Tuesday, 2016-09-20

*** uck has quit IRC00:08
*** spzala has quit IRC01:02
*** spzala has joined #openstack-heat-translator03:02
*** spzala has quit IRC03:07
*** tbh has joined #openstack-heat-translator03:24
*** tbh has quit IRC03:49
*** spzala has joined #openstack-heat-translator11:34
*** tbh has joined #openstack-heat-translator11:57
*** bobh has joined #openstack-heat-translator11:59
*** bobh has quit IRC12:06
*** spzala has quit IRC12:28
*** vishwanathj has joined #openstack-heat-translator12:57
*** tbh has quit IRC13:05
*** tbh has joined #openstack-heat-translator13:06
*** spzala has joined #openstack-heat-translator13:11
*** spzala has quit IRC13:11
*** spzala has joined #openstack-heat-translator13:11
*** mvelten has joined #openstack-heat-translator13:22
*** openstackstatus has joined #openstack-heat-translator13:36
*** ChanServ sets mode: +v openstackstatus13:36
-openstackstatus- NOTICE: OpenStack Infra now has a Twitter bot, follow it at https://twitter.com/openstackinfra13:43
spzalamvelten: tbh: hi13:50
spzalawish if Kanagaraj was here on the channel13:50
mveltenhi there13:51
tbhspzala, hi13:51
tbhspzala, mvelten the approach you mentioned sounds good13:53
spzalahi guys13:53
spzalatbh: cool, thanks13:53
spzalamvelten: did you look at my latest comment on the patch?13:53
mveltenyes fine for me13:54
spzalamvelten: awesome, thanks!!13:55
mveltenwith "test with a real heat deployment" added :)13:55
spzalamvelten: :) absolutely +2 for that13:55
spzalatbh: please test it once you remove files section. I will test it too - only thing I don't have in my env is Aodh/Ceilometer.13:56
spzalatbh: mvelten: I actually tested real heat deployment with tbh's previous patch (before he added files section) and that's how I found out two problems13:56
spzalaI ended up this two errors:13:57
spzalaERROR: Property error: : resources.asg_group.properties: : Unknown Property resources13:57
spzalaERROR: Property error: : resources.asg_group.properties: : Unknown Property default_instances13:57
spzalaand the fix is now added in the tbh's patch13:57
spzalaso totally agree that unless we try real test deployment, it's always shaky13:58
spzalathose two errors weren't part of tbh's changes though, it's original bad code but I was told that 'real test deployment' was done in the initial patch on scaling .. oh well13:59
spzala(trying to see if we can have Kanagaraj join us and see if he is fine from Tacker's perspective )14:02
tbhspzala, will test and push14:02
spzalatbh: prefect! thanks!14:03
*** bobh has joined #openstack-heat-translator14:42
*** KanagarajM has joined #openstack-heat-translator15:00
spzalaKanagarajM: Hi15:01
spzalayou made it :-)15:01
spzalathanks for joining15:01
KanagarajMspzala, :)15:01
KanagarajMspzala, hey, nope.15:02
KanagarajMlooking into15:02
KanagarajMspzala, for generating the nested template, may you could give an option for the output dir and generate the templ and nested tmp in that dir15:04
spzalatbh: just fyi, KanagarajM is here15:04
spzalaKanagarajM: that's what I actually suggested but we thought for quick fix let's use PWD?15:04
KanagarajMand while using like sdk in tacker, may be you could give addition python api, to get the nested tmpl(s) and main tmpl15:04
spzala KanagarajM: we have --output-file option for CLI15:05
spzala KanagarajM: and if that option is provided my thought was to save the translated and nested templates there15:05
spzalaif that option is not provided than we simply store in the $PWD15:05
KanagarajMyes, that sounds better15:05
spzalabut with that patch, we assumed the second scenario only.15:06
KanagarajMand for tacker, pls address the option suggested here, it would be great :)15:06
spzala KanagarajM: Cool15:06
tbhKanagarajM, which option you are talking about?15:06
spzala KanagarajM: sure, so hopefully we should be able to cover both this options15:06
mveltenmy patch adresses the needs so we can merge that with $PWD and see later15:06
spzalamvelten: that's awesome15:06
spzala KanagarajM: so for now, can we just go with $PWD for heat-translator upcoming 0.6.0 release if that fits in Tacker's case?15:07
KanagarajMtbh, in tacker, we use translator to get the HOT tmpl, as part of it, i would like to get the nested tmpl in python api15:07
mveltenit also allows to switch to only output one template, with substack embedded as text inside (the output is ugly but it works)15:07
spzala KanagarajM: and then with mvelten's patch both options will be available with master repo15:08
KanagarajMspzala, but is files in the HOT template  valid one ?15:08
tbhKanagarajM, for that purpose, we have added "files" action, which can't be deployed directly15:08
spzala KanagarajM: yes, absolutely15:08
tbhKanagarajM, it is not valid15:08
spzala KanagarajM: 'files' is not valid15:08
spzala KanagarajM: so we removing it15:08
KanagarajMyes, i am concerned on it15:09
spzala KanagarajM: sorry, i thought you mean translated files :-)15:09
spzala KanagarajM: yup, totally agree and you can see comments related to your comments.. and the final approach (the last comment) that we will take out 'files' section15:09
tbhKanagarajM, for tacker, if we need both templates using translator we have to output both yaml files15:09
spzala KanagarajM: so with that, it will generate nested templates (e.g. asg_reg.yaml on it's own)15:10
KanagarajMspzala, 1 sec,15:10
spzalamvelten: that's awesome, so are you saying that with one template, it will be two 'stacks' like two stack heat resources?15:10
spzala KanagarajM: sure15:10
KanagarajMspzala, pls refer here https://github.com/openstack/tacker/blob/master/tacker/vnfm/infra_drivers/openstack/openstack.py#L279-L28115:12
KanagarajMi would expect to provide something like translator.get_main_template() ... translator.get_nested_templates()15:12
KanagarajM2nd method would give me the map15:13
spzalaKanagarajM: in that case,15:13
KanagarajMwhile first one give the pyhton dic15:13
mveltenmine does that :D the output is a map15:13
KanagarajMand make the backward compatible on existing api, so that it does not break tacker15:13
KanagarajMmvelten, 1 sec15:14
mveltenI think the api didnt change ? with tbh patch I think ?15:14
spzalamvelten: that's true, no api change with tbh's patch15:14
KanagarajMmvelten, https://review.openstack.org/#/c/371063/8/translator/hot/tosca_translator.py here there is no api provided as suggested above15:14
spzalamvelten: but it won't satisfy the need for Kanagaraj15:15
tbhspzala, mvelten right no change in api15:15
spzalatbh: thanks, yup agree15:15
KanagarajMspzala, i mean, when translator.translate() is called, let it behave like earlier15:16
spzalaKanagarajM: with the current approach is like, you need to look into $PWD and get the nested template but I agree with you that for your need to programatically using nested template we need APIs15:16
tbhKanagarajM, spzala  why can't we pass output-dir arg to Translator and then read the templates from the dir15:16
KanagarajMspzala, yes, for tacker that is important15:16
mveltenwe need tacker integration for now ?15:17
tbhKanagarajM, spzala  probably we can follow the naming convention to name "main.yaml" for parent template and add remaining files to "files" in tacker code15:17
KanagarajMtbh, it would be an workarround than a solution. and for tacker, translator is like sdk and not a cli/tool15:17
mveltenbecause we were planning to introduce this properly with my patch (multi files support)15:17
mveltenthis is a workaround15:18
mveltenso if you dont need it right now in tacker I would wait for my patch to come in before doing anything15:19
tbhKanagarajM, and at the same time, heat-translator has to keep other projects requirements also in mind15:19
spzalaKanagarajM: true, but can we go with workaround for now? and as mvelten: said with his patch which is up for review it will be a proper solution15:19
KanagarajMtbh, yes, so i would prefer to keep as sdk , which would help tacker, heat-translaotr cli and other dependent15:20
spzalathe reason we are pushing for workaround vs. mvelten: patch is review time needed for his patch etc.15:20
spzalaKanagarajM: Tacker's not using heat-translator from master right?15:20
spzalai.e. it uses pypi of translator?15:21
mveltenI need to change a lot of stuffs in my patch to handle correctly the various scaling possibilities added in the last 2 months, I can't do that properly in a rush for newton15:21
spzalafor Tacker to use new pypi of translator with fix (proper fix or workaround) we only have little time for newton cycle it seems before tacker's newton release15:21
tbhspzala, yes pypi15:22
KanagarajMspzala, i am ok to go without this patch in place, as 1. tacker does the translator job now and it works as expected 2. it does not break any projects. so if we are all fine, we could push this support to ocata.15:22
spzalamvelten: agree, rushing may break other things and we shouldn't rush15:22
mveltenit will only not work if you pass a tosca that would trigger the generation of multiples files15:22
KanagarajMmvelten, spzala  yes. exactly +115:23
spzalaKanagarajM: that's great15:23
spzalayes +2 from me15:23
spzalawell, that's desired conclusion of our discussion15:23
spzala:-) thank you so much guys KanagarajM: tbh: mvelten:15:24
spzalatbh: so it sounds good with you too?15:24
KanagarajMspzala, tbh, mvelten thank you.15:24
tbhspzala, you mean to abandon the current patch?15:25
spzalatbh: no no15:25
spzalatbh: your patch is valuable15:25
spzalaKanagarajM: yrw, thank you15:25
spzalatbh: your patch is great to fix the bug15:25
tbhspzala, I mean to get it in ocata15:26
spzalatbh: Tacker probably can't use it as such so they will keep their workaround for now15:26
spzalatbh: no, we will merge it today for only after get mvelten: and kanagaraj review15:26
spzalaand then we will do heat-translator pypi so the patch changes go in it15:26
KanagarajMspzala, tbh , yes, at this moment, i don't want to enable anything related to OS::Heat::AutoScaling, OS::Heat::ScalingPolicy in translaotr15:27
KanagarajMas it might regret tacker.15:27
spzalasee, even though tacker can't use the patch the bug Kanagaraj reported is valid15:27
spzalaand your patch provides a fix for it, it's just that Tacker won't be able to use the fix15:27
tbhspzala, agree15:27
spzalaKanagarajM: hmmm.. little confused now15:28
spzalaKanagarajM: we already have some autoscaling support in translator which is broken as you know, and tbh: patch fixes it15:28
spzalaKanagaraj: so that's ok to have in newer release of translator for tacker right? it's just Tacker won't use it as such15:29
spzalaKanagarajM: as this at least enables user to use scaling non-programatically15:29
KanagarajMspzala, sure, as far as it does not break the tacker functioanlity, i am very happy to enable it in translator,15:30
spzalaKanagarajM: sure, with this patch no change in API or anything so it shouldn't break anything on Tacker's side15:31
spzalaKanagarajM: I usually get bobh: help to verify that translator/parser not breaking Tacker before releasing new PyPI15:31
spzalabobh: hi  .. hope I can count on your help this time too for verification :-)15:32
KanagarajMnow tacker uses earlier version of pypi translator15:32
mveltenok I get it now, the problem is that the fix cant be used right now by tacker because it needs to handle the multiple files case, what I didnt get is that the problem was actually reported by tacker :)15:32
spzalaKanagarajM: so if there is anything that breaks Tacker we will undo those changes15:32
spzalamvelten: :) hehe15:32
mveltenit should not break, it will not just fix the pb for them15:32
spzalamvelten: exactly15:32
KanagarajMspzala, i think its other way, translator would pull out :)15:33
spzalaKanagarajM: sorry didn't get it but what I was saying is we will remove any changes in translator if it breaks Tacker - that sounds good? I guess we are on same page there? :-)15:34
KanagarajMspzala, yeah ;)15:34
spzalaKanagarajM: ;) OK15:34
spzalaAlright, so cool. The conclusion is:15:35
spzala1. Go with tbh: patch once he remove 'files' section, enable $PWD nested template etc. per discussion15:35
spzala2. Create release of heat-translator pypi 0.6.0 and if it break tacker in anyway we will revert15:36
spzala3. mvelten: patch during ocata will provide a proper fix for Tacker15:36
KanagarajMspzala, +1 :)15:37
tbhspzala, yup15:37
spzalaKanagarajM: mvelten: tbh: thanks!! tbh: way to go :-)15:37
spzalaPlease review the patch once tbh: make changes15:38
KanagarajMspzala, yrw. thanks.15:39
spzalaKanagarajM: :) :)15:40
KanagarajMspzala, :)15:40
bobhspzala: I'll try :-)16:00
spzalabobh: :-) THANKS!16:03
spzalabobh: it's been a big help every time with release. once tbh: patch merge I will reach out to you16:03
bobhspzala: no problem - I'm still trying to test the heat template version update16:10
spzalabobh: ah, nice that you are looking at it. The patch that's under review, I think that one is important one but we can push it next release16:10
spzalabobh: as it's last minute now and not sure if it can break anything16:11
bobhspzala: might be a good idea16:11
spzalabobh: cool, thanks16:11
openstackgerritbharaththiruveedula proposed openstack/heat-translator: Add nested template support for auto scaling  https://review.openstack.org/37106317:02
*** KanagarajM has quit IRC17:02
spzalasridhar_ram: thanks, sorry missed the ping due to lunch but I am glad that you are fine with the current approach we are taking17:08
sridhar_ramspzala: yes, can't take risk at this stage17:21
*** uck has joined #openstack-heat-translator17:38
*** tbh has quit IRC17:52
*** bobh_ has joined #openstack-heat-translator18:33
*** bobh has quit IRC18:33
*** bobh_ has quit IRC20:24
*** bobh has joined #openstack-heat-translator20:24
*** bobh has quit IRC20:49
*** bobh has joined #openstack-heat-translator20:50
*** bobh has quit IRC21:23
*** spzala has quit IRC21:43
*** bobh has joined #openstack-heat-translator23:05
*** bobh has quit IRC23:43
*** bobh has joined #openstack-heat-translator23:47

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