| *** cloudnull has joined #openstack-ansible | 00:02 | |
| *** jamesden_ has quit IRC | 00:23 | |
| *** jamesdenton has joined #openstack-ansible | 00:25 | |
| *** djhankb has quit IRC | 00:27 | |
| *** djhankb has joined #openstack-ansible | 00:27 | |
| *** jamesden_ has joined #openstack-ansible | 00:45 | |
| *** jamesdenton has quit IRC | 00:47 | |
| *** cshen has joined #openstack-ansible | 00:49 | |
| *** cshen has quit IRC | 00:53 | |
| *** gyee has quit IRC | 01:32 | |
| *** yolanda has quit IRC | 02:07 | |
| *** djhankb has quit IRC | 02:28 | |
| *** djhankb has joined #openstack-ansible | 02:29 | |
| *** cshen has joined #openstack-ansible | 02:49 | |
| *** cshen has quit IRC | 02:54 | |
| *** cshen has joined #openstack-ansible | 04:29 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-ansible | 04:33 | |
| *** cshen has quit IRC | 04:33 | |
| *** dave-mccowan has quit IRC | 04:42 | |
| *** shyamb has joined #openstack-ansible | 04:52 | |
| *** shyamb has quit IRC | 05:14 | |
| *** shyamb has joined #openstack-ansible | 05:21 | |
| *** shyam89 has joined #openstack-ansible | 05:24 | |
| *** shyamb has quit IRC | 05:27 | |
| BlackFX | I am about ready to give up lol | 05:30 |
|---|---|---|
| *** shyam89 has quit IRC | 05:49 | |
| jrosser | BlackFX: there is a playbook playbooks/lxc-containers-destroy.yml | 05:49 |
| jrosser | BlackFX: did you try an all-in-one in the end? | 05:52 |
| jrosser | noonedeadpunk: senlin has unusual properties, like heavy use of openstacksdk inside the service itself, like here https://opendev.org/openstack/senlin/src/branch/master/senlin/drivers/sdk.py#L109-L126 | 05:56 |
| jrosser | there does not appear to be a way to put connection parameters for openstacksdk into the config file, so things we might usually template into a config file can't be set | 05:57 |
| *** cshen has joined #openstack-ansible | 06:00 | |
| BlackFX | Yeah I did, that's how I worked out the indentation error. | 06:02 |
| *** pcaruana has joined #openstack-ansible | 06:04 | |
| BlackFX | I am gonna clean up all the containers and try a fresh run, I can't really see why it'd be upset. The only thing I am doing that isn't in one of the seplates is haveing 2 LVM backends for cinder. | 06:04 |
| *** cshen has quit IRC | 06:05 | |
| BlackFX | *templates | 06:05 |
| *** pcaruana has quit IRC | 06:06 | |
| *** pcaruana has joined #openstack-ansible | 06:06 | |
| jrosser | being able to completely clean the nodes is very useful for testing | 06:14 |
| jrosser | otherwise it's never clear if something is residual from previous things | 06:14 |
| BlackFX | do I need to do anything other than the container destroy to clean up? | 06:15 |
| jrosser | that will remove the containers | 06:15 |
| *** pcaruana has quit IRC | 06:15 | |
| jrosser | but there is very much config done to the host itself | 06:15 |
| jrosser | i'm not sure you've shared the probem you currently have? | 06:15 |
| BlackFX | Still stuck with the galera cluster member issue when running setup-infrastructure | 06:16 |
| jrosser | sorry i cannot find it in the scollback | 06:18 |
| noonedeadpunk | jrosser: I'm not sure about this tbh... as eventually it's passing creds https://opendev.org/openstack/senlin/src/branch/master/senlin/drivers/os/keystone_v3.py#L93-L97 | 06:18 |
| noonedeadpunk | and creds are taken from config | 06:19 |
| jrosser | noonedeadpunk: ok sure, but endpoint, insecure flag...... | 06:19 |
| jrosser | i found it all goes wrong in an AIO | 06:19 |
| noonedeadpunk | hm, that's true.... | 06:20 |
| jrosser | standard failure pattern of defaulting to public endpoint then failing on cert verify | 06:20 |
| jrosser | and no way apparently to configure that | 06:20 |
| noonedeadpunk | let's suggest patch to senlin?:) | 06:20 |
| noonedeadpunk | as you can pass endpoint to openstacksdk connect | 06:21 |
| noonedeadpunk | (not 100% about insecure) | 06:21 |
| BlackFX | TASK [galera_server : Fail when the host is not in galera_cluster_members] ********************************************************************************************** | 06:21 |
| BlackFX | fatal: [infra1_utility_container-8d88b1ab]: FAILED! => {"changed": false, "msg": "The host infra1_utility_container-8d88b1ab must be in galera_cluster_members."} | 06:21 |
| jrosser | noonedeadpunk: in create_connection i hacked in params['identity_interface']='internal' | 06:21 |
| jrosser | and then it started working with cli from utility container | 06:22 |
| jrosser | i was wondering if the normal OS_<...> environment vars in the systemd unit could be used, if the openstacksdk library reads those.... | 06:23 |
| noonedeadpunk | we also can kind of add `'interface': cfg.CONF.authentication.interface` here https://opendev.org/openstack/senlin/src/branch/master/senlin/drivers/os/keystone_v3.py#L120 | 06:23 |
| noonedeadpunk | but need to check if they're probably passing smth with kwargs... | 06:23 |
| jrosser | BlackFX: so that is the same problem from before, somehow the utility container is targetted as one of the galera hosts | 06:24 |
| BlackFX | Yes, so I am just going to try destroying and regenerating the containers | 06:25 |
| jrosser | well | 06:25 |
| BlackFX | but I can see no reason for it to be behaving that way | 06:25 |
| jrosser | can you paste more of the context, just the error message itself is difficult to spot things | 06:26 |
| jrosser | playbooks/galera_insall.yml and just paste the whole output to paste.openstack.org / | 06:26 |
| BlackFX | ok | 06:27 |
| jrosser | i would suspect the deleting / re-creating the containers will behave the same, becasue the inventory information about which tasks run against which hosts, and ansible groups is all coming from outside those | 06:27 |
| *** pcaruana has joined #openstack-ansible | 06:28 | |
| noonedeadpunk | and verify can go there as well.... | 06:28 |
| jrosser | the whole code is interesting - not even sure it is using the normal keystone authtoken section at all right now | 06:30 |
| noonedeadpunk | it's using authenticate section I guess | 06:30 |
| noonedeadpunk | so yeah | 06:30 |
| noonedeadpunk | but I like using sdk tbh | 06:30 |
| jrosser | first time i see the sdk used like this in a service, which is kind of cool | 06:31 |
| jrosser | yes | 06:31 |
| noonedeadpunk | if you was not going to, I'll try to make patch to senlin to cover this | 06:32 |
| jrosser | please - i probably take short notice holiday next week so wont have opportunity to chase it | 06:32 |
| jrosser | wanted to get a test in with an AIO before that to look for issues - i already had someone ask for this kind of function here so very interested | 06:33 |
| jrosser | spatel patches in an AIO will deploy fine but then 'openstack cluster build info' in utility container will fail straight away with a stack trace | 06:34 |
| jrosser | the template for the [authentication] section is currently wrong but i put comments on the review | 06:35 |
| *** shyamb has joined #openstack-ansible | 06:35 | |
| BlackFX | Doing that has given totally different errors: http://paste.openstack.org/show/797450/ | 06:40 |
| jrosser | ok so tht suggests ansible cannot connect to your loadblanacer node at 192.168.2.2 | 06:43 |
| jrosser | hmm maybe not quite.... | 06:45 |
| jrosser | BlackFX: here is the task that fails https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/common-tasks/haproxy-endpoint-manage.yml#L16 | 06:48 |
| jrosser | and it looks like 'failed to connect' is maybe the haproxy control socket | 06:48 |
| jrosser | so perhaps haproxy is failing to start? | 06:48 |
| noonedeadpunk | jrosser: I think this should do the trick https://review.opendev.org/749874 | 06:49 |
| BlackFX | Hmm - okay - I will dig in that direction :) | 06:49 |
| jrosser | BlackFX: maybe check out what is actually inside all the playbooks/setup-<....>.yml | 06:49 |
| jrosser | it's just lists of other playbooks, you can call them individually just fine | 06:49 |
| jrosser | work methodically through | 06:49 |
| BlackFX | Okay so playbooks/galera_insall.yml now completes cleanly. setup-infrastructure still fails in exactly the same way | 06:53 |
| jrosser | noonedeadpunk: also this https://opendev.org/openstack/senlin/src/branch/master/senlin/drivers/os/keystone_v3.py#L151 | 06:56 |
| noonedeadpunk | ah damn it | 06:56 |
| noonedeadpunk | yeah | 06:56 |
| jrosser | noonedeadpunk: i expect this is actually involved in the functionality more http://codesearch.openstack.org/?q=get_senlin_endpoint&i=nope&files=&repos= | 06:57 |
| jrosser | so not really sure about what that actually does | 06:57 |
| jrosser | maybe tempest shows for that? | 06:57 |
| jrosser | public may be appropriate if it needs to find it's own endpoint from the perspective of VM | 06:58 |
| noonedeadpunk | I'm wondering about keystone_authtoken section as it's not mentioned anywhere except docs | 06:59 |
| jrosser | yes, it was kind of confusing | 06:59 |
| *** shyamb has quit IRC | 07:10 | |
| jrosser | BlackFX: i have to head off now but context for any error messages is really important | 07:10 |
| BlackFX | Okay - thanks | 07:10 |
| jrosser | this is so unusual "The host infra1_utility_container-8d88b1ab must be in galera_cluster_members" that i think folk will struggle to understand/help without the output from the play in a paste | 07:11 |
| BlackFX | it's failing in utility-install | 07:11 |
| BlackFX | rather than the galera install... | 07:11 |
| BlackFX | I'll see where I get to | 07:11 |
| jrosser | ok, important info :) | 07:11 |
| jrosser | which branch/tag is this? | 07:12 |
| BlackFX | On branch stable/ussuri | 07:13 |
| jrosser | here in the utlility playbook is where it should install the galera *client* https://github.com/openstack/openstack-ansible/blob/stable/ussuri/playbooks/utility-install.yml#L64-L67 | 07:14 |
| *** shyamb has joined #openstack-ansible | 07:14 | |
| noonedeadpunk | maybe it's just for keystonemiddleware to verify tokens.... | 07:14 |
| jrosser | your error suggests that maybe the galera server instead is being installed into the utility container, which is very wrong | 07:15 |
| BlackFX | Yeah thats what I am seeing locally too | 07:15 |
| jrosser | BlackFX: paste the utility play output and maybe noonedeadpunk can make a suggestion, really i must go now | 07:15 |
| *** shyamb has quit IRC | 07:16 | |
| *** shyamb has joined #openstack-ansible | 07:16 | |
| BlackFX | Okay thanks for your help | 07:16 |
| BlackFX | http://paste.openstack.org/show/797452/ this is the output if your keen @noonedeadpunk :) | 07:18 |
| *** cshen has joined #openstack-ansible | 07:18 | |
| *** shyamb has quit IRC | 07:18 | |
| *** shyamb has joined #openstack-ansible | 07:19 | |
| noonedeadpunk | iirc you don't have utility container in galera group? | 07:21 |
| *** tosky has joined #openstack-ansible | 07:25 | |
| BlackFX | for some reason it's attempting to install the server role on the utility container. Cant see any reason why it is, the playbook says not to | 07:26 |
| *** shyamb has quit IRC | 07:29 | |
| *** andrewbonney has joined #openstack-ansible | 07:40 | |
| noonedeadpunk | haven't you override galera_install_server? | 07:42 |
| noonedeadpunk | ie in user_variables or somewhere? | 07:42 |
| ebbex | noonedeadpunk: could you cast a vote on this https://review.opendev.org/#/c/749282/ ? | 07:45 |
| *** shyamb has joined #openstack-ansible | 07:47 | |
| *** shyamb has quit IRC | 07:48 | |
| *** shyamb has joined #openstack-ansible | 07:48 | |
| *** shyam89 has joined #openstack-ansible | 07:49 | |
| noonedeadpunk | ebbex: hm... have some concerns... | 07:53 |
| *** shyamb has quit IRC | 07:53 | |
| *** mmethot has joined #openstack-ansible | 07:54 | |
| *** mmethot_ has quit IRC | 07:56 | |
| noonedeadpunk | (commented) | 08:11 |
| BlackFX | Hmmmm - what listens on port 8181 - MaxRetryError: HTTPConnectionPool(host='192.168.2.2', port=8181) | 08:40 |
| *** shyamb has joined #openstack-ansible | 08:40 | |
| noonedeadpunk | used to be repo_server iirc | 08:42 |
| noonedeadpunk | but nothing nowadays | 08:42 |
| *** shyam89 has quit IRC | 08:42 | |
| BlackFX | this step in utility-install.yml is using it : TASK [python_venv_build : Upgrade pip/setuptools/wheel to the versions we want] * | 08:43 |
| noonedeadpunk | you should check that /root/.pip.conf is absent | 08:44 |
| *** SecOpsNinja has joined #openstack-ansible | 08:50 | |
| BlackFX | nope, no sign of it. | 09:02 |
| SecOpsNinja | regaring lxc containers, how is the network configured in them? the host has default route but the containers don't (even if i try to recreate them). this happend after i tried to create the external flat network in haproxy and replaced the haproxy configuration | 09:02 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_glance master: Fix native service path https://review.opendev.org/749907 | 09:12 |
| noonedeadpunk | SecOpsNinja: so lxcbr should have srcnat from the host | 09:12 |
| noonedeadpunk | BlackFX: hm, weird. maybe smth in /etc/pip.conf or dunno.. I think that package source is overriden, but we don't have such thing since queens or rocky I guess... | 09:14 |
| jrosser | port 8181 is still the repo server (via the loadbalancer) for wheels and venv build stuff | 09:17 |
| noonedeadpunk | indeed, sory... | 09:20 |
| jrosser | and missing default route is probably broken dnsmasq on the host for lxcbr0 | 09:22 |
| *** aedc has joined #openstack-ansible | 09:23 | |
| SecOpsNinja | noonedeadpunk, i was checking the containers and they dont have the default rule and because of that i can install software in them. is this definied in any tasks? what is normally the default network of them eth0 (lxbrc0) or eth1 (br-mgmt)? | 09:27 |
| jrosser | SecOpsNinja: eth0 is normally the default route | 09:28 |
| jrosser | which is NAT from the host by dnsmasq running on the host | 09:28 |
| SecOpsNinja | yep fro some reason the eth0 doesn have any ip in the containers in eth0 interface | 09:29 |
| SecOpsNinja | even trying to recreate it... | 09:29 |
| noonedeadpunk | hm, dnsmasq on the host should care about ips | 09:29 |
| SecOpsNinja | the eth0 is connected to lxcbr0 right? | 09:29 |
| jrosser | thats all configured here https://github.com/openstack/openstack-ansible-lxc_hosts/blob/master/defaults/main.yml#L101-L121 | 09:29 |
| jrosser | in the lxc_hosts role | 09:30 |
| BlackFX | haproxy was sick, thanks for the pointer | 09:30 |
| SecOpsNinja | how can i force the recreating of lxcbr0 in the infra host? | 09:35 |
| SecOpsNinja | becuase im seasing the that brctl only shows 1 container eth0 in lxcbr0 | 09:36 |
| jrosser | SecOpsNinja: https://github.com/openstack/openstack-ansible-lxc_hosts/blob/master/files/lxc-veth-check.sh | 09:37 |
| SecOpsNinja | jrosser, thanks | 09:38 |
| jrosser | that script should be installed by lxc_hosts role | 09:39 |
| SecOpsNinja | yes a lot of them is reporting has a broken veth d2e27537_eth0 and an automated fix can not be found | 09:40 |
| SecOpsNinja | yep and even in the recreated container thathave its eth0 in lxcbr0 is not able to resolve names (it doesn't have anything in /etc/resolv.conf) so i supose the prolem should be in dnsmasq? | 09:44 |
| *** shyamb has quit IRC | 09:54 | |
| BlackFX | So close lol, now failing setting up keystone when adding the service project | 09:59 |
| BlackFX | I think that will be a task for tomorrow | 09:59 |
| *** shyamb has joined #openstack-ansible | 10:01 | |
| BlackFX | http://paste.openstack.org/show/797456/ | 10:23 |
| *** djhankb has quit IRC | 10:28 | |
| *** djhankb has joined #openstack-ansible | 10:29 | |
| openstackgerrit | Erik Berg proposed openstack/openstack-ansible-ceph_client stable/ussuri: Remove trailing '/' from ceph_apt_repo_url https://review.opendev.org/749282 | 10:57 |
| ebbex | noonedeadpunk, jrosser : would this be better? Removing the repo with trailing slash line before we add a new ceph-repo (which someone could possibly override to add back with a trailing slash) | 11:04 |
| *** aedc has quit IRC | 11:04 | |
| noonedeadpunk | perwsonally I like it more that way) | 11:08 |
| noonedeadpunk | worth adding note that we can remove that task in wallaby release | 11:08 |
| jrosser | ebbex: we should probably revisit the master patch? | 11:46 |
| *** djhankb has quit IRC | 11:53 | |
| *** lkoranda has joined #openstack-ansible | 11:54 | |
| *** djhankb has joined #openstack-ansible | 11:54 | |
| *** omegapoint has joined #openstack-ansible | 11:56 | |
| *** dave-mccowan has joined #openstack-ansible | 11:59 | |
| mgariepy | can we ask for centos8 repos to be mirrored in CI ? | 12:03 |
| mgariepy | ho.. different fail on the other patch. | 12:05 |
| mgariepy | ho. not really, same error only not at the same place. | 12:06 |
| mgariepy | 2020-09-04 07:33:01.800559 | centos-8 | fatal: [aio1_glance_container-d3d2f29b]: FAILED! => {"changed": false, "msg": "Failed to download metadata for repo 'epel-networkd': Yum repo downloading error: Downloading error(s): repodata/c984028c6055d55d0d54c2f4324c00445148b90451a2e0bd4c545447c10c4a01-primary.xml.gz - Cannot download, all mirrors were already tried without success; repodata/2094cdfc0278a8e6fb7dd86f308943f3751496c0bbba3245b | 12:06 |
| mgariepy | db2c48366166636-filelists.xml.gz - Cannot download, all mirrors were already tried without success; repodata/e2c3195544a9b831b711df514ee9f0558ca6c2455e8e144b3c0b2d61aed1e624-prestodelta.xml.xz - Cannot download, all mirrors were already tried without success", "rc": 1, "results": []} | 12:06 |
| *** redrobot has quit IRC | 12:10 | |
| noonedeadpunk | eah, epel instanly brokem :( | 12:15 |
| noonedeadpunk | have no idea how to work around it | 12:16 |
| noonedeadpunk | Not sure if mirrioring will do the trick as it seems the issue is that during runtime they change these uids | 12:17 |
| noonedeadpunk | we had the same for centos 7 rdo repos and were doing smth complicated stuff there | 12:17 |
| noonedeadpunk | (trying to avoid uid usage) | 12:17 |
| noonedeadpunk | but I didn't find the way for epel.... | 12:18 |
| mgariepy | not sure why ppl like centos.. | 12:19 |
| *** cshen has quit IRC | 12:21 | |
| *** csmart has quit IRC | 12:21 | |
| *** fridtjof[m] has quit IRC | 12:21 | |
| *** masterpe has quit IRC | 12:22 | |
| *** ioni has quit IRC | 12:22 | |
| *** shyamb has quit IRC | 12:24 | |
| watersj | centos == old =~ stable | 12:30 |
| *** masterpe has joined #openstack-ansible | 12:31 | |
| mgariepy | well. sure it can keep running but when you do use centos you always need some extra unstable repos | 12:34 |
| *** mmethot has quit IRC | 13:01 | |
| *** fridtjof[m] has joined #openstack-ansible | 13:03 | |
| *** ioni has joined #openstack-ansible | 13:03 | |
| *** csmart has joined #openstack-ansible | 13:03 | |
| *** mmethot has joined #openstack-ansible | 13:03 | |
| *** mmethot has quit IRC | 13:08 | |
| *** spatel has joined #openstack-ansible | 13:08 | |
| *** mmethot has joined #openstack-ansible | 13:08 | |
| *** spatel has quit IRC | 13:12 | |
| *** lkoranda has quit IRC | 13:18 | |
| openstackgerrit | James Denton proposed openstack/openstack-ansible-os_ironic master: [WIP] Add iPXE support to Ironic Conductor https://review.opendev.org/736336 | 13:28 |
| openstackgerrit | James Denton proposed openstack/openstack-ansible-os_ironic master: Add iPXE support to Ironic Conductor https://review.opendev.org/736336 | 13:31 |
| *** cshen has joined #openstack-ansible | 13:32 | |
| ebbex | jrosser: if https://review.opendev.org/#/c/748412/1 goes through, it's almost like the other patch in master never happened. | 13:37 |
| jrosser | oh of course! | 13:38 |
| noonedeadpunk | ebbex: are you sure it needs depends-on? | 13:40 |
| ebbex | noonedeadpunk: not anymore, as it's already in master. they were submitted together | 13:46 |
| ebbex | the cleanup-logic was mostly for convenience to cherry-pick down to stable branches. | 13:47 |
| noonedeadpunk | it's worth I think to specify not patch id, but link to the patch. as now we're waiting for ussuri patch to merge to merge master... | 13:47 |
| omegapoint | hey, i'm running into this bug trying to upgrade from pike to queens, how can i fix this? https://review.opendev.org/#/c/605601/2/tasks/ceilometer_db_setup.yml (the parameter doesn't exist so my playbook run fails, but i don't know where to find the file that parameterizes this locally) | 13:48 |
| ebbex | noonedeadpunk: I can resubmit it without the depends-on. gimme two minutes :) | 13:48 |
| openstackgerrit | Erik Berg proposed openstack/openstack-ansible-ceph_client master: Remove cleanup logic for trailing '/' https://review.opendev.org/748412 | 13:51 |
| *** d34dh0r53 has joined #openstack-ansible | 14:03 | |
| *** dave-mccowan has quit IRC | 14:08 | |
| openstackgerrit | Satish Patel proposed openstack/openstack-ansible-os_senlin master: Adjusted senlin_pip_packages: and changed senlin_container_mysql_password to senlin_galera_password https://review.opendev.org/749365 | 14:21 |
| *** dave-mccowan has joined #openstack-ansible | 14:24 | |
| *** cshen has quit IRC | 14:34 | |
| *** cshen has joined #openstack-ansible | 15:01 | |
| *** rpittau is now known as rpittau|afk | 15:18 | |
| omegapoint | is it possible that the docs for using the calico neutron plugin is incomplete? https://docs.openstack.org/openstack-ansible-os_neutron/queens/app-calico.html i performed the steps in the docs but it appears to yield an invalid neutron-server config when switching from linuxbridge to calico (it still contains vlan and vxlan type drivers but they aren't available). | 15:21 |
| jamesden_ | it is possible, yes. | 15:31 |
| *** jamesden_ is now known as jamesdenton | 15:31 | |
| jamesdenton | what is valid there? flat? | 15:32 |
| jamesdenton | if ml2.calico is set as the neutron_plugin_type (not neutron_plugin_types), then this should be invoked: https://github.com/openstack/openstack-ansible-os_neutron/blob/stable/queens/vars/main.yml#L203-L208 | 15:33 |
| omegapoint | yeah i did set neutron_plugin_type to ml2.calico but it didn't render it as "flat,local" as described in that file... strange | 15:38 |
| omegapoint | i just resorted to setting neutron_ml2_drivers_type to "flat,local", which should have the same effect i guess? | 15:39 |
| jamesdenton | it should, yes | 15:44 |
| *** gyee has joined #openstack-ansible | 15:44 | |
| jamesdenton | did the mechanism_drivers get set to calico in ml2_conf.ini? | 15:45 |
| omegapoint | yep | 15:47 |
| *** cshen has quit IRC | 15:48 | |
| jamesdenton | can you share the ml2 conf file? | 15:49 |
| omegapoint | http://paste.openstack.org/show/797481/ | 15:54 |
| jamesdenton | ok, so your type drivers got set OK | 15:57 |
| jamesdenton | but yeah, tenant network types need to be adjusted accordingly | 15:57 |
| jamesdenton | and that might be coming from your provider network definition in openstack_user_config | 15:58 |
| jamesdenton | or user_variables if you're overriding that | 15:58 |
| NobodyCam | Good morning OSA folks! | 16:08 |
| *** MickyMan77 has joined #openstack-ansible | 16:11 | |
| NobodyCam | Off the wall question, I ended up with a IP conflict between a compute host and a lxc-container it is there a way to edit the Ip in the inventory, I attempted to remove the offending host with: | 16:13 |
| NobodyCam | https://www.irccloud.com/pastebin/8PJUsZUJ/ | 16:13 |
| NobodyCam | but it is finding the lxc-container | 16:14 |
| MickyMan77 | I gettings this kind of error on one of the controller nodes, all nodes is installed with same hardware/software, the other two nodes is installed without any problem.--------------------fatal: [controller03-for_utility_container-13eded48]: FAILED! => {"changed": false, "checksum": "4c93a25d01ca2ab694e06c7d7697df48b47fdbf2", "msg": "Destination | 16:14 |
| MickyMan77 | directory /etc/sudoers.d does not exist"}controller03-for_utility_container-13eded48 : ok=63 changed=3 unreachable=0 failed=1 skipped=3 rescued=0 ignored=0------------------openstack-ansible version 21.0.0.0rc1-108-g72a39d652CentOS 8.2 | 16:14 |
| MickyMan77 | i will repost that. | 16:15 |
| MickyMan77 | I gettings this kind of error on one of the controller nodes, all nodes is installed with same hardware/software, the other two nodes is installed without any problem. | 16:15 |
| MickyMan77 | fatal: [controller03-for_utility_container-13eded48]: FAILED! => {"changed": false, "checksum": "4c93a25d01ca2ab694e06c7d7697df48b47fdbf2", "msg": "Destination directory /etc/sudoers.d does not exist"} | 16:15 |
| MickyMan77 | controller03-for_utility_container-13eded48 : ok=63 changed=3 unreachable=0 failed=1 skipped=3 rescued=0 ignored=0 | 16:15 |
| MickyMan77 | --- | 16:16 |
| MickyMan77 | openstack-ansible version 21.0.0.0rc1-108-g72a39d652CentOS 8.2 | 16:16 |
| ebbex | MickyMan77: do you know which step this in the playbooks setup-hosts, setup-infrastructure, setup-openstack? | 16:19 |
| MickyMan77 | setup-hosts | 16:19 |
| *** cshen has joined #openstack-ansible | 16:19 | |
| *** cshen has quit IRC | 16:24 | |
| ebbex | I can't really tell where this goes wrong, what are the tasks preceeding the failure? | 16:31 |
| *** cshen has joined #openstack-ansible | 17:20 | |
| *** SecOpsNinja has left #openstack-ansible | 17:38 | |
| *** omegapoint has quit IRC | 17:49 | |
| openstackgerrit | Merged openstack/openstack-ansible-ceph_client stable/ussuri: Remove trailing '/' from ceph_apt_repo_url https://review.opendev.org/749282 | 18:07 |
| *** andrewbonney has quit IRC | 18:22 | |
| *** cshen has quit IRC | 18:29 | |
| *** MickyMan77 has quit IRC | 18:33 | |
| jrosser | NobodyCam: if you have your /etc/openstack_deploy in git then you could try editing and references to the IP of the compute host in the inventory json | 18:44 |
| jrosser | just make sure you have a way to back out if something goes bad | 18:45 |
| NobodyCam | Ah thank you | 18:45 |
| jrosser | moving the compute host IP is probably less work than fixing an lxc container IP | 18:46 |
| *** dave-mccowan has quit IRC | 18:46 | |
| NobodyCam | Yea that’s what I’m attempting | 18:49 |
| *** cshen has joined #openstack-ansible | 18:59 | |
| *** alvinstarr has quit IRC | 19:10 | |
| *** cshen has quit IRC | 19:40 | |
| *** spatel has joined #openstack-ansible | 19:52 | |
| *** spatel has quit IRC | 19:56 | |
| *** cshen has joined #openstack-ansible | 20:03 | |
| *** cshen has quit IRC | 20:07 | |
| *** renich has joined #openstack-ansible | 20:22 | |
| *** renich has quit IRC | 21:21 | |
| *** rh-jelabarre has quit IRC | 21:29 | |
| BlackFX | Does anyone know what ansible is doing during the "os_keystone: Add service project" step? It is failing but the debug output doesn't show what it's actually running through python on the host. http://paste.openstack.org/show/797456/ | 21:44 |
| BlackFX | Hmm now it's changed to a 504 error http://paste.openstack.org/show/797490/ | 21:53 |
| *** cshen has joined #openstack-ansible | 22:04 | |
| *** cshen has quit IRC | 22:08 | |
| BlackFX | Anyone know where the keystone container sends it's logs? | 22:14 |
| BlackFX | UFW!!!!! | 22:39 |
| *** tosky has quit IRC | 23:07 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!