| *** fnaval_ has joined #openstack-metering | 00:38 | |
| *** fnaval_ has quit IRC | 00:39 | |
| *** dina_belova has joined #openstack-metering | 00:44 | |
| *** haomeng has quit IRC | 00:45 | |
| *** dina_belova has quit IRC | 00:49 | |
| *** shanewang has joined #openstack-metering | 00:49 | |
| *** shanewang has left #openstack-metering | 00:49 | |
| *** zul has quit IRC | 00:57 | |
| *** zul has joined #openstack-metering | 00:57 | |
| *** Ruetobas has quit IRC | 01:27 | |
| *** flwang has quit IRC | 01:27 | |
| *** Ruetobas has joined #openstack-metering | 01:28 | |
| *** dina_belova has joined #openstack-metering | 01:45 | |
| *** dina_belova has quit IRC | 01:49 | |
| *** d34dh0r53 has joined #openstack-metering | 01:59 | |
| *** d34dh0r53 has quit IRC | 02:01 | |
| *** shang has joined #openstack-metering | 02:20 | |
| *** flwang has joined #openstack-metering | 02:20 | |
| *** shakayumi has quit IRC | 02:21 | |
| *** Ruetobas has quit IRC | 02:23 | |
| *** dina_belova has joined #openstack-metering | 02:45 | |
| *** anteaya has quit IRC | 02:47 | |
| *** dina_belova has quit IRC | 02:50 | |
| *** terriyu has joined #openstack-metering | 02:52 | |
| *** flwang1 has joined #openstack-metering | 02:56 | |
| *** flwang has quit IRC | 02:59 | |
| *** flwang1 has quit IRC | 03:03 | |
| *** flwang has joined #openstack-metering | 03:27 | |
| *** SergeyLukjanov has joined #openstack-metering | 03:40 | |
| *** dina_belova has joined #openstack-metering | 03:46 | |
| *** dina_belova has quit IRC | 03:50 | |
| *** dina_belova has joined #openstack-metering | 04:47 | |
| *** evanjfraser has quit IRC | 04:50 | |
| *** dina_belova has quit IRC | 04:51 | |
| *** boris-42 has joined #openstack-metering | 04:54 | |
| *** SergeyLukjanov has quit IRC | 05:19 | |
| *** tasdomas_afk is now known as tasdomas | 05:34 | |
| *** eglynn has joined #openstack-metering | 06:19 | |
| *** eglynn has quit IRC | 06:54 | |
| *** eglynn has joined #openstack-metering | 07:06 | |
| *** boris-42 has quit IRC | 07:23 | |
| *** shardy_afk is now known as shardy | 07:30 | |
| *** dina_belova has joined #openstack-metering | 07:33 | |
| *** SergeyLukjanov has joined #openstack-metering | 07:33 | |
| *** Fegnqian has joined #openstack-metering | 07:56 | |
| *** nsaje has quit IRC | 08:11 | |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Add testing of migrations from oslo https://review.openstack.org/43872 | 08:17 |
|---|---|---|
| *** Alexei_987 has joined #openstack-metering | 08:20 | |
| *** jbuckley has joined #openstack-metering | 08:20 | |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Implement testing of migrations https://review.openstack.org/44809 | 08:34 |
| *** SergeyLukjanov has quit IRC | 08:36 | |
| *** SergeyLukjanov has joined #openstack-metering | 08:41 | |
| *** boris-42 has joined #openstack-metering | 08:50 | |
| lsmola | jd__, I have a quick question, do you have small amount of time now? | 08:51 |
| jd__ | lsmola: go ahead | 08:52 |
| lsmola | jd__, I am wondering whether there is a blueprint for limit and ordering statistics results | 08:59 |
| lsmola | jd__, e.g. eg when I want top5 cpu consumers | 08:59 |
| lsmola | jd__, I would go for some cpu meter, group by Project | 09:00 |
| lsmola | jd__, then I need to order by that meter + limit 5 | 09:01 |
| jd__ | lsmola: I don't think we have that | 09:01 |
| lsmola | jd__, ok | 09:01 |
| lsmola | jd__, could it be a blueprint below the sample-api we've created yesterday? | 09:02 |
| lsmola | jd__, cause it would be really helpful feature | 09:02 |
| jd__ | I don't know about below, but it could be a new blueprint | 09:04 |
| lsmola | jd__, like it could depend on sample-api blueprint | 09:05 |
| jd__ | I don't think it has a dependency on that | 09:06 |
| *** dina_belova has quit IRC | 09:06 | |
| jd__ | the blueprints can be implemented independently | 09:06 |
| lsmola | jd__, not sure here about the general re-usability for statistics etc. | 09:06 |
| lsmola | jd__, yeah, if it would be a general solution | 09:06 |
| lsmola | jd__, should I start blueprint for this or you will? | 09:07 |
| lsmola | jd__, also this kind of depend on Group By blueprint | 09:08 |
| jd__ | lsmola: go ahead | 09:08 |
| jd__ | lsmola: yes it depends on group by IMHO | 09:09 |
| *** flwang has quit IRC | 09:13 | |
| lsmola | jd__, ok, will create that after lunch, thank you very much | 09:15 |
| openstackgerrit | A change was merged to openstack/ceilometer: Plug alarm history logic into the API https://review.openstack.org/43849 | 09:19 |
| *** Fegnqian has quit IRC | 09:23 | |
| EmilienM | Metering guys, I would love to have one more review from you on Ceilometer Central Agent in HA documentation > https://review.openstack.org/#/c/44584/ | 09:38 |
| jd__ | EmilienM: I've added ceilometer-core to the review FWIW | 09:43 |
| EmilienM | jd__: that's fair, thx a lot | 09:43 |
| *** dina_belova has joined #openstack-metering | 09:45 | |
| *** lexx has joined #openstack-metering | 09:49 | |
| nijaba | any english native speaker around to review a little doc patch? https://review.openstack.org/#/c/44774/ | 09:55 |
| nijaba | just a few lines really ;) | 09:55 |
| *** jbuckley has quit IRC | 10:02 | |
| openstackgerrit | Fengqian.gao proposed a change to openstack/ceilometer: Add next link in Meter/Sample/Resource/Alarm https://review.openstack.org/44838 | 10:05 |
| openstackgerrit | Fengqian.gao proposed a change to openstack/ceilometer: Add pagination query support for get_samples https://review.openstack.org/44492 | 10:05 |
| openstackgerrit | Fengqian.gao proposed a change to openstack/ceilometer: Add pagination support for sqlalchemy database https://review.openstack.org/35454 | 10:05 |
| openstackgerrit | Fengqian.gao proposed a change to openstack/ceilometer: Add pagination support to selected APIs https://review.openstack.org/37454 | 10:05 |
| openstackgerrit | Fengqian.gao proposed a change to openstack/ceilometer: Change pagination query method in mongodb https://review.openstack.org/41869 | 10:05 |
| openstackgerrit | Fengqian.gao proposed a change to openstack/ceilometer: Add a fake UUID to Meter on API level https://review.openstack.org/44278 | 10:05 |
| *** lexx has quit IRC | 10:07 | |
| nijaba | EmilienM: review of your patch done. a bit more work for you, sorry | 10:09 |
| *** SergeyLukjanov has quit IRC | 10:21 | |
| eglynn | dhellmann: re. https://review.openstack.org/44840 only one blessing is required as the backporter (me) is also on stable-maint | 10:29 |
| eglynn | dhellmann: (so you can both +2 and +1 ...) | 10:29 |
| dhellmann | eglynn: ok, doing that now | 10:29 |
| eglynn | dhellmann: thank you sir! | 10:29 |
| openstackgerrit | Haomeng,Wang proposed a change to openstack/ceilometer: fix measurements.rst typos https://review.openstack.org/44844 | 10:33 |
| *** SergeyLukjanov has joined #openstack-metering | 10:34 | |
| *** flwang has joined #openstack-metering | 10:51 | |
| *** SergeyLukjanov has quit IRC | 10:54 | |
| EmilienM | nijaba: no worry at all, thanks for review! | 11:02 |
| *** SergeyLukjanov has joined #openstack-metering | 11:02 | |
| openstackgerrit | Haomeng,Wang proposed a change to openstack/ceilometer: Fixes two typos in this measurements.rst https://review.openstack.org/44848 | 11:04 |
| *** boris-42 has quit IRC | 11:04 | |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Fix wrong migrations https://review.openstack.org/44539 | 11:11 |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Add testing of migrations from oslo https://review.openstack.org/43872 | 11:13 |
| openstackgerrit | Eoghan Glynn proposed a change to openstack/ceilometer: Alarm history storage implementation for mongodb https://review.openstack.org/43850 | 11:19 |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Implement testing of migrations https://review.openstack.org/44809 | 11:20 |
| *** shang has quit IRC | 11:27 | |
| *** eglynn has quit IRC | 11:39 | |
| *** shakayumi has joined #openstack-metering | 11:48 | |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Implement testing of migrations https://review.openstack.org/44809 | 11:51 |
| openstackgerrit | A change was merged to openstack/ceilometer: Add a fake UUID to Meter on API level https://review.openstack.org/44278 | 11:52 |
| *** anteaya has joined #openstack-metering | 11:53 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Append /usr/sbin:/sbin to the path for searching mongodb https://review.openstack.org/44747 | 11:56 |
| openstackgerrit | A change was merged to openstack/ceilometer: Sync rpc from oslo-incubator https://review.openstack.org/44561 | 11:56 |
| lsmola | jd__, https://blueprints.launchpad.net/ceilometer/+spec/statistics-order-by-and-limit-for-grouped-query | 12:02 |
| lsmola | jd__, I have added a brief description and the dependencies, feel free to hijack it again :-) | 12:02 |
| lsmola | jd__, thank you for your help | 12:03 |
| *** SergeyLukjanov has quit IRC | 12:09 | |
| *** SergeyLukjanov has joined #openstack-metering | 12:09 | |
| *** dina_belova has quit IRC | 12:17 | |
| *** shakayumi has quit IRC | 12:20 | |
| *** eglynn has joined #openstack-metering | 12:21 | |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Implement testing of migrations https://review.openstack.org/44809 | 12:27 |
| *** dina_belova has joined #openstack-metering | 12:31 | |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Fix migration with fkeys https://review.openstack.org/38607 | 12:34 |
| *** annegentle_away is now known as annegentle | 12:46 | |
| *** thomasm has joined #openstack-metering | 12:57 | |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Fix wrong migrations https://review.openstack.org/44539 | 12:58 |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Add testing of migrations from oslo https://review.openstack.org/43872 | 12:59 |
| openstackgerrit | Svetlana Shturm proposed a change to openstack/ceilometer: Implement testing of migrations https://review.openstack.org/44809 | 13:00 |
| openstackgerrit | Cyril Roelandt proposed a change to openstack/ceilometer: Network: process metering reports from Neutron https://review.openstack.org/43892 | 13:17 |
| openstackgerrit | Eoghan Glynn proposed a change to openstack/ceilometer: Alarm history storage implementation for mongodb https://review.openstack.org/43850 | 13:22 |
| *** gordc has joined #openstack-metering | 13:24 | |
| thomasm | Hey all! | 13:27 |
| thomasm | dhellmann, ping | 13:27 |
| *** sandywalsh has joined #openstack-metering | 13:32 | |
| *** thomasm has quit IRC | 13:33 | |
| *** thomasm has joined #openstack-metering | 13:36 | |
| thomasm | back | 13:38 |
| *** changbl has quit IRC | 13:43 | |
| *** dina_belova has quit IRC | 13:44 | |
| *** dina_belova has joined #openstack-metering | 13:46 | |
| *** dina_belova has quit IRC | 13:46 | |
| *** boris-42 has joined #openstack-metering | 13:49 | |
| openstackgerrit | Julien Danjou proposed a change to openstack/ceilometer: doc: fix storage backend features status https://review.openstack.org/44453 | 14:11 |
| *** dina_belova has joined #openstack-metering | 14:12 | |
| sandywalsh | dhellmann, ping? | 14:20 |
| dhellmann | sandywalsh: pong | 14:20 |
| sandywalsh | :) question about stevedore .map() | 14:20 |
| sandywalsh | I need the return values from each dispatcher | 14:20 |
| sandywalsh | is that possible? | 14:21 |
| dhellmann | map() returns the values returned by the plugins | 14:21 |
| dhellmann | so yeah, have each dispatcher return a list, and then map will return a list of lists | 14:21 |
| sandywalsh | hmm, what if each dispatcher returns different things? | 14:22 |
| sandywalsh | (if the database dispatcher is different than the log dispatcher for example) | 14:22 |
| dhellmann | well, the dispatchers need to behave consistently | 14:23 |
| dhellmann | so they should all return a list of the duplicates, if that's what we say the API is | 14:23 |
| dhellmann | sometimes the lists can be empty | 14:23 |
| dhellmann | but the return value should be a list, and the contents should always be the same type | 14:23 |
| sandywalsh | right, ok ... I think I can mess with this ... thanks (may have more questions :) | 14:23 |
| dhellmann | I'll be in and out today, but ping me if you need | 14:24 |
| sandywalsh | np. thanks | 14:24 |
| anteaya | dhellmann: welcome back | 14:31 |
| openstackgerrit | Cyril Roelandt proposed a change to openstack/ceilometer: Network: process metering reports from Neutron https://review.openstack.org/43892 | 14:39 |
| dhellmann | hi, anteaya, thanks :-) | 14:41 |
| anteaya | :D | 14:41 |
| thomasm | dhellmann, jd__, ping? | 14:41 |
| jd__ | ? | 14:41 |
| *** changbl has joined #openstack-metering | 14:42 | |
| thomasm | So, I'd like to have a quick conversation around that DB2 driver issue. In the review it looks like we have two opposing ideas and I don't want us to end up holding back a bug fix over controversy with how we enforce drivers support. :\ | 14:42 |
| *** sdake has joined #openstack-metering | 14:43 | |
| *** sdake has joined #openstack-metering | 14:43 | |
| thomasm | We can't please everyone, it seems, so I'd like for us to figure out the standard way we'll handle a fundamental functionality, like get_resources, lacking full support in some drivers. I thought we had with raise NotImplementedError, but this is a case that will probably come up more frequently as the project gains adoption. | 14:45 |
| thomasm | dhellmann, jd__, So, should we address that now or in the CM meeting this week? | 14:46 |
| thomasm | I just don't want to remove the NotImplementedError raise and get -1 for that, and then add it back in and get -1 for that from another side. That's not helping anyone, heh. | 14:47 |
| jd__ | thomasm: patch URL? | 14:48 |
| thomasm | https://review.openstack.org/#/c/44277 | 14:48 |
| jd__ | hum | 14:50 |
| jd__ | I don't see how dhellmann comment can help actually | 14:50 |
| jd__ | I agree that a bug should be open if the patch is merged as it is | 14:50 |
| dhellmann | the patch as it is completely breaks the db2 driver, though | 14:52 |
| dhellmann | that's what I was objecting to | 14:52 |
| jd__ | dhellmann: I don't think forcing contributors to fix bugs in HBase and DB2 is going to work | 14:53 |
| dhellmann | that's not what I want | 14:53 |
| dhellmann | I was just saying take the raise out | 14:53 |
| dhellmann | getting some resources with the wrong data is better than not being able to query at all | 14:53 |
| jd__ | so you prefer we ship drivers that returns data that are not compliant to the API and what others drivers return? | 14:53 |
| openstackgerrit | A change was merged to openstack/ceilometer: Fixes two typos in this measurements.rst https://review.openstack.org/44848 | 14:53 |
| jd__ | ok, I tend to disagree with that actually | 14:54 |
| dhellmann | :-/ | 14:54 |
| jd__ | it seems a terrible idea to me to let people use drivers returning something else that we the Ceilometer API is supposed to | 14:54 |
| dhellmann | I classify this as a bug in this driver. Do we break a driver, or remove it, every time we find a bug? Or do we document the bugs and fix them? | 14:55 |
| jd__ | plus that's a slippery slope to special casing storage tests to each backend driver | 14:55 |
| jd__ | dhellmann: well my answer is "fix it, if you can't fix it yourself ask the maintainer, if he doesn't fix it, mark it at broken" | 14:55 |
| jd__ | (that's what thomasm did so far) | 14:55 |
| dhellmann | "first do no harm" | 14:56 |
| jd__ | and now anyway litong is proposing patches that disable some tests entirely for DB2 | 14:56 |
| dhellmann | that may be a different issue | 14:56 |
| dhellmann | maybe the db2 driver just isn't ready to ship and we should remove it entirely? | 14:57 |
| thomasm | dhellmann, In this patch I added two tests to check the default scenario for latest metadata out of unordered samples. | 14:57 |
| jd__ | "do no harm", that can be read as "don't provide wrong data to the users in your driver, otherwise they will build application on a wrong API" too :) | 14:57 |
| thomasm | That breaks in DB2. After a long discussion with litong, we determined that the DB2 driver couldn't really support that functionality yet due to spotty support for aggregation. | 14:57 |
| *** SergeyLukjanov has quit IRC | 14:57 | |
| dhellmann | the method is minimally useful in that it does return a list of resources, but if it raises NotImplementedError then the db2 driver is not useful at all | 14:58 |
| jd__ | dhellmann: I wouldn't mind | 14:58 |
| jd__ | dhellmann: but I prefer to start with a NotImplemented and let someone a chance to fix it first | 14:58 |
| dhellmann | thomasm: so are you saying there's no way to fix this problem? or just that we don't know it? | 14:59 |
| thomasm | dhellmann, I don't know it, litong said there wasn't a way. | 14:59 |
| *** briancline has joined #openstack-metering | 14:59 | |
| dhellmann | jd__: in the other cases where we raise NotImplemented, we are only doing it in a limited case (we don't support sorting, or pagination, or whatever) | 15:00 |
| dhellmann | in this case, though, it completely disables a key method of the driver | 15:00 |
| dhellmann | thomasm: ok, if db2 cannot produce the required data then we shouldn't ship the driver at all | 15:00 |
| dhellmann | it seems a little odd that there's no way to do it, but it probably just means we need a different schema or something | 15:01 |
| dragondm | I suspect the NoSQl support in db2 itself is rather experimental. Thus the lack of aggregation, etc. | 15:01 |
| dhellmann | dragondm: good point | 15:01 |
| dhellmann | do we mark the driver as experimental? or leave it out of the documentation? | 15:02 |
| jd__ | dhellmann: don't blame me, I wish there would be a way to only raise NotImplemented on a subset, but the basic feature is broken | 15:02 |
| thomasm | dhellmann, dragondm jd__, So, if the driver is experiemental (because the backend it is interacting with is) maybe it needs to be in an experimental branch, not master? | 15:02 |
| dhellmann | jd__: yeah, I'm starting to think we should just pull the driver entirely | 15:02 |
| jd__ | dhellmann: haha, litong is going to hate you more than me | 15:03 |
| dragondm | It would be nice to have a way of marking the driver as experimental. | 15:03 |
| thomasm | dragondm, that'd be neat, yeah | 15:03 |
| dhellmann | jd__: let's leave it for h3 but pull it by the rc if it is not fixed | 15:04 |
| *** SergeyLukjanov has joined #openstack-metering | 15:04 | |
| thomasm | We could easily accomplish that with a naming convention, methinks. Any other thoughts? | 15:04 |
| jd__ | dhellmann: fine with me | 15:04 |
| dhellmann | jd__: I'll add something to the agenda for our meeting this week | 15:04 |
| dhellmann | jd__, thomasm : where's the bug for this? | 15:05 |
| dragondm | thomasm: perhaps a decorator. | 15:05 |
| thomasm | dhellmann, lemme get that for you | 15:05 |
| *** flwang has quit IRC | 15:05 | |
| thomasm | dhellmann, https://bugs.launchpad.net/ceilometer/+bug/1208547 | 15:05 |
| thomasm | dhellmann, jd__: So, shall I remove the raise and we just remove the driver if it's not passing those tests by release? Or shall I leave the raise? | 15:06 |
| *** dina_belova has quit IRC | 15:07 | |
| dhellmann | thomasm: leave the raise | 15:07 |
| thomasm | dhellmann, Okey dokey | 15:07 |
| dhellmann | if we're going to pull the driver, and I think we are, then it's fine | 15:07 |
| thomasm | dhellmann, jd__, dragondm: Thanks for your time! =] | 15:08 |
| dragondm | (this db2 issue reminds me of the HyperV driver kerfuffle in Nova.) | 15:10 |
| *** dina_belova has joined #openstack-metering | 15:12 | |
| dhellmann | dragondm: yeah, it seems like there are people contributing in this case, though :-) | 15:13 |
| *** litong has joined #openstack-metering | 15:15 | |
| *** dina_belova has quit IRC | 15:16 | |
| dragondm | Heh, true. | 15:17 |
| *** flwang has joined #openstack-metering | 15:20 | |
| openstackgerrit | A change was merged to openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/44405 | 15:23 |
| thomasm | Oh, right - I remember you telling me about that. | 15:25 |
| *** dina_belova has joined #openstack-metering | 15:27 | |
| thomasm | dragondm, Are you thinking a decorator that would cause the tests not to fail from an experiemental driver and log that it's experimental? | 15:27 |
| thomasm | "Use with caution" | 15:28 |
| thomasm | etc. etc. | 15:28 |
| dragondm | something along those lines. Possibly allowing driver methods to be marked with 'knownbug' or somesuch, too. | 15:29 |
| dragondm | May be usefull idea. May be more trouble than it's worth. | 15:29 |
| thomasm | Hmmm, depends on how often we find ourselves at this stage. :P | 15:30 |
| dragondm | true. | 15:30 |
| thomasm | But, yeah - Hmmm we could add to the documentation from there? Not sure how that automation works yet. | 15:31 |
| dragondm | hmm.. dunno. | 15:33 |
| *** shaneduan[afk] is now known as shaneduan | 15:35 | |
| openstackgerrit | Eoghan Glynn proposed a change to openstack/ceilometer: Add query support to alarm history API https://review.openstack.org/44908 | 15:42 |
| *** shang has joined #openstack-metering | 15:56 | |
| *** shaneduan is now known as shaneduan[afk] | 15:56 | |
| *** jergerber has joined #openstack-metering | 16:12 | |
| eglynn | jd__: just working thru' your review comments, thanks! | 16:14 |
| *** boris-42 has quit IRC | 16:14 | |
| eglynn | jd__: https://review.openstack.org/#/c/44751/1/ceilometer/alarm/service.py ... "If you are looking for a particular extension, go ahead and just use the class" | 16:14 |
| eglynn | jd__: basically do you mean just instatiate the class directly, by-passing the extensions mechanism completely? | 16:14 |
| jd__ | eglynn: yes | 16:14 |
| eglynn | jd__: ok | 16:15 |
| *** shaneduan[afk] is now known as shaneduan | 16:27 | |
| *** tasdomas is now known as tasdomas_afk | 16:34 | |
| *** Bada has joined #openstack-metering | 16:36 | |
| *** dina_belova has quit IRC | 16:45 | |
| briancline | what are the general rules around submitting patches to support additional publishers that have dependencies like protobufs? (dependencies deemed stable by the community at large) | 16:52 |
| briancline | I know other projects are extremely stringent on dependencies to the point of duplicating their code and renaming them, but it seems to vary by project | 16:53 |
| *** flwang has quit IRC | 16:55 | |
| *** Alexei_987 has quit IRC | 16:58 | |
| *** flwang has joined #openstack-metering | 17:07 | |
| *** anteaya has quit IRC | 17:08 | |
| *** dina_belova has joined #openstack-metering | 17:09 | |
| *** SergeyLukjanov has quit IRC | 17:12 | |
| *** dina_belova has quit IRC | 17:13 | |
| *** eglynn has quit IRC | 17:24 | |
| *** SergeyLukjanov has joined #openstack-metering | 17:45 | |
| *** shakayumi has joined #openstack-metering | 17:52 | |
| litong | @eglynn, ping, can you please review this patch, very few lines. https://review.openstack.org/#/c/44514/. | 17:57 |
| *** yjiang5_away is now known as yjiang5 | 17:57 | |
| *** boris-42 has joined #openstack-metering | 17:58 | |
| *** shang has quit IRC | 18:03 | |
| *** shaneduan is now known as shaneduan[afk] | 18:03 | |
| *** shaneduan[afk] is now known as shaneduan | 18:04 | |
| *** eglynn has joined #openstack-metering | 18:06 | |
| *** dina_belova has joined #openstack-metering | 18:13 | |
| *** shaneduan is now known as shaneduan[afk] | 18:13 | |
| *** dina_belova has quit IRC | 18:18 | |
| *** dina_belova has joined #openstack-metering | 18:23 | |
| *** dina_belova has quit IRC | 18:28 | |
| *** shaneduan[afk] is now known as shaneduan | 18:34 | |
| openstackgerrit | Mathieu Gagné proposed a change to openstack/ceilometer: Handle volume.resize.* notifications https://review.openstack.org/44935 | 18:43 |
| openstackgerrit | Mathieu Gagné proposed a change to openstack/ceilometer: Handle volume.resize.* notifications https://review.openstack.org/44935 | 19:03 |
| *** zul has quit IRC | 19:21 | |
| openstackgerrit | A change was merged to openstack/ceilometer: db2 distinct call results are different from mongodb call https://review.openstack.org/44514 | 19:21 |
| *** anteaya has joined #openstack-metering | 19:23 | |
| *** dina_belova has joined #openstack-metering | 19:24 | |
| *** dina_belova has quit IRC | 19:29 | |
| *** zul has joined #openstack-metering | 19:33 | |
| gordc | thomasm: i'm speaking with Tong on how to properly address the db2 issue. after your patch lands, https://review.openstack.org/#/c/44277/, will the tests be correct? need to give the db2 folk the exact differences between what db2 is giving us and what we're expecting. | 19:34 |
| litong | @gordc, the link does not seem to be right. | 19:35 |
| *** eglynn has quit IRC | 19:36 | |
| litong | @gordc, the period at the end mess things up. | 19:36 |
| thomasm | gordc, Yes, the two tests I added specifically address samples being out of order to ensure sorting when returning a resource. | 19:37 |
| gordc | thomasm: awesome. will follow up on that. hope you had fun figuring out db2.lol | 19:38 |
| thomasm | gordc, Ohhh, it was a trip. I think it's good that we have all of these conversations now. =] I think it's helping us really understand how we want to approach these situations going forward. | 19:39 |
| thomasm | gordc, Did you see all of the conversations around how we're handling more experimental drivers? | 19:40 |
| gordc | thomasm: yep. definitely helps if we ever add more backends. i'm surprised HBase is holding up to be honest. | 19:41 |
| thomasm | gordc, Me too. I was able to fix that one in the Python. Reading the HappyBase documentation, it essentially says, "Sorry, you have to sort in your Python". | 19:41 |
| gordc | thomasm: was the discussion earlier today? maybe i'll make use of our new irc logs for once. | 19:41 |
| thomasm | gordc, There was one last Friday and another earlier this morning. | 19:41 |
| thomasm | gordc, There have been several. | 19:42 |
| gordc | thomasm: i'll glance through for your name. | 19:42 |
| thomasm | gordc, I think the reason I addressed DB2 the way I did (deferring to the leads and asking you/litong) was because of my experience with HBase. | 19:43 |
| gordc | thomasm: haha. "Sorry, you have to sort in your Python". wasn't enough of a hint for you to get started on? | 19:44 |
| thomasm | gordc, It was, but that was after a lot of digging. | 19:44 |
| gordc | thomasm: i made the mistake of fixing the backend tests once. HBase was a trip. luckily it was pretty minor fix. | 19:44 |
| thomasm | gordc, You see, being so new to this, I was thinking HappyBase would provide similar querying capability to like PyMongo with Mongo or SQLAlchemy. I was dead wrong and felt as such about a day and a half into digging through documentation. | 19:45 |
| gordc | thomasm: you got a new skill now. add that to your resume: HBase expert. :) | 19:46 |
| thomasm | gordc, LOL | 19:46 |
| thomasm | gordc, I think they'll see it in my face. | 19:46 |
| thomasm | gordc, Nah, HBase looks really interesting. I am curious to see what it's strengths are relative to other solutions. | 19:47 |
| gordc | thomasm: just don't break gaze..lol | 19:47 |
| thomasm | hahaha | 19:48 |
| gordc | thomasm: yeah, i only started to look at mongo and hbase after joining ceilometer. definitely good to learn about them. | 19:48 |
| thomasm | Gotcha. gordc, I honestly didn't know we had specific teams to maintain certain drivers, though it sort of makes sense to allow domain expertise to drive the implementation. Otherwise we end up with a collaborative effort of junior skillsets. | 19:50 |
| litong | @thomasm, just added some code which distingushes between db2 and mongodb. | 19:50 |
| litong | in your patch, you probably can use that. so you can mark the method this. | 19:50 |
| litong | if self._using_mongodb: notImplemented. | 19:51 |
| litong | @thomasm, then I can fix things in db2 driver. | 19:51 |
| gordc | thomasm: backend stuff definitely has gaps, even sql backend is getting fixed constantly. | 19:52 |
| litong | @thomasm, that way, your patch does not have to hold up because you do not know how db2 works. | 19:52 |
| thomasm | gordc, Definitely. We need to get tests around that too. I think we have a ton of blind spots and that's where this pain comes from. | 19:52 |
| litong | @thomasm, unless we get all the right backend on jenkins, and on all develop env. which is not easy. | 19:53 |
| thomasm | But, that's growing pains for you. | 19:53 |
| thomasm | yeah | 19:53 |
| gordc | agreed, definitely need to look a fixing/beefing tests up in Icehouse. | 19:54 |
| litong | @thomasm, did you get the suggestion above? | 19:54 |
| litong | @thomasm, using self._using_mongodb which I just added, it can be used to get around these problems. | 19:55 |
| thomasm | litong, Yes, I saw it - it's the patch from last week; I remember the code. :\ The problem is that this driver seems to be in a more experimental phase right now... | 19:56 |
| litong | @thomasm, when you do rebase again you should have it in. | 19:57 |
| litong | @thomasm, looks like you will need another patchset since there are still some comments. | 19:57 |
| litong | @thomasm, it is in this patchset. https://review.openstack.org/#/c/44514/ | 19:58 |
| thomasm | litong, The comments were resolved this morning, though. The decision from leadership was to continue raising the error. We will be addressing this in the meeting this week. | 19:58 |
| litong | @thomasm, I know, but that is 9/5, will be too late for a patchset to make ot h3. | 19:59 |
| thomasm | litong, Then we should probably have this conversation sooner with them. The problem is manually passing tests for a driver undermines the purpose. | 20:00 |
| litong | @thomasm, at this point, there is no way around it, there is no db2 on jenkins. | 20:00 |
| litong | the db2 driver was meant for db2, not for mongodb. | 20:01 |
| thomasm | litong, Of course, and if there was we'd still have to pass the tests because db2 can't support the operations, according to your points last week. | 20:01 |
| litong | @thomasm, I bet you do not have db2 in your env. right? | 20:01 |
| litong | @thomasm, my point was that we can not use aggregate method to do that same you did for mongodb. | 20:01 |
| litong | @thomasm, I did not say that the entire method can not be done using other methods. | 20:02 |
| litong | @thomasm, you actually did not quote me correctly in earlier discussion. | 20:02 |
| thomasm | litong, That's what I thought you had said last week. You said that if I wanted historical data we could not do that. | 20:03 |
| thomasm | with the current DB2 functionality | 20:03 |
| thomasm | 2013-08-30T15:33:24 <litong> @thomasm, oh, wow, you are looking for a history. I did not get that, | 20:03 |
| thomasm | 2013-08-30T15:33:34 <litong> @thomasm, if that is the case, I do not think you can do that. | 20:03 |
| litong | @thomasm, I said since resource collection does not have that, you can not do it against resource collection. | 20:03 |
| litong | @thomasm, need to use the meter collection, not the resource collection, that is what I am saying. | 20:04 |
| thomasm | litong, That was the point I was trying to make during most of that conversation. We have to use samples since that's the only place for a historical representation. | 20:04 |
| litong | by samples you mean from meter collection, right? | 20:04 |
| thomasm | yes | 20:04 |
| litong | probably some misunderstanding, if you only look at resource collection, then you can not. | 20:05 |
| thomasm | Yeah, because that's only the latest *stored* representation which we also can't trust. | 20:05 |
| litong | if you bring in meter, then you have timestamp, | 20:05 |
| litong | that is what the patch I first proposed did. | 20:06 |
| thomasm | Yes, but I can't aggregate/group, so I guess we'd have to do it in the Python, or over several DB calls to pick out a sample and min(timestamp) for each resource. | 20:06 |
| litong | @thomasm, ok, I think that is where you lost me. I feel we are going around and around. | 20:07 |
| litong | up for a call ? | 20:07 |
| thomasm | litong, I do too. Sure, I can do that. | 20:07 |
| thomasm | Let me brb really fast. | 20:07 |
| litong | let's use my conf #. | 20:08 |
| litong | call this # 888-426-6840 | 20:08 |
| litong | and use this code 63840224 | 20:09 |
| litong | @thomasm, I am in the call now. | 20:10 |
| *** dina_belova has joined #openstack-metering | 20:24 | |
| litong | https://review.openstack.org/#/c/44514/ | 20:28 |
| *** dina_belova has quit IRC | 20:29 | |
| litong | https://review.openstack.org/#/c/44390/ | 20:32 |
| EmilienM | Hi here, if some people from ceilometer team could review this small patch in the doc, I would pay a beer to them > https://review.openstack.org/#/c/44584/ | 20:37 |
| thomasm | litong, https://review.openstack.org/#/c/44277/6 | 20:45 |
| *** evanjfraser has joined #openstack-metering | 20:50 | |
| *** boris-42 has quit IRC | 20:59 | |
| *** SergeyLukjanov has quit IRC | 21:12 | |
| *** sandywalsh has quit IRC | 21:12 | |
| openstackgerrit | Monsyne Dragon proposed a change to openstack/ceilometer: Add configuration-driven conversion to Events https://review.openstack.org/42713 | 21:20 |
| *** sandywalsh has joined #openstack-metering | 21:24 | |
| *** dina_belova has joined #openstack-metering | 21:25 | |
| *** dina_belova has quit IRC | 21:30 | |
| *** eglynn has joined #openstack-metering | 21:30 | |
| *** Bada has quit IRC | 21:30 | |
| *** litong has quit IRC | 21:37 | |
| *** openstackgerrit has quit IRC | 21:48 | |
| *** openstackgerrit has joined #openstack-metering | 21:48 | |
| *** thomasm has quit IRC | 22:02 | |
| *** thomasm has joined #openstack-metering | 22:03 | |
| openstackgerrit | Alessandro Pilotti proposed a change to openstack/ceilometer: Updates oslo-incubator files https://review.openstack.org/44961 | 22:17 |
| *** dina_belova has joined #openstack-metering | 22:25 | |
| *** changbl has quit IRC | 22:28 | |
| *** dina_belova has quit IRC | 22:30 | |
| *** jergerber has quit IRC | 22:48 | |
| *** eglynn has quit IRC | 22:48 | |
| *** evanjfraser has quit IRC | 23:07 | |
| *** evanjfraser has joined #openstack-metering | 23:08 | |
| *** gordc has quit IRC | 23:17 | |
| *** shardy is now known as shardy_afk | 23:21 | |
| *** changbl has joined #openstack-metering | 23:42 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!