issues but then reordered and set to reload https://github.com/openstack/openstack-ansible/commit/4b7527917dd18d6d0d3f98e2eceb7264e10cfd41 i'm noticing after a N2Q upgrade that cinder scheduler is spinning at 100% complaining about a connection loss to db which | 16:35 |
antonym | seems to settle after a final restart | 16:35 |
*** vnogin has quit IRC | 16:38 | |
openstackgerrit | Manuel Buil proposed openstack/openstack-ansible-rabbitmq_server master: Start rabbitmq service when config is in place https://review.openstack.org/613104 | 16:39 |
*** ansmith has quit IRC | 16:39 | |
*** pcaruana has joined #openstack-ansible | 16:46 | |
openstackgerrit | Merged openstack/openstack-ansible-tests stable/rocky: Do not use master branch for role tests https://review.openstack.org/612983 | 16:50 |
jrosser | odyssey4me: i made a stab at reducing all those -dev packages - https://review.openstack.org/#/q/topic:reduce-packages+(status:open+OR+status:merged) | 17:05 |
jrosser | i'm a bit stuck on cinder when installing packages into the cinder venv | 17:05 |
*** irclogbot_0 has quit IRC | 17:08 | |
*** pvradu has joined #openstack-ansible | 17:23 | |
*** ansmith has joined #openstack-ansible | 17:27 | |
*** pvradu has quit IRC | 17:27 | |
*** reineruhry has joined #openstack-ansible | 17:33 | |
*** spatel has quit IRC | 17:36 | |
reineruhry | Hi anybody there? | 17:43 |
*** ThiagoCMC has joined #openstack-ansible | 17:47 | |
*** reineruhry has quit IRC | 17:47 | |
*** hamzaachi has quit IRC | 17:48 | |
*** brent_u has joined #openstack-ansible | 17:50 | |
*** reineruhry has joined #openstack-ansible | 17:52 | |
reineruhry | I wanna rebuild a node, is there a playbook where I can clean everything? | 17:54 |
brent_u | what type of node? | 18:01 |
openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Stop arbitrary shutting down services https://review.openstack.org/606221 | 18:02 |
reineruhry | It's one of the hosts running mixed services, want to be sure that I can remove it and all configurations pointing to it, will also be removed! | 18:02 |
*** electrofelix has quit IRC | 18:05 | |
*** aredinger has quit IRC | 18:07 | |
openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: Break out config files for each octavia service https://review.openstack.org/603228 | 18:08 |
jrosser | reineruhry: I don’t think there is such a thing as “remove configurations pointing to it” | 18:10 |
jrosser | If you have a HA deployment then in principle the services should be resilient to a node being removed | 18:10 |
reineruhry | Yes, but the configuration files will keep a little "dirt", as it will have the old config, in it, right? | 18:11 |
jrosser | There are instructions for “major upgrades” which quite likely involve something similar, I woulld give those a read | 18:11 |
reineruhry | Ok, I'll, thank you! | 18:12 |
jrosser | I assume that when you say rebuild a node you want to bring it back online again doing the same function? | 18:12 |
reineruhry | no, I wan't it doing only one function... | 18:13 |
jrosser | Ah well that is a bit different I guess | 18:13 |
reineruhry | before I had the node running, horizon, galera, cinder... | 18:14 |
reineruhry | we decided to dedicate the node to cinder... | 18:14 |
reineruhry | but after "cleaning" as much as possible, I started having issues with repo, and many others... | 18:14 |
jrosser | Right, so that sounds more like you want to redistribute the existing containers | 18:14 |
reineruhry | btw... it's a "fresh" setup | 18:15 |
jrosser | You shouldn’t need to do any cleaning, deleting containers from one host, updating the inventory to change the physical host and redeploying the container should require no config changes | 18:15 |
reineruhry | but as we're facing this kind of situation now, we would like to understand what to do later | 18:16 |
reineruhry | right... | 18:16 |
reineruhry | so I'll bring the error that I'm having now... | 18:16 |
reineruhry | the repository is failing with file not found... | 18:17 |
reineruhry | calling http://IP:8181/venvs/18.0.0/ubuntu-18.04-x86_64/keystone-18.0.0-x86_64.chec | 18:17 |
*** ivve has quit IRC | 18:18 | |
reineruhry | I've tried to rebuild the repos containers, but keeps the same error | 18:18 |
openstackgerrit | tonytan4ever proposed openstack/openstack-ansible-ops master: Parameterize ansible version in embedded runtime https://review.openstack.org/613133 | 18:18 |
openstackgerrit | tonytan4ever proposed openstack/openstack-ansible-ops master: Parameterize ansible version in embedded runtime https://review.openstack.org/613133 | 18:19 |
jrosser | reineruhry: sorry I have to go now - there will be folks in better time zones who can help too.... | 18:20 |
reineruhry | thank you! | 18:21 |
openstackgerrit | Victor Palma proposed openstack/openstack-ansible-ops master: update to osquery https://review.openstack.org/613136 | 18:23 |
*** reineruhry has quit IRC | 18:26 | |
*** brent_u has quit IRC | 18:28 | |
*** tridde has joined #openstack-ansible | 18:43 | |
*** jhesketh_ has joined #openstack-ansible | 18:44 | |
*** FrankZhang has quit IRC | 18:48 | |
*** ThiagoCMC has quit IRC | 18:48 | |
*** priteau has quit IRC | 18:48 | |
*** aludwar has quit IRC | 18:48 | |
*** raginbajin has quit IRC | 18:48 | |
*** jbadiapa has quit IRC | 18:48 | |
*** trident has quit IRC | 18:48 | |
*** sum12 has quit IRC | 18:48 | |
*** dcdamien has quit IRC | 18:48 | |
*** mcarden has quit IRC | 18:48 | |
*** spotz has quit IRC | 18:48 | |
*** radez has quit IRC | 18:48 | |
*** errr has quit IRC | 18:48 | |
*** jhesketh has quit IRC | 18:48 | |
*** jbadiapa has joined #openstack-ansible | 18:49 | |
*** sum12 has joined #openstack-ansible | 18:49 | |
*** dcdamien has joined #openstack-ansible | 18:49 | |
*** errr has joined #openstack-ansible | 18:50 | |
*** spotz has joined #openstack-ansible | 18:52 | |
*** aludwar has joined #openstack-ansible | 18:54 | |
*** openstackgerrit has quit IRC | 19:06 | |
*** DanyC has joined #openstack-ansible | 19:14 | |
*** ivve has joined #openstack-ansible | 19:20 | |
*** hamzaachi has joined #openstack-ansible | 19:24 | |
*** irclogbot_0 has joined #openstack-ansible | 19:27 | |
*** irclogbot_0 has quit IRC | 19:27 | |
*** irclogbot_0 has joined #openstack-ansible | 19:32 | |
*** DanyC has quit IRC | 19:45 | |
*** openstackgerrit has joined #openstack-ansible | 19:54 | |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_keystone master: Remove paste references in Keystone https://review.openstack.org/613030 | 19:54 |
openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_keystone master: Remove paste references in Keystone https://review.openstack.org/613030 | 19:58 |
*** strattao has quit IRC | 20:17 | |
*** ivve has quit IRC | 20:19 | |
*** irclogbot_0 has quit IRC | 20:21 | |
*** DanyC has joined #openstack-ansible | 20:23 | |
*** ansmith has quit IRC | 20:34 | |
*** mmercer has joined #openstack-ansible | 20:43 | |
*** DanyC has quit IRC | 20:49 | |
*** pcaruana has quit IRC | 20:50 | |
*** openstack has quit IRC | 21:03 | |
*** openstack has joined #openstack-ansible | 21:07 | |
*** ChanServ sets mode: +o openstack | 21:07 | |
*** fghaas has left #openstack-ansible | 21:10 | |
*** Leo_m has quit IRC | 21:10 | |
*** spsurya has quit IRC | 21:21 | |
*** vnogin has joined #openstack-ansible | 21:25 | |
*** vnogin has quit IRC | 21:29 | |
*** ansmith has joined #openstack-ansible | 21:30 | |
openstackgerrit | Nicolas Bock proposed openstack/openstack-ansible-os_nova master: Add `cpu_model` configuration option https://review.openstack.org/613168 | 21:37 |
*** vakuznet has quit IRC | 21:40 | |
*** spatel has joined #openstack-ansible | 21:44 | |
*** isitirctime has joined #openstack-ansible | 21:46 | |
isitirctime | Hope everyone is well | 21:46 |
isitirctime | Seem to have an issue going from pike to queens with ansible bootstrap | 21:46 |
isitirctime | ailed: [localhost] (item={'scm': u'git', 'src': u'https://git.openstack.org/openstack/openstack-ansible-os_tacker', 'version': u'30bd74292021e5507cdfad4b9a7f827a5f8452f6', 'name': u'os_tacker'}) => {"attempts": 2, "changed": false, "cmd": "/usr/bin/git checkout --force 30bd74292021e5507cdfad4b9a7f827a5f8452f6", "failed": true, "item": {"name": "os_tacker", "scm": "git", "src": "https://git.openstack.org/openstack/openstack-a | 21:46 |
isitirctime | nsible-os_tacker", "version": "30bd74292021e5507cdfad4b9a7f827a5f8452f6"}, "msg": "Failed to checkout 30bd74292021e5507cdfad4b9a7f827a5f8452f6", "rc": 128, "stderr": "fatal: reference is not a tree: 30bd74292021e5507cdfad4b9a7f827a5f8452f6\n", "stderr_lines": ["fatal: reference is not a tree: 30bd74292021e5507cdfad4b9a7f827a5f8452f6"], "stdout": "", "stdout_lines": []} | 21:46 |
isitirctime | the repo exist and I can pull it down on my desk and git check out. Kind of perplexed | 21:47 |
*** irclogbot_0 has joined #openstack-ansible | 21:48 | |
openstackgerrit | Victor Palma proposed openstack/openstack-ansible-ops master: update to osquery https://review.openstack.org/613136 | 21:49 |
*** hamzaachi has quit IRC | 21:50 | |
*** irclogbot_0 has quit IRC | 21:50 | |
*** tosky has quit IRC | 21:53 | |
*** spatel has quit IRC | 22:21 | |
*** cshen has quit IRC | 22:37 | |
*** aredinger has joined #openstack-ansible | 23:07 | |
isitirctime | deleting /etc/ansible/roles/os_tacker and rerunning the setup worked for me | 23:09 |
isitirctime | My test openstack updating to the same version of queens did not run into that problem so probably just an issue with the preexisting pike repo | 23:16 |
*** openstackgerrit has quit IRC | 23:20 | |
*** spatel has joined #openstack-ansible | 23:33 | |
*** spatel has quit IRC | 23:37 | |
*** ansmith has quit IRC | 23:38 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!