Thursday, 2016-06-30

*** bobh has joined #openstack-mistral00:11
*** bobh has quit IRC00:25
*** gyee has quit IRC00:50
*** bobh has joined #openstack-mistral00:53
*** bobh has quit IRC00:57
*** rrecio_ has joined #openstack-mistral01:26
*** cheneydc has joined #openstack-mistral01:27
*** rrecio_ has quit IRC01:34
*** cheneydc has quit IRC02:08
*** bobh has joined #openstack-mistral02:53
*** bobh has quit IRC03:07
*** Ravikiran_K has joined #openstack-mistral03:51
*** saphi has joined #openstack-mistral04:11
saphihi rakhmerov04:12
rakhmerovhi04:12
saphihow to input an object to a action?04:12
saphiexample: nova.servers_evacuate action04:13
rakhmerovsaphi: I don't understand your question, what do you mean by "object?04:15
saphisorry. I have been confused04:19
rakhmerovconfused by what?04:31
saphibecause when I use nova.server_migrate action I can't input parameters to execute. But with servers_evacuate I can input parameters04:44
*** Ravikiran_K has quit IRC04:54
rakhmerovsaphi: can you please be more specific so that I don't have to ask you a million questions. What do you mean by "I can't input parameters"?04:54
rakhmerovgive the steps to reproduce the problem and the stack trace04:54
rakhmerovotherwise, it's hard to understand what we're talking about04:54
rakhmerovhparekh, kong, nmakhotkin, akuznetsova: hi guys, can you please be slightly more active on reviewing patches (if time allows)? There's a bunch of patches on review now but not enough eyes to look at them04:57
hparekhrakhmerov, Ok sure. Actually I was busy with someother stuff .05:04
rakhmerovhparekh: ok :)05:04
rakhmerovI'm just asking05:04
rakhmerovgenerally, we're lacking core reviewers, I'd like to promote 1-2 more members to core reviewers05:05
rakhmerovbut I don't see who else at this point05:05
kongrakhmerov: got you05:07
rakhmerovkong: hi05:10
rakhmerovbtw, your patch failed for some reason05:10
rakhmerovit seems like something is wrong with service coordinator05:10
rakhmerovkong: can you look at it please?05:10
kongrakhmerov: yeah, I planed to to do that tonight05:11
kongmaybe it's not related to my patch05:11
kongsome module update isssue I think05:11
kongdidn't get a chance to dig it05:11
*** saphi has quit IRC05:33
openstackgerritXavier Hardy proposed openstack/mistral: Add support for custom YAQL functions  https://review.openstack.org/33480405:50
openstackgerritMerged openstack/python-mistralclient: Use osc_lib instead of cliff  https://review.openstack.org/33544206:07
*** Ravikiran_K has joined #openstack-mistral06:08
xavierhardyhow does output for actions work ?06:28
xavierhardymore exactly, is there anything in YAQL to get the original action execution result?06:29
xavierhardyLet's say I do an HTTP call, I want to return only its content, because the rest of the result superfluous in my case, how do I do that from the action execution (I know how to do it from a task)?06:30
xavierhardyenv and execution are None06:32
xavierhardyin an action execution (no task)06:32
xavierhardyOK, i found it in the context06:33
xavierhardy(in $)06:33
xavierhardyI think "output" for actions is not documented06:38
xavierhardybut it's super useful06:39
xavierhardyI noticed a bug with with-items and default values in actions. The default values of the action are not taken into account and considered missing when ommitted. This should be fixed; I guess it's just a matter of updating the task YAQL context.06:46
rakhmerovxavierhardy: hi, give me a sec..06:51
*** shardy has joined #openstack-mistral06:51
rakhmerovas for the bug, can you pls file a bug so that we don't forget about it?06:52
xavierhardySure06:53
rakhmerovon your question, I'm not sure I fully understood it. Action output is always accessible in YAQL as "task(task_name).result"06:53
rakhmerovbut seems like you mean something different06:54
rakhmerovwhat do you mean by "in an action execution"?06:54
xavierhardyIn an action06:54
xavierhardyWhen you run an action alone06:55
rakhmerovin Python code that implements action?06:55
rakhmerovrun() method?06:55
xavierhardyNo06:55
xavierhardyIn the DSL06:55
rakhmerov:)06:55
xavierhardyI found out that the "output" field is also available for actions06:55
rakhmerovso you're talking about ad-hoc actions then?06:55
xavierhardythough it seems undocumented06:55
xavierhardyyes06:55
rakhmerovooooh, yes06:55
xavierhardyI found out that $ contains the result06:55
rakhmerovthat's important piece of info :)06:55
rakhmerovyes, true06:56
xavierhardyYes, it means I can run an aciton and do some regex on it06:56
rakhmerovit may not be documented, yes06:56
xavierhardyit's documented only for workflows I think06:56
rakhmerovlet me see..06:56
xavierhardyI am also expriencing deadlock with with-items when not setting a reasonable limit06:56
xavierhardy*deadlocks06:56
rakhmerovlimit of what?06:57
xavierhardyit's documented actually now06:57
xavierhardyeven the result, ok mea culpa06:57
xavierhardyFor with-items, I am talking about concurrency06:57
rakhmerovit's in http://docs.openstack.org/developer/mistral/dsl/dsl_v2.html#attributes06:57
xavierhardyI did not set it06:57
rakhmerovok06:58
rakhmerovand what did you get?06:58
rakhmerovDB deadlock?06:58
xavierhardyMessagingTimeout: Timed out waiting for a reply to message ID f417d8f9520144a2823ed62696a18d4406:58
xavierhardyFor the DB deadlock, I cannot confirm it yet.06:58
rakhmerovI am inclined to think that it's not a deadlock probably06:59
xavierhardybut this appeared a few minutes after some of the action executions got stuck06:59
rakhmerovit might be purely a messaging problem06:59
rakhmerovhm... can you file a bug pls?06:59
xavierhardyWill do06:59
rakhmerovthanks a lot, that's a good stuff06:59
xavierhardyIt's very easy to reproduce: with-items with HTTP06:59
xavierhardylet's say 25 threads (I have 3 workers)07:00
rakhmerovmeaning 25 items in the collection07:00
xavierhardyor more yes07:00
*** mgershen has joined #openstack-mistral07:01
rakhmerovdo you use std.http or std.mistral_http?07:03
xavierhardystd.http07:03
xavierhardyah... weird07:05
xavierhardyI did not reproduce it by simply using std.http07:05
xavierhardyon example.com07:05
rakhmerovok07:06
rakhmerovI know that people use with-items actively07:06
rakhmerovwith hundreds of items07:06
rakhmerovso pls as much as possible into the bug description, I'd like to look at it soon07:07
rakhmerovprobably next week07:07
xavierhardyWell, I can't put everything, that's the problem.07:07
xavierhardymaybe it has something to do with std.http reaching the timeout07:09
rakhmerovI see07:10
rakhmerovmaybe HTTP timeout is greater than a timeout set for messaging07:10
rakhmerovsomething like this..07:10
xavierhardyyes, I remember my colleagues saying they had problem with the timeout option and the messages we return can be quite big, the API might take a while to answer + it might be overloaded.07:11
*** mgershen has quit IRC07:14
*** mgershen has joined #openstack-mistral07:14
xavierhardyhttps://bugs.launchpad.net/mistral/+bug/159763707:17
openstackxavierhardy: Error: malone bug 1597637 not found07:17
*** mgershen has quit IRC07:18
*** saphi has joined #openstack-mistral07:19
rakhmerovthanks a bunch07:21
xavierhardyThis is the "with-items default value" error: https://bugs.launchpad.net/mistral/+bug/159764007:22
openstackLaunchpad bug 1597640 in Mistral "with-items does not process ad-hoc action default values" [Undecided,New]07:22
xavierhardyThe action and workflow given should reproduce it for sure, at least it does here.07:22
*** mgershen has joined #openstack-mistral07:29
rakhmerovxavierhardy: ok07:35
*** jpich has joined #openstack-mistral07:39
openstackgerritRenat Akhmerov proposed openstack/mistral: Add zake into dependencies  https://review.openstack.org/33583707:41
saphiI want to write a workflow include some task: task 1. check list hypervisors have state down. task 2: list all instance in list hypervisor. And task 3: perform evacuate those instances When I write task 2, I use with-items I think It only run action not run publish So I Can't get id those instances08:00
rakhmerovhparekh, kong, nmakhotkin: can you please approve https://review.openstack.org/#/c/335837/ ?08:19
nmakhotkinhi rakhmerov08:22
rakhmerovhi )08:22
nmakhotkinsec..08:22
nmakhotkinwhy it should be added?08:22
rakhmerovtooz now requires it08:23
rakhmerovbut through stevedore08:23
rakhmerovthat's why we explicitly need to add it08:23
rakhmerovbecaouse it's not resolved through dependencies automatically08:23
rakhmerovit's used in a plugin in tooz08:23
nmakhotkinok08:23
nmakhotkinI can't approve it though08:24
rakhmerovall gates are failing now because of it08:24
rakhmerovyes, I mean just review08:24
rakhmerovas usually08:24
nmakhotkinwaiting for hparekh or kong08:24
rakhmerovyes, right08:24
rakhmerovthanks08:24
hparekhdone08:26
rakhmerovappreciate08:28
hparekhrakhmerov, But I think these decencies was automatically resolved previously08:29
hparekhI can see http://logs.openstack.org/38/330438/1/gate/gate-mistral-python27/56973e9/console.html#_2016-06-28_23_39_17_21911008:29
rakhmerovhparekh: I guess they changed impl in tooz08:29
rakhmerovI looked at the log and saw what happens there. They load modules through stevedore08:30
rakhmerovdrivers08:31
rakhmerovthis is because it's not resolved08:31
hparekhhmm looks like they removed zake from requirements.txt in 1.40 version08:31
rakhmerovbecause it's a dynamic dependency08:31
hparekhOk, Got it08:31
rakhmerovyes, that's the point08:31
rakhmerovright08:31
openstackgerritMerged openstack/mistral: Add zake into dependencies  https://review.openstack.org/33583708:33
*** Yaminikanth has joined #openstack-mistral08:38
kongrakhmerov: thanks for your patch to solve that problem08:50
rakhmerovnp08:50
openstackgerritLingxian Kong proposed openstack/mistral: Add db models for event trigger  https://review.openstack.org/32049708:55
openstackgerritLingxian Kong proposed openstack/mistral: Add event engine service  https://review.openstack.org/32050008:55
openstackgerritLingxian Kong proposed openstack/mistral: Add event trigger REST API  https://review.openstack.org/32050908:55
*** Kiall_ has joined #openstack-mistral09:03
*** ruhe2 has joined #openstack-mistral09:04
openstackgerritMerged openstack/mistral: Use client credentials to retrieve service list  https://review.openstack.org/33046909:04
*** igormarnat__ has joined #openstack-mistral09:04
*** ruhe has quit IRC09:08
*** FL1SK has quit IRC09:08
*** Kiall has quit IRC09:08
*** igormarnat has quit IRC09:08
*** hparekh has quit IRC09:08
*** ruhe2 is now known as ruhe09:08
*** igormarnat__ is now known as igormarnat09:08
*** hparekh has joined #openstack-mistral09:09
*** FL1SK has joined #openstack-mistral09:16
openstackgerritMerged openstack/mistral: Fix SSHActionsTestsV2 failure  https://review.openstack.org/33494509:19
saphihi everybody. How to random an item in a list with YAQL? Please help me.09:20
rakhmerovsaphi: sorry, I won't be available for the next couple of hours09:20
rakhmerovnmakhotkin, hparekh, kong: guys, is anybody available to help saphi?09:21
saphioh ... I'm waiting09:21
kongsorry, I'm not expert with YAQL, I'm not sure if it's mentioned in YAQL documentation.09:32
kongsaphi: did you take a look at the doc?09:32
saphiYes. But It was not mentioned in doc.09:33
kong:-(09:33
kongsaphi: https://github.com/openstack/yaql/blob/master/yaql/tests/test_math.py#L15609:41
kongsaphi: hope that will help you09:41
kongI think it will do what you need09:42
xavierhardyThe YAQL documentation is extremely scarce, but they are working on it09:54
xavierhardyWhy it takes so much time to release a documentation is because they want to have most of it automated (e.g. using the existing types in the code to document the type...)09:55
xavierhardyThe best documentation to date:09:56
xavierhardyhttps://github.com/openstack/yaql/tree/master/yaql/standard_library09:56
xavierhardy;)09:56
kongxavierhardy is right, the code is the best doc for developers :-)09:57
kongand sometimes, unit tests will tell more secrets09:58
xavierhardyit shouldn't be09:58
saphithank kong and xavierhardy10:00
saphiI will try10:00
kongsaphi: np10:01
*** Yaminikanth has quit IRC10:05
saphihey. I try with: action: nova.servers_evacuate server=<% $.instance %> host=<% with(random() -> $ $.all_hypervisor_active %>10:07
saphibut it wrong. note: $.all_hypervisor_active is a list10:07
nmakhotkinhi saphi10:08
nmakhotkinI get it worked with [10, 12, 20][random(0, len([10 ,12, 20]) - 1)]10:08
saphihi nmakhotkin10:08
nmakhotkinwhere [10, 12, 20] is your list10:08
*** mgershen has quit IRC10:09
nmakhotkinrandom(0, 10) gives a number from 0 to 1010:09
kongsaphi: nmakhotkin is YAQL expert10:10
kong:-)10:10
nmakhotkinso we can get random index of the list10:10
nmakhotkinkong, haha :)10:10
saphiwith my action. is it? action: nova.servers_evacuate server=<% $.instance %> host=<% with(random($.all_hypervisor_active) %>10:10
nmakhotkinno, I'm not an expert too :)10:11
saphibut when I update to my workflow. Mistral notice: Parse error: unexpected end of statement10:11
nmakhotkinsaphi: try this:  host=<% $.all_hypervisor_active.random(0, len($.all_hypervisor_active)) %>10:12
nmakhotkinoh, no it is wrong10:12
nmakhotkinhost=<% $.all_hypervisor_active.random(0, len($.all_hypervisor_active) - 1) %>10:12
nmakhotkinwrong again :)10:12
nmakhotkinhost=<% $.all_hypervisor_active[random(0, len($.all_hypervisor_active) - 1)] %>10:13
nmakhotkin$.all_hypervisor_active should be a list10:13
kong<% $.all_hypervisor_active[random((0, len($.all_hypervisor_active) - 1) %>10:13
nmakhotkinkong: yes, exactly :) but extra parentheses10:14
kongyes10:14
saphiyes10:14
kongsaphi: please have a try10:14
kongand let us know the lucky result10:15
saphiI think It work. nmakhotkin10:15
kongawesome!10:15
saphiBut I have with nova.servers_evacuate action10:15
saphiI try run this action. But when this action successfull. my instance error10:16
kongsaphi: it should be a nova problem then10:16
saphiI try run with cli. `nova evacuate <server> <host>` It work10:16
saphiI don't know which problem?10:16
saphibecause I try with nova command. It work10:17
kongsaphi: you randomly choose a host in your workflow10:17
kongbut in CLI, you specify a host10:17
kongI think that matters10:17
kongif you can take a look at the log to see which host was choosen in your workflow10:18
kongand run CLI with that host10:18
saphiyes. because when use nova evacuate We have to specify a host. But we have many host. And In the host fail has many instance. I have to evacuate those instances to others hosts10:19
nmakhotkinsaphi: also, try to publish this host value (which is evaluated with YAQL)10:19
nmakhotkinand see what it gives10:19
kongsaphi: you must be doing that instance HA thing10:20
saphiI have only 2 compute-node to test10:20
* kong goes to write 'funny' UT for event trigger10:20
saphiI want to check a compute node down. And after that evacuate all instances in that host to other hosts10:21
ddejahey saphi, are you working on instance HA?10:23
ddejasaphi: oh, ok, didn't see your latest post10:24
ddejasaphi: I'm working on same thing :) You can take a look at this PoC: https://github.com/gryf/mistral-evacuate10:25
*** mgershen has joined #openstack-mistral10:29
saphigood job ddeja .I will try it. but I want to evacuate all instance in a compute node10:30
saphiI think my workflow can work. But The  nova.servers_evacuate I don't know it doesn't work. My workflow here: http://pastebin.com/DnidiF5U10:32
saphiplease review for me. Thanks. If ddeja can test it, You could report for me. :D10:33
ddejasaphi: One thing: why you choosing randomly where to put evacuated instance?10:34
ddejayou can let the nova scheduler decide, which will be better IMO10:34
ddejasaphi: and something not related directly with your workflow, but you are relating on nova information about the state of the host, and then evacuate all the vms, which may produce so called, 'zombie vms'10:35
saphibecause the nova.servers_evacuate have to need server and host. I can't find other action.10:37
ddejasaphi: you can watch the presentation that I coauthored with aspiers about the VM HA https://www.youtube.com/watch?v=lddtWUP_IKQ if you would like to gather more info :)10:38
ddejasaphi: oh, really? 'host' param should be optional10:38
ddejahm..10:38
ddejalet me check10:38
saphiI have been confused again :(10:43
saphiThat parameter is optional10:43
ddejaoh10:44
ddejaOK10:44
saphiOne problem still haven't solved. How to get all instances in all_tenants10:44
saphiwhen I input "search_opts" = {"all_tenant":1} to nova.servers_list action. It failed10:45
ddejasaphi: you want to list all tenants or all tenants from given host?10:46
saphiI want to get all instance from a host. because I have many tenants, So I have to evacuate all instances in other tenants10:47
ddejaoh, I see10:48
ddejasaphi: do you use admin privileges?10:50
saphiyes. I use admin account :D10:51
ddejahm, from what i remember admin just see vms from all tenants10:52
ddejadid you try without this option?10:52
saphihttp://pastebin.com/YYPsea5C. this is an error10:52
ddejasaphi: looking10:57
saphiyes11:00
openstackgerritMerged openstack/mistral-specs: Custom actions API specification  https://review.openstack.org/32576911:04
*** Ravikiran_K has quit IRC11:07
ddejasaphi: regarding vm listing problem, this works for me http://paste.openstack.org/show/524241/11:14
ddejaneed to drop, I'll be avaiable in an hour11:14
openstackgerritRenat Akhmerov proposed openstack/python-mistralclient: WIP: Add KeyCloak authentication  https://review.openstack.org/33594411:14
saphiI solved :D11:22
*** saphi has quit IRC11:25
*** mgershen has quit IRC11:35
*** mgershen has joined #openstack-mistral11:35
openstackgerritRenat Akhmerov proposed openstack/python-mistralclient: WIP: Add KeyCloak authentication  https://review.openstack.org/33594411:36
*** mgershen has quit IRC11:40
d0ugalrakhmerov: Hey11:42
d0ugalMistral has started using osc-lib already, when Ironic started to use it they were warned against doing so. You may want to read Dean's comment here: https://review.openstack.org/#/c/333311/11:43
d0ugalThe key bit being that they don't consider osc-lib ready yet for consumption: "So I think we are way ahead of ourselves in pushing osc-lib out yet.  It isn't finished."11:44
d0ugalIt might be okay, but I don't think it is wise for us to be using it yet.11:45
rakhmerovd0ugal: hi11:50
rakhmerovhm.. ok11:50
rakhmerovI'll read it11:50
d0ugalrakhmerov: I found out because I seen it in Mistral and started to add it to tripleoclient, then an ironic dev warned me11:50
d0ugalso I thought I should pass the warning along :)11:51
rakhmerovyeah, I see11:51
rakhmerovthanks11:51
rakhmerovwhat was the topic where you discussed that?11:51
rakhmerovooh, sorry11:51
rakhmerovyou said it's a comment11:51
openstackgerritXavier Hardy proposed openstack/mistral: Add support for custom YAQL functions  https://review.openstack.org/33480411:51
d0ugalYeah, Dean's comment on the ironic patch ^11:51
d0ugalDean is the PTL for python-openstackclient11:52
d0ugal(I assume he still is, I've not checked for a while)11:52
openstackgerritXavier Hardy proposed openstack/mistral: Add support for custom YAQL functions  https://review.openstack.org/33480411:53
*** vishwanathj has joined #openstack-mistral12:17
*** Ravikiran_K has joined #openstack-mistral12:32
*** jistr is now known as jistr|mtg14:31
*** tonytan4ever has joined #openstack-mistral14:45
*** rbrady has quit IRC14:50
*** rbrady has joined #openstack-mistral14:55
xavierhardyhi, is there any doc on ACL?14:56
*** rrecio has joined #openstack-mistral15:16
*** jistr|mtg is now known as jistr15:23
*** bobh has joined #openstack-mistral15:35
*** tonytan4ever has quit IRC16:06
*** gyee has joined #openstack-mistral16:36
*** jpich has quit IRC16:36
*** saphi has joined #openstack-mistral17:19
*** tonytan4ever has joined #openstack-mistral17:31
*** bobh has quit IRC17:36
*** tonytan4ever has quit IRC17:41
*** tonytan4ever has joined #openstack-mistral17:41
*** gyee has quit IRC17:48
*** Ravikiran_K has quit IRC17:57
*** Ravikiran_K has joined #openstack-mistral18:16
*** shardy has quit IRC18:41
*** saphi has quit IRC18:43
*** bobh has joined #openstack-mistral18:48
*** tonytan4ever has quit IRC19:37
*** tonytan4ever has joined #openstack-mistral19:50
*** tonytan4ever has quit IRC20:05
*** tonytan4ever has joined #openstack-mistral20:06
*** tonytan4ever has quit IRC20:12
*** tonytan4ever has joined #openstack-mistral20:13
*** tonytan4ever has quit IRC20:17
*** tonytan4ever has joined #openstack-mistral20:18
*** gyee has joined #openstack-mistral20:26
*** tonytan4ever has quit IRC21:21
*** vishwanathj has quit IRC21:22
*** pkoniszewski_ has joined #openstack-mistral21:28
*** pkoniszewski has quit IRC21:29
*** bobh has quit IRC21:45
*** _gryf has quit IRC21:45
*** catintheroof has joined #openstack-mistral22:03
*** Ravikiran_K has quit IRC22:10
openstackgerritShaik Apsar proposed openstack/mistral: [WIP] Add tacker actions in mistral  https://review.openstack.org/33182722:44
*** catintheroof has quit IRC23:18

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