| *** Qiming has joined #openstack-mistral | 01:54 | |
| Qiming | morning/evening | 01:54 |
|---|---|---|
| Qiming | any mistral cores online? | 01:55 |
| kong_ | Qiming: afternoon :-) | 01:55 |
| Qiming | hi, kong_ | 01:55 |
| Qiming | I'm looking at mistral support from openstacksdk | 01:55 |
| Qiming | it is not there | 01:55 |
| kong_ | yeah, as i expected :-) | 01:55 |
| Qiming | we want in enabled from openstacksdk, so that people can interact with mistal without having to install its client | 01:56 |
| Qiming | s/in/it | 01:56 |
| Qiming | when looking at the mistral code, it looks different from other services | 01:56 |
| Qiming | the first thing is about version discovery | 01:57 |
| Qiming | according to the guideline from API-wg: http://git.openstack.org/cgit/openstack/api-wg/tree/guidelines/microversion_specification.rst#n140 | 01:57 |
| Qiming | a service is supposed to return a {'versions': [...]} JSON | 01:57 |
| Qiming | and the 'link' key should be plural | 01:57 |
| Qiming | is it okay for me to try fix it? | 01:58 |
| kong_ | Qiming: feel free to do that | 01:59 |
| Qiming | alright | 01:59 |
| kong_ | you are always welcomed to fix anything | 01:59 |
| Qiming | just want to ensure that I'm not working on something that team here already has a decision of NOT doing it, :) | 01:59 |
| kong_ | the other thing i need to mention is mistral client doesn't support session, not sure it's another roadblock for you | 02:00 |
| Qiming | if we support mistral resources directly from openstacksdk, we won't depend on mistralclient | 02:00 |
| kong_ | Qiming: that's great | 02:01 |
| Qiming | the goal of the sdk is to "kill" all service client libraries | 02:01 |
| Qiming | so that the CLI can go to openstackclient, the library can go to openstacksdk | 02:01 |
| Qiming | :) | 02:01 |
| kong_ | Qiming: just curious, what's the difference between oepnstacksdk and openstackclient? | 02:01 |
| Qiming | an ideal vision which may never come true | 02:01 |
| kong_ | you already answered that | 02:01 |
| kong_ | openstackclient will call openstacksdk, is that the roadmap of openstackclient team? | 02:02 |
| kong_ | Qiming ^^ | 02:02 |
| Qiming | yes, but openstackclient moved very fast | 02:03 |
| Qiming | for most services, it is still calling their clients | 02:03 |
| kong_ | ok | 02:03 |
| kong_ | i see | 02:03 |
| Qiming | there are some migration works recently, e.g. rebasing neutron resources onto sdk | 02:03 |
| kong_ | the first step is to improve sdk, and if it's mature enough, you can convince openstackclient to use it | 02:04 |
| Qiming | yes | 02:04 |
| kong_ | mature, i mean it contains most of openstack service resources | 02:04 |
| Qiming | right, that means a lot of resources to be modeled, :) | 02:04 |
| kong_ | Qiming: sounds like a long way to go :-) | 02:05 |
| Qiming | yep | 02:05 |
| Qiming | and the right thing to do | 02:05 |
| kong_ | Qiming: btw, what's the motivation of you doing that? | 02:06 |
| kong_ | adding mistral support in openstacksdk? | 02:06 |
| Qiming | a user story of VM HA | 02:07 |
| kong_ | Qiming: awesome | 02:07 |
| Qiming | you know I'm working on senlin | 02:07 |
| kong_ | haha，我知道 | 02:07 |
| Qiming | so far we have failure detection and failure recovery implemented already | 02:07 |
| Qiming | a missing link is about recovery of user apps | 02:08 |
| Qiming | and we don't know whatelse we can do ... | 02:08 |
| Qiming | along the recovery process, we want to insert a mistral workflow | 02:08 |
| kong_ | yeah, mistral is good candidate for that job | 02:08 |
| Qiming | users can define and manage their custom workflows and then tell us | 02:08 |
| Qiming | "when you are recovering my VMs in this cluster, please invoke this workflow when a VM is back to life" | 02:09 |
| Qiming | it is an extension to the builtin recovery process | 02:09 |
| kong_ | which means, users should install mistral first, right? | 02:10 |
| kong_ | sorry, the service provider | 02:10 |
| Qiming | yes, they have to install mistral to enable this feature | 02:10 |
| kong_ | ok, in order to interact with mistral, you choose to use openstacksdk in senlin, rather than all kinds of clients of other opensatck services? | 02:11 |
| Qiming | yes, we interact with nova, keystone, neutron, ceilometer, heat via sdk | 02:12 |
| kong_ | ok, fair enough | 02:12 |
| Qiming | and openstacksdk then becomes the single dependency for us | 02:12 |
| Qiming | the management is drastically simplified | 02:12 |
| Qiming | we don't need to chase the client version changes | 02:13 |
| kong_ | sorry for the inconvenience brought by mistral, hah | 02:13 |
| Qiming | and..., in an ideal case, they ALL should be bead in near future | 02:13 |
| Qiming | tox -e p27 produced 13 errors ... | 02:15 |
| kong_ | Qiming: sorry was distracted by internal stuff | 02:18 |
| kong_ | Qiming: can you paste the error? | 02:18 |
| kong_ | it shouldn't be that case since py27 jenkins job is green :-) | 02:19 |
| Qiming | FAIL: mistral.tests.unit.api.test_cors_middleware.TestCORSMiddleware.test_invalid_cors_get_request | 02:21 |
| Qiming | File "mistral/tests/unit/engine/test_cron_trigger.py", line 112, in test_workflow_without_auth | 02:23 |
| Qiming | self.assertEqual(1, len(next_triggers)) | 02:23 |
| Qiming | testtools.matchers._impl.MismatchError: 1 != 0 | 02:23 |
| Qiming | Traceback (most recent call last): | 02:23 |
| Qiming | File "mistral/tests/unit/services/test_trigger_service.py", line 202, in test_oneshot_trigger_create | 02:23 |
| Qiming | trigger.next_execution_time | 02:23 |
| Qiming | testtools.matchers._impl.MismatchError: !=: | 02:23 |
| Qiming | reference = datetime.datetime(4242, 12, 25, 13, 37) | 02:23 |
| Qiming | actual = datetime.datetime(4242, 12, 25, 18, 37) | 02:23 |
| kong_ | Qiming: http://paste.openstack.org/ is your friend | 02:25 |
| Qiming | it is too big, :) | 02:26 |
| kong_ | hmm...you can run it again to try your luck | 02:27 |
| Qiming | trying | 02:27 |
| Qiming | so ... I have to get mistral installed and configured to run unit tests? | 02:27 |
| Qiming | I'm not familar with pecan | 02:27 |
| kong_ | i don't thing so, it should work only if you git clone mistral repo | 02:28 |
| *** catintheroof has quit IRC | 02:28 | |
| Qiming | okay, will keep trying | 02:29 |
| Qiming | current summary: | 02:29 |
| Qiming | Ran 838 (-3) tests in 242.164s (+0.685s) | 02:29 |
| Qiming | FAILED (id=1, failures=13, skips=13) | 02:29 |
| *** catintheroof has joined #openstack-mistral | 02:30 | |
| *** catintheroof has quit IRC | 02:34 | |
| openstackgerrit | Qiming Teng proposed openstack/mistral: Fix configuration generator https://review.openstack.org/412256 | 02:47 |
| *** FL1SK has quit IRC | 03:00 | |
| *** FL1SK has joined #openstack-mistral | 03:01 | |
| *** sharatss has quit IRC | 03:16 | |
| *** sharatss has joined #openstack-mistral | 03:16 | |
| *** sharatss has quit IRC | 03:21 | |
| *** sharatss has joined #openstack-mistral | 03:22 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral: Exclude .tox folder from coverage report https://review.openstack.org/411137 | 03:33 |
| kong_ | Qiming: sorry, was distracted by some internal stuff | 03:52 |
| kong_ | Qiming: i ran the tox on my dev env, all passed | 03:52 |
| kong_ | Qiming: http://paste.openstack.org/show/592751/ | 03:53 |
| kong_ | Qiming: I have to head home now, if you have any other questions, please leave message in this channel | 03:54 |
| * kong_ is way | 03:54 | |
| Qiming | thanks, kong_, will do | 04:38 |
| *** hparekh has joined #openstack-mistral | 05:06 | |
| openstackgerrit | Qiming Teng proposed openstack/mistral: Fix docs in README.rst https://review.openstack.org/412284 | 05:11 |
| openstackgerrit | Merged openstack/mistral-extra: Add clone vm example https://review.openstack.org/396840 | 05:13 |
| *** jaosorior has joined #openstack-mistral | 06:06 | |
| *** pkoniszewski has left #openstack-mistral | 06:56 | |
| *** ist has joined #openstack-mistral | 06:56 | |
| *** sharatss has quit IRC | 07:24 | |
| *** sharatss has joined #openstack-mistral | 07:25 | |
| openstackgerrit | Merged openstack/mistral: Updated from global requirements https://review.openstack.org/412137 | 07:36 |
| *** jaosorior has quit IRC | 07:58 | |
| openstackgerrit | Renat Akhmerov proposed openstack/mistral: Fix 'with-items' task completion condition https://review.openstack.org/412341 | 07:59 |
| *** shardy has joined #openstack-mistral | 08:02 | |
| openstackgerrit | Merged openstack/mistral: Add more tests to mistral rally https://review.openstack.org/408709 | 08:13 |
| openstackgerrit | Qiming Teng proposed openstack/mistral: Fix docs in README.rst https://review.openstack.org/412284 | 08:16 |
| rakhmerov | ddeja, d0ugal, hparekh: guys, can you pleaaase review my 'with-items' patches? | 08:23 |
| rakhmerov | need them badly | 08:23 |
| hparekh | ok sure | 08:24 |
| *** jaosorior has joined #openstack-mistral | 08:24 | |
| ddeja | rakhmerov: yup, sure thing | 08:28 |
| rakhmerov | :) | 08:28 |
| rakhmerov | thx | 08:28 |
| rakhmerov | ddeja: btw, oslo team fixe the breakage (https://review.openstack.org/#/c/411555/1/global-requirements.txt) | 08:31 |
| rakhmerov | but they didn't send a patch to make a release | 08:31 |
| rakhmerov | although I don't like the fix | 08:31 |
| rakhmerov | they again simply specified 4.0.2 | 08:31 |
| rakhmerov | why can't they use a range? Do you know? | 08:32 |
| rakhmerov | kombu 4.0.3 will break us again | 08:32 |
| *** sharatss has quit IRC | 08:33 | |
| *** sharatss has joined #openstack-mistral | 08:34 | |
| *** ^Gal^ has joined #openstack-mistral | 08:45 | |
| *** jaosorior has quit IRC | 08:45 | |
| *** jaosorior has joined #openstack-mistral | 08:46 | |
| *** tuan_ has joined #openstack-mistral | 08:47 | |
| tuan_ | Hi guys | 08:47 |
| tuan_ | hopefully there is someone online since Christmas is coming | 08:47 |
| tuan_ | :) | 08:47 |
| *** jpich has joined #openstack-mistral | 08:50 | |
| tuan_ | d0ugal: Hi ddeja, sorry to disturb you again | 08:54 |
| tuan_ | i would like to know that do we have any benchmark for performance of mistral when running actions | 08:54 |
| tuan_ | how much rams, cpus, etc. (i mean resources) needed for running actions in mistral? | 08:54 |
| ddeja | rakhmerov: I'm not sure... maybe they hope that next version would be OK | 08:56 |
| ddeja | we may try to add it to our requirements | 08:56 |
| ddeja | but I'm not sure if it's good idea | 08:56 |
| ddeja | tuan_: Hi | 08:56 |
| rakhmerov | :)) | 08:57 |
| *** jtomasek has joined #openstack-mistral | 08:57 | |
| ddeja | tuan_: There is no such mechanism | 08:57 |
| tuan_ | hi ddeja | 08:57 |
| rakhmerov | ddeja: yeah | 08:57 |
| ddeja | rakhmerov: or, on the other hand, we may add the package that is missing for kombu to our requirements | 08:57 |
| tuan_ | in our environment, it seems that mistral running with quite pretty much resources | 08:58 |
| tuan_ | but i do not think it is because of mistral itself | 08:58 |
| rakhmerov | tuan_: what version of Mistral? | 08:59 |
| rakhmerov | from master? | 08:59 |
| tuan_ | liberty | 08:59 |
| rakhmerov | ddeja: I guess there should be a patch soon added by a bot to update our requirements | 08:59 |
| rakhmerov | I wonder why we still don't have it | 08:59 |
| rakhmerov | tuan_: :)) can you switch to a newer version | 09:01 |
| ddeja | rakhmerov: there must be next release of oslo to pip, so we can have it | 09:02 |
| rakhmerov | since then we improved it by orders of magnitude | 09:02 |
| tuan_ | let's me check again about the version, i am not sure | 09:02 |
| tuan_ | :D | 09:02 |
| rakhmerov | ok | 09:02 |
| tuan_ | but why liberty has this problem? | 09:02 |
| rakhmerov | many reasons | 09:02 |
| rakhmerov | we did a huge number of optimizations since then | 09:02 |
| tuan_ | OK, give me a sec | 09:03 |
| rakhmerov | ddeja: yes, right | 09:03 |
| rakhmerov | even if it's fixed in global requirements it needs to be fixed and released in oslo | 09:03 |
| rakhmerov | I wrote it to their channel | 09:04 |
| ddeja | rakhmerov: BTW, I'm not sure if I'll be able to participate in todays meeting | 09:04 |
| d0ugal | liberty is really old for Mistral :) | 09:04 |
| rakhmerov | but seems like there's still a weekend for most of them | 09:04 |
| rakhmerov | ddeja: that's ok | 09:04 |
| ddeja | I have some kind of xmas party with my high school class mates :) | 09:04 |
| rakhmerov | xmas ? | 09:04 |
| rakhmerov | what's that? | 09:04 |
| rakhmerov | ooh, I see | 09:05 |
| rakhmerov | Christmas | 09:05 |
| ddeja | yup ;) | 09:05 |
| rakhmerov | sure, np | 09:05 |
| rakhmerov | I actually don't believe we'll gather a bunch of people during the next 2-3 weeks ) | 09:05 |
| ddeja | I can try to jump using my phone at the begginign | 09:05 |
| rakhmerov | ok, but don't worry about it | 09:06 |
| rakhmerov | enjoy your party man :) | 09:06 |
| ddeja | oh, and also you will have the Christmas on the beggining of January in Russia, am I right? | 09:06 |
| ddeja | instead of end of December | 09:07 |
| rakhmerov | yeah | 09:07 |
| rakhmerov | Jan 7th | 09:07 |
| rakhmerov | but usually Christmas is not so greatly celebrated as NY in Russia | 09:07 |
| rakhmerov | it's kind of quite religios holiday | 09:07 |
| ddeja | rakhmerov: Oh, I see | 09:07 |
| rakhmerov | yes | 09:07 |
| rakhmerov | many simply forget about it at all :) | 09:08 |
| rakhmerov | but New Year is a big thing here | 09:08 |
| ddeja | well, In Poland we got both 25 and 26 of December free days | 09:08 |
| rakhmerov | everyone is drunk :) | 09:08 |
| rakhmerov | :) | 09:08 |
| ddeja | and a lot of people get vacation till new year ;) | 09:08 |
| rakhmerov | ooh, by the way | 09:09 |
| ddeja | but the main "focus" is on Cristmas itself | 09:09 |
| ddeja | rakhmerov: hm? | 09:09 |
| rakhmerov | ddeja, d0ugal, hparekh, kong_ and others: I'll be mostly unavailable from Dec 31 till Jan 9th | 09:09 |
| d0ugal | I'll be unavailable from 22nd until the 4th :) | 09:10 |
| rakhmerov | every year we have long holidays in Russia in the beginning of January | 09:10 |
| rakhmerov | 8-9 days | 09:10 |
| d0ugal | Nice! | 09:10 |
| rakhmerov | it is! | 09:10 |
| rakhmerov | :) | 09:10 |
| ddeja | I'll be less available starting 24 of December up to, hmm, 2 of January | 09:10 |
| rakhmerov | ddeja, d0ugal: ok, guys, having that said :) | 09:11 |
| rakhmerov | I wanna ask you to review patches :) | 09:11 |
| ddeja | rakhmerov: sure, as always ;) | 09:11 |
| rakhmerov | before you go on holidays | 09:11 |
| rakhmerov | :) | 09:11 |
| d0ugal | what ddeja said :) | 09:11 |
| ddeja | I'll look on the with-items today, but probably it would be my afternoon (in 2-3 ours) | 09:11 |
| rakhmerov | np | 09:12 |
| tuan_ | sorry guys | 09:12 |
| tuan_ | it is 3.0.0 rc2 | 09:12 |
| rakhmerov | this one is specifically important: https://review.openstack.org/#/c/412341/ | 09:12 |
| rakhmerov | due to it the combination of 'retry' and 'with-items' simply didn't work | 09:12 |
| d0ugal | tuan_: so Liberty, I think? | 09:12 |
| tuan_ | damm it, it took me time sicne we use it in virtualenv | 09:12 |
| tuan_ | no no | 09:12 |
| tuan_ | 3.0.0 | 09:12 |
| tuan_ | it was released on Oct, 2016 | 09:13 |
| rakhmerov | newton | 09:13 |
| tuan_ | yeah | 09:13 |
| tuan_ | newton | 09:13 |
| d0ugal | right, sorry | 09:13 |
| rakhmerov | ok, why do you think it uses too much resources? | 09:13 |
| * d0ugal is still waking up. I'll remember how counting works soon. | 09:13 | |
| rakhmerov | can you pls provide more details? | 09:13 |
| rakhmerov | some reports maybe etc. | 09:13 |
| rakhmerov | monitoring info | 09:13 |
| tuan_ | i do not think the resouces it takes is quite much | 09:14 |
| tuan_ | since we had problem of resources leak | 09:14 |
| tuan_ | and after checking, what i saw is that the most consumable process is mistral | 09:14 |
| rakhmerov | tuan_: so you are not sure atm that exactly Mistral is guilty? :) | 09:14 |
| rakhmerov | how consumable? | 09:15 |
| tuan_ | i just want to know some more info about mistral in consuming resources | 09:15 |
| tuan_ | yeap | 09:15 |
| tuan_ | it is around 20% RAM | 09:15 |
| rakhmerov | I mean what amount of memory may be? | 09:15 |
| rakhmerov | what is the approximate value? In megabytes? | 09:15 |
| tuan_ | 1GB around :D, haha | 09:15 |
| rakhmerov | ok, and what do you run on Mistral? | 09:16 |
| tuan_ | woa, so many actions including so many components | 09:16 |
| tuan_ | that is why i think it may depend on the type of actions i run | 09:16 |
| rakhmerov | tuan_: it might not be a memory actively used all the time. It may be a peak value that Mistral used on peak loads | 09:16 |
| tuan_ | but mistral architect itself | 09:17 |
| tuan_ | ah ha | 09:17 |
| tuan_ | interesting point | 09:17 |
| rakhmerov | peak memory usually gets freed when memory is not enough for other processes | 09:17 |
| rakhmerov | or over time | 09:17 |
| rakhmerov | by GC | 09:17 |
| rakhmerov | yeah | 09:17 |
| rakhmerov | and yes, it certainly depends on specific actions you run | 09:18 |
| tuan_ | yeap, and peak memory happens in mistral at the very beginning of loading actions | 09:19 |
| tuan_ | ? | 09:19 |
| *** jtomasek has quit IRC | 09:19 | |
| rakhmerov | yes, and when you run some heavy workflows | 09:20 |
| tuan_ | but what about this case: | 09:21 |
| tuan_ | when mistral loads workloads, the memory is not sufficient for peak memory then it needs the swap in this case | 09:21 |
| tuan_ | then it frees memory but swap is not | 09:21 |
| tuan_ | then i wander that do we have some mechanism or based on experiences about how much RAM we should provide for running mistral | 09:22 |
| tuan_ | to advoid this case | 09:23 |
| tuan_ | for instance, in porduction what i saw that the controller node is around 32GB | 09:23 |
| tuan_ | of course based on experiments and experiences | 09:23 |
| tuan_ | otherwise we do not have any benchmark to state that | 09:24 |
| d0ugal | tuan_: it sounds like you are using tripleo? | 09:24 |
| tuan_ | nope | 09:25 |
| d0ugal | oh, nevermind then :) | 09:26 |
| openstackgerrit | Sharat Sharma proposed openstack/mistral-dashboard: Testing fix for dashboard docs-gate https://review.openstack.org/412362 | 09:26 |
| tuan_ | may be later on we will test with tripleO | 09:26 |
| tuan_ | but we deploy mistral in a VM running on top of openstack | 09:27 |
| tuan_ | :( | 09:27 |
| d0ugal | tuan_: I just guessed beause you said 32gb and controller - that is what we recommend :) | 09:27 |
| d0ugal | tuan_: aha, I see | 09:27 |
| d0ugal | tuan_: anyway, ignore that bit. | 09:27 |
| d0ugal | tuan_: I think it is probably a hard question to answer, because it depends how heavily your Mistral is used. | 09:27 |
| tuan_ | yeah, i think so and i do not believe on that it is because of mistral architect | 09:28 |
| tuan_ | since we jsut have 4GB for RAM of vm where mistral is running | 09:28 |
| tuan_ | that sounds stupid | 09:28 |
| tuan_ | :( | 09:28 |
| d0ugal | tuan_: I can tell you how much my Mistral is using if you can wait about 30 mins | 09:29 |
| tuan_ | 4GB is just for OS of 14.04 of Ubuntu, how does it fit with other processes like mistral when it loads actions | 09:29 |
| tuan_ | well, after 30mins, it still uses around 20GB | 09:29 |
| d0ugal | You should only load actions when you first start? | 09:29 |
| tuan_ | 20% RAM | 09:29 |
| tuan_ | sorry | 09:29 |
| d0ugal | haha, 20gb would be bad :-D | 09:30 |
| tuan_ | yeah :D | 09:30 |
| tuan_ | well, you know i am not the person who did this experiment | 09:30 |
| d0ugal | right | 09:30 |
| tuan_ | but i am now the man who have to dig into this case to find out what happens | 09:30 |
| d0ugal | actions only need to be loaded when you first install Mistral or when you add a new custom action. | 09:31 |
| tuan_ | and the guy who did it this spawned a vm with 4GB of RAM | 09:31 |
| *** Ravikiran_K has joined #openstack-mistral | 09:31 | |
| tuan_ | yeah, that is right | 09:31 |
| tuan_ | you can imagine like that | 09:31 |
| tuan_ | after spawning a vm with mistral installed | 09:31 |
| tuan_ | i let it run in couple of hours | 09:32 |
| hparekh | rakhmerov: I will be unavailable from 22nd DEC to 1st of JAN | 09:32 |
| tuan_ | then mistral daemon still running with 20% RAM | 09:32 |
| rakhmerov | hparekh: ok, thanks for letting us know | 09:32 |
| tuan_ | free memory is about 1GB over 4GB | 09:32 |
| tuan_ | but SWAP is full | 09:32 |
| openstackgerrit | Sharat Sharma proposed openstack/mistral-dashboard: Fix docs-gate for mistral-dashboard https://review.openstack.org/412362 | 09:33 |
| d0ugal | tuan_: interesting, that does sound bad :) | 09:33 |
| tuan_ | it seems like: it may happen that mistral use peak memory for loading actions from very beginning of spawning vm | 09:33 |
| tuan_ | that peak memory needs swap too | 09:33 |
| tuan_ | and then it frees out the memory | 09:33 |
| tuan_ | that is the only case what i can see | 09:33 |
| tuan_ | :( | 09:33 |
| tuan_ | so that 4GB of flavor is stupid | 09:34 |
| d0ugal | lol | 09:34 |
| sharatss | rakhmerov: d0ugal finally above patch fixes the docs gate | 09:34 |
| d0ugal | sharatss: great. | 09:34 |
| tuan_ | d0ugal: btw, thank you :D | 09:35 |
| tuan_ | i will ask that guy to test it again | 09:35 |
| tuan_ | merry christmas to you guys and happy new year | 09:35 |
| d0ugal | tuan_: no problem, I am just waiting for my openstack to install and then I'll take a look at this | 09:35 |
| tuan_ | haha | 09:35 |
| tuan_ | yeap | 09:35 |
| tuan_ | :) | 09:35 |
| rakhmerov | sharatss: great! | 09:36 |
| rakhmerov | tuan_: I doubt that the problem is swap | 09:37 |
| rakhmerov | I really do | 09:37 |
| rakhmerov | tuan_: I started Mistral on my local machine now w/o running anything yet, it is taking 62 mb | 09:42 |
| tuan_ | rakhmerov: Awesome, and i also did that in my devstack without any actions | 09:43 |
| tuan_ | since we have many actions so that as you mentioned before the peak memory of loading workflows | 09:44 |
| rakhmerov | tuan_: so please collect more info: what workflows you run, what actions, when consumed RAM goes up etc. | 09:44 |
| rakhmerov | tuan_: what do you do with Mistral? | 09:44 |
| rakhmerov | :) | 09:44 |
| rakhmerov | so that I knew the context | 09:44 |
| tuan_ | well, we have some our own workflows that mistral will take to talk to HEAT to deploy some vm on openstack | 09:45 |
| tuan_ | and other workflows to run ansible to configure them | 09:45 |
| tuan_ | it consumes RAM of course | 09:45 |
| tuan_ | and now i think mistral works well as expected | 09:46 |
| rakhmerov | ok, good to know | 09:48 |
| rakhmerov | interesting | 09:48 |
| openstackgerrit | Lingxian Kong proposed openstack/mistral: Support openstack context for workflow and task https://review.openstack.org/407240 | 10:06 |
| sharatss | rakhmerov: thanks. whenever u r free pls review that patch. we can start working on dashboard if it gets merged | 10:06 |
| openstackgerrit | Qiming Teng proposed openstack/mistral: Fix version response from root controller https://review.openstack.org/412389 | 10:22 |
| *** hparekh has quit IRC | 10:51 | |
| *** szaher has joined #openstack-mistral | 11:16 | |
| *** szaher has quit IRC | 11:23 | |
| *** szaher has joined #openstack-mistral | 11:27 | |
| *** szaher has quit IRC | 11:35 | |
| *** szaher has joined #openstack-mistral | 11:37 | |
| *** sharatss has quit IRC | 11:40 | |
| *** sharatss has joined #openstack-mistral | 11:41 | |
| d0ugal | Getting a strange Mistral error here: http://logs.openstack.org/92/408992/3/check-tripleo/gate-tripleo-ci-centos-7-ovb-nonha/b55fdf1/console.html#_2016-12-19_06_27_31_502513 | 11:47 |
| d0ugal | The action is missing the uuid parameter for some reason. | 11:48 |
| d0ugal | https://github.com/openstack/python-ironic-inspector-client/blob/master/ironic_inspector_client/client.py#L33 | 11:48 |
| d0ugal | ^ that is what it should call. | 11:48 |
| d0ugal | actually, it calls this: https://github.com/openstack/python-ironic-inspector-client/blob/master/ironic_inspector_client/v1.py#L90 | 11:49 |
| d0ugal | but same issue. | 11:49 |
| jaosorior | d0ugal: last time it was the case that it couldn't instantiate the ironic-inspector client instance | 11:57 |
| jaosorior | d0ugal: which is why the keystone session was added here https://github.com/openstack/mistral/blob/master/mistral/actions/openstack/actions.py#L405 | 11:58 |
| jaosorior | d0ugal: ironic inspector client also requires the server to be running already in order to retrieve it's own version | 11:59 |
| jaosorior | so, effectively, that db-populate call requires keystone and ironic-inspector server to be running and ready before it can be called | 11:59 |
| jaosorior | and in tripleo, it needs haproxy to be available as well | 11:59 |
| jaosorior | it could also be some other issue with creating that client instance... so we really need those logs to see what happened | 12:01 |
| d0ugal | jaosorior: right | 12:03 |
| d0ugal | jaosorior: I find it intersting that it was half added, the action is there but the parameters are not | 12:04 |
| d0ugal | so it suggests that it is getting further than creating the client. | 12:04 |
| * d0ugal reads more code | 12:04 | |
| jaosorior | d0ugal: yeah, IIRC that was pretty much the behavior when I was seeing that error. | 12:05 |
| d0ugal | https://github.com/openstack/mistral/blob/master/mistral/actions/openstack/action_generator/base.py#L91-L96 | 12:06 |
| d0ugal | That seems dangerous to me - if we can't get the method, we shouldn't add the action. it should fail more loudly. | 12:06 |
| d0ugal | and like 87 needs to be LOG.exception | 12:06 |
| d0ugal | jaosorior: the command output might not tell us anything | 12:06 |
| jaosorior | d0ugal: at least if it's an issue creating the client it will go here https://github.com/openstack/mistral/blob/master/mistral/actions/openstack/action_generator/base.py#L87 | 12:07 |
| d0ugal | jaosorior: true, so we will see that message but have no idea what the exception was :-D | 12:08 |
| jaosorior | d0ugal: do you have any idea why this is here https://github.com/openstack/mistral/blob/master/mistral/actions/openstack/action_generator/base.py#L94 ? why don't we just skip adding the action if we couldn't get the method | 12:08 |
| jaosorior | d0ugal: it does print the exception | 12:08 |
| d0ugal | jaosorior: no idea, I just linked to that above ^ | 12:09 |
| d0ugal | jaosorior: true, it should be LOG.error tho' | 12:09 |
| jaosorior | it's the third parameter | 12:09 |
| d0ugal | I am making a patch | 12:09 |
| jaosorior | that's what the 'e' is for here https://github.com/openstack/mistral/blob/master/mistral/actions/openstack/action_generator/base.py#L88 | 12:09 |
| jaosorior | but yeah | 12:09 |
| jaosorior | log.error would be better | 12:09 |
| d0ugal | I see it :-P | 12:09 |
| openstackgerrit | Dougal Matthews proposed openstack/mistral: Use LOG.error when adding an OpenStack action fails https://review.openstack.org/412430 | 12:11 |
| openstackgerrit | Qiming Teng proposed openstack/mistral: Fix configuration generator https://review.openstack.org/412256 | 12:12 |
| openstackgerrit | Dougal Matthews proposed openstack/mistral: Use LOG.error when adding an OpenStack action fails https://review.openstack.org/412430 | 12:12 |
| *** shardy is now known as shardy_lunch | 12:15 | |
| openstackgerrit | Dougal Matthews proposed openstack/mistral: Don't create actions with empty arg_lists https://review.openstack.org/412433 | 12:15 |
| *** sharatss has quit IRC | 12:18 | |
| *** catintheroof has joined #openstack-mistral | 12:18 | |
| *** sharatss has joined #openstack-mistral | 12:18 | |
| openstackgerrit | Qiming Teng proposed openstack/mistral: Fix docs in README.rst https://review.openstack.org/412284 | 12:19 |
| openstackgerrit | Sharat Sharma proposed openstack/python-mistralclient: Wrap the contents of the input and description fields https://review.openstack.org/403341 | 12:21 |
| *** dprince has joined #openstack-mistral | 12:22 | |
| openstackgerrit | Sharat Sharma proposed openstack/python-mistralclient: Initial commit for python-mistralclient document https://review.openstack.org/404736 | 12:26 |
| openstackgerrit | Sharat Sharma proposed openstack/mistral: Added congress actions into Mistral https://review.openstack.org/410208 | 12:40 |
| *** jtomasek has joined #openstack-mistral | 12:56 | |
| openstackgerrit | Sharat Sharma proposed openstack/python-mistralclient: Wrap the contents of the input and description fields https://review.openstack.org/403341 | 12:59 |
| *** tuan_ has quit IRC | 12:59 | |
| *** tuan_ has joined #openstack-mistral | 13:04 | |
| openstackgerrit | Dougal Matthews proposed openstack/mistral: Don't create actions with empty arg_lists https://review.openstack.org/412433 | 13:21 |
| openstackgerrit | Dougal Matthews proposed openstack/mistral: Use LOG.exception when adding an OpenStack action fails https://review.openstack.org/412430 | 13:21 |
| d0ugal | jpich: ^ thanks. | 13:21 |
| jpich | d0ugal: Np! | 13:22 |
| *** weshay_lata is now known as weshay | 13:31 | |
| *** zaneb has quit IRC | 13:36 | |
| *** tuan_ has quit IRC | 13:46 | |
| *** clenimar has joined #openstack-mistral | 13:46 | |
| openstackgerrit | Sharat Sharma proposed openstack/mistral: Added congress actions into Mistral https://review.openstack.org/410208 | 13:58 |
| *** bobh has joined #openstack-mistral | 14:09 | |
| *** chlong has joined #openstack-mistral | 14:14 | |
| *** shardy_lunch is now known as shardy | 14:16 | |
| *** bobh has quit IRC | 14:41 | |
| *** jtomasek has quit IRC | 14:50 | |
| *** jamielennox is now known as jamielennox|away | 14:56 | |
| *** jtomasek has joined #openstack-mistral | 14:59 | |
| *** bobh has joined #openstack-mistral | 15:06 | |
| *** bobh has quit IRC | 15:07 | |
| *** bobh has joined #openstack-mistral | 15:07 | |
| *** bobh has quit IRC | 15:12 | |
| *** jtomasek has quit IRC | 15:13 | |
| *** weshay is now known as weshay_afk | 15:41 | |
| *** weshay_afk is now known as weshay | 15:45 | |
| *** jtomasek has joined #openstack-mistral | 16:12 | |
| *** jtomasek has quit IRC | 16:16 | |
| openstackgerrit | Gal Margalit proposed openstack/mistral-dashboard: Workflow list - added missing fields https://review.openstack.org/412517 | 16:20 |
| *** weshay is now known as weshay_bbiab | 16:34 | |
| *** jaosorior has quit IRC | 16:48 | |
| *** jaosorior has joined #openstack-mistral | 16:49 | |
| openstackgerrit | Pradeep Kilambi proposed openstack/mistral: Add aodh actions to mistral https://review.openstack.org/408684 | 16:51 |
| *** ^Gal^ has quit IRC | 17:03 | |
| *** jpich has quit IRC | 17:04 | |
| *** weshay_bbiab is now known as weshay | 17:51 | |
| *** shardy has quit IRC | 18:08 | |
| *** openstack has joined #openstack-mistral | 19:15 | |
| *** chlong has quit IRC | 19:34 | |
| *** chlong has joined #openstack-mistral | 19:36 | |
| *** jaosorior has quit IRC | 19:52 | |
| *** Ravikiran_K has quit IRC | 19:54 | |
| *** thrash is now known as thrash|biab | 20:16 | |
| *** weshay is now known as weshay_relocatin | 20:42 | |
| *** jamielennox|away is now known as jamielennox | 21:04 | |
| *** thrash|biab is now known as thrash | 21:49 | |
| *** bobh has joined #openstack-mistral | 21:56 | |
| *** bobh has quit IRC | 21:58 | |
| *** dprince has quit IRC | 22:02 | |
| *** enykeev has quit IRC | 23:28 | |
| *** enykeev has joined #openstack-mistral | 23:28 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!