Thursday, 2017-02-09

*** adisky_ has quit IRC00:49
*** zhurong has joined #openstack-watcher01:29
*** zhurong has quit IRC01:35
*** thorst_ has joined #openstack-watcher01:54
*** thorst_ has quit IRC01:58
*** Yumeng has joined #openstack-watcher02:06
*** yuanying has quit IRC03:05
*** yuanying has joined #openstack-watcher03:07
*** thorst_ has joined #openstack-watcher03:08
*** thorst_ has quit IRC03:08
*** yuanying has quit IRC03:11
*** thorst_ has joined #openstack-watcher03:38
*** thorst_ has quit IRC03:38
*** thorst_ has joined #openstack-watcher04:38
*** thorst_ has quit IRC04:44
*** adisky_ has joined #openstack-watcher05:10
openstackgerritHidekazu Nakamura proposed openstack/watcher master: Add checking audit state  https://review.openstack.org/42999505:23
*** hidekazu has joined #openstack-watcher05:48
openstackgerritMerged openstack/puppet-watcher master: Fix typo in api.pp  https://review.openstack.org/42310206:00
*** yuanying has joined #openstack-watcher06:07
*** thorst_ has joined #openstack-watcher06:40
*** thorst_ has quit IRC06:45
*** diga has joined #openstack-watcher07:02
adisky_Hi can anybody tell me on which tool watcher state machine diagram are made?? I need to edit them to propose the specs for action plan stop07:07
hidekazuadisky_: i believe https://github.com/openstack/watcher/blob/master/doc/source/image_src/plantuml/audit_state_machine.txt07:14
hidekazusorry, this is audit. action plan is here: https://github.com/openstack/watcher/blob/master/doc/source/image_src/plantuml/action_plan_state_machine.txt07:15
adisky_hidekazu:07:15
adisky_thanks :)07:15
hidekazuadisky_: how to build: https://github.com/openstack/watcher/blob/master/doc/source/image_src/plantuml/README.rst07:16
*** diga has quit IRC07:16
hidekazuadisky_: i don't know you should include uml in spec though.07:17
adisky_hidekazu: I wont't include UML in stack, I will just upload it somwhere and give the links, is that fine??07:18
adisky_sorry *stack->specs07:18
hidekazuadisky_: it helps readers.07:19
adisky_hidekazu: yes, it will be helpful to understand.07:29
*** diga has joined #openstack-watcher07:54
hidekazulicanwei: here?08:07
*** alexchadin has joined #openstack-watcher08:20
licanweiyes08:35
*** vincentfrancoise has joined #openstack-watcher08:36
licanweihidekazu ?08:36
alexchadinvincentfrancoise: hi08:40
*** thorst_ has joined #openstack-watcher08:40
*** thorst_ has quit IRC08:46
*** alexchadin has quit IRC08:47
*** alexchadin has joined #openstack-watcher08:47
*** alexchadin has quit IRC08:54
*** alexchadin has joined #openstack-watcher08:54
*** openstackgerrit has quit IRC09:03
alexchadinvincentfrancoise: acabot: ping09:18
*** openstackgerrit has joined #openstack-watcher09:35
openstackgerritlicanwei proposed openstack/watcher master: Optimize audit process  https://review.openstack.org/43140809:35
vincentfrancoisealexchadin: morning09:37
*** alexchadin_ has joined #openstack-watcher09:38
*** alexchadin has quit IRC09:41
hidekazulicanwei: sorry i was in meeting, now iam free. still here?09:41
alexchadin_vincentfrancoise: could you please discuss action notifications with me?09:44
vincentfrancoiseyup09:48
vincentfrancoisealexchadin_: did you understand my comments ?09:48
vincentfrancoisealexchadin_: on the review09:48
alexchadin_vincentfrancoise: yeah, but if we flow_task already has execute method and it runs this only after pre_execute.09:51
alexchadin_I mean, if we execute pre_execute in execute, we will have 2 calls for pre_execute09:51
alexchadin_pre_execute->execute->pre_execute->do_execute->post_execute->post_execute09:52
*** hidekazu has quit IRC09:52
vincentfrancoisealexchadin_: I don't understand why you say there are 2 pre_execute09:53
alexchadin_https://github.com/openstack/taskflow/blob/master/taskflow/atom.py#L31009:54
vincentfrancoisealexchadin_: because execute is the "public" function and pre_execute/do_execute/post_execute are "private" and abstract09:54
alexchadin_vincentfrancoise: ^09:54
alexchadin_vincentfrancoise: https://github.com/openstack/taskflow/blob/df10cb5acc7d38c7bbe6f06f42393969b40e31ac/taskflow/engines/action_engine/executor.py#L5209:55
vincentfrancoisealexchadin_: oh ok I see, they already have that kind of separation09:56
alexchadin_vincentfrancoise: yeap, that's why I just wrapped them09:56
vincentfrancoisealexchadin_: IMHO, our execute should still wrap taskflow's pre/do/post09:57
vincentfrancoisebut because we have a naming issue09:57
alexchadin_vincentfrancoise: we can call self.pre_execute in child's pre_execute09:58
vincentfrancoisealexchadin_: yeah we could do a super()09:58
alexchadin_vincentfrancoise: that's the good way09:59
vincentfrancoisealexchadin_: should be working just fine :)10:03
alexchadin_vincentfrancoise: then I need to get exception block up to child class10:27
alexchadin_vincentfrancoise: cause there will be different workflow child.pre_execute -> base.pre_execute10:28
alexchadin_vincentfrancoise: I had base.pre_execute -> child.pre_execute and child.pre_execute was wrapped by try/except block10:29
vincentfrancoisealexchadin_: base.pre_execute -> child.pre_execute sounds good to me10:31
vincentfrancoisealexchadin_: why do you want to make it the other way around?10:31
alexchadin_cause call starts from plugin10:32
vincentfrancoisealexchadin_: capy paste you code on https://codeshare.io/ampvVw if that helps10:32
vincentfrancoisecopy*10:33
*** yuanying has quit IRC10:36
*** thorst_ has joined #openstack-watcher10:41
*** thorst_ has quit IRC10:46
*** danpawlik has joined #openstack-watcher11:36
*** diga has quit IRC12:26
*** thorst_ has joined #openstack-watcher12:41
*** thorst_ has quit IRC12:42
*** thorst_ has joined #openstack-watcher12:43
alexchadin_licanwei: hi13:07
*** alexchadin_ has quit IRC13:29
*** brunograz1 has joined #openstack-watcher13:48
*** brunograz has quit IRC13:51
*** brunograz has joined #openstack-watcher13:53
*** brunograz has quit IRC14:14
*** brunograz1 has quit IRC14:22
*** brunograz has joined #openstack-watcher14:37
*** vincentfrancoise has quit IRC14:42
*** acabot has quit IRC14:42
*** vincentfrancoise has joined #openstack-watcher14:44
chrisspencero/ hi15:06
vincentfrancoisechrisspencer: hi15:07
chrisspencervincentfrancoise: I updated the notification bug: https://review.openstack.org/#/c/422295 and was able to get it to print out both the event and the publisher id in the context.15:09
vincentfrancoisechrisspencer: ah nice! What did you do to make it work then?15:10
chrisspencerJust as you said, needed to update that context that is passed in.  But in order to display both (on our gate check/devstack) I modified that devstack watcher initialization of logging_context_format_string15:12
chrisspencerso it currently only solves the issue (by default) for devstack envs, however, I think it just a conf setting for production envs15:13
vincentfrancoiseok so now I see it (http://logs.openstack.org/95/422295/7/check/gate-watcher-dsvm-multinode-ubuntu-xenial-nv/23251d6/logs/screen-watcher-decision-engine.txt.gz#_2017-02-07_21_29_43_612)15:13
vincentfrancoisebut the information is not exactly the one I expected :p15:13
vincentfrancoisechrisspencer: [nova-compute:ubuntu-xenial-2-node-osic-cloud1-s3500-7167403 instance.update admin admin] => [publisher_id event_type XXX XXX]15:14
chrisspencerah, ack, that info bit should be easier to change :P15:14
vincentfrancoisechrisspencer: wait15:14
vincentfrancoisechrisspencer: what I sent you is what is currently being displayed15:15
vincentfrancoisechrisspencer: but the pusblisher_id is not the same as request_id15:15
vincentfrancoisechrisspencer: in the context of notifications, request_id is the message_id15:18
chrisspencervincentfrancoise: hm, where is that message_id generated?15:19
chrisspencerbecause it looks like we were already displaying request ID as part of that context, just needed to set it properly15:19
vincentfrancoisechrisspencer: actually yes15:20
*** wootehfoot has joined #openstack-watcher15:21
vincentfrancoisemessage_id is the rabbitmq message uuid I guess15:21
vincentfrancoisechrisspencer: but that's the only attribute that is actually unique to a particular notification15:22
vincentfrancoisechrisspencer: so that's why that's the attribute we need to track notifications in the logs15:22
chrisspencerah, vincentfrancoise where is that attribute passed into watcher? is it part of the payload or metadata?15:27
vincentfrancoisechrisspencer: part of metadata --> you can see it in the log link from above15:28
vincentfrancoisehttp://logs.openstack.org/95/422295/7/check/gate-watcher-dsvm-multinode-ubuntu-xenial-nv/23251d6/logs/screen-watcher-decision-engine.txt.gz#_2017-02-07_21_29_43_61215:28
vincentfrancoisechrisspencer: with metadata {'timestamp': u'2017-02-07 21:29:43.609934', 'message_id': u'495baf33-109f-470e-9aed-39472707b042'}15:29
chrisspencervincentfrancoise:yes got it.  cool. will make that change and push15:29
vincentfrancoisechrisspencer: wait15:29
vincentfrancoisechrisspencer: one more thing :p15:29
vincentfrancoisechrisspencer: now that we know that it's due to the context formatting15:29
vincentfrancoisechrisspencer: can we re-use the notation from https://review.openstack.org/#/c/422295/3/watcher/decision_engine/model/notification/nova.py@224 or https://review.openstack.org/#/c/422295/4/watcher/decision_engine/model/notification/nova.py@223 instead?15:30
vincentfrancoisechrisspencer: so that the whole thing gets more concise15:31
chrisspenceryup!15:31
vincentfrancoisechrisspencer: cool :)15:31
*** adisky_ has quit IRC15:39
*** acabot has joined #openstack-watcher15:48
*** brunograz has quit IRC16:03
*** brunograz has joined #openstack-watcher16:09
*** vincentfrancoise has quit IRC17:29
openstackgerritChris Spencer proposed openstack/watcher master: Adding additional details to notification logs  https://review.openstack.org/42229518:27
openstackgerritKen'ichi Ohmichi proposed openstack/watcher master: Switch to use test_utils.call_until_true  https://review.openstack.org/43171218:50
*** harlowja_ has joined #openstack-watcher20:35
*** harlowja has quit IRC20:37
*** thorst_ has quit IRC21:57
*** harlowja_ has quit IRC21:58
*** thorst_ has joined #openstack-watcher21:58
*** thorst_ has quit IRC22:02
*** harlowja has joined #openstack-watcher22:05
*** thorst_ has joined #openstack-watcher22:23
*** thorst_ has quit IRC22:28
*** gabor_antal has quit IRC23:05
*** gabor_antal has joined #openstack-watcher23:07
*** yuanying has joined #openstack-watcher23:12
*** gabor_antal has quit IRC23:15
*** gabor_antal has joined #openstack-watcher23:17
*** grumpycatt has quit IRC23:29
*** grumpycatt has joined #openstack-watcher23:30
*** wootehfoot has quit IRC23:56

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!