Thursday, 2017-07-27

*** wanghao has joined #openstack-mogan00:50
wanghaomorning!00:59
*** wanghao has quit IRC01:00
*** wanghao has joined #openstack-mogan01:00
liushengmorning01:03
openstackgerritliusheng proposed openstack/mogan master: Improve resources updating process  https://review.openstack.org/48735101:03
*** wanghao_ has joined #openstack-mogan01:04
zhenguomorning!01:06
liushengzhenguo: hi, I just updated the patch ^, add a condition check :D01:07
*** wanghao has quit IRC01:07
zhenguoliusheng: ok01:07
zhenguoliusheng: I added a patch to bump the placement api version to latest if not specified to be able to use aggregate01:08
liushengzhenguo: seems currently, the version can be inputted when call an interface01:10
zhenguoliusheng: yes,01:11
zhenguoliusheng: but I would perfer to add a default API version01:12
zhenguoliusheng: instead of make different version of interfaces01:12
liushengzhenguo: I don't prever to set the default version to latest, because latest is unfixed, it will be increased by placement changes, maybe the latest version is not stable in the future01:12
liushengzhenguo: that is ok01:12
liushengzhenguo: what do you think about my question ?01:12
liushengzhenguo: how about just set the default version to currect biggest one ?01:13
zhenguoliusheng: yes, we can change to a fixed version not latest01:13
liushengzhenguo: ok01:13
zhenguoliusheng: let me check what the current latest version is01:14
liushengzhenguo: 1.1001:15
liushengzhenguo: nova/api/openstack/placement/rest_api_version_history.rst01:15
zhenguoliusheng: ok, thanks01:15
*** harlowja has quit IRC01:19
openstackgerritZhenguo Niu proposed openstack/mogan master: Add default placement api version  https://review.openstack.org/48737201:21
openstackgerritMerged openstack/mogan master: Updated from global requirements  https://review.openstack.org/48747001:38
openstackgerritMerged openstack/mogan master: Remove support for py3.3  https://review.openstack.org/48732101:58
wanghao_meeting?01:59
zhenguoyeah02:00
wanghao_zhenguo: Do we need to decide the tasks which move to next cycle?02:52
liushengwanghao_: hi, do you have the eterpad link for chinese presentation ?02:52
wanghao_liusheng: https://etherpad.openstack.org/p/china-sydney-summit-sessions02:53
liushengwanghao_: thanks a lot02:53
wanghao_liusheng: np02:54
wanghao_zhenguo: Seems we forgot this topic at meeting....02:54
zhenguowanghao_: we can decide which things must be done in Pike :D03:04
wanghao_zhenguo: I think aggregates, adopt existing node,  quota for keypairs could be don in Pike03:05
zhenguowanghao_: yes, and we will try to add server groups as well03:05
zhenguowanghao_: as it's just based on aggregate03:06
wanghao_zhenguo: yeah,  it also be here03:06
wanghao_zhenguo: do we can merge thing until end of august?03:06
zhenguowanghao_: I think so, hah03:06
wanghao_zhenguo: since we didn't follow the offical sceduler of Pike03:07
wanghao_hah03:07
zhenguowanghao_: yes, we just need one release that named pike03:07
wanghao_zhenguo: en, sure03:07
*** harlowja has joined #openstack-mogan03:31
zhenguoliusheng: do you know why we cache the rps and rp_agg_map?03:53
liushengzhenguo: seems in Nova it can reduce the unnecessary request for rps and rp_aggs03:54
zhenguoliusheng: ok, maybe we need to refactor it03:55
liushengzhenguo: ok03:56
zhenguoliusheng: but now I will just follow the cache implementation, refactoring can be hanppend later03:56
* zhenguo brb03:57
openstackgerritTao Li proposed openstack/mogan master: Manage existing BMs: Part-2  https://review.openstack.org/48154404:08
*** wanghao has joined #openstack-mogan04:12
*** wanghao_ has quit IRC04:13
*** harlowja has quit IRC04:52
openstackgerritZhenguo Niu proposed openstack/mogan master: [WIP] Add aggregate nodes API  https://review.openstack.org/48728405:26
openstackgerritZhenguo Niu proposed openstack/mogan master: Clean up some unused codes  https://review.openstack.org/48770105:43
*** bani has joined #openstack-mogan05:50
*** litao__ has joined #openstack-mogan05:53
openstackgerritTao Li proposed openstack/mogan master: Fix the stuck in *ing in powering servers.  https://review.openstack.org/47686206:27
zhenguowanghao: maybe we should also add cinder integration before Pike06:31
wanghaozhenguo: en,  introduce the cinder client and support to create and attach volume for node?06:32
zhenguowanghao: only support boot from volume like nova06:32
zhenguowanghao: as Ironic will basically get boot-from-volume support, that's a cool thing if we can also land06:32
wanghaozhenguo: en sure, that could be done06:33
zhenguowanghao: seems only need a few API06:33
zhenguowanghao: the new V3 is enough for us06:33
wanghaozhenguo: yeah, I aslo suggest to use v3 api in cinder06:34
zhenguowanghao: cool06:34
wanghaozhenguo: since cinder can support create volume,  create volume from image , snapshot, backup, volume.06:35
wanghaozhenguo: so I think we may just support to create volume from image first.06:35
zhenguowanghao: yeah, it's ok, we only have image now06:36
wanghaozhenguo: since that will create a bootable volume in cinder06:36
zhenguowanghao: no snapshot support yet06:36
wanghaozhenguo: sure06:36
openstackgerritTao Li proposed openstack/mogan master: Rollback to the original status when server powering action failed  https://review.openstack.org/47686206:36
wanghaozhenguo: let'm checkt the ironic API to see the detail about boot-from-volume.06:36
zhenguowanghao: not sure if it's finished or not06:37
wanghaozhenguo: any review link?06:37
zhenguowanghao: but I find the task will be pushing in Pike, so it should be able to get landed06:37
zhenguowanghao: the workflow seems like this https://specs.openstack.org/openstack/ironic-specs/specs/not-implemented/volume-connection-information.html06:37
wanghaozhenguo: cool,  since we need to know the API format.06:38
zhenguowanghao: yes06:38
wanghaozhenguo: so maybe we implement it earyler than Ironic?06:38
wanghaozhenguo: never mind,  we can do the high layer coding. :)06:39
zhenguowanghao: yes, it's almost done in ironic06:39
wanghaozhenguo: okay06:40
zhenguoliusheng: do you know how to run py35 with tox now?06:44
liushengzhenguo: did't try, tox -epy35 ?06:45
zhenguoliusheng: seems can't work06:45
liushengzhenguo: let me try06:46
zhenguoliusheng: thanks06:46
* zhenguo brb06:47
openstackgerritTao Li proposed openstack/mogan master: Manage existing BMs: Part-2  https://review.openstack.org/48154406:47
openstackgerritTao Li proposed openstack/mogan master: Rollback to the original status when server powering action failed  https://review.openstack.org/47686206:50
zhenguowanghao: I remember you mentioned there should be a following up patch for gmr after uwsgi06:59
wanghaozhenguo: yes06:59
wanghaozhenguo: we have finished the uwsgi job?07:00
zhenguowanghao: seems yes, liusheng^^07:00
wanghaozhenguo: okay, great,  I will add it ASAP07:01
zhenguowanghao: ok, thanks07:01
zhenguoliusheng: just found, we still left some unused codes after moving to placement https://review.openstack.org/#/c/487701/07:03
wanghaozhenguo: I found the API has merged: https://review.openstack.org/#/c/214586/07:04
wanghaozhenguo: in Ironic for boot from volume07:04
zhenguowanghao: cool07:04
liushengzhenguo: thanks for cleanup, I remember I removed that... hah07:05
zhenguoliusheng: hah07:05
wanghaoguys,  is there any question for keypairs quota?  Need your review : )07:25
zhenguowanghao: will test it after my env is ready07:38
wanghaozhenguo: cool, thanks again!07:38
zhenguowanghao: hah, sorry for the delay07:38
openstackgerritwanghao proposed openstack/mogan master: Support Guru in uwsgi way  https://review.openstack.org/48773808:03
wanghaozhenguo: np :_08:10
wanghao:)08:10
wanghaoliusheng: ping08:10
wanghaoliusheng: hava a question for uwsgi08:11
liushengwanghao: pong08:11
wanghao2. On deb-based systems copy or symlink the file to08:11
wanghao   ``/etc/apache2/sites-available/mogan.conf``08:11
wanghaoyou mean which file?08:11
wanghaoI think it's not /etc/mogan/mogan.conf right?08:11
liushengzhenguo: sorry, seems there is a problem, let me check the doc08:17
wanghaoyou mean me, right?  hahaha08:20
openstackgerritliusheng proposed openstack/mogan master: Fix uwsgi doc to avoid misunderstanding  https://review.openstack.org/48775708:36
liushengwanghao: I just uploaded a patch to update the doc08:37
wanghaoliusheng: okay08:37
liushengzhenguo: sorry for misleading :(08:37
wanghaoliusheng: btw, do you meet this error: open("./python_plugin.so"): No such file or directory [core/utils.c line 3686]?08:38
wanghaoliusheng: I can't  start the uwsgi /etc/mogan/mogan-uwsgi.ini  since this error in my unbuntu08:38
liushengwanghao: no, the "pip install uwsgi" is ok ?08:39
wanghaoyes08:40
liushengwanghao: please try: uwsgi --ini /etc/mogan/mogan-uwsgi.ini08:42
wanghaoliusheng: I add this plugins-dir = /usr/lib/uwsgi/plugins08:42
wanghaoliusheng: that works08:43
wanghaoliusheng: seems you need to specify the plugins-dir in mogan-uwsgi.ini08:43
openstackgerritliusheng proposed openstack/mogan master: Fix uwsgi doc to avoid misunderstanding  https://review.openstack.org/48775708:43
liushengwanghao: you mean in mogan-uwsgi.ini ?08:44
wanghaoyes08:44
wanghaoand I also can't find: /var/run/uwsgi/mogan-api-wsgi.socket08:44
wanghaoso there is one more error: bind(): No such file or directory [core/socket.c line 230]08:45
liushengwanghao: the mogan-uwsgi.ini is genrated by devstack scripts, it can work in my env without plugins-dir08:45
wanghaoliusheng: I didn't use devstack,  but just follow your guide: 4. Copy mogan/etc/mogan-uwsgi.ini.sample to /etc/mogan/mogan-uwsgi.ini.08:46
wanghaoliusheng:  so I copy this file to /etc/mogan,  but it has some error.08:47
wanghaoliusheng: could you find /var/run/uwsgi/mogan-api-wsgi.socket?08:47
liushengwanghao: do you have config file under /etc/apache2/sites-available/ and /etc/apache2/sites-enable/ ?08:47
liushengzhenguo: and restart apache208:48
zhenguohah,08:48
liushengwanghao: and restart apache208:48
liushenghah, sent to wrong person08:48
wanghaoliusheng: wait a second08:49
zhenguosorry guys, I'm in a internal meeting08:49
wanghaoliusheng: could not restart apache2 sucessful08:52
wanghaoliusheng: AH00526: Syntax error on line 2 of /etc/apache2/sites-enabled/mogan.conf08:52
wanghaoProxyPass URL must be absolute!: unix:/var/run/uwsgi/mogan-api-wsgi.socket08:52
wanghaoliusheng: I can't find the mogan-api-wsgi.socket08:53
wanghaoSetEnv proxy-sendcl 108:55
wanghaoProxyPass "/baremetal_compute" "unix:/var/run/uwsgi/mogan-api-wsgi.socket" retry=008:55
wanghaoI copy those to mogan.conf, but it failed08:55
liushengwanghao: how about: ProxyPass "/baremetal_compute" "unix:/var/run/uwsgi/mogan-api-wsgi.socket|uwsgi://uwsgi-uds-mogan-api-wsgi/" retry=008:56
wanghaoliusheng: one line?08:58
wanghaoliusheng: it works08:59
liushengwanghao: the doc was copied from other project :(09:00
wanghao...09:00
wanghaoseems the last one is okay09:01
wanghaoliusheng: but still can't start uwsgi: uwsgi /etc/mogan/mogan-uwsgi.ini09:02
wanghaoit failed09:02
wanghaobind(): No such file or directory [core/socket.c line 230]09:02
liushengwanghao: no /var/run/uwsgi/mogan-api-wsgi.socket file ?09:03
openstackgerritliusheng proposed openstack/mogan master: Fix uwsgi doc to avoid misunderstanding  https://review.openstack.org/48775709:03
wanghaono09:03
liushengwanghao: that is strange, seems, the socket file can be created automatically09:08
wanghaohow?09:08
wanghaouwsgi --ini /etc/mogan/mogan-wsgi.ini ?09:08
wanghaoseems couldn't09:09
liushengwanghao: yes09:10
wanghaoGot it why09:10
wanghaoI didn't have /var/run/uwsgi09:10
wanghaoso it failed09:11
wanghaobetter to check it in our doc, I think09:11
liushengwanghao: I just tried in my env, it can autumatically generate a socket file09:13
wanghaoliusheng: yes, since you have /var/run/uwsgi09:13
wanghaoliusheng: I'm not why I didn't have,  maybe devstack will build it09:13
liushengwanghao: yes :(,it is created by devstack09:14
liushengzhenguo: how if you create the dir and chonw to stack.stack, then try ?09:14
wanghaoliusheng: but I think we'd better add a note in our doc,  since user may not use it by devstack,09:14
liushengwanghao: sure09:14
liushengzhenguo: will update the patch, thank for check this!09:15
wanghaoliusheng: np09:16
wanghaoliusheng: and guru for supporting this have tested, it works fine.09:17
wanghao========================================================================09:17
wanghao====                         Green Threads                          ====09:17
wanghao========================================================================09:17
liushengwanghao: ok, will review, thanks09:17
wanghaowe can get the guru report by touch a file under uwsgi09:17
wanghaoliusheng: sure09:17
openstackgerritliusheng proposed openstack/mogan master: Fix uwsgi doc to avoid misunderstanding  https://review.openstack.org/48775709:28
liushengwanghao: I have updated the patch :)09:30
wanghaoliusheng: sure09:31
*** wanghao has quit IRC09:32
*** wanghao has joined #openstack-mogan09:35
openstackgerritliusheng proposed openstack/mogan master: Fix uwsgi doc to avoid misunderstanding  https://review.openstack.org/48775709:54
*** wanghao has quit IRC09:55
*** wanghao has joined #openstack-mogan09:56
*** wanghao has quit IRC09:56
*** wanghao has joined #openstack-mogan09:57
zhenguoliusheng: seems it's not easy to delete aggregate, as the rp may have allocations09:58
liushengzhenguo: that is sad :(09:58
zhenguoliusheng: but we must provide that09:59
*** wanghao has quit IRC10:01
liushengzhenguo: but Nova now is FFE :(10:03
liushengzhenguo: s/FFE/FF10:03
zhenguoliusheng: seems we misused aggregates set, it will replace all aggregates of the rp10:20
openstackgerritZhenguo Niu proposed openstack/mogan master: Add aggregate nodes API  https://review.openstack.org/48728411:01
openstackgerritZhenguo Niu proposed openstack/mogan master: Add aggregate nodes API  https://review.openstack.org/48728411:49
* zhenguo away11:50
*** litao__ has quit IRC11:51
*** bani has quit IRC13:26
*** harlowja has joined #openstack-mogan17:09
*** openstack has joined #openstack-mogan22:46

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!