*** rakhmerov has quit IRC | 00:03 | |
*** gokrokve has quit IRC | 00:04 | |
*** rakhmerov has joined #murano | 00:12 | |
*** rakhmerov has quit IRC | 00:13 | |
*** rakhmerov has joined #murano | 00:14 | |
*** rakhmerov has quit IRC | 00:18 | |
*** stanlagun has quit IRC | 00:19 | |
*** gokrokve has joined #murano | 00:26 | |
*** stanlagun has joined #murano | 00:29 | |
*** IlyaE has quit IRC | 00:33 | |
*** rakhmerov has joined #murano | 01:14 | |
*** rakhmerov has quit IRC | 01:19 | |
*** rakhmerov has joined #murano | 02:15 | |
*** rakhmerov has quit IRC | 02:20 | |
*** killer_prince has quit IRC | 03:08 | |
*** chandan_kumar has joined #murano | 03:11 | |
*** rakhmerov has joined #murano | 03:16 | |
*** rakhmerov has quit IRC | 03:20 | |
*** chandan_kumar has quit IRC | 03:35 | |
*** rakhmerov has joined #murano | 03:58 | |
*** chandan_kumar has joined #murano | 03:59 | |
*** chandan_kumar has quit IRC | 04:01 | |
*** chandan_kumar has joined #murano | 04:06 | |
*** lazy_prince has joined #murano | 04:11 | |
*** lazy_prince is now known as killer_prince | 04:11 | |
*** gokrokve_ has joined #murano | 04:50 | |
*** gokrokve has quit IRC | 04:53 | |
*** IlyaE has joined #murano | 04:57 | |
*** IlyaE has quit IRC | 05:05 | |
*** killer_prince has quit IRC | 05:14 | |
*** IlyaE has joined #murano | 05:49 | |
*** saju_m has joined #murano | 06:21 | |
openstackgerrit | Stan Lagun proposed a change to stackforge/murano-api: Use native oslo.messaging notification dispatcher https://review.openstack.org/84265 | 06:31 |
---|---|---|
*** IlyaE has quit IRC | 06:51 | |
igormarnat | Howdy | 07:01 |
*** IlyaE has joined #murano | 07:17 | |
ogelbukh | igormarnat: hi | 07:27 |
ogelbukh | how're you doing? | 07:27 |
*** IlyaE has quit IRC | 07:37 | |
*** gokrokve_ has quit IRC | 07:40 | |
*** gokrokve has joined #murano | 07:41 | |
*** gokrokve has quit IRC | 07:45 | |
igormarnat | It depends on aspect. Speaking of weather, it's not Hawaii | 07:55 |
igormarnat | Though speaking of work, it's a lot of fun:) | 07:55 |
*** killer_prince has joined #murano | 08:02 | |
*** killer_prince has quit IRC | 08:23 | |
*** saju_m has quit IRC | 08:43 | |
*** saju_m has joined #murano | 08:56 | |
*** gokrokve has joined #murano | 09:16 | |
*** gokrokve has quit IRC | 09:21 | |
openstackgerrit | A change was merged to stackforge/murano-api: Use native oslo.messaging notification dispatcher https://review.openstack.org/84265 | 09:24 |
*** chandankumar_ has joined #murano | 09:26 | |
*** gokrokve has joined #murano | 09:47 | |
*** gokrokve has quit IRC | 09:52 | |
*** chandankumar_ has quit IRC | 09:54 | |
openstackgerrit | Timur Sufiev proposed a change to stackforge/murano-dashboard: Implement creating new App from other App Creation dialog. https://review.openstack.org/83039 | 10:06 |
*** gokrokve has joined #murano | 10:48 | |
*** gokrokve has quit IRC | 10:53 | |
*** ruhe_ has joined #murano | 11:13 | |
igormarnat | lol:) | 11:17 |
dteselkin | mm ? :) | 11:18 |
*** saju_m has quit IRC | 11:44 | |
*** gokrokve has joined #murano | 11:47 | |
igormarnat | freenode+ sounds good | 11:47 |
*** gokrokve has quit IRC | 11:52 | |
*** ruhe_ has left #murano | 11:56 | |
*** saju_m has joined #murano | 11:56 | |
*** saju_m has quit IRC | 12:02 | |
*** asalkeld has joined #murano | 12:17 | |
*** rakhmerov has quit IRC | 12:17 | |
*** asalkeld has left #murano | 12:19 | |
*** saju_m has joined #murano | 12:21 | |
*** saju_m has quit IRC | 12:23 | |
*** sergmelikyan has quit IRC | 12:38 | |
*** saju_m has joined #murano | 12:39 | |
openstackgerrit | Ekaterina Fedorova proposed a change to stackforge/murano-api: Implement upload call to the repository API https://review.openstack.org/84099 | 12:41 |
*** gokrokve has joined #murano | 12:47 | |
*** gokrokve has quit IRC | 12:52 | |
ruhe | katyafervent: ping | 12:54 |
katyafervent | Hi) | 12:54 |
katyafervent | what's up | 12:54 |
ruhe | hi :) | 12:54 |
ruhe | i broke my brains in https://review.openstack.org/#/c/84099/3/muranoapi/api/v1/catalog.py :) | 12:54 |
katyafervent | *^_^* | 12:55 |
ruhe | maybe there is an indentation problem? shouldn't try/catch block where you create package_to_upload belong to the with() block above? | 12:56 |
katyafervent | I updated it btw. Please leave comments how make it better) | 12:56 |
ruhe | it's still the same | 12:57 |
ruhe | ok, i'll leave my comments, but... | 12:57 |
ruhe | i just wanted to say that touching tempf outside of the block where it was created is not the right way to go | 12:58 |
katyafervent | ok. but I handing it's deletion | 12:59 |
katyafervent | * I'm | 12:59 |
*** saju_m has quit IRC | 13:15 | |
*** saju_m has joined #murano | 13:28 | |
openstackgerrit | Stan Lagun proposed a change to stackforge/murano-api: Adoption of pre-0.5 environment API to MuranoPL format https://review.openstack.org/83047 | 13:38 |
*** asalkeld has joined #murano | 13:41 | |
*** asalkeld has left #murano | 13:41 | |
*** gokrokve has joined #murano | 13:47 | |
openstackgerrit | Sergey Murashov proposed a change to stackforge/murano-api: Refactor murano functional tests https://review.openstack.org/84435 | 13:49 |
*** gokrokve_ has joined #murano | 13:49 | |
*** gokrokve has quit IRC | 13:52 | |
*** gokrokve_ has quit IRC | 13:54 | |
openstackgerrit | Ekaterina Fedorova proposed a change to stackforge/murano-api: Implement upload call to the repository API https://review.openstack.org/84099 | 14:01 |
openstackgerrit | Ekaterina Fedorova proposed a change to stackforge/murano-api: Implement upload call to the repository API https://review.openstack.org/84099 | 14:03 |
*** saju_m has quit IRC | 14:14 | |
*** rakhmerov has joined #murano | 14:19 | |
*** sjmc7 has joined #murano | 14:24 | |
*** lazy_prince has joined #murano | 14:27 | |
*** lazy_prince is now known as killer_prince | 14:27 | |
*** rakhmerov has quit IRC | 14:29 | |
*** rakhmero_ has joined #murano | 14:29 | |
ruhe | i'd be happy to get some reviews in https://review.openstack.org/#/c/83601/ | 14:31 |
*** killer_prince has quit IRC | 14:34 | |
*** lazy_prince has joined #murano | 14:34 | |
*** lazy_prince is now known as killer_prince | 14:35 | |
*** rakhmero_ has quit IRC | 14:43 | |
*** gokrokve has joined #murano | 14:47 | |
*** gokrokve has quit IRC | 14:51 | |
*** gokrokve has joined #murano | 14:53 | |
*** IlyaE has joined #murano | 14:55 | |
*** TravT has joined #murano | 15:04 | |
*** chandankumar_ has joined #murano | 15:12 | |
openstackgerrit | A change was merged to stackforge/murano-api: Adoption of pre-0.5 environment API to MuranoPL format https://review.openstack.org/83047 | 15:33 |
openstackgerrit | A change was merged to stackforge/murano-dashboard: Implement creating new App from other App Creation dialog. https://review.openstack.org/83039 | 15:39 |
openstackgerrit | Timur Sufiev proposed a change to stackforge/murano-dashboard: Horizontally align tiles with different App description length. https://review.openstack.org/84481 | 15:52 |
*** julienve_ has joined #murano | 15:55 | |
*** julienve_ is now known as julienvey | 15:55 | |
*** stanlagun has quit IRC | 16:37 | |
*** stanlagun has joined #murano | 16:39 | |
*** rakhmerov has joined #murano | 16:40 | |
*** rakhmerov has quit IRC | 16:53 | |
*** sergmelikyan has joined #murano | 16:59 | |
*** TravT has quit IRC | 17:40 | |
*** TravT has joined #murano | 17:44 | |
*** chandankumar_ has quit IRC | 17:46 | |
*** julienvey has quit IRC | 17:47 | |
*** rakhmerov has joined #murano | 17:49 | |
*** rakhmerov has quit IRC | 17:54 | |
gokrokve | Agent installation is good but we need to understand how to deliver this agent code to VM. | 18:01 |
sergmelikyan | gokrokve, regarding installation - most simple way is pip and python eggs | 18:02 |
sergmelikyan | but also we may have specs for deb/rpm as we have for other murano components | 18:02 |
gokrokve | For linux. You will need to install python on windows :-) | 18:02 |
sergmelikyan | gokrokve, it does not matter how many software are you installing with cloud-init, we can install python and then install our agent | 18:03 |
* sergmelikyan talking about windows | 18:03 | |
ruhe | i think that cloud-installable agent should be just an option | 18:04 |
sergmelikyan | And I think it will be really nice, opposing heavy windows bootstrapping described in our manual | 18:04 |
gokrokve | ok. I suppose Murano will still support both prebacked image and bootstrap via cloud-init | 18:04 |
gokrokve | ok | 18:04 |
sergmelikyan | ruhe, sure, but i think it will be most popular use-case | 18:04 |
ruhe | and pre-built images should be the main way to speak with VMs | 18:04 |
sergmelikyan | ruhe ? | 18:05 |
ruhe | sergmelikyan: from my expirience in Sahara/Savanna, most of users prefer to have as much software to be installed as possible | 18:05 |
ruhe | * pre-installed | 18:05 |
ruhe | they want the deployment process to be as fast as possible | 18:06 |
ruhe | and they don't see any problem in having pre-built custom images | 18:06 |
sjmc7 | as a relative newcomer to murano, how much crossover is there between the murano agent and heat's software config functionality? | 18:07 |
sergmelikyan | ruhe, it is true when you are talking about savanna that deploys Hadoop. But in Murano same image may be used for deployment of thouthands of applications. You could not have so many images pre-built, or one with all pre-requisites | 18:07 |
sergmelikyan | sjmc7, they solve one problem but with slightly different approach. Application developer can choose what way to use, since it is possible to do provisioning using software config or murano-agent at the same time/ | 18:09 |
sjmc7 | ok. for our use cases, custom images is a problem, so we'd either be looking at using heat or installing the agent via cloudinit. will the murano agent still require its own rabbitmq? | 18:11 |
sergmelikyan | One important thing about murano-agent vs software-config is that software config is iterative, while software-config defines 'state of the world' | 18:12 |
sergmelikyan | sjmc7, currently we use own rabbit mq for security reasons (we don't want to expose rabbitmq used by OpenStack to VM's, to protect our Cloud from abuse) | 18:13 |
sjmc7 | in many cases it feels like cloudinit on its own would be sufficient | 18:13 |
sjmc7 | yes sergmelikyan - the security aspect makes good sense, but it's another moving part | 18:14 |
sergmelikyan | sjmc7, there is no problem to share one instance of rabbit mq and leave security and abuse issues on rabbitmq features (vhosts, certificates) and networking level | 18:16 |
sergmelikyan | At this moment we don't have clean solution/vision about how to not use rabbit mq in case of murano-agent involved. It may change in next release | 18:17 |
sjmc7 | it's more that we'd rather not have the direct communication with rabbitmq | 18:17 |
sjmc7 | ok, got it | 18:17 |
ruhe | sjmc7: will you install/provision local package mirror (or maybe just web server) to host agent packages? | 18:17 |
sjmc7 | ruhe - i believe so, yes | 18:18 |
ruhe | ok, i'll make a note about it | 18:18 |
sjmc7 | at least for debian packages, not sure about a pypi mirror | 18:18 |
ruhe | deb/rpm is a better option to distribute software compared to pypi | 18:19 |
sergmelikyan | sjmc7, we evaluating Marconi, but Marconi does not satisfy our requirements yet. Changing murano <-> agent transport is not so hard, so we can move to another transpport once we found it | 18:19 |
sjmc7 | got it. i'm just trying to get a feel for how complicated deployment's likely to be | 18:19 |
sjmc7 | extra maintenance == extra work for me | 18:19 |
sjmc7 | heat's all REST, right? | 18:20 |
sergmelikyan | sjmc7, what do you mean by 'all REST'? | 18:20 |
sjmc7 | over http? | 18:20 |
sjmc7 | i may be mistaken | 18:20 |
ruhe | Heat just sends pre-defined list of commands to VM via cloud-init | 18:21 |
ruhe | but in some cases it's not enough where software requires interactive configuration | 18:21 |
sjmc7 | ah, ok, my bad. i was under the impression it communicated more than that | 18:21 |
sjmc7 | how often is it necessary for murano to communicate with VMs after creation? | 18:22 |
sjmc7 | and please forgive my stupid questions | 18:22 |
ruhe | sjmc7: we will not forgive your questions. we encourage you to do so :) | 18:22 |
sergmelikyan | sjmc7, there is no nessesary communication between murano and VM, all communication commands explicitly issues in MuranoPL by Application Writer | 18:23 |
sergmelikyan | *issued | 18:23 |
sergmelikyan | So there is may be an application that does not require any communication beyond cloud-init done by Heat (for example) | 18:23 |
gokrokve | sergmelikyan: That is not the real case. Murano engine has to communicate with VM. | 18:24 |
sergmelikyan | gokrokve & | 18:24 |
sergmelikyan | gokrokve, ? | 18:24 |
gokrokve | The typical scenario is when you are installing a complex application like MS SQL cluster. | 18:24 |
ruhe | guys, i have to leave to catch the last bus home, ttl | 18:24 |
sjmc7 | thanks gokrokve, that was what i was getting at - what cases are there where the agent receives messages after creation that couldn't have been in cloudinit | 18:24 |
sjmc7 | thanks ruhe | 18:25 |
gokrokve | Murano engine sends first commands to cluster nodes then waits for reply from nodes and continue to send new commands in proper sequence. | 18:25 |
sergmelikyan | gokrokve, of we are talking about possibility - definitely Murano does not require communication with VM, but in real case - almost all our existing application (on 0.4) require more then one calls to VM (changing password, installing software1, then soft2 and so on) | 18:26 |
gokrokve | Like Active directory setup. You have to install Primary domain controller then join the domain on secondary controller and setup a secondary controller AD. | 18:26 |
gokrokve | Yes. For limited set of simple app you don't need to send different commands multiple time. | 18:27 |
sergmelikyan | One of the major gain with murano and murano-agent that you can done cross-instance orchestration | 18:27 |
sergmelikyan | For example do something on VM2 depending on result from VM1 | 18:27 |
gokrokve | sergmelikyan: Yes. This is the valuable part. | 18:27 |
sjmc7 | ok, that makes sense. it seems though that you've then got a communication channel via rabbitmq that won't be used beyond the initial setup (which as you've said may be timing dependent and complex) | 18:27 |
sjmc7 | heat has the concept of signalling which seems related | 18:28 |
gokrokve | sjmc7: No. Murano will use this channel any time it need to execute something on VM. | 18:28 |
stanlagun | sjmc7: it may be used after initial setup | 18:28 |
sergmelikyan | sjmc7, yes, but with signaling you don't have 'output' | 18:28 |
sjmc7 | ok gokrokve - so where you later add other services for instance that require work on existing ones? | 18:29 |
gokrokve | Murano application can have actions like "backup", "scaleUP". User can call this action on depoyed app and Murano will use existing communication channel to invoke some comand on VM. | 18:30 |
gokrokve | sjmc7: Technically you can add some new components on deployed VM instance, yes. | 18:30 |
gokrokve | The goal is to be able to communicate with VM at any time while it is up and running. | 18:31 |
sergmelikyan | or reconfigure existing, depending on adding new service to environment | 18:31 |
gokrokve | Yes. | 18:32 |
sjmc7 | ok, i see, thanks | 18:32 |
sergmelikyan | Regarding how easy is to deploy murano - murano is OpenStack service and does not require nothing more than other OpenStack service, for example glance. Second RabbitMQ is more like improvement than requirement (i could no asses how it is valuable from security point of view) | 18:34 |
sjmc7 | for us we'd have to deploy the second rabbitMQ, we wouldn't be allowed direct access to the core one | 18:35 |
sjmc7 | but i get your point | 18:36 |
*** tsufiev_ has joined #murano | 18:36 | |
*** tsufiev_ has quit IRC | 18:36 | |
sergmelikyan | sjmc7, it is reasonable, Fuel does it too. And I understand that less moving parts - easy deployment and support | 18:36 |
sjmc7 | i'm not trying to argue, i just know someone's gonna ask me all this stuff :) | 18:37 |
sergmelikyan | :) | 18:37 |
sergmelikyan | And we trying to answer this question from all perspective that we know :) | 18:37 |
sergmelikyan | *questions | 18:37 |
sergmelikyan | *perspectives | 18:38 |
stanlagun | :) | 18:41 |
stanlagun | this may be useful: https://wiki.openstack.org/wiki/Murano/UnifiedAgent | 18:42 |
sergmelikyan | stanlagun, sjmc7 btw great document about how our agent work, and why we done it in this way | 18:46 |
* sergmelikyan totally forgot about that page | 18:47 | |
stanlagun | Anyway agent is an optional part of Murano. One may want to use Marconi-based agent or HOT software components or cloud init and this is perfectly fine with Murano | 18:48 |
sjmc7 | yep, that page is a good description, thanks | 18:49 |
*** rakhmerov has joined #murano | 18:50 | |
*** julienvey has joined #murano | 18:58 | |
*** julienvey has quit IRC | 19:02 | |
*** rakhmerov has quit IRC | 19:02 | |
*** gokrokve_ has joined #murano | 19:10 | |
*** gokrokve has quit IRC | 19:14 | |
*** gokrokve_ has quit IRC | 19:20 | |
*** julienvey has joined #murano | 19:34 | |
*** TravT has quit IRC | 19:45 | |
*** gokrokve has joined #murano | 19:46 | |
*** TravT has joined #murano | 19:54 | |
*** rakhmerov has joined #murano | 19:59 | |
*** julienvey has quit IRC | 19:59 | |
*** rakhmerov has quit IRC | 20:03 | |
*** julienvey has joined #murano | 20:08 | |
openstackgerrit | Stan Lagun proposed a change to stackforge/murano-api: statistics renamed to request_statistics https://review.openstack.org/84540 | 20:16 |
*** IlyaE has quit IRC | 20:44 | |
*** rakhmerov has joined #murano | 21:00 | |
*** rakhmerov has quit IRC | 21:05 | |
*** IlyaE has joined #murano | 21:07 | |
*** ruhe has quit IRC | 21:54 | |
*** openstackgerrit has quit IRC | 21:54 | |
*** ruhe has joined #murano | 21:55 | |
*** rakhmerov has joined #murano | 22:01 | |
*** rakhmerov has quit IRC | 22:05 | |
*** julienvey has quit IRC | 22:12 | |
*** TravT has quit IRC | 22:36 | |
*** rakhmerov has joined #murano | 23:01 | |
*** rakhmerov has quit IRC | 23:06 | |
*** gokrokve has quit IRC | 23:12 | |
*** sjmc7 has quit IRC | 23:35 | |
*** asalkeld has joined #murano | 23:40 | |
*** asalkeld has left #murano | 23:52 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!