Tuesday, 2017-02-14

*** catintheroof has quit IRC00:25
*** `_^gk`_^1wm`_^ has joined #openstack-mistral00:38
*** `_^gk`_^1wm`_^ has quit IRC00:39
*** jamielennox is now known as jamielennox|away01:09
*** jamielennox|away is now known as jamielennox01:17
*** gongysh has joined #openstack-mistral01:49
*** jrist has quit IRC01:58
rakhmerovkong: hi02:06
kongrakhmerov: hi02:07
rakhmerovhow are you?02:08
kongnot bad, how are you?02:09
rakhmerovnot bad at all )02:09
rakhmerovkong: please wait till Fri morning with the release02:09
rakhmerovwe'll need a few more patches02:09
kongrakhmerov: ok, sure02:09
rakhmerovthanks02:09
*** jrist has joined #openstack-mistral02:11
*** dtturner has quit IRC02:27
zhenguokong: hi, could you please help +A the dashboard patches if you have time, thanks!02:36
kongzhenguo: ok02:37
zhenguokong: thanks02:37
kongzhenguo: np02:38
openstackgerritMerged openstack/mistral-dashboard master: Replace github with git.openstack.org  https://review.openstack.org/43281603:41
openstackgerritMerged openstack/mistral-dashboard master: Fix 'tox -ereleasenotes' Error  https://review.openstack.org/43282703:46
openstackgerritMerged openstack/mistral-dashboard master: Update .gitignore  https://review.openstack.org/43287903:48
*** Kevin_Zheng has quit IRC04:43
*** sharatss has joined #openstack-mistral05:20
*** jaosorior has joined #openstack-mistral06:09
*** jtomasek has quit IRC06:12
openstackgerritMerged openstack/mistral master: Updated from global requirements  https://review.openstack.org/43312306:17
openstackgerritRenat Akhmerov proposed openstack/mistral master: Refactor RPC serialization: add polymophic serializer  https://review.openstack.org/42964306:25
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add Keycloak authentication doc for server side  https://review.openstack.org/43346306:59
*** sharatss has quit IRC07:17
*** sharatss has joined #openstack-mistral07:17
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add Keycloak authentication doc for client side  https://review.openstack.org/43346707:21
d0ugalNew logo looks good :)07:24
d0ugalI am sad that I wont be at the PTG to get some of the stickers07:25
*** shardy has joined #openstack-mistral07:58
*** shardy has quit IRC08:24
*** jpich has joined #openstack-mistral08:41
rakhmerovd0ugal: yeah, logo is very cool IMO08:43
d0ugalMaybe I can get somebody local-ish to bring me some back08:43
rakhmerovand it is not like many others in OpenStack08:44
d0ugal(some stickers back)08:44
rakhmerovd0ugal: yes, why not08:44
d0ugaljpich: can you collect stickers? :-D08:44
jpichd0ugal: Sure, does that mean you're gonna come speak at an Irish conference soon? ;)08:45
d0ugalhaha, why not!08:45
jpichmwa, ha, ha08:45
d0ugaljpich: except the only one I am aware of likely to be in November ... ages away!08:46
jpichd0ugal: Don't worry, we got loadsa conferences, I'll find you one no problems08:46
jpichalso the local Python meetup has monthly talks. Sorted!08:46
d0ugalyay08:46
jpichSo... the stickers are the fancy new official ones which I'm guessing will be readily available? Or are those special stickers requiring a special handshake?08:47
d0ugaljpich: http://lists.openstack.org/pipermail/openstack-dev/2017-February/112091.html08:48
d0ugal"At the PWG, we’ll have stickers for your team of the mascot, plus signage on your room."08:49
d0ugalI assume that should be PTG08:49
rakhmerovjpich: new official ones08:49
jpichlol08:49
jpichGotcha!08:49
rakhmerov:)08:49
d0ugalrakhmerov: I wonder why the Mistral logo is so different08:49
d0ugalMost of them seem to be blue08:49
d0ugalactually, that isn't true :)08:50
rakhmerovd0ugal: it's different because of me :)08:55
d0ugalhaha, cool - good job!08:55
rakhmerovI gave them an idea that was kinda crazy08:55
rakhmerovI decided to experiment last minute, because other options were already occupied by other projects (e.g. octopus)08:56
rakhmerovbut the result seems cool08:56
rakhmerovunexpectedly :)08:56
d0ugal:-D08:56
d0ugalWho has the octopus?08:56
rakhmerovaah, gosh.. don't remember now08:57
rakhmerovmaybe UI08:57
rakhmerovUX08:58
rakhmerovhttps://www.openstack.org/project-mascots/08:59
rakhmerovI like what Jay Pipes wrote about Mistral mascot :)09:00
rakhmerovQuote: "To me, it kind of looks like people jumping joyously off the top of a ferris wheel"09:00
rakhmerov:)) funny09:00
rakhmerovd0ugal: I seem to have addressed your comments in https://review.openstack.org/#/c/429643/09:01
rakhmerovwould appreciate if you could review it09:02
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add Keycloak authentication doc for client side  https://review.openstack.org/43346709:19
*** gongysh has quit IRC09:32
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add Keycloak authentication doc for server side  https://review.openstack.org/43346309:32
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add Keycloak authentication doc for client side  https://review.openstack.org/43346709:32
*** gongysh has joined #openstack-mistral09:34
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add details into docs about semantics of 'on-XXX' clauses.  https://review.openstack.org/43355710:07
openstackgerritMerged openstack/mistral master: Refactor RPC serialization: add polymophic serializer  https://review.openstack.org/42964310:35
rakhmerovd0ugal, ddeja: guys, please take a look at my doc patches: https://review.openstack.org/433463, https://review.openstack.org/433467 and https://review.openstack.org/43355710:38
*** gongysh has quit IRC10:39
*** shardy has joined #openstack-mistral10:48
ddejarakhmerov: I will as soon as I finish fixin tests :)10:59
rakhmerovok10:59
rakhmerovddeja: I noticed one suspicious thing. With Kombu RPC when I start a big workflow (~ 1000 tasks) the workflow starts normally (although it takes 10-15 seconds just to start it) but the client says that request failed due to RPC timeout11:04
rakhmerovit doesn't happen with o.m11:05
rakhmerovmaybe we really need to increase some default timeouts11:05
rakhmerovnot sure where11:05
rakhmerovddeja: btw, another good thing about Kombu RPC: it reduces memory usage pretty significantly11:07
rakhmerovtens of megabytes on my test workflows11:07
ddejarakhmerov: OK, I can investigate it next week11:09
rakhmerovyes, please11:10
rakhmerovddeja: I also wanted to talk to you about what's missing in Kombu RPC from your perspective11:10
rakhmerovthat stops us from using it in production11:10
ddejarakhmerov: Right now I don't see any obvious feature that is missing11:11
rakhmerovdo we need to implement some heartbeats?11:12
ddejaso from my perspective it's only 2 things: test it & brake it ;)11:12
ddejarakhmerov: I don't think so11:12
ddejawhy do we need them?11:12
*** sharatss has quit IRC11:12
rakhmerovI know that there used to be a problem in o.m, if a connection with Rabbit was idle during some time it would drop11:12
*** sharatss has joined #openstack-mistral11:12
ddejarakhmerov: I've naver faced such issue11:13
rakhmerovas far as I remember, RabbitMQ drops connections if they are idle but kombu doesn't know about it11:13
ddejaoh, I see11:13
rakhmerovyes11:13
ddejasoo some mechanism that would send a message from time to time 'I'm alive, don't drop connection' ;)11:13
rakhmerovwe need to think what it will cost to implement some simple heartbeat mechanism (if it's really needed)11:14
rakhmerovyes11:14
rakhmerovbut test first, of course11:14
ddejaOK, I can think something11:14
rakhmerovok, so I got your opinion :)11:14
ddejayes, sure, I'm running tox right now11:14
rakhmerovnothing else is missing11:14
rakhmerovok11:14
ddejaI think so11:15
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add Keycloak authentication doc for client side  https://review.openstack.org/43346711:23
rakhmerovd0ugal: fixed https://review.openstack.org/43346711:24
d0ugalI spotted :)11:25
*** jrist has quit IRC11:27
*** jkilpatr has quit IRC11:36
*** jrist has joined #openstack-mistral11:40
*** gongysh has joined #openstack-mistral11:46
openstackgerritRenat Akhmerov proposed openstack/mistral master: Add Keycloak authentication doc for client side  https://review.openstack.org/43346711:48
*** Kevin_Zheng_ has joined #openstack-mistral11:57
*** jkilpatr has joined #openstack-mistral12:07
openstackgerritDawid Deja proposed openstack/mistral master: Fix for coverage job showing 0% coverage for kombu  https://review.openstack.org/43307412:32
*** thrash|g0ne is now known as thrash12:38
openstackgerritDawid Deja proposed openstack/mistral master: Fix for coverage job showing 0% coverage for kombu  https://review.openstack.org/43307412:41
ddejarakhmerov: It should fix the unit tests: https://review.openstack.org/43307412:41
rakhmerovok, let's wait till they pass12:42
d0ugalWhat does it mean for an action to have system=True?12:44
d0ugalrakhmerov: ^12:44
d0ugalhttps://github.com/openstack/mistral/blob/master/mistral/db/v2/sqlalchemy/models.py#L10412:44
ddejad0ugal: AFAIK it's an action that cannot be changed by user12:44
rakhmerovall std.xxx actions are considered 'system' since they are provided initially12:45
rakhmerovyes12:45
rakhmerovthey can't be changed12:45
d0ugalI see12:45
d0ugalall the openstack actions are also True12:45
d0ugalas are our custom TripleO actions :)12:45
d0ugalthis is the only non-system action I have: https://github.com/openstack/mistral/blob/master/mistral/resources/actions/wait_ssh.yaml12:47
d0ugal(I didn't even know that one existed before!)12:47
d0ugalso it seems like is_system=False actions are ad-hoc actions.12:48
rakhmerovd0ugal: I recalled something else12:53
rakhmerov:)12:53
rakhmerovI think now only ad-hoc actions are not system12:54
rakhmerovthose that we can define right in YAML12:54
rakhmerovhence we can change only these actions through API12:54
rakhmerovddeja: something failed again :(12:54
d0ugalrakhmerov: I see, thanks.12:55
ddejarakhmerov: Oh, not good12:56
rakhmerovyep12:56
ddejaOK, looking back what may happen...12:57
d0ugalrakhmerov: What do you think about allowing expressions within engine commands?13:04
d0ugalrakhmerov: i.e. like 16 here wont work as I want: http://paste.openstack.org/show/598812/13:04
rakhmerovit should work13:04
rakhmerov1 sec13:04
d0ugaloh, it is because I made a typo13:04
d0ugallol13:04
d0ugalsorry13:04
rakhmerovwhat typo?13:05
rakhmerovdon't see anything bad..13:05
d0ugalIt should be $.my_var13:05
rakhmerovooh, right13:05
rakhmerov:)13:05
rakhmerovshould work13:05
d0ugalyeah, it does13:05
d0ugalsorry for the noise13:05
d0ugalI am surprised YAQL doesn't complain.13:05
rakhmerov:)13:05
rakhmerovd0ugal: one thing that you might missed while on vacation13:06
rakhmerovI talked about it here in IRC13:06
rakhmerovit turns out that Jinja2 works much faster than YAQL13:06
d0ugaloh, cool13:06
rakhmerov3-4 times at least13:06
*** catintheroof has joined #openstack-mistral13:07
d0ugalGood to know.13:07
rakhmerovif you have some structures in wf context with many items (hundreds or thousands) it may make a big difference13:07
rakhmeroveven on simple expressions like <% $.my_var %> if my_var is a dict say with 10000 entries13:07
rakhmerovyeah, keep in mind, I think we'll have to mention that in the docs too13:08
openstackgerritMichal Gershenzon proposed openstack/mistral master: Update docs for tasks function  https://review.openstack.org/43309613:08
rakhmerovI asked YAQL folks to look at it but they didn't look at it yet as far as I know13:08
d0ugalokay, cool13:09
rakhmerovd0ugal: also, Kombu based RPC that Dawid has been working on is faster than oslo.messaging around 1.5 times13:09
rakhmerov50%13:09
d0ugalnice13:09
rakhmerovyep13:09
d0ugalHow do we switch to that? Just a config change?13:09
rakhmerovjust add one config property13:10
rakhmerovrpc_implementation = kombu13:10
rakhmerovthat's it13:10
d0ugalrakhmerov: I'll try the TripleO gate if I can figure out where I would need to set that...13:10
rakhmerov:)13:10
rakhmerovis it not easy to find a config?13:11
d0ugalrakhmerov: well, I think I need to update it in puppet-mistral and check that triggers a tripleo gate13:11
d0ugalotherwise I would need to make a tripleo repo depend on that puppet change, I guess13:11
d0ugalthere might be an easier way13:11
rakhmerov:)13:14
rakhmerovI see13:14
d0ugalIn theory making CI do it should be easier than testing it myself13:15
*** jamielennox is now known as jamielennox|away13:21
*** shardy has quit IRC13:33
*** shardy has joined #openstack-mistral13:34
ddejarakhmerov: about the failing unit tests: "it works for me"™13:37
ddejaI need to wait for jenkins report13:37
*** sharatss has quit IRC13:50
openstackgerritMerged openstack/mistral master: Add Keycloak authentication doc for server side  https://review.openstack.org/43346313:51
openstackgerritMerged openstack/mistral master: Add details into docs about semantics of 'on-XXX' clauses.  https://review.openstack.org/43355713:51
*** AlexeyAbashkin has joined #openstack-mistral14:06
*** toure|biab is now known as toure14:23
ddejarakhmerov: OK, I see why tests are failing - master got updated in the meantime14:33
*** Kevin_Zheng_ has quit IRC14:43
*** Kevin_Zheng_ has joined #openstack-mistral14:47
*** tuan_756 has joined #openstack-mistral14:51
tuan_756ddeja, renat: Could you guy take a look to the idea of pagination for this patch14:52
tuan_756https://review.openstack.org/#/c/424621/8/mistral/services/expiration_policy.py14:52
*** tuan_756 has quit IRC15:00
d0ugalWhy would this ever happen? https://github.com/openstack/mistral/blob/master/mistral/utils/expression_utils.py#L145-L14615:03
d0ugalI thought these would be the same, but wf2 outputs None.15:04
openstackgerritDawid Deja proposed openstack/mistral master: Fix for coverage job showing 0% coverage for kombu  https://review.openstack.org/43307415:05
ddejarakhmerov: ^^ hopefully now it's OK15:05
d0ugalI would update the commit message, kinda true but there is a bit more to it than that :)15:05
*** AlexeyAbashkin has quit IRC15:08
ddejad0ugal: OK, I'll do it but firstly I'd like Jenkins to test it ;) (Or just updating the commit message wouldn't make jenkins to run the test again?)15:10
d0ugalddeja: I am not sure, it might re-start it15:10
ddejad0ugal: whatever, let's check it :D15:12
openstackgerritDawid Deja proposed openstack/mistral master: Fix for coverage job showing 0% coverage for kombu  https://review.openstack.org/43307415:13
ddejad0ugal: it does restart it15:14
d0ugalhah, now we know. thanks ddeja15:14
d0ugalLOL I broke mistral. http://paste.openstack.org/show/598843/15:15
d0ugalThat is the full output :/15:15
d0ugalwow, nothing in the logs either.15:16
ddejad0ugal: nice15:18
openstackgerritDawid Deja proposed openstack/mistral master: Fix for coverage job showing 0% coverage for kombu  https://review.openstack.org/43307415:18
ddejaI see you are doing what I said today's morning15:19
ddeja'test it & brake it'15:19
ddejanow comes part 3 - fix it :D15:19
d0ugal:)15:19
d0ugalddeja: if you are curious: http://paste.openstack.org/show/598845/15:23
d0ugalthe issue is that on-complete should be indented.15:24
d0ugalI am not sure if the problem is the "-" or if it is because the task is a list...15:24
d0ugalI think it is a combination of the two15:26
d0ugalkinda weird :)15:26
ddejad0ugal: I wonder if on complete: fail may be the problem15:26
ddejaI saw on-success: fail15:26
ddejait works fine15:26
ddejabut how about on-complete15:26
ddejacan you submit a bug for that15:27
ddejayou can assigne me to it :)15:27
d0ugalddeja: haha, sure15:27
ddejaif you don't have time/don't want to do it ;)15:27
d0ugalddeja: on-complete: fail works otherwise15:27
d0ugalI think this is a vailidation issue.15:28
ddejaprobably yeah15:28
*** gongysh has quit IRC15:28
d0ugalfor some reason the name on-complete causes it to have different validation rules15:28
d0ugaloh, so I think the issue is the dash.15:29
d0ugalMaybe I caused a regression here: https://github.com/openstack/mistral/commit/b8d94ee4b832f52008d80d72b4b15bd804e738f015:30
d0ugalddeja: https://bugs.launchpad.net/mistral/+bug/166461215:33
openstackLaunchpad bug 1664612 in Mistral "tasks with a hyphen in their name don't work" [Undecided,Confirmed] - Assigned to Dawid Deja (dawid-deja-0)15:33
ddejad0ugal: OK, thanks15:35
d0ugalI would really like a run-workflow CLI command that creates a workflow from a file and runs it but doesn't save the workflow15:36
d0ugalwould be super useful15:36
ddejaagree15:40
rbradyd0ugal: could it have an arg for save-on-error or something?  might be useful for troubleshooting15:41
d0ugalrbrady: yeah, maybe - but I'd be concerned about testing workbooks being left behind15:42
d0ugalworkflows15:43
d0ugalbut actually, workbooks make it a but more complicated :)15:43
d0ugalI imagine something like: mistral run-workflow workbook.yaml workflowname 'input...'15:44
* d0ugal shrugs15:44
d0ugalit would also be useful for testing new versions of workflows without updating the workflow that is in Mistral (or changing the name)15:45
d0ugalOkay, I am going to open a blueprint.15:45
d0ugalrbrady: the action-executions could all be saved. that would be fine15:48
d0ugalbut exposing a testing workflow would be potentially dangerous or just messy15:49
d0ugalI guess it could be quite tricky to do15:49
rbradyd0ugal: I'm just worried that a testing workflow is likely to run into an error at somepoint and if it's gone, can I reasonably figure out what's wrong15:50
rbradywhether that is an action execution or something wrong with how I created the task in the workflow itself15:50
d0ugalrbrady: yeah, so the execution and action-exection could save all that15:50
d0ugalrbrady: I just don't want the testing workflow to ever show up in "mistral workflow-list" etc.15:51
rbradyd0ugal: cool15:51
*** jaosorior has quit IRC15:54
d0ugalhttps://blueprints.launchpad.net/mistral/+spec/mistral-run-workflow15:56
d0ugalrakhmerov: ^ :)15:56
* d0ugal reports another bug16:10
d0ugalddeja: I have gotten stuck in a test and break loop ;)16:10
ddejad0ugal: OK, I need to drop16:11
d0ugalHave a good evening :)16:12
*** bobh has joined #openstack-mistral16:31
*** jpich has quit IRC16:58
*** jkilpatr_ has joined #openstack-mistral17:00
*** jkilpatr has quit IRC17:02
*** catintheroof has quit IRC17:07
*** catintheroof has joined #openstack-mistral17:07
openstackgerritMichal Gershenzon proposed openstack/mistral master: Reove output from list action executions API  https://review.openstack.org/43383117:54
*** jkilpatr has joined #openstack-mistral18:01
*** jkilpatr_ has quit IRC18:01
*** thrash is now known as thrash|biab18:28
*** bobh has quit IRC18:34
*** thrash|biab is now known as thrash18:58
*** shardy has quit IRC19:00
*** bobh_ has joined #openstack-mistral20:14
*** jamielennox|away is now known as jamielennox20:17
*** openstackgerrit has quit IRC20:18
*** jkilpatr has quit IRC21:20
*** jkilpatr has joined #openstack-mistral22:03
*** kozhukalov_ has joined #openstack-mistral22:40
*** histrio_ has joined #openstack-mistral22:41
*** jkilpatr has quit IRC22:42
*** kozhukalov has quit IRC22:42
*** histrio has quit IRC22:42
*** jamielennox has quit IRC22:42
*** histrio_ is now known as histrio22:42
*** kozhukalov_ is now known as kozhukalov22:44
*** zigo has quit IRC22:46
*** brunograz has quit IRC22:46
*** chlong has quit IRC22:46
*** jkilpatr has joined #openstack-mistral22:51
*** zigo has joined #openstack-mistral22:52
*** brunograz has joined #openstack-mistral22:52
*** chlong has joined #openstack-mistral22:52
*** zigo is now known as Guest2705722:54
*** bobh_ has quit IRC23:36

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