| *** rhsu has quit IRC | 00:26 | |
| *** mriedem has joined #openstack-qa | 01:01 | |
| *** prekarat has joined #openstack-qa | 02:48 | |
| *** prekarat has quit IRC | 02:53 | |
| *** SergeyLukjanov has joined #openstack-qa | 03:18 | |
| *** mriedem has quit IRC | 03:22 | |
| *** SergeyLukjanov has quit IRC | 05:15 | |
| *** afazekas has joined #openstack-qa | 05:17 | |
| *** raies has joined #openstack-qa | 05:21 | |
| raies | afazekas: hi | 05:21 |
|---|---|---|
| raies | afazekas: please check https://review.openstack.org/#/c/39023/ | 05:25 |
| afazekas | raies: I did not seen a newly added issue, +2. Thank you | 05:27 |
| raies | afazekas: I have seen your +2 Thank you very much :) | 05:28 |
| afazekas | Anju: FYI https://review.openstack.org/#/c/40070/2 | 05:28 |
| afazekas | Anju: this change aims the fix 422 issue and replacing the error code to 404 when you try an action on a deleted VM | 05:29 |
| raies | dkranz: please check https://review.openstack.org/#/c/39023/ it is reviewed | 05:32 |
| afazekas | I considered the +2 on Friday, but I got an interrupt and I had to leave.. | 05:37 |
| raies | afazekas: its ok | 05:39 |
| raies | waiting for approval | 05:39 |
| raies | it took so long to be approve | 05:40 |
| afazekas | raies: It is a weekend in many country | 05:45 |
| afazekas | raies: try to create an another change | 05:46 |
| raies | afazekas: hmm | 05:46 |
| raies | afazekas: xml support for network API is ready | 05:47 |
| raies | should I make a blueprint for this first ? then to submit code ? | 05:47 |
| afazekas | raies: how big will be the total xml support, and how many patches ? | 05:48 |
| raies | arround 200+ loc | 05:49 |
| raies | for now it is one | 05:49 |
| afazekas | raies: IMHO not necessary to create a blueprint for one patch | 05:50 |
| raies | afazekas: ok i.e i should submit ? | 05:50 |
| afazekas | you can submit it as dependent patch to your previous one, or you can submit as an individual patch, and you may need to rebase the other patch after the first one merged | 05:52 |
| afazekas | but in both way, you can get review responses for both patch | 05:53 |
| raies | ok | 05:53 |
| raies | still it is dependent on what you review for me i.e. https://review.openstack.org/#/c/39023/ | 05:53 |
| raies | I thought that once https://review.openstack.org/#/c/39023/ is merged then I will submit xml support | 05:54 |
| afazekas | rebase you patch the new patch on the 39023 and the 'git review' will submit it as second change which depends on the previous one | 05:55 |
| afazekas | As I remember just the second change really changes even if the 'git review' says otherwise | 05:56 |
| * afazekas not sure | 05:56 | |
| raies | ok | 05:56 |
| raies | thanks for yor input | 05:56 |
| afazekas | if you accidentally lose the +2 because of git review command issue let me know. | 05:56 |
| raies | ok | 05:57 |
| raies | afazeas: talk to you later and once again thnks for your input | 05:58 |
| afazekas | raies: you are welcome | 05:59 |
| *** raies has left #openstack-qa | 06:00 | |
| *** SergeyLukjanov has joined #openstack-qa | 06:36 | |
| *** tkammer has joined #openstack-qa | 07:31 | |
| *** jhenner has joined #openstack-qa | 07:52 | |
| *** jhenner has quit IRC | 07:57 | |
| *** SergeyLukjanov has quit IRC | 07:59 | |
| *** SergeyLukjanov has joined #openstack-qa | 08:00 | |
| *** jhenner has joined #openstack-qa | 08:49 | |
| *** jhenner has quit IRC | 08:53 | |
| *** tkammer has quit IRC | 09:02 | |
| *** tkammer has joined #openstack-qa | 09:28 | |
| mkoderer | afazekas: ping | 09:46 |
| afazekas | mkoderer: pong | 09:47 |
| mkoderer | Good morning ;) | 09:47 |
| afazekas | Good morning ;) | 09:47 |
| mkoderer | I think my maillist setup is somehow messed up | 09:48 |
| mkoderer | how do I write to the new mailing list? | 09:48 |
| mkoderer | we use the general one right? | 09:48 |
| afazekas | AFAIK you need to write to the dev list with [QA] tag in the subject | 09:48 |
| mkoderer | ok so I just add this tag to the subject | 09:49 |
| mkoderer | we have a lot of discussions in the reviews about the stress test | 09:50 |
| afazekas | http://lists.openstack.org/pipermail/openstack-dev/2013-August/012912.html like this | 09:50 |
| mkoderer | I think we need to clarify what exactly is the purpose of the stress test as such | 09:50 |
| afazekas | I think we should focus on the integration points first | 09:51 |
| mkoderer | I see two use cases | 09:51 |
| afazekas | floating ip with real end-to-end check (I started something on Friday, but I got too many interrupts ..) | 09:51 |
| afazekas | the other should be the volume issue what we seen | 09:52 |
| afazekas | is the volume real visible on the guest | 09:52 |
| afazekas | With an optional write | 09:52 |
| afazekas | it can be confugurable | 09:52 |
| mkoderer | mhh yes sounds good | 09:52 |
| mkoderer | if you have a look here https://review.openstack.org/#/c/39752/ | 09:52 |
| afazekas | the snapshot creation also an interesting point | 09:52 |
| mkoderer | there is a more general discussion about "what should we stress test" | 09:52 |
| afazekas | cinder can mova images to glance and swift | 09:53 |
| afazekas | mostly the things what can have potential race issues.. | 09:53 |
| mkoderer | I want to use the stress tests for a different purpose .. | 09:53 |
| afazekas | For example security rule apply / remove effect time .. | 09:53 |
| afazekas | What is your use case ? | 09:54 |
| mkoderer | I want to use it in a production like environment to indentify potential problems before there occur | 09:54 |
| mkoderer | my target are more real-life use cases | 09:54 |
| afazekas | Just the end user activity ? | 09:55 |
| mkoderer | yes, I want to have a mix of traffic that can occur in a production system | 09:56 |
| afazekas | Or including admin events , or random system failures ? | 09:56 |
| mkoderer | yes maybe everything - I need to plan some scenarios | 09:56 |
| afazekas | so probably multiple tempest run with good filtering and scenario test can help in this | 09:56 |
| afazekas | mkollaro also has this repo for fail injection https://github.com/mkollaro/destroystack | 09:57 |
| mkoderer | yes but I need to have a high load, so I like the multiprocessing thing in the stress test | 09:57 |
| afazekas | She focused on swift, but I hope it can be extended for other serices | 09:57 |
| afazekas | mkoderer: what limits you in the number of tempest process to start ? | 09:58 |
| mkoderer | DestoryStack sounds interesting ;) | 09:58 |
| afazekas | I have bunch of idea for creating a DebugStack repo, .. but time is less :( | 10:00 |
| mkoderer | afazekas: I think the handling would be a bit hard | 10:00 |
| afazekas | We will have many <foobar>Stack project | 10:00 |
| mkoderer | putting all together to one statistic is exactly what the stress test runner can do for me | 10:00 |
| afazekas | mkoderer: so the UnitTest runner plugin is actually useful for you ? | 10:01 |
| mkoderer | yes for my point of view this is what I need | 10:01 |
| mkoderer | but I feel that we have different opinions within Qa.. | 10:02 |
| afazekas | FYI: the subuit has some kind of scheduler plug ability, I did not checked yet, but it can be another option.. | 10:02 |
| afazekas | mkoderer: I do not think we have different opinion | 10:03 |
| afazekas | I see even see more use cases sometimes :) | 10:03 |
| mkoderer | ok - I suggest to create a HACKING.rst for the stress test | 10:03 |
| *** SergeyLukjanov has quit IRC | 10:03 | |
| mkoderer | and there we put the use cases we want to have | 10:03 |
| afazekas | for me usually not enough to see it is not working, I want to know in fastest way why | 10:04 |
| mkoderer | afazekas: that true - maybe a open point | 10:04 |
| mkoderer | for me it's often hard to determin what went wrong in a openstack deployment | 10:05 |
| afazekas | I am considering to implement full systemtap based tracer designed for openstack type apps.. | 10:06 |
| afazekas | The first issue is important for me, because I usually do not really care about the happenings after the first issue | 10:07 |
| mkoderer | afazekas: cool.. my master thesis was about systemtap.. long time ago ;) | 10:07 |
| afazekas | So maybe you know even better it | 10:07 |
| mkoderer | question is how you want to indentify the first problem? | 10:08 |
| afazekas | looks like the awk style does not really lets you to process everything in the stap script, basically I will need a lot of post processing and pre processing, and may be I will have some small blinds spots | 10:08 |
| afazekas | mkoderer: The first test case failure | 10:09 |
| mkoderer | I have to think about it | 10:10 |
| afazekas | http://fedoraproject.org/wiki/Features/SystemtapStaticProbes#Python python is very noisy ( no real threading workarounds), so first I would like correlate the syscalls with python stack traces | 10:12 |
| afazekas | The second usual issue is the process and network traffic connection | 10:12 |
| mkoderer | what it the current state of systemtap - the time I used it it wasn't really stable | 10:13 |
| afazekas | with neutron / openvswich / iptables / namspaces the in kernel pocket flow also interesting .. | 10:13 |
| afazekas | mkoderer: AFAIK it is stable | 10:13 |
| mkoderer | cool... I used dtrace very extensivly on Solaris, and this was really stable even in the kernel | 10:15 |
| mkoderer | afazekas: I need to have a look the tap and pyhton | 10:15 |
| mkoderer | afazekas: I would really like to help you with it | 10:16 |
| afazekas | mkoderer: the other option would be python remote debuger on every python process (you can modify some weak symbols + LD_PRELOAD to have it started where it is needed ) | 10:16 |
| afazekas | after the neutron smoke gate restored we can talk more about this | 10:18 |
| mkoderer | afazekas: yes sure I will already have a look about it | 10:18 |
| mkoderer | it seems that I have more time next week... our system is now up and running | 10:19 |
| mkoderer | so I won't be that busy .. and I will use the time for upstream | 10:20 |
| *** psedlak has joined #openstack-qa | 10:20 | |
| afazekas | cool | 10:26 |
| *** psedlak has quit IRC | 10:26 | |
| mkoderer | afazekas: when will you add more core reviewers? already next qa meeting? | 10:27 |
| *** SergeyLukjanov has joined #openstack-qa | 10:33 | |
| afazekas | mkoderer: I have no idea. last time when we considered, nobody seamed close to the core reviewer's review count/month | 10:35 |
| mkoderer | ok I started this weekend - I will try to review what I can during the week ;) | 10:36 |
| mkoderer | but sure it's hard to get close since I don't have your experience with tempest | 10:38 |
| mkoderer | I need long for a review.. | 10:38 |
| afazekas | Please try to do better than usually I , so say 'thank you' and 'please' more frequently, and try to be as much informative as possible when you -1 something | 10:38 |
| mkoderer | yes sure and I am always open for feedback ;) | 10:41 |
| afazekas | jaypipes usually has nice/polite review comments, but unfortunate nowadays I rarely see them. | 10:44 |
| mkoderer | afazekas: is there a script somewhere where you see the stats out of gerrit | 10:52 |
| afazekas | yes, but I do not know where sdague used it , afaik it is from russelb . You can see some references in the QA mailing list archive | 10:53 |
| mkoderer | ok I will search it | 10:54 |
| mkoderer | I am off for today - have a nice sunday | 10:54 |
| afazekas | btw the unittest runner thing should be extended by json generator, when you can create unittest pairs. | 10:55 |
| afazekas | The use case us you have 1 racy test case, and you would like to run it together with all possible conflicting test case | 10:56 |
| afazekas | the aggragates test case may conflict with the host clients.. | 10:56 |
| mkoderer | ok will have a look at it | 11:01 |
| afazekas | And we should think about what should be 10-20 minute stress test run as gate job | 11:07 |
| afazekas | now 2 complex test case skipped because of the volume issue, and we lose coverage in another places | 11:08 |
| afazekas | mkoderer: and usually if this happens, when the other bug is fixed, a new one introduced. In this very bad shape, does not really makes sense to attack with real life patters, since we know, it fails even with a mini size pattern. | 11:11 |
| *** SergeyLukjanov has quit IRC | 11:26 | |
| *** mkollaro has joined #openstack-qa | 11:37 | |
| afazekas | mkoderer: <evilmode> would be better to not skip the test when there is "race" issue in one of the openstack components , it would enforce more people to have a look </evilmode> | 12:04 |
| mkoderer | afazekas: I think this we cause a lot of trouble ;) | 12:14 |
| mkoderer | afazekas: I think there should be a process to escalate such issues | 12:14 |
| afazekas | I think stress voting gate job, could prevent many of those issues | 12:18 |
| mkoderer | yes think so | 12:19 |
| *** psedlak has joined #openstack-qa | 12:22 | |
| *** giulivo has joined #openstack-qa | 12:50 | |
| *** giulivo has quit IRC | 12:55 | |
| *** giulivo has joined #openstack-qa | 13:09 | |
| *** mriedem has joined #openstack-qa | 13:36 | |
| *** mriedem has quit IRC | 13:39 | |
| *** Anju_ has joined #openstack-qa | 13:42 | |
| *** Anju_ has quit IRC | 13:46 | |
| *** psedlak_ has joined #openstack-qa | 13:54 | |
| *** psedlak has quit IRC | 13:55 | |
| *** Anju_ has joined #openstack-qa | 13:57 | |
| jaypipes | afazekas: sorry man, I have been swamped in deployments :( | 14:05 |
| jaypipes | afazekas: will try to do some reviews this week. apologies. | 14:05 |
| afazekas | jaypipes: No problem, thank you | 14:10 |
| mordred | jaypipes: we miss you! | 14:12 |
| jaypipes | mordred: I miss free time! | 14:12 |
| mordred | jaypipes: hire more interns? ;) | 14:13 |
| jaypipes | mordred: lol. fire more internals. | 14:13 |
| mordred | jaypipes: ++ | 14:13 |
| Anju_ | afazekas:do u devstack environment this time? | 14:17 |
| afazekas | Anju_ : Now I am playing with another thing | 14:18 |
| Anju_ | afazekas:ok thanku | 14:18 |
| afazekas | HL2 :) | 14:19 |
| Anju_ | HL2????? | 14:19 |
| afazekas | https://en.wikipedia.org/wiki/Half-Life_2 | 14:34 |
| afazekas | BTW: if you have install issues on an rpm system you can git revert the braking patch | 14:37 |
| Anju_ | okk...afazekas:nice to hear that you r doing something else(other than openstack) | 14:46 |
| Anju_ | :P | 14:46 |
| mkoderer | giulivo: ping | 15:01 |
| giulivo | pong mkoderer ; | 15:02 |
| giulivo | here but "not very much" | 15:02 |
| *** raies has joined #openstack-qa | 15:02 | |
| mkoderer | me either | 15:02 |
| mkoderer | giulivo: I think would be good to start the discussion in the mailing list before the meeting | 15:03 |
| *** SergeyLukjanov has joined #openstack-qa | 15:03 | |
| giulivo | np , it's a good idea | 15:03 |
| giulivo | if you prefer to send the message is okay with me | 15:03 |
| openstackgerrit | A change was merged to openstack/tempest: add image tests v2 https://review.openstack.org/39212 | 15:04 |
| mkoderer | giulivo: I am totally fine to have a discussion on the meeting | 15:04 |
| mkoderer | but maybe we can already start thinking before | 15:04 |
| giulivo | yeah that was my same intention with the "long" gerrit comment | 15:04 |
| raies | jaypipes: please check for approval https://review.openstack.org/#/c/39023/ | 15:05 |
| *** SergeyLukjanov has quit IRC | 15:05 | |
| mkoderer | but anyway .. I really like the code now ;) | 15:05 |
| *** SergeyLukjanov has joined #openstack-qa | 15:05 | |
| mkoderer | I think we can find a good way to get it in | 15:06 |
| *** SergeyLukjanov has quit IRC | 15:06 | |
| mkoderer | I just have a problem with a huge work flow without a real use case | 15:06 |
| giulivo | mkoderer, that would be in case we merge all of them, right? | 15:07 |
| mkoderer | for me it looks like random actions | 15:07 |
| giulivo | btw, note that we have two sections one for stress tests and one for scenario tests | 15:07 |
| giulivo | this is supposed to be a stress test, not a scenario test | 15:07 |
| mkoderer | right | 15:07 |
| giulivo | will you send something to the list? | 15:09 |
| mkoderer | but if you have a combined action that you have to call "sever_actions" just because you don't find a real topic for them | 15:09 |
| mkoderer | giulivo: yes I will do it tomorrow morning | 15:09 |
| jaypipes | raies: done. | 15:09 |
| giulivo | mkoderer, also true, so let's see if anything pops up from the list | 15:10 |
| giulivo | jaypipes, thanks for reviewing the scenario thing | 15:10 |
| giulivo | actually the response is yes and no | 15:10 |
| jaypipes | giulivo: no probs. sorry I've been so busy :( | 15:10 |
| giulivo | jaypipes, I thought it wasn't useful to check for the api return codes because I wanted to make sure the flow completes | 15:11 |
| giulivo | so didn't have much to "assert" except from actual failures/timeouts | 15:11 |
| raies | jaypipes: thanks for your quick response | 15:11 |
| giulivo | the particular single actions are tested by the api tests already | 15:12 |
| giulivo | so with the scenario tests, what is that we can check apart from the successful completion of the steps? | 15:12 |
| giulivo | mkoderer, ^^ | 15:12 |
| giulivo | maybe we can open up the question to both scenario and stress tests | 15:12 |
| jaypipes | giulivo: OK, no worries. Was just double-checking to see if that was intentional. | 15:13 |
| jaypipes | giulivo: I will change to approve. | 15:13 |
| giulivo | jaypipes, wait, no no | 15:13 |
| jaypipes | heh, ok :) | 15:13 |
| giulivo | I mean, I agree with your concerns | 15:13 |
| giulivo | there is something wrong in my whole approach to that test probabl | 15:13 |
| giulivo | like, I could ssh on the systems and check for the actual disk contents, not sure | 15:14 |
| jaypipes | giulivo: well, after thinking about what kinds of assertions would make sense there, I don't see much other than perhaps doing a get on the instance info and checking statuses... but that's what wait_for_server_status() does already. | 15:14 |
| jaypipes | giulivo: and you are not testing for correctness of the get instance return or anything, since that is already tested in many places | 15:14 |
| mkoderer | giulivo: at the end "stress tests" aren't a real test group. I mean every test whether api test or scenario test can be executed as a stress test | 15:15 |
| jaypipes | giulivo: so I think it's probably best just to put a quick comment in there saying "hey, we aren't doing other assertions here, since we assert for the status in the wait loops and that is all this test is validating" | 15:15 |
| giulivo | yeah os please allow me for a few more time, if I can get anything better in I will otherwise I'll just comment on that | 15:15 |
| jaypipes | no probs. | 15:16 |
| mkoderer | giulivo: if you have a look here https://review.openstack.org/#/c/38980/ | 15:16 |
| jaypipes | giulivo: just ping me when you push a new change. | 15:16 |
| *** jaypipes has left #openstack-qa | 15:16 | |
| *** jaypipes has joined #openstack-qa | 15:16 | |
| *** ChanServ sets mode: +o jaypipes | 15:16 | |
| jaypipes | raies: need a rebase for https://review.openstack.org/39023 | 15:17 |
| raies | japipes: ok i will do this | 15:18 |
| giulivo | mkoderer, thanks for pointing to that | 15:18 |
| raies | jaypipes: In patch2 "Zhi Kun Liu" and "Sean Dague" suggested not to use assertTrue() as it is not debugable. Thats why I added an alernative way to do this. | 15:18 |
| giulivo | mkoderer, how about we try to make the question broader so we try to find out what is a good approach for both stress and scenario patterns? | 15:18 |
| raies | jaypipes: I see here that build is beig failed for https://review.openstack.org/#/c/39023/ | 15:19 |
| jaypipes | raies: ok, understood. | 15:19 |
| mkoderer | giulivo: yes at the end it is a very general question | 15:20 |
| jaypipes | raies: it is saying you need to rebase: "This change was unable to be automatically merged with the current state of the repository and the following changes which were enqueued ahead of it: 39730, 40046. Please rebase your change and upload a new patchset." | 15:20 |
| raies | jaypipes: ok thanks for your input | 15:20 |
| mkoderer | I see that we have different viewpoints all over QA about stress test, api tests and scenario test | 15:21 |
| jaypipes | raies: are you familiar with how to rebase? | 15:23 |
| *** adalbas has joined #openstack-qa | 15:23 | |
| raies | jaypipes: yes I am | 15:24 |
| jaypipes | raies: k, lemme know if you need any assistance. thx! | 15:24 |
| raies | jaypipes: I will do rebase and will let you know soon :) | 15:24 |
| jaypipes | cheers | 15:24 |
| raies | jaypipes: to u to sir once again tanks for your kind support | 15:26 |
| jaypipes | any time. | 15:27 |
| giulivo | afazekas, so hl2 is one of my favourites by far and I enjoyed episode two _very_ much | 15:37 |
| raies | jaypipes: merging of 39730, 40046 is under process (as jenkins is showng), So rebase will not be possible now i think. Any alternative way to this ? | 15:38 |
| jaypipes | raies: well, you could premerge those, but that's dangerous... I'd just wait until they hit trunk. | 15:39 |
| afazekas | giulivo: I just got the Orange Box, I started with HL2, Looks like I do not remember to many things, so it is half new again | 15:40 |
| * afazekas hunt for food | 15:40 | |
| raies | jaypipes: I will wait till them finished and then will rebase it . can jenkins automatically merge my code after 39730, 40046 code merging or I have to it manually using rebase? | 15:41 |
| * afazekas My be I should turn of my 2th monitor :) | 15:41 | |
| jaypipes | raies: gerrit is alerting you that it was not able to rebase your branch onto the other ones, so you will need to manually rebase and fix merge conflicts. | 15:42 |
| giulivo | afazekas, so that game is great because you get into the story and won't stop when it is finished | 15:42 |
| giulivo | let me know when you finished with episode two what you think | 15:42 |
| raies | jaypipes: ok thank you. I think better be to wait :) | 15:43 |
| jaypipes | raies: yes, will save you a headache ;) | 15:43 |
| raies | jaypipes: yep ! | 15:43 |
| afazekas | giulivo: I try to rest before entering into the battle with the neutron issue :) , So probably next week (or 2), I will disable all other interrupt .. | 15:49 |
| *** raies has quit IRC | 15:52 | |
| openstackgerrit | A change was merged to openstack-dev/devstack: Update to PLUMgrid plugin configuration https://review.openstack.org/39182 | 16:13 |
| openstackgerrit | A change was merged to openstack/tempest: White space after # in tempest/api files https://review.openstack.org/39730 | 16:13 |
| openstackgerrit | A change was merged to openstack/tempest: Negative attribute added in server negative test https://review.openstack.org/40046 | 16:13 |
| *** tkammer has quit IRC | 16:20 | |
| openstackgerrit | anju Tiwari proposed a change to openstack/tempest: Negative tests added for network https://review.openstack.org/40049 | 16:31 |
| Anju_ | afazekas :https://review.openstack.org/#/c/40049/ please have look | 16:36 |
| *** mkollaro has quit IRC | 16:53 | |
| *** mkollaro has joined #openstack-qa | 16:55 | |
| *** mkollaro has quit IRC | 16:55 | |
| afazekas | Anju_: LGTM, let's see the jenkins longs after it finished | 17:15 |
| Anju_ | afazekas:thanks | 17:22 |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Handle the credentials as object https://review.openstack.org/38641 | 17:22 |
| *** Anju_ has quit IRC | 17:30 | |
| *** raies has joined #openstack-qa | 17:32 | |
| *** raies has quit IRC | 17:37 | |
| *** Anju_ has joined #openstack-qa | 17:48 | |
| *** giulivo has quit IRC | 17:58 | |
| *** psedlak_ has quit IRC | 18:26 | |
| *** Anju_ has left #openstack-qa | 18:29 | |
| *** SergeyLukjanov has joined #openstack-qa | 19:13 | |
| *** SergeyLukjanov has quit IRC | 20:02 | |
| openstackgerrit | andrea-frittoli proposed a change to openstack/tempest: Whitebox server expects deleted > 0. https://review.openstack.org/40129 | 23:27 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!