| *** jtriley has quit IRC | 00:17 | |
| *** unicell1 has quit IRC | 00:17 | |
| *** unicell has joined #kolla | 00:17 | |
| *** jtriley has joined #kolla | 00:18 | |
| *** Kennan has quit IRC | 00:20 | |
| *** Kennan has joined #kolla | 00:21 | |
| *** dims has joined #kolla | 00:23 | |
| *** Marga__ has quit IRC | 00:27 | |
| *** Marga_ has joined #kolla | 00:28 | |
| *** dims has quit IRC | 00:32 | |
| *** jtriley has quit IRC | 00:37 | |
| *** jasonsb has joined #kolla | 00:37 | |
| *** ssurana has quit IRC | 00:42 | |
| *** iceyao has joined #kolla | 00:53 | |
| *** sdake has joined #kolla | 00:55 | |
| *** dims has joined #kolla | 00:56 | |
| *** sdake has quit IRC | 01:10 | |
| *** unicell has quit IRC | 01:13 | |
| *** unicell1 has joined #kolla | 01:13 | |
| *** jtriley has joined #kolla | 01:23 | |
| *** tzn has joined #kolla | 01:24 | |
| *** tzn has quit IRC | 01:29 | |
| *** iceyao has quit IRC | 01:37 | |
| *** unicell1 has quit IRC | 01:54 | |
| *** salv-orlando has joined #kolla | 02:02 | |
| *** salv-orl_ has quit IRC | 02:04 | |
| *** rhallisey has quit IRC | 02:05 | |
| *** tfukushima has joined #kolla | 02:05 | |
| *** rhallisey has joined #kolla | 02:08 | |
| *** vhosakot has joined #kolla | 02:12 | |
| *** jtriley has quit IRC | 02:23 | |
| *** jtriley has joined #kolla | 02:28 | |
| openstackgerrit | Merged openstack/kolla: Use ubuntu Mitaka repo https://review.openstack.org/279910 | 02:35 |
|---|---|---|
| *** rhallisey has quit IRC | 02:35 | |
| *** dims has quit IRC | 02:38 | |
| *** jtriley has quit IRC | 02:46 | |
| *** klint has joined #kolla | 02:48 | |
| sbezverk | SamYaple, are you still here? | 02:52 |
| *** unicell has joined #kolla | 02:54 | |
| *** jtriley has joined #kolla | 03:07 | |
| *** jtriley has quit IRC | 03:12 | |
| openstackgerrit | Daniel Gonzalez Nothnagel proposed openstack/kolla: Unify vagrant bootstrap.sh scripts https://review.openstack.org/288821 | 03:17 |
| *** dave-mccowan has quit IRC | 03:17 | |
| *** vhosakot has quit IRC | 03:18 | |
| *** vhosakot has joined #kolla | 03:18 | |
| *** yuanying has quit IRC | 03:22 | |
| *** Kennan has quit IRC | 03:23 | |
| *** Kennan has joined #kolla | 03:26 | |
| *** phuongnh has joined #kolla | 03:37 | |
| *** yuanying has joined #kolla | 03:38 | |
| *** yuanying has quit IRC | 03:45 | |
| SamYaple | sbezverk: off and on | 03:46 |
| *** ayoung has quit IRC | 03:57 | |
| *** Marga_ has quit IRC | 03:59 | |
| *** Marga_ has joined #kolla | 04:00 | |
| openstackgerrit | Vikram Hosakote proposed openstack/kolla: Implement binary Ubuntu container for Kibana https://review.openstack.org/291986 | 04:03 |
| *** sdake has joined #kolla | 04:06 | |
| *** yuanying has joined #kolla | 04:06 | |
| openstackgerrit | Merged openstack/kolla: Unify vagrant bootstrap.sh scripts https://review.openstack.org/288821 | 04:09 |
| *** sdake_ has joined #kolla | 04:09 | |
| *** sdake has quit IRC | 04:11 | |
| *** tfukushima has quit IRC | 04:14 | |
| *** sdake_ has quit IRC | 04:19 | |
| *** ssurana has joined #kolla | 04:21 | |
| *** Marga_ has joined #kolla | 04:21 | |
| *** Marga_ has quit IRC | 04:25 | |
| openstackgerrit | MD NADEEM proposed openstack/kolla: Manila ubuntu binary container https://review.openstack.org/286379 | 04:26 |
| mdnadeem | vhosakot, hi, i have tested the ^^ patch, it successfully build and deploy images except manila-share as expected | 04:28 |
| vhosakot | mdnadeem: cool, so all manila containers built and deployed fine, right ? | 04:28 |
| mdnadeem | vhosakot, yes | 04:28 |
| vhosakot | mdnadeem: cool, thanks for testing and confirming! | 04:28 |
| mdnadeem | vhosakot, I face UnicodeEncodeError while deploying manila. to fix it a minor update required in manila.conf.j2, i will upload a patch to fix it soon | 04:31 |
| vhosakot | cool | 04:31 |
| *** Marga_ has joined #kolla | 04:33 | |
| *** Marga_ has quit IRC | 04:34 | |
| *** Marga_ has joined #kolla | 04:34 | |
| *** SiRiuS_ has joined #kolla | 04:39 | |
| openstackgerrit | MD NADEEM proposed openstack/kolla: Fix UnicodeEncodeError for Manila https://review.openstack.org/292707 | 04:40 |
| *** SiRiuS_ has quit IRC | 04:41 | |
| *** SiRiuS_ has joined #kolla | 04:41 | |
| *** tzn has joined #kolla | 05:01 | |
| *** tzn has quit IRC | 05:01 | |
| mandre | SamYaple: would it make sense to add more dependencies for our ansible roles beside the 'common' role? | 05:03 |
| SamYaple | mandre: if needed. i think nova has a memcache dep | 05:04 |
| SamYaple | but we should be focused on stablizing right now | 05:04 |
| SamYaple | too many new features being pumped out right now.... | 05:04 |
| mandre | SamYaple, agreed, now is not the good time to work on new feature, right now I wanted to know if that's a good idea or not | 05:05 |
| SamYaple | i think so | 05:05 |
| SamYaple | if it makes sense | 05:05 |
| SamYaple | what are you thinkign about specifically? | 05:06 |
| mandre | well, that's mostly for so that deploying with a tag will start the required containers without having to think about all dependent tags that are missing from the command line | 05:06 |
| SamYaple | oh well i dont know about that | 05:06 |
| SiRiuS_ | does kolla include the LXD nova compute driver? | 05:06 |
| SamYaple | i mean nova has a dep on keystone then | 05:06 |
| SamYaple | SiRiuS_: kolla has only been tested with kvm, qemu, and xen | 05:07 |
| SamYaple | and xen support is partial for mitaka | 05:07 |
| SiRiuS_ | SamYaple: ok, thanks | 05:07 |
| mandre | yeah, so if I kolla-ansible deploy -t nova, I expect to have keystone deployed as well | 05:07 |
| SamYaple | mandre: thats not a reasonable expectation | 05:12 |
| SamYaple | we have built these from the beginning to make is so you dont have to deploy the full suite | 05:12 |
| SamYaple | (you can run nova without kolla-keystone) | 05:12 |
| mandre | will it be functional though? | 05:12 |
| SamYaple | but more importantly, if i run --tags nova, i only want the nova playbooks to run | 05:12 |
| SamYaple | if its setup right sure | 05:13 |
| mandre | didn't you just say nova has a dep on keystone? | 05:13 |
| SamYaple | not kolla-keystone | 05:15 |
| SamYaple | if we do what you purpose how can i run only the nova code? | 05:15 |
| SamYaple | nova upgrade playbooks me keystone upgrades first | 05:15 |
| SamYaple | you cant _not_ do that | 05:15 |
| *** sdake has joined #kolla | 05:15 | |
| mandre | oh ok, I see what you mean | 05:16 |
| mandre | that's right, we certainly don't want to upgrade anything beside what's specified on the command line | 05:17 |
| *** sdake has quit IRC | 05:33 | |
| *** SiRiuS_ has quit IRC | 05:47 | |
| SamYaple | mandre: i have pfsense replacing my neutron routers | 05:48 |
| SamYaple | its pretty sweet | 05:48 |
| SamYaple | its got transparent caching! | 05:48 |
| SamYaple | rpm and debs | 05:49 |
| SamYaple | all at 10gb speed (after initial pull) | 05:49 |
| *** SiRiuS_ has joined #kolla | 05:49 | |
| mandre | SamYaple: sweet! how long does it take for a full rebuild? | 05:50 |
| openstackgerrit | Merged openstack/kolla: Fix UnicodeEncodeError for Manila https://review.openstack.org/292707 | 05:53 |
| *** salv-orl_ has joined #kolla | 05:54 | |
| *** pbourke has quit IRC | 05:55 | |
| *** pbourke has joined #kolla | 05:55 | |
| *** salv-orlando has quit IRC | 05:57 | |
| *** SiRiuS_ has quit IRC | 06:04 | |
| openstackgerrit | Vikram Hosakote proposed openstack/kolla: Implement Ubuntu container for Kibana https://review.openstack.org/291986 | 06:14 |
| *** ssurana has quit IRC | 06:38 | |
| *** coolsvap has quit IRC | 06:48 | |
| *** Serlex has joined #kolla | 07:04 | |
| *** mikelk has joined #kolla | 07:10 | |
| openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Remove deployment_id_prefix for per-service commands https://review.openstack.org/291054 | 07:11 |
| openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Move common config functions into a new file https://review.openstack.org/286995 | 07:11 |
| openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Move Runner classes to service.py https://review.openstack.org/286994 | 07:11 |
| openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: CLI: implement service commands https://review.openstack.org/286996 | 07:11 |
| openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: CLI: implement config commands https://review.openstack.org/289295 | 07:11 |
| openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Convert chronos-cli to kolla-mesos chronos https://review.openstack.org/292738 | 07:11 |
| openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Move write_openrc to a common file https://review.openstack.org/292739 | 07:11 |
| openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Move apply_deployment_vars to a common file https://review.openstack.org/292740 | 07:11 |
| openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: CLI: implement deployment commands https://review.openstack.org/292741 | 07:11 |
| *** SiRiuS_ has joined #kolla | 07:39 | |
| *** athomas has joined #kolla | 07:54 | |
| *** iceyao has joined #kolla | 07:55 | |
| *** iceyao has quit IRC | 07:56 | |
| *** stvnoyes has quit IRC | 08:00 | |
| *** stvnoyes has joined #kolla | 08:00 | |
| *** jmccarthy has quit IRC | 08:03 | |
| *** jmccarthy has joined #kolla | 08:04 | |
| *** akwasnie has joined #kolla | 08:10 | |
| *** vhosakot has quit IRC | 08:11 | |
| *** salv-orl_ has quit IRC | 08:18 | |
| *** mbound has joined #kolla | 08:56 | |
| *** tzn has joined #kolla | 09:01 | |
| *** phuongnh has quit IRC | 09:04 | |
| *** salv-orlando has joined #kolla | 09:21 | |
| *** klint has quit IRC | 09:28 | |
| *** klint has joined #kolla | 09:28 | |
| *** salv-orlando has quit IRC | 09:29 | |
| *** salv-orlando has joined #kolla | 09:30 | |
| *** tzn has quit IRC | 09:33 | |
| openstackgerrit | MD NADEEM proposed openstack/kolla: Fix race condition in HAproxy https://review.openstack.org/292782 | 09:41 |
| nihilifer | can we get this stream of patches reviewed? https://review.openstack.org/#/c/291114 | 09:48 |
| nihilifer | akwasnie: pbourke ? ;) | 09:48 |
| *** kproskurin has joined #kolla | 09:53 | |
| *** cristicalin has joined #kolla | 10:01 | |
| pbourke | nihilifer: i'll try take a look soon | 10:05 |
| *** pbourke has quit IRC | 10:09 | |
| *** kman has joined #kolla | 10:11 | |
| *** cristicalin has quit IRC | 10:11 | |
| *** pbourke has joined #kolla | 10:12 | |
| *** kman has quit IRC | 10:13 | |
| *** dave-mccowan has joined #kolla | 10:24 | |
| *** athomas has quit IRC | 10:25 | |
| *** kproskurin has quit IRC | 10:26 | |
| *** athomas has joined #kolla | 10:31 | |
| *** shardy has joined #kolla | 10:39 | |
| *** tzn has joined #kolla | 10:44 | |
| *** kproskurin has joined #kolla | 10:55 | |
| *** cristicalin has joined #kolla | 11:06 | |
| *** v1k0d3n has quit IRC | 11:07 | |
| *** v1k0d3n has joined #kolla | 11:08 | |
| *** dims has joined #kolla | 11:12 | |
| *** rhallisey has joined #kolla | 11:12 | |
| -openstackstatus- NOTICE: Gerrit is going to be restarted | 11:12 | |
| *** kproskurin has quit IRC | 11:14 | |
| SiRiuS_ | I have some suggestions for the Vagrant bootstrap script | 11:16 |
| SiRiuS_ | For CentOS the package net-tools should be installed | 11:16 |
| SiRiuS_ | Because the cleanup-host script uses ifconfig which is not present on CentOS by default | 11:17 |
| *** cristicalin has quit IRC | 11:19 | |
| *** akwasnie has quit IRC | 11:20 | |
| *** dims has quit IRC | 11:20 | |
| *** dims has joined #kolla | 11:21 | |
| *** dims has quit IRC | 11:21 | |
| *** dims has joined #kolla | 11:21 | |
| SiRiuS_ | Also, on line 166 of the bootstrap.sh "tox -e genconfig" won't work because there is no tox.ini in the home directory | 11:22 |
| SiRiuS_ | There should be a "cd koala" followed by "tox -e genconfig" followed by "cd .." | 11:23 |
| SiRiuS_ | Actually it should be "cd ${KOLLA_PATH}" followed by "tox -e genconfig" followed by "cd -" | 11:27 |
| SiRiuS_ | SamYaple: ping | 11:29 |
| -openstackstatus- NOTICE: Gerrit had to be restarted because was not responsive. As a consequence, some of the test results have been lost, from 08:30 UTC to 10:30 UTC approximately. Please recheck any affected jobs by this problem. | 11:32 | |
| *** akwasnie has joined #kolla | 11:32 | |
| *** tzn has quit IRC | 11:32 | |
| *** akwasnie has quit IRC | 11:33 | |
| *** tzn has joined #kolla | 11:33 | |
| -openstackstatus- NOTICE: Gerrit had to be restarted because was not responsive. As a consequence, some of the test results have been lost, from 09:30 UTC to 11:30 UTC approximately. Please recheck any affected jobs by this problem. | 11:34 | |
| *** akwasnie has joined #kolla | 11:34 | |
| *** alyson_ has joined #kolla | 11:34 | |
| *** mlima has joined #kolla | 11:35 | |
| *** kproskurin has joined #kolla | 11:36 | |
| *** prithiv has joined #kolla | 11:40 | |
| prithiv | have anyone tried kolla multi-node setup | 11:44 |
| openstackgerrit | Michal Rostecki proposed openstack/kolla-mesos: CLI: implement deployment commands https://review.openstack.org/292741 | 11:45 |
| *** dwalsh has joined #kolla | 11:49 | |
| *** akwasnie has quit IRC | 11:51 | |
| *** salv-orl_ has joined #kolla | 11:54 | |
| SiRiuS_ | prithiv: The multi-node does not work properly, see https://bugs.launchpad.net/kolla/+bug/1546789 | 11:54 |
| openstack | Launchpad bug 1546789 in kolla "neutron failed to deploy in a multi-node deployment" [High,Triaged] | 11:54 |
| ccesario_ | SiRiuS_, prithiv ... you can try setup this for a workaround ... http://sprunge.us/cjCR?diff | 11:57 |
| ccesario_ | http://puu.sh/nGBB9/a087164b69.png http://puu.sh/nGBCl/97e4c2e0fa.png | 11:57 |
| *** salv-orlando has quit IRC | 11:57 | |
| *** fthiagogv has joined #kolla | 11:59 | |
| SiRiuS_ | ccesario_: thanks | 11:59 |
| *** akwasnie has joined #kolla | 12:01 | |
| akwasnie | nihilifer: hi, sorry I had some meetings | 12:01 |
| akwasnie | which patch? | 12:02 |
| *** dencaval has quit IRC | 12:02 | |
| *** salv-orl_ has quit IRC | 12:08 | |
| ccesario_ | SiRiuS_, if you test it, please let me know if works as expected :) | 12:14 |
| *** dencaval has joined #kolla | 12:15 | |
| SiRiuS_ | ccesario_: sure thing | 12:15 |
| *** shardy has quit IRC | 12:16 | |
| *** iceyao has joined #kolla | 12:17 | |
| *** prithiv has quit IRC | 12:22 | |
| *** prithiv has joined #kolla | 12:22 | |
| *** ayoung has joined #kolla | 12:23 | |
| *** dims has quit IRC | 12:25 | |
| *** dims has joined #kolla | 12:26 | |
| fthiagogv | SamYaple, about this https://review.openstack.org/#/c/291888/ this flavor id option is needed to user do not create a new flavor each time that deploy. For easy use this option is needed. | 12:39 |
| *** shardy has joined #kolla | 12:44 | |
| *** shardy has quit IRC | 12:47 | |
| *** dwalsh has quit IRC | 12:48 | |
| *** shardy has joined #kolla | 12:48 | |
| sbezverk | rhallisey ping | 12:52 |
| rhallisey | hey | 12:53 |
| sbezverk | check this out: http://paste.openstack.org/show/490499/ | 12:53 |
| rhallisey | nice | 12:54 |
| sbezverk | :-) yeah, finally | 12:54 |
| rhallisey | nice job | 12:55 |
| sbezverk | rhallisey thank you. I have a list of about 10 hacks I had to do to make it working. Need advice. | 12:55 |
| rhallisey | ok | 12:56 |
| sbezverk | it is in addition to the patch | 12:56 |
| sbezverk | I am concern that it would make main patch too big | 12:56 |
| rhallisey | sbezverk, split it up then | 12:56 |
| *** mgoddard_ has joined #kolla | 12:57 | |
| sbezverk | ok should I put these 10 hack in one patch or patch per hack.. | 12:58 |
| rhallisey | depends if they're related | 12:58 |
| rhallisey | try and group them together | 12:58 |
| *** mgoddard_ has quit IRC | 12:59 | |
| *** mgoddard__ has joined #kolla | 12:59 | |
| *** mgoddard has quit IRC | 13:00 | |
| sbezverk | rhallisey ok I will try to split and not break anything ;-) | 13:01 |
| SamYaple | fthiagogv: that option is the exact time of option that the augment files are for | 13:01 |
| *** klint has quit IRC | 13:03 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Test https://review.openstack.org/292869 | 13:04 |
| fthiagogv | SamYaple, I did not understand what did you mean. | 13:06 |
| SamYaple | fthiagogv: are you familiar with the augment files? | 13:06 |
| SiRiuS_ | I found a bug in the Vagrantfile | 13:07 |
| SamYaple | prithiv: i run it exclusively in multinode nowadays | 13:07 |
| SamYaple | openstack.yaple.net is a full HA multinode cluster with native kolla | 13:07 |
| *** sdake has joined #kolla | 13:08 | |
| prithiv | @Sam | 13:08 |
| prithiv | i dont see my list of hypervisors | 13:08 |
| prithiv | in multi node setup with fake libvirt driver | 13:08 |
| SiRiuS_ | `when "virtualbox_ubuntu"` and `when "virtualbox_centos"` .... should be replaced by just one line `when "virtualbox"` | 13:08 |
| *** inc0 has joined #kolla | 13:08 | |
| inc0 | good morning | 13:08 |
| SiRiuS_ | otherwise the host manager plugin won't function properly | 13:09 |
| SamYaple | prithiv: i have not used the fake libvirt driver myself | 13:09 |
| SamYaple | if i recall hui implemented that | 13:09 |
| SamYaple | hey inc0 ! | 13:09 |
| prithiv | Hui ? | 13:10 |
| SamYaple | prithiv: i dont know his irc name | 13:10 |
| prithiv | no problem i will find it out. cheers | 13:10 |
| SamYaple | prithiv: he is not in US and i dont think he is on IRC often | 13:11 |
| *** dwalsh has joined #kolla | 13:11 | |
| prithiv | I will find him in the launchpad and will try to email him. I think I know him he is from Intel if am not wrong. | 13:11 |
| *** salv-orlando has joined #kolla | 13:16 | |
| *** ayoung has quit IRC | 13:26 | |
| *** mgoddard_ has joined #kolla | 13:32 | |
| openstackgerrit | Proskurin Kirill proposed openstack/kolla-mesos: run_once commands locking rework https://review.openstack.org/291693 | 13:33 |
| openstackgerrit | Proskurin Kirill proposed openstack/kolla-mesos: Add timestamp to logging and remove useless log string https://review.openstack.org/291692 | 13:33 |
| openstackgerrit | Proskurin Kirill proposed openstack/kolla-mesos: set_state now could use exact path as an arg https://review.openstack.org/292876 | 13:33 |
| openstackgerrit | Proskurin Kirill proposed openstack/kolla-mesos: Create separate ZK lock nodes https://review.openstack.org/292877 | 13:33 |
| openstackgerrit | Proskurin Kirill proposed openstack/kolla-mesos: Add hostname of a node as a lock name. https://review.openstack.org/292878 | 13:33 |
| *** jtriley has joined #kolla | 13:34 | |
| *** mgoddard__ has quit IRC | 13:35 | |
| *** sdake has quit IRC | 13:38 | |
| fthiagogv | SamYaple, I am not. But I will read about it to undestand what you mean. :) | 13:40 |
| openstackgerrit | Proskurin Kirill proposed openstack/kolla-mesos: Add hostname of a node as a lock name. https://review.openstack.org/292878 | 13:40 |
| *** ayoung has joined #kolla | 13:40 | |
| *** dwalsh has quit IRC | 13:42 | |
| SamYaple | fthiagogv: you cna created a file at /etc/kolla/config/manila.conf with content [generic]\service_instance_flavor_id = <id> | 13:43 |
| SamYaple | it will inject it | 13:43 |
| SamYaple | or you can setup the flavor with id 100 in your cloud | 13:43 |
| *** dwalsh has joined #kolla | 13:43 | |
| SamYaple | thats _your_ call, but it might be worth documentation | 13:44 |
| *** kproskurin has quit IRC | 13:47 | |
| *** Jeffrey4l has joined #kolla | 13:53 | |
| sbezverk | SamYaple multiple process in openstack access /etc/iscsi/initiatorname.conf, a target host might not have iscsi-initiator utils installed, so we cannot depend on it. I made iscsi-initatos part of base image so processes with right volume mapping can access it, but here comes a problem, this name gets generated when we build container, if the same image gets deployed on multiple targets, they | 13:54 |
| sbezverk | will all have the same initiator id. would it be ok during the deployment substitute generated id with some randonly generated id? | 13:54 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Move sysctl out of haproxy container https://review.openstack.org/292892 | 13:55 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Switch to with items for sysctl https://review.openstack.org/292893 | 13:55 |
| SamYaple | sbezverk: im sorry can you ask me later today? im a bit musy at the moment and havent processed your question | 13:56 |
| fthiagogv | SamYaple, so will be it similar to ceph? Just will say to create a config file before to deploy? | 13:56 |
| sbezverk | SamYaple sure thing.. | 13:56 |
| fthiagogv | in a doc | 13:56 |
| openstackgerrit | Serguei Bezverkhi proposed openstack/kolla: [WIP] iscsi container with lvm2 support https://review.openstack.org/291285 | 13:58 |
| *** kproskurin has joined #kolla | 13:59 | |
| *** Guest41345 has quit IRC | 14:03 | |
| *** jmccarthy1 has joined #kolla | 14:04 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Test https://review.openstack.org/292869 | 14:06 |
| *** Marga_ has quit IRC | 14:08 | |
| *** spisarski has joined #kolla | 14:09 | |
| *** salv-orlando has quit IRC | 14:12 | |
| *** salv-orlando has joined #kolla | 14:13 | |
| *** sdake has joined #kolla | 14:25 | |
| *** tfukushima has joined #kolla | 14:25 | |
| *** tfukushima has quit IRC | 14:26 | |
| *** jtriley has quit IRC | 14:26 | |
| sdake | morning | 14:33 |
| akwasnie | hi sdake | 14:34 |
| sdake | hey akwasnie | 14:34 |
| sdake | did you see the bug i reported about authenticating kibana | 14:34 |
| *** salv-orlando has quit IRC | 14:35 | |
| *** salv-orl_ has joined #kolla | 14:35 | |
| akwasnie | yes, I was working yesterday on a POC and have a working solution based on nginx but would like to discuss it on the next meeting | 14:35 |
| sdake | i think apache would be better | 14:35 |
| sdake | but whatever works | 14:35 |
| sdake | re the dashboard thing | 14:36 |
| sdake | i think i can find you some helpers | 14:36 |
| sdake | do you have a handle on how the json plugin stuff works by looking at thefuel plugin? | 14:36 |
| akwasnie | I tried Apache also but configuration for nginx was easier | 14:36 |
| sdake | apache will be easier to add a tls endpoint to | 14:37 |
| sdake | maybe that is already handled by haproxy though | 14:37 |
| sdake | what do you need to discuss inthe meeting | 14:37 |
| sdake | just go ahead and submit patches | 14:38 |
| akwasnie | whether we want to create separate container for 'kibana-proxy' or sth like that? or do it in kibana container? | 14:38 |
| SamYaple | akwasnie: im not so sure i like that approach | 14:39 |
| sdake | akwasnie this is why apache would be better | 14:39 |
| SamYaple | akwasnie: does kibana access need to be autheneticated | 14:39 |
| SamYaple | sdake: no it wouldnt | 14:39 |
| sdake | apache could serve up the web pages rather then ngix run as a proxy | 14:39 |
| SamYaple | its the same issue | 14:39 |
| akwasnie | so do we want to install apache in kibana container and do everything there? | 14:39 |
| SamYaple | akwasnie: no | 14:40 |
| sdake | akwasnie you know how kibana runs its own "web server" app now | 14:40 |
| SamYaple | thats not a good approach | 14:40 |
| sdake | what we want is to run apache instead | 14:40 |
| akwasnie | SamYaple: yes, we would like to add authentication | 14:40 |
| SamYaple | akwasnie: what about shield? is there a reason we dont want that (ie is it commercial or free)? | 14:40 |
| sdake | commercial pay | 14:40 |
| sdake | shield is auth per row of data | 14:41 |
| sdake | we just need auth of kibana in general | 14:41 |
| SamYaple | shield looks free to me | 14:41 |
| SamYaple | and protests elasticsearch and kibana | 14:41 |
| sdake | my wife worked there she said it definately wasn't free | 14:41 |
| sdake | but i guess if it is, we could use it | 14:41 |
| SamYaple | idk im checking now | 14:41 |
| SamYaple | might be hidden pay | 14:42 |
| akwasnie | looks like it is free | 14:42 |
| sdake | i dont mind commercial pay in our solutions but we can't mandate them | 14:42 |
| sdake | i dont know if shield does password auth or row locking only | 14:42 |
| *** rtmie has quit IRC | 14:43 | |
| sdake | it would be nice to know if hsield was a 30 day eval thing or soething liek that | 14:43 |
| sdake | because thats not ideal - we can't mandate proprietary softwrae implementations in a FOSS project | 14:43 |
| *** jmccarthy1 has quit IRC | 14:44 | |
| sdake | i'm not really sure that shield solves the problem either | 14:44 |
| akwasnie | I don't know how shield exactly works but can look at it | 14:44 |
| *** jtriley has joined #kolla | 14:44 | |
| sdake | serving kibana pages via apache is a easy win | 14:45 |
| sdake | but if shield works and is free then that also works | 14:45 |
| SamYaple | akwasnie: if shield doesnt work for pay reasons nginx proxy is what well have to do. | 14:45 |
| SamYaple | where we set it up is the question | 14:46 |
| sdake | SamYaple you can serve kibana pages via apache | 14:46 |
| SamYaple | would it be infront of haproxy, or in front of each node | 14:46 |
| SamYaple | sdake: nginx is the recommended approach | 14:46 |
| sdake | then put an htpaswd file in apache | 14:46 |
| sdake | recommended by who? | 14:46 |
| SamYaple | feel free to do your own research sdake | 14:47 |
| sdake | i did, the only thing i saw is "use nginx" threads on various websites | 14:47 |
| sdake | but clearly no recommendation from elastic! | 14:47 |
| SamYaple | why would they recommend that when they have shield | 14:48 |
| SamYaple | they wont make a recomendation | 14:48 |
| sdake | precisely | 14:48 |
| sdake | serving up via apache is simple and easy and easy to authenticate | 14:48 |
| sdake | i pasted an apache file right in the bug | 14:48 |
| sdake | i do not know what nginx is, my understanding is it is a proxy server | 14:49 |
| sdake | we really dont want proxy servers if we can just serve content directly | 14:49 |
| *** kproskurin has quit IRC | 14:50 | |
| * elemoine_ reading the Kibana discussion… | 14:50 | |
| akwasnie | sdake: nginx configuration looks like this http://paste.openstack.org/show/490525/ | 14:50 |
| SamYaple | dude we arent using .htpasswd that stuff is awful | 14:50 |
| SamYaple | nginx is much better for auth | 14:50 |
| SamYaple | .htpasswd is prone to security issues by messed up perms | 14:50 |
| sdake | so your suggeting running an entire nginx proxy container just for kibana auth? | 14:51 |
| *** haplo37 has joined #kolla | 14:51 | |
| sdake | how is kibana secured in that model? | 14:51 |
| SamYaple | how will we secure elasticsearch then? | 14:51 |
| SamYaple | what happens when we have more things to secure? | 14:52 |
| sdake | well that is actually a good point | 14:52 |
| SamYaple | im hoping we can use it for memcached too but im not sure yet | 14:52 |
| sdake | hadn't thought about the es angle | 14:52 |
| sdake | although i dont know if it can be used for es auth | 14:53 |
| sdake | akwasnie try out shield first, learn how it works - see if it gets the job done | 14:53 |
| SamYaple | even without the other stuff, its still faster than apache and not prone to permission issues | 14:53 |
| akwasnie | sdake: ok | 14:54 |
| -openstackstatus- NOTICE: Launchpad OpenID SSO is currently experiencing issues preventing login. The Launchpad team is working on the issue | 14:55 | |
| *** ChanServ changes topic to "Launchpad OpenID SSO is currently experiencing issues preventing login. The Launchpad team is working on the issue" | 14:55 | |
| openstackgerrit | Jeffrey Zhang proposed openstack/kolla: add Ubuntu binary build and deploy gate https://review.openstack.org/292930 | 14:57 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Test https://review.openstack.org/292869 | 14:58 |
| *** jmccarthy1 has joined #kolla | 14:58 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Move sysctl out of haproxy container https://review.openstack.org/292892 | 14:59 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Switch to with items for sysctl https://review.openstack.org/292893 | 14:59 |
| *** mgoddard_ has quit IRC | 14:59 | |
| *** mgoddard has joined #kolla | 14:59 | |
| elemoine_ | akwasnie: I am pretty sure Shield would do the job. Shield also provides a Kibana plugin for authentication/access control. | 15:00 |
| *** ssurana has joined #kolla | 15:00 | |
| akwasnie | elemoine_: good to hear that, I will try it | 15:01 |
| akwasnie | thanks | 15:01 |
| *** jmccarthy1 has quit IRC | 15:01 | |
| *** jmccarthy1 has joined #kolla | 15:02 | |
| elemoine_ | akwasnie: and I really don't know how the problem can be solved w/o Shield | 15:03 |
| akwasnie | elemoine_: another idea is to use nginx as a proxy or serving Kibana via apache | 15:06 |
| elemoine_ | serving Kibana via Apache: not sure what this means | 15:07 |
| elemoine_ | but Kibana 4 runs in NodeJS, so it cannot run in Apache | 15:07 |
| akwasnie | it was possible for Kibana 3 | 15:07 |
| akwasnie | to serve it via Apache | 15:08 |
| elemoine_ | Kibana 3 and 4 are very different | 15:08 |
| akwasnie | yes, that is why I used nginx | 15:08 |
| elemoine_ | Kibana 3 was a client-side app | 15:08 |
| elemoine_ | Kibana 4 has a server part | 15:08 |
| sbezverk | what is the best way to create a folder which is required by the service but which is not created during the installation of this service? config.yml?? | 15:09 |
| elemoine_ | akwasnie: using a proxy in front of Kibana might be an option | 15:10 |
| elemoine_ | but I am not sure right now | 15:10 |
| akwasnie | elemoine_: we decided to try Shield first and see how it works and what it can do | 15:11 |
| elemoine_ | yeah saw that | 15:11 |
| SamYaple | sbezverk: can you expand? which folder (in a container, out?) | 15:11 |
| *** vhosakot has joined #kolla | 15:13 | |
| *** vhosakot has quit IRC | 15:13 | |
| *** vhosakot has joined #kolla | 15:14 | |
| sdake | elemoine_ thanks for that piece of info about kibana 4 not servable in a apache container | 15:16 |
| sdake | that is helpful | 15:16 |
| *** dwalsh has quit IRC | 15:17 | |
| sbezverk | SamYaple cinder_volume process stores targets in /var/lib/cinder/volumes | 15:19 |
| sdake | SamYaple did you see sbezverk got cinder working with lvm2+iscsi | 15:19 |
| sbezverk | tgtd process has access to this folder as well to read these targets. tgtd | 15:20 |
| SamYaple | sdake: i did | 15:20 |
| sdake | sbezverk use a named vollume | 15:20 |
| sbezverk | needs to include this folder in tgtd.conf | 15:20 |
| SamYaple | sbezverk: yea all those targets have to be shared | 15:20 |
| sbezverk | I want to add tgtd.conf as you normally do it with config.yml | 15:20 |
| SamYaple | there are two directories if i recall correctly | 15:20 |
| SamYaple | named_volume will be the right approach | 15:21 |
| sbezverk | but I want also to create /var/lib/cinder/volumes | 15:21 |
| SamYaple | inc0: i broke infra :P | 15:21 |
| inc0 | SamYaple, congrats | 15:21 |
| inc0 | how so? | 15:21 |
| SamYaple | https://review.openstack.org/#/c/292187/ followed by https://review.openstack.org/292924 | 15:21 |
| sdake | lol | 15:23 |
| sdake | best not to try to get fancy with zuul at this point | 15:23 |
| *** jtriley has quit IRC | 15:23 | |
| sdake | our gates are not nearly as invasive as other projects | 15:23 |
| sdake | what infra cares about is long running gates | 15:23 |
| SamYaple | no infra is working on it, that patch should have worked and i will be restoring it as soon as they fix the problem | 15:24 |
| sdake | they dont care about how many gates you have | 15:24 |
| SamYaple | they are actually freaked out agout it | 15:24 |
| sdake | freaked about what | 15:24 |
| sbezverk | SamYaple what about folder creation question? | 15:24 |
| sdake | sbezverk whe nyou mount hte named volume the folkder aluto-creates | 15:24 |
| sdake | if there are many gate jobs, they run in aprallel until all resources are consumed | 15:25 |
| sdake | but if gate jobs take a long time to run indivudually one project could hog all the resources for long periods | 15:25 |
| SamYaple | sdake: that i can break the gate with a bad regex | 15:25 |
| sdake | but if it has 20 jobs for example, thats no big deal if they take 10 mins about | 15:25 |
| sdake | SamYaple ya that is bad | 15:26 |
| SamYaple | (its actually a valid regex thats whats freaking them out) | 15:26 |
| *** blahRus has joined #kolla | 15:26 | |
| inc0 | yeah regex looks good | 15:26 |
| inc0 | what was the error> | 15:26 |
| *** salv-orl_ has quit IRC | 15:27 | |
| inc0 | maybe it's too inclusive? | 15:27 |
| *** ChanServ changes topic to "Kolla IRC meetngs on Wednesdays @ 16:30 UTC even weeks, 23:00 UTC odd weeks - see agenda @ https://goo.gl/OXB0DL - IRC channel is *LOGGED* @ http://goo.gl/VKpPzA" | 15:28 | |
| -openstackstatus- NOTICE: Launchpad SSO is back to normal - happy hacking | 15:28 | |
| sbezverk | sdake SamYaple iscsi initiatorname get created when container is getting built. in case of a multinode deployment, if the same container is used, all nodes will have the same initiatorname and it is not good. WOuld you be ok to modify initiator name during the deployment to make it unique? | 15:30 |
| sdake | my point is if we had 20 jobs it wouldnt matter to infra - zuul manages the gear via nodepool and het gating | 15:30 |
| sdake | sbezverk damn dude stop asking or permission | 15:31 |
| sdake | :) | 15:31 |
| sdake | just do it | 15:31 |
| sdake | and we will review and bitch appropriately tehre :) | 15:31 |
| sbezverk | sdake I just do not want to do something and then you tell me it is not good and I have to re-do it | 15:31 |
| sdake | sbezverk that is a normal part of open source | 15:31 |
| sdake | what isn't normal is asking permission for everything ;) | 15:32 |
| sdake | in 90% of cases where you ask for permission it doesn't matter | 15:32 |
| sdake | asking for advice is a different thing of course ;) | 15:32 |
| sdake | so if you want advice would lit be ok to modify the initator name during deployment - i think that is fine | 15:33 |
| sdake | but its hard to tell without seeing the implementation if its actually fine | 15:33 |
| sdake | hence asking permission in the other 10% of cases is wasted effort - the work could still require rework :) | 15:33 |
| *** jtriley has joined #kolla | 15:36 | |
| *** iceyao has quit IRC | 15:44 | |
| *** absubram has joined #kolla | 15:50 | |
| *** Jeffrey4l has quit IRC | 15:50 | |
| sdake | inc0 rabbitmq coming along? | 15:56 |
| *** salv-orlando has joined #kolla | 15:57 | |
| *** Serlex has quit IRC | 15:58 | |
| sdake | vhosakot are you working on fixing the tar thing with kibana an es? | 16:08 |
| vhosakot | sdake: yes | 16:08 |
| sdake | can you file a bug plz | 16:08 |
| sdake | so i can add it othe mielstone tracker | 16:09 |
| sdake | is it kibana-4? | 16:09 |
| *** jasonsb has quit IRC | 16:09 | |
| vhosakot | sdake: bug already there - https://bugs.launchpad.net/kolla/+bug/1553203 | 16:09 |
| openstack | Launchpad bug 1553203 in kolla "Ubuntu Container implementation missing kibana" [Critical,In progress] - Assigned to Vikram Hosakote (vhosakot) | 16:09 |
| vhosakot | sdake: and, https://bugs.launchpad.net/kolla/+bug/1554934 | 16:10 |
| openstack | Launchpad bug 1554934 in kolla "CentOS Container implementation missing kibana" [Critical,Confirmed] - Assigned to Vikram Hosakote (vhosakot) | 16:10 |
| *** jasonsb has joined #kolla | 16:10 | |
| sdake | ok changed it to high | 16:10 |
| vhosakot | cool | 16:10 |
| *** jasonsb has quit IRC | 16:10 | |
| inc0 | sdake, SamYaple we need to figure out these hostnames for rabbit | 16:12 |
| sdake | inc0 need more details pzl expand | 16:12 |
| inc0 | so we don't expect people to have hostnames in for nodes now | 16:13 |
| inc0 | and yet, we use hostnames in rabbitmq configs and such | 16:13 |
| inc0 | not IPS | 16:13 |
| inc0 | IPs | 16:13 |
| inc0 | my upgrade requires hostname | 16:13 |
| inc0 | Sam already turned ceph to IP, but rabbitmq is still hostname based | 16:13 |
| sdake | can we make rabbitmq ip based? | 16:14 |
| sdake | I'd like the whole infrastructure ip'able if possible ;) | 16:14 |
| inc0 | sure, just that's meddling with initial deploy | 16:14 |
| inc0 | and we're pretty late | 16:14 |
| sdake | we are good on time at our current volicty | 16:14 |
| sdake | velocity | 16:14 |
| inc0 | so we need to decide if we move rabbitmq to ip this release, then I need to change upgrades to IP | 16:15 |
| inc0 | simple change | 16:15 |
| sdake | i think if we are goin to bite the bullet , we need to do it now | 16:15 |
| sdake | later will cause all kinds of havok i suspect | 16:15 |
| inc0 | I agree, but I can't do it myself, I'm overloaded currently | 16:15 |
| SamYaple | inc0: i suggest ips otherwise hostnames are requirement | 16:15 |
| sdake | vhosakot do you have any bandwidth to help inc0? | 16:16 |
| SamYaple | im moving mariadb | 16:16 |
| SamYaple | i alreayd moved ceph | 16:16 |
| SamYaple | thats all of them i believe | 16:16 |
| inc0 | SamYaple, I agree with IP, but we need to do the work | 16:16 |
| SamYaple | if your regex is right then its no trouble | 16:16 |
| sdake | april 8th is when we ship this ball of fun :) | 16:16 |
| inc0 | SamYaple, regex is hostname based but it's easy change | 16:17 |
| SamYaple | sdake: we have to wait for all other projects to tag first | 16:17 |
| SamYaple | i doubt we can shit april 8th | 16:17 |
| sdake | SamYaple understood | 16:17 |
| SamYaple | shop**** | 16:17 |
| sdake | shit :) | 16:17 |
| SamYaple | ship* | 16:17 |
| inc0 | and play needs to use ip rather than hostname | 16:17 |
| SamYaple | ugh | 16:17 |
| SamYaple | inc0: it shouldnt be based on any of that. it should be based on structure of output | 16:17 |
| SamYaple | it _shouldnt_ have to change for each | 16:17 |
| inc0 | SamYaple, depends, if it's too inclusive it's error-prone | 16:17 |
| inc0 | anyway, I'll make it work for both, not problem | 16:18 |
| inc0 | but it will be hostname and ip, not any structure | 16:18 |
| SamYaple | if its based on the pinned output it shouldbe error proned | 16:18 |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Add missing tar command for oraclelinux binary https://review.openstack.org/293027 | 16:18 |
| SamYaple | it should be whatever rabbitmq has | 16:18 |
| inc0 | yeah, so now hostname | 16:18 |
| SamYaple | rabbitmq does the validattion not us | 16:18 |
| SamYaple | so it should be whatever it returns, we shouldnt be filtering | 16:18 |
| SamYaple | just using | 16:18 |
| inc0 | sdake, look at email, our little panel looks better now | 16:19 |
| inc0 | SamYaple, but error prone on our side | 16:19 |
| SamYaple | inc0: how so? | 16:19 |
| inc0 | if it takes too much, it's not good | 16:19 |
| inc0 | anyway, I'll make it work for both ip and hostname | 16:20 |
| inc0 | there is no other option | 16:20 |
| inc0 | and actually I'd use just ip | 16:20 |
| inc0 | as we will use it later on to compare to IP | 16:20 |
| SamYaple | inc0: rabbitmqctl eval 'rabbit_clusterer:status().' | awk -F'[@}]' '/gospel/ {print $2}' | 16:20 |
| SamYaple | at some point we want to have hostname OR ip stuff, and that measn it should work for both | 16:21 |
| SamYaple | but i trust you | 16:21 |
| SamYaple | you know what youre doig | 16:21 |
| inc0 | yup | 16:21 |
| inc0 | I'll make it regex based | 16:21 |
| vhosakot | sdake: inc0: yeah, I can help... rabbit upgrade ? I reviewed the code and aware of the getent issue... so, how can I help ? | 16:22 |
| *** tzn has quit IRC | 16:23 | |
| inc0 | vhosakot, it's rather issue of rabbit deploy | 16:23 |
| inc0 | in rabbitmq-clusterer config we use hostnames | 16:23 |
| *** mbound has quit IRC | 16:23 | |
| inc0 | we need to use IP | 16:23 |
| sdake | SamYaple one option is to convert hostnaames to ips on the fly | 16:24 |
| vhosakot | inc0: does rabbit allow that (IP instead of hostname in rabbitmq-clusterer.config) ? or, we need to check ? | 16:24 |
| inc0 | vhosakot, https://github.com/openstack/kolla/blob/master/ansible/roles/rabbitmq/templates/rabbitmq_clusterer.config.j2#L4 this one | 16:24 |
| *** tzn has joined #kolla | 16:24 | |
| inc0 | sdake, no-good, if we have hostname resolution we can use hostname | 16:24 |
| inc0 | vhosakot, it should but we need to confirm ofc | 16:25 |
| sdake | inc0 ok - this just goes bck to needing both | 16:25 |
| sdake | which I agree, we should be able to specify a hostname and have the thing work ;) | 16:25 |
| SamYaple | this storage_interface/tunnel/etc and hostname stuff can be worked out in newton | 16:25 |
| *** dwalsh has joined #kolla | 16:25 | |
| SamYaple | for now its are we requiring resolution or not, and the answer was "not" | 16:25 |
| sdake | ya a clear decision wasn't relaly madeon whether we require resolution or not pretty much until today | 16:26 |
| sdake | but I think eveyrone is on the same page, we don't want to mandate dns resolution because many labs don't have such things | 16:26 |
| SamYaple | we can get it optional in newton, i have some ideas | 16:26 |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Increase the docker parition for rht bases in gate https://review.openstack.org/293031 | 16:27 |
| sdake | i think having resolution on one node via /etchosts is better then having resollution on al nodes via /etc/hosts file | 16:27 |
| vhosakot | inc0: I see IP is used in rabbitmq.config.j2 as "{{ hostvars[inventory_hostname]['ansible_' + api_interface]['ipv4']['address'] }}"... is this about using a similar call in rabbitmq_clusterer.config.j2 ? | 16:27 |
| sdake | during deploy do the resolution if there are names | 16:27 |
| vhosakot | inc0: plus it is a loop over hosts | 16:27 |
| openstackgerrit | Paul Bourke proposed openstack/kolla: Increase the docker parition for rht bases in gate https://review.openstack.org/293031 | 16:28 |
| sdake | pbourke why not ust convert to using overlayfs | 16:28 |
| sdake | then you wont have these problems at all :) | 16:29 |
| sdake | and we get one more data point about overlayfs vs btrfs | 16:29 |
| pbourke | sdake: what kind of saving do we get? | 16:29 |
| pbourke | space wise | 16:29 |
| *** tzn has quit IRC | 16:29 | |
| sdake | a whole bunch | 16:29 |
| sdake | moment i'll prodce #s | 16:29 |
| vhosakot | pbourke: FYI, let me know if you want to make docker daemon use a different dir with bigger space (other than /var/lib/docker)... | 16:30 |
| vhosakot | inc0: I will dig more about the rabbit config and let you know | 16:30 |
| inc0 | vhosakot, thanks | 16:30 |
| pbourke | vhosakot: thanks but I dont think that's necessary as we control the loop device creation that backs /var/lib/docker | 16:30 |
| *** mikelk has quit IRC | 16:31 | |
| SamYaple | pbourke: none space savings | 16:31 |
| vhosakot | ah ok ok | 16:31 |
| SamYaple | overlayfs is not smaller and you cant switch to it | 16:31 |
| vhosakot | zfs :) | 16:31 |
| SamYaple | lol vhosakot also no | 16:31 |
| *** sdake_ has joined #kolla | 16:31 | |
| SamYaple | zfs is actually the slower one on linux as far as backend goes | 16:31 |
| vhosakot | why, gate does not have zfs ? docker supports zfs storage driver | 16:32 |
| vhosakot | oh ok | 16:32 |
| sdake_ | pbourke dfof /var/lib/docker | 16:32 |
| sdake_ | 13141404. | 16:32 |
| SamYaple | vhosakot: zfs is not in the linux kernel | 16:32 |
| sdake_ | 13gb | 16:32 |
| sdake_ | for one distro | 16:32 |
| sdake_ | from source | 16:32 |
| vhosakot | yep | 16:32 |
| sdake_ | (centos from source) | 16:32 |
| SamYaple | you are not calculating that correct sdake | 16:32 |
| SamYaple | shared files duping space | 16:32 |
| sdake_ | how shall i calculate it | 16:32 |
| SamYaple | there is no good way to (no user space tools exist) | 16:33 |
| sdake_ | how did pbourke calculate the 80gb? | 16:33 |
| pbourke | SamYaple: I thought if you have a clean /var/lib/docker and do a build then the size of that mount is the real size of the images | 16:33 |
| sdake_ | pbourke it depends on the storage backend driver | 16:33 |
| SamYaple | pbourke: it _is_ but you cant use du, you have to use df | 16:34 |
| SamYaple | we do have the output of that info in every gate run btw | 16:34 |
| sdake_ | bummer my /var/lib/docker is not in df | 16:34 |
| *** sdake has quit IRC | 16:34 | |
| sdake_ | is not in a specific partition | 16:34 |
| vhosakot | pbourke: I checked that... summation of all my images in "docker images" is around 64GB, but my /var/lib/docker is just ~9GB | 16:34 |
| pbourke | vhosakot: the size in docker images is not useful | 16:35 |
| SamYaple | pbourke: according to our gates going back a while, centos source is consuming 6.5GB with 2GB+ to spare | 16:35 |
| sdake_ | pbourke when you build oracle linux, how large is /var/lib/docker via du? | 16:35 |
| pbourke | vhosakot: due to layer sharing. 9GB sounds about right | 16:35 |
| *** Marga_ has joined #kolla | 16:35 | |
| vhosakot | cool | 16:35 |
| pbourke | my numbers are on the ML | 16:35 |
| sdake_ | we only ahve 8gb, I thought 80gb | 16:35 |
| pbourke | 10.01GB for binary oraclelinux, 8.89GB for binary centos | 16:36 |
| pbourke | source is probably smaller as SamYaple says, I haven't checked those | 16:36 |
| sdake_ | pbourke via df? | 16:37 |
| pbourke | yes | 16:37 |
| SamYaple | pbourke: you are right we are chewing ~8.6GB for centos binary | 16:38 |
| SamYaple | i checked going back a few | 16:39 |
| sdake_ | let me build centos binary in a overlayfs | 16:39 |
| SamYaple | i commented in your review | 16:39 |
| sdake_ | see if there is a difference | 16:39 |
| SamYaple | sdake_: since you didnt accurately measure btrfs i dont know how youll compare | 16:39 |
| sdake_ | the gate measures btrfs right? | 16:39 |
| pbourke | SamYaple: I forgot to put this link in the msg http://docs.openstack.org/infra/system-config/contribute-cloud.html | 16:39 |
| pbourke | SamYaple: so I guess what you're saying is the above is incorrect | 16:40 |
| pbourke | that's why I posted to the ML to try and get this kind of info | 16:40 |
| SamYaple | pbourke: if we can be assured they all have 80GB (and i think rax doesnt on the root disk) then im fine | 16:40 |
| SamYaple | its just about testing and probing the gates at this point | 16:40 |
| SamYaple | sdake_: yes but your userspace tools wont help you with overlayfs either | 16:41 |
| pbourke | i'll do a handle of check experimentals and see if we can get a good selection of gate output | 16:41 |
| sdake_ | df wont help? | 16:41 |
| pbourke | *handful | 16:41 |
| SamYaple | sdake_: you said it wasnt on a seperate partition | 16:41 |
| sdake_ | I've got a partition now for /var/lib/docker | 16:41 |
| sdake_ | i just made one | 16:41 |
| sdake_ | going to build overlay now | 16:41 |
| SamYaple | should be about the same then yes | 16:41 |
| pbourke | wouldn't life be grand if we just supported one base image type | 16:41 |
| SamYaple | no then you would have to use ubuntu | 16:42 |
| pbourke | kolla OS anyone? | 16:42 |
| SamYaple | that would suck for you | 16:42 |
| sdake_ | pbourke we support multiple bsae imagews because ubuntu is defective ;) | 16:42 |
| SamYaple | pbourke: you joke, but i love the idea of using arch or alpinelinux or something to do an all from source kolla deploy | 16:42 |
| SamYaple | and yet rdo is never going to be a voting gate | 16:43 |
| SamYaple | but its ubuntu thats defective.... | 16:43 |
| SamYaple | the one that builds _all_ containers right now, and is the only one to do so | 16:43 |
| sdake_ | centos source could be a voting gate | 16:44 |
| sdake_ | anywy SamYaple you missed the joke | 16:44 |
| SamYaple | sdake_: i got it :P | 16:44 |
| SamYaple | sdake_: no it wont. it still has rdo stuff | 16:44 |
| SamYaple | if you remove the rdo stuff, sure | 16:44 |
| *** akwasnie has quit IRC | 16:47 | |
| *** tzn has joined #kolla | 16:47 | |
| sdake_ | SamYaple the only way we get any gate voting is by mirroring | 16:49 |
| sdake_ | which means any distro to vote needs to mirror all the things | 16:49 |
| sdake_ | i am hesitent to make rdo voting but centos shoudl be fine | 16:49 |
| sdake_ | centos source | 16:49 |
| SamYaple | centos source uses rdo mirrors which are not infra controlled | 16:50 |
| pbourke | SamYaple: whats holding us back from gating ubuntu source atm? | 16:52 |
| *** unicell has quit IRC | 16:52 | |
| SamYaple | pbourke: some non-infra things needed | 16:52 |
| pbourke | ah yes | 16:52 |
| SamYaple | so the external deps | 16:52 |
| pbourke | but... will infra ever mirror those deps? | 16:53 |
| SamYaple | once infra mirrors some repos outside, we will be solid (minus the curls that we do..., so maybe only deploy gates until thats fixed | 16:53 |
| SamYaple | pbourke: i got a solid yes out of clarkb, but i havent done the legwork yet | 16:53 |
| SamYaple | probably wont until april | 16:53 |
| SamYaple | its complicated | 16:53 |
| *** tzn has quit IRC | 16:54 | |
| sdake_ | pbourke the word from infra is yes, they will mirror what we need | 16:54 |
| pbourke | good to know | 16:54 |
| sdake_ | pbourke they are doing it in a global order that is best for everyone | 16:54 |
| sdake_ | not necessarily best for kolla | 16:55 |
| sdake_ | so the global order next up is centos + epel | 16:55 |
| sdake_ | we are free to contribute our time to make it go faster | 16:55 |
| sdake_ | but none of us i think has the expertise to add mirrors and whatnot | 16:55 |
| sdake_ | from what I can tell its not self-service | 16:55 |
| *** mbound has joined #kolla | 16:55 | |
| sdake_ | or requires intense domain expertise | 16:56 |
| sdake_ | pbourke still building | 17:06 |
| pbourke | thanks sdake_ | 17:06 |
| openstackgerrit | Thiago Gomes proposed openstack/kolla: [WIP] Add Manila deployment document https://review.openstack.org/292588 | 17:06 |
| pbourke | keep in mind SamYaple had some reason we need to use btrfs.. | 17:06 |
| pbourke | maybe he will elaborate :) | 17:06 |
| *** Slower has joined #kolla | 17:08 | |
| *** openstackgerrit has quit IRC | 17:14 | |
| *** openstackgerrit_ is now known as openstackgerrit | 17:14 | |
| *** openstackgerrit has quit IRC | 17:15 | |
| *** openstackgerrit_ has joined #kolla | 17:15 | |
| SamYaple | sdake_: infra isnt doing some order for mirrors | 17:15 |
| SamYaple | i can add the ubuntu mirrors right now | 17:15 |
| SamYaple | centos + epel is.... well its not happening right this moments | 17:15 |
| *** openstackgerrit_ is now known as openstackgerrit | 17:16 | |
| SamYaple | apparently its difficult | 17:16 |
| SamYaple | we arent actually waiting on infra for our package mirrors for ubuntu | 17:16 |
| *** openstackgerrit_ has joined #kolla | 17:16 | |
| *** unicell has joined #kolla | 17:19 | |
| *** spisarski has quit IRC | 17:32 | |
| *** dwalsh has quit IRC | 17:34 | |
| *** prithiv has quit IRC | 17:36 | |
| *** jmccarthy1 has quit IRC | 17:47 | |
| *** salv-orl_ has joined #kolla | 17:54 | |
| *** salv-orlando has quit IRC | 17:56 | |
| sdake_ | pbourke /dev/sda1 3845557232 10910188 3639279828 1% /var/lib/docker | 17:58 |
| sdake_ | that is with overlayfs | 17:58 |
| sdake_ | Filesystem 1K-blocks Used Available Use% Mounted on | 17:58 |
| sdake_ | 10.9GB | 17:58 |
| *** spisarski has joined #kolla | 18:02 | |
| *** inc0 has quit IRC | 18:02 | |
| *** dims_ has joined #kolla | 18:03 | |
| *** dims has quit IRC | 18:03 | |
| openstackgerrit | Thiago Gomes proposed openstack/kolla: [WIP] Add Manila deployment document https://review.openstack.org/292588 | 18:05 |
| openstackgerrit | Andrey Pavlov proposed openstack/kolla-mesos: CLI: implement commands list command https://review.openstack.org/293081 | 18:20 |
| *** dims_ has quit IRC | 18:29 | |
| *** sdake_ has quit IRC | 18:34 | |
| vhosakot | SamYaple: per the comments I received, I will combine both CentOS and Ubuntu packages for Kibana and dupe one bug to the other.. does this sound good ? | 18:36 |
| SamYaple | vhosakot: yes please | 18:36 |
| vhosakot | SamYaple: and will submit a single patch set so it is eady to review | 18:36 |
| vhosakot | SamYaple: cool, thanks | 18:36 |
| *** inc0 has joined #kolla | 18:50 | |
| *** vhosakot has quit IRC | 18:50 | |
| openstackgerrit | Merged openstack/kolla-mesos: CLI: Implement service definition https://review.openstack.org/286992 | 18:57 |
| *** vhosakot has joined #kolla | 18:59 | |
| *** sdake has joined #kolla | 19:07 | |
| *** fthiagogv has quit IRC | 19:07 | |
| *** dims has joined #kolla | 19:12 | |
| *** dims has quit IRC | 19:18 | |
| *** sdake_ has joined #kolla | 19:21 | |
| *** sdake has quit IRC | 19:23 | |
| *** dims has joined #kolla | 19:26 | |
| sbezverk | SamYaple, I saw your comment, unfortunately to make lvm working in container, these two setting must be turned off, otherwise lvm is no go.. | 19:31 |
| SamYaple | sbezverk: you are wrong | 19:32 |
| *** athomas has quit IRC | 19:32 | |
| SamYaple | but either way, they are dangeous and wont be enabled | 19:32 |
| SamYaple | it needs udev info, remember the mounts /run/udev (and maybe /lib/udev) | 19:32 |
| SamYaple | as well as /dev/ of course | 19:32 |
| *** dwalsh has joined #kolla | 19:38 | |
| *** athomas has joined #kolla | 19:40 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Change libvirt volume name to libvirtd https://review.openstack.org/293106 | 19:41 |
| SamYaple | inc0: ^ | 19:41 |
| *** salv-orl_ has quit IRC | 19:43 | |
| *** dims has quit IRC | 19:45 | |
| inc0 | SamYaple, do we change nova_compute volume as well? | 19:45 |
| inc0 | name | 19:45 |
| inc0 | because it's changing container;) | 19:45 |
| sbezverk | I am trying to provide a service config during the deployment, but I am getting this error: fatal: [192.168.80.25] => {'msg': "File contains no section headers.\nfile: <???>, line: 1\n'include /var/lib/cinder/volumes/*\\n'", 'failed': True} I suspect I am missing so formating, any pointers what installed expects in here? | 19:49 |
| vhosakot | sbezverk: merge_config not working for non-ini files.. known issue | 19:49 |
| sbezverk | vhosakot what are my options then? | 19:50 |
| vhosakot | sbezverk: "File contains no section headers" means the file does not have [section] | 19:50 |
| vhosakot | blah=value | 19:50 |
| vhosakot | sbezverk: well, you want reconfigure ? or new config in container | 19:51 |
| vhosakot | sbezverk: I just update non-ini config file manually and restart the container and check the logs | 19:51 |
| *** sdake_ has quit IRC | 19:51 | |
| vhosakot | sbezverk: which file is this ? can you pastebin ? | 19:51 |
| sbezverk | vhosakot I need to add 1 line to tgtd.conf in iscsi container | 19:51 |
| vhosakot | can you pastebin the file | 19:52 |
| sbezverk | this line is required for tgtd to use cinder_volume generated targets | 19:52 |
| sbezverk | which file? | 19:52 |
| vhosakot | sbezverk: now, the workaround is... run the iscsi container without your chance in tgtd.conf, then, change the file inside using "docker exec iscsi ..." and restart the container | 19:53 |
| vhosakot | change* | 19:53 |
| *** tummy has joined #kolla | 19:53 | |
| sbezverk | vhosakot man, I would not eve dare to think of a workaround like that, I thought I would be shot on site | 19:53 |
| vhosakot | sbezverk: yep, there will be a solution for non-ini file soon.. we can talk in tomorrow's meeting as well... I have started some implementation... fully not working yet | 19:54 |
| vhosakot | sbezverk: but the workaround worked for me for rabbit and mariadb fine | 19:54 |
| sbezverk | vhosakot another way I thought would be to change this file when the container gets built, I see you guys use it all the time.. | 19:55 |
| sbezverk | but I will wait.. | 19:55 |
| vhosakot | yes, "when the container gets built" is reconfiguring the container's non-ini file without going inside the container... cool, I will keep you posted the issue | 19:56 |
| *** zaneb has joined #kolla | 19:58 | |
| *** haplo37 has quit IRC | 20:08 | |
| *** inc0 has quit IRC | 20:08 | |
| *** dims has joined #kolla | 20:11 | |
| *** sdake has joined #kolla | 20:11 | |
| *** prithiv has joined #kolla | 20:16 | |
| *** Marga_ has quit IRC | 20:19 | |
| *** tummy has quit IRC | 20:23 | |
| *** jtriley has quit IRC | 20:30 | |
| *** vhosakot has quit IRC | 20:33 | |
| *** vhosakot has joined #kolla | 20:34 | |
| *** salv-orlando has joined #kolla | 20:37 | |
| *** tummy has joined #kolla | 20:41 | |
| *** tummy has quit IRC | 20:44 | |
| *** sdake has quit IRC | 20:50 | |
| *** SamYaple is now known as NotSamYaple | 20:52 | |
| *** sdake has joined #kolla | 20:53 | |
| *** NotSamYaple is now known as SamYaple | 20:53 | |
| *** fthiagogv has joined #kolla | 20:53 | |
| fthiagogv | sdake, hi. Should all clients be installed in quick start guide? | 20:55 |
| sdake | i think if its clear that is meant for the client node | 20:56 |
| sdake | rather then the deployment targets or deployment node | 20:56 |
| sdake | then that would be fine | 20:56 |
| sdake | fthiagogv make the client install a separate patch | 20:57 |
| sdake | so it can be evaluted separtely | 20:57 |
| sdake | i dont like big changes to docs in one go | 20:58 |
| sdake | i like to be surgical with docs, since they are pretty much the most important thing the oeprator has to go with with kolla | 20:58 |
| *** sdake_ has joined #kolla | 21:02 | |
| *** rajathagasthya has joined #kolla | 21:02 | |
| *** sdake has quit IRC | 21:05 | |
| *** shardy has quit IRC | 21:10 | |
| sdake_ | hey SamYaple i have one between the lines left from that box i bought in greenville | 21:10 |
| *** dwalsh has quit IRC | 21:13 | |
| openstackgerrit | Thiago Gomes proposed openstack/kolla: Add Manila deployment document https://review.openstack.org/292588 | 21:20 |
| SamYaple | sdake_: what now? | 21:21 |
| sdake_ | oh i bought a box of super rare cigars in greenville | 21:21 |
| sdake_ | smoking one now | 21:21 |
| sdake_ | and have 1 left | 21:21 |
| sdake_ | btww i tried to send my address to you but your email is not functional forsome rason | 21:22 |
| sdake_ | it bounced | 21:22 |
| *** sdake_ is now known as sdake | 21:23 | |
| SamYaple | email is working fine here. just got mail. try it again? | 21:24 |
| sdake | lemme know if you got that | 21:24 |
| SamYaple | yup | 21:25 |
| *** inc0 has joined #kolla | 21:29 | |
| *** ccesario has joined #kolla | 21:32 | |
| *** ccesario has quit IRC | 21:34 | |
| *** ccesario has joined #kolla | 21:34 | |
| *** SiRiuS_ has quit IRC | 21:35 | |
| *** ccesario has quit IRC | 21:35 | |
| *** AndChat|241401 has joined #kolla | 21:35 | |
| *** rajathagasthya has quit IRC | 21:36 | |
| *** ccesario has joined #kolla | 21:55 | |
| *** Marga_ has joined #kolla | 21:56 | |
| *** fthiagogv has quit IRC | 21:59 | |
| *** AndChat|241401 has quit IRC | 22:00 | |
| *** kbyrne has quit IRC | 22:01 | |
| *** alisonh has quit IRC | 22:04 | |
| *** sdake has quit IRC | 22:09 | |
| *** inc0 has quit IRC | 22:10 | |
| *** dims_ has joined #kolla | 22:11 | |
| *** jtriley has joined #kolla | 22:12 | |
| *** dims has quit IRC | 22:14 | |
| *** dims has joined #kolla | 22:16 | |
| *** dims_ has quit IRC | 22:17 | |
| *** absubram has quit IRC | 22:21 | |
| *** alisonh has joined #kolla | 22:23 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: MariaDB lights out recovery https://review.openstack.org/293161 | 22:26 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Mariadb upgrade https://review.openstack.org/293162 | 22:26 |
| *** jtriley has quit IRC | 22:28 | |
| *** jtriley has joined #kolla | 22:31 | |
| *** shakamunyi has joined #kolla | 22:32 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Mariadb upgrade https://review.openstack.org/293162 | 22:33 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: MariaDB lights out recovery https://review.openstack.org/293161 | 22:33 |
| *** jtriley has quit IRC | 22:36 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Mariadb upgrade https://review.openstack.org/293162 | 22:37 |
| openstackgerrit | Sam Yaple proposed openstack/kolla: MariaDB lights out recovery https://review.openstack.org/293161 | 22:37 |
| SamYaple | there ugh. hate when patches get mixed together | 22:37 |
| *** jtriley has joined #kolla | 22:37 | |
| *** prithiv has quit IRC | 22:40 | |
| openstackgerrit | Sam Yaple proposed openstack/kolla: Change libvirt volume name to libvirtd https://review.openstack.org/293106 | 22:41 |
| *** jtriley has quit IRC | 22:49 | |
| *** alisonh has quit IRC | 22:49 | |
| *** mbound has quit IRC | 22:50 | |
| *** alisonh has joined #kolla | 22:53 | |
| *** dave-mccowan has quit IRC | 22:54 | |
| *** jtriley has joined #kolla | 22:56 | |
| *** blahRus has quit IRC | 23:00 | |
| *** dims has quit IRC | 23:06 | |
| *** ssurana1 has joined #kolla | 23:08 | |
| *** ssurana has quit IRC | 23:11 | |
| *** jtriley has quit IRC | 23:14 | |
| *** sdake has joined #kolla | 23:45 | |
| *** salv-orlando has quit IRC | 23:49 | |
| *** jasonsb has joined #kolla | 23:50 | |
| *** mbound has joined #kolla | 23:51 | |
| *** vhosakot has quit IRC | 23:54 | |
| *** mbound has quit IRC | 23:56 | |
| *** jasonsb has quit IRC | 23:59 | |
| *** sdake_ has joined #kolla | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
