| *** _cjones_ has joined #openstack-ceilometer | 00:04 | |
| openstackgerrit | Merged openstack/ceilometer: ensure unique pipeline names https://review.openstack.org/139203 | 00:06 |
|---|---|---|
| *** sbfox has quit IRC | 00:14 | |
| openstackgerrit | Julien Danjou proposed stackforge/gnocchi: indexer: simplify resource ↔ metric relationship https://review.openstack.org/142264 | 00:15 |
| *** ccrouch1 has joined #openstack-ceilometer | 00:19 | |
| *** ccrouch has quit IRC | 00:20 | |
| *** exploreshaifali has quit IRC | 00:27 | |
| *** _cjones_ has quit IRC | 00:30 | |
| *** _cjones_ has joined #openstack-ceilometer | 00:39 | |
| *** ccrouch1 has quit IRC | 00:47 | |
| *** ddieterly has joined #openstack-ceilometer | 00:51 | |
| *** _cjones_ has quit IRC | 00:51 | |
| *** Longgeek has joined #openstack-ceilometer | 00:54 | |
| *** _cjones_ has joined #openstack-ceilometer | 00:55 | |
| *** Longgeek has quit IRC | 00:58 | |
| *** promulo_ has joined #openstack-ceilometer | 01:12 | |
| *** nellysmitt has joined #openstack-ceilometer | 01:13 | |
| *** promulo has quit IRC | 01:16 | |
| *** nellysmitt has quit IRC | 01:18 | |
| *** _cjones_ has quit IRC | 01:20 | |
| *** _cjones_ has joined #openstack-ceilometer | 01:29 | |
| *** _cjones_ has quit IRC | 01:38 | |
| *** gordc has joined #openstack-ceilometer | 01:46 | |
| *** mitz has quit IRC | 01:47 | |
| *** mitz has joined #openstack-ceilometer | 01:52 | |
| *** jaypipes has quit IRC | 01:52 | |
| *** _cjones_ has joined #openstack-ceilometer | 01:58 | |
| openstackgerrit | Merged openstack/ceilometer-specs: Add spec for more power and thermal data https://review.openstack.org/139537 | 02:01 |
| *** _cjones_ has quit IRC | 02:03 | |
| *** ccrouch has joined #openstack-ceilometer | 02:08 | |
| *** ajc___ has joined #openstack-ceilometer | 02:19 | |
| *** gordc has quit IRC | 02:21 | |
| *** sbfox has joined #openstack-ceilometer | 02:26 | |
| *** LinstatSDR has joined #openstack-ceilometer | 02:27 | |
| *** mitz has quit IRC | 02:37 | |
| *** mitz has joined #openstack-ceilometer | 02:37 | |
| *** amalagon has quit IRC | 02:42 | |
| *** sbfox has quit IRC | 02:55 | |
| *** LinstatSDR has quit IRC | 03:14 | |
| *** nellysmitt has joined #openstack-ceilometer | 03:14 | |
| *** changbl has joined #openstack-ceilometer | 03:17 | |
| *** nellysmitt has quit IRC | 03:19 | |
| *** harlowja is now known as harlowja_away | 03:25 | |
| *** fnaval has quit IRC | 03:30 | |
| *** ryanpetrello has quit IRC | 03:39 | |
| *** ddieterly has quit IRC | 03:42 | |
| *** fnaval has joined #openstack-ceilometer | 03:43 | |
| *** mitz has quit IRC | 03:50 | |
| *** amalagon has joined #openstack-ceilometer | 03:52 | |
| *** amalagon has quit IRC | 03:57 | |
| *** _cjones_ has joined #openstack-ceilometer | 03:57 | |
| *** mitz has joined #openstack-ceilometer | 03:59 | |
| openstackgerrit | liusheng proposed openstack/ceilometer: Avoid NoneType error when inspect mem usage failed https://review.openstack.org/142329 | 04:02 |
| *** adriant has quit IRC | 04:03 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements https://review.openstack.org/140049 | 04:04 |
| *** mitz has quit IRC | 04:15 | |
| *** mitz has joined #openstack-ceilometer | 04:16 | |
| *** _cjones_ has quit IRC | 04:23 | |
| *** _cjones_ has joined #openstack-ceilometer | 04:24 | |
| *** ryanpetrello has joined #openstack-ceilometer | 04:36 | |
| *** mitz has quit IRC | 04:42 | |
| *** mitz has joined #openstack-ceilometer | 04:45 | |
| *** promulo_ has quit IRC | 05:01 | |
| *** promulo has joined #openstack-ceilometer | 05:02 | |
| *** _cjones_ has quit IRC | 05:10 | |
| *** mitz has quit IRC | 05:12 | |
| *** mitz has joined #openstack-ceilometer | 05:13 | |
| *** mitz has quit IRC | 05:14 | |
| *** nellysmitt has joined #openstack-ceilometer | 05:15 | |
| *** mitz has joined #openstack-ceilometer | 05:16 | |
| *** nellysmitt has quit IRC | 05:19 | |
| *** mitz has quit IRC | 05:20 | |
| *** mitz has joined #openstack-ceilometer | 05:21 | |
| *** ryanpetrello has quit IRC | 05:22 | |
| *** cmyster has quit IRC | 05:27 | |
| *** mitz has quit IRC | 05:31 | |
| *** _cjones_ has joined #openstack-ceilometer | 05:32 | |
| *** sbfox has joined #openstack-ceilometer | 05:33 | |
| *** mitz has joined #openstack-ceilometer | 05:34 | |
| *** mitz has quit IRC | 05:51 | |
| *** mitz has joined #openstack-ceilometer | 05:52 | |
| *** sbfox has quit IRC | 05:59 | |
| *** sbfox has joined #openstack-ceilometer | 06:00 | |
| *** mitz has quit IRC | 06:00 | |
| *** mitz has joined #openstack-ceilometer | 06:00 | |
| *** sbfox1 has joined #openstack-ceilometer | 06:02 | |
| *** sbfox has quit IRC | 06:02 | |
| *** mitz has quit IRC | 06:03 | |
| *** cmyster has joined #openstack-ceilometer | 06:05 | |
| *** cmyster has joined #openstack-ceilometer | 06:05 | |
| *** sbfox has joined #openstack-ceilometer | 06:06 | |
| *** sbfox1 has quit IRC | 06:06 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ceilometer: Imported Translations from Transifex https://review.openstack.org/139527 | 06:10 |
| *** sbfox has quit IRC | 06:10 | |
| *** sbfox has joined #openstack-ceilometer | 06:10 | |
| *** fnaval has quit IRC | 06:11 | |
| *** fnaval has joined #openstack-ceilometer | 06:12 | |
| *** fnaval has quit IRC | 06:16 | |
| *** sbfox has quit IRC | 06:17 | |
| *** sbfox has joined #openstack-ceilometer | 06:19 | |
| *** _cjones_ has quit IRC | 06:48 | |
| *** _cjones_ has joined #openstack-ceilometer | 06:49 | |
| *** _cjones_ has quit IRC | 06:53 | |
| *** ryanpetrello has joined #openstack-ceilometer | 07:08 | |
| *** ryanpetrello has quit IRC | 07:13 | |
| *** yatin has joined #openstack-ceilometer | 07:16 | |
| *** nellysmitt has joined #openstack-ceilometer | 07:16 | |
| *** pcaruana has joined #openstack-ceilometer | 07:16 | |
| *** nellysmitt has quit IRC | 07:21 | |
| *** Ala has joined #openstack-ceilometer | 07:33 | |
| *** _nadya_ has joined #openstack-ceilometer | 07:46 | |
| *** _nadya_ has quit IRC | 07:52 | |
| *** ajc___ has quit IRC | 07:58 | |
| *** nellysmitt has joined #openstack-ceilometer | 08:23 | |
| *** sbfox has quit IRC | 08:25 | |
| *** sbfox has joined #openstack-ceilometer | 08:25 | |
| *** sbfox1 has joined #openstack-ceilometer | 08:27 | |
| *** sbfox has quit IRC | 08:29 | |
| *** sbfox1 has quit IRC | 08:32 | |
| *** Longgeek has joined #openstack-ceilometer | 08:33 | |
| *** sbfox has joined #openstack-ceilometer | 08:36 | |
| *** _nadya_ has joined #openstack-ceilometer | 08:42 | |
| *** sbfox has quit IRC | 08:45 | |
| *** safchain has joined #openstack-ceilometer | 08:45 | |
| *** eglynn has joined #openstack-ceilometer | 08:48 | |
| *** ryanpetrello has joined #openstack-ceilometer | 08:57 | |
| *** _nadya_ has quit IRC | 09:01 | |
| *** ajc___ has joined #openstack-ceilometer | 09:01 | |
| *** ryanpetrello has quit IRC | 09:01 | |
| *** _nadya_ has joined #openstack-ceilometer | 09:05 | |
| *** sudipto has joined #openstack-ceilometer | 09:11 | |
| sudipto | eglynn: Hi | 09:11 |
| eglynn | sudipto: morning | 09:11 |
| eglynn | sudipto: 'sup? | 09:11 |
| sudipto | eglynn: Apologies for dragging a conversation from a few days back. | 09:12 |
| eglynn | sudipto: please remind me | 09:12 |
| sudipto | eglynn: I actually figured out the resource tracker in nova-compute which gives a plugin based model to gather host state data - that in turn can be used for placement decisions via a filter. | 09:12 |
| eglynn | sudipto: yes, that's the way the nova scheduler works | 09:13 |
| eglynn | sudipto: per-compute resource tracker feeds into the filtering/weighting logic in the scheduler | 09:13 |
| sudipto | eglynn: This resource tracker follows a plugin based model and hence it can be extended into external services too i suppose. | 09:13 |
| sudipto | eglynn: An external service like Ceilometer... | 09:14 |
| sudipto | eglynn: It appears that the nova-compute resource tracker offers two kinds of data that could be collected and reported - static and dynamic. | 09:14 |
| sudipto | englynn: static data seems to have been implemented using a resource - like 'vcpu' | 09:15 |
| sudipto | eglynn: And the 'dynamic' data seems to be reported via a monitor - via a virt driver like 'libvirt' | 09:15 |
| sudipto | eglynn: I want to report data that is based on ceilometer DB. it won't be collected via 'libvirt'...but can be obtained via the Ceilometer DB. | 09:16 |
| *** _nadya_ has quit IRC | 09:16 | |
| eglynn | sudipto: do you mean, (a) ceilometer could provide an alternative nova-resource tracker implementation | 09:16 |
| eglynn | sudipto: ... or (b) ceilometer could listen into the resource-tracker=>scheduler interaction? | 09:17 |
| sudipto | eglynn: https://etherpad.openstack.org/p/IcehouseNovaExtensibleSchedulerMetrics I read this spec for the details and speaks that the external service can extend this resource tracker... | 09:17 |
| eglynn | sudipto: IIRC the resource-tracker=>scheduler interaction is mediated via the nova DB | 09:17 |
| sudipto | eglynn: sort of like the latter. Yeah. The resource tracker currently is operational via the nova DB... But the resource tracker spec talks about future extensibility into external services like Ceilometer. | 09:18 |
| sudipto | eglynn: A data like - 'network bandwidth' may not be obtained via the nova virt drivers. In such a case, we might want to rely on an external service to update that info...I am into that kind of realm of things. | 09:19 |
| eglynn | sudipto: one sec, I need to deal with something else for a minute, back shortly | 09:20 |
| sudipto | eglynn: Sure. | 09:21 |
| eglynn | sudipto: sorry, back | 09:30 |
| eglynn | sudipto: OK, so where were we? | 09:30 |
| sudipto | eglynn: Yeah sure, so basically i am thinking about writing a filter on 'memory bandwidth' and the data for that is gathered by Ceilometer. I thought extending the resource tracker is the best approach. But there has been none of the implementations outside of Nova. | 09:31 |
| eglynn | sudipto: basically the notion of ceilometer providing an alternative nova resource-tracker implementation | 09:31 |
| eglynn | sudipto: OK, so two problems occur to me | 09:31 |
| eglynn | sudipto: 1. the nature of the nova resource-tracker=>scheduler interaction is not via a public mechanism | 09:32 |
| sudipto | eglynn: Yeah you mean via REST or some of sort of de-coupled service... | 09:33 |
| eglynn | sudipto: (i.e. not via a REST API or AMQP notifications, instead it goes via the nova DB IIUC, a private part of the nova implementation) | 09:33 |
| eglynn | sudipto: 2. there are plans afoot to extend the resource-tracker to be more explicitly aware of scheduler concerns such as over-commitment policies | 09:34 |
| DinaBelova | eglynn, ityaptin - fyi http://grafana.org | 09:34 |
| eglynn | (again an internal nova concern, not relevant to ceilometer) | 09:34 |
| DinaBelova | eglynn, here is the playground as well - http://play.grafana.org/#/dashboard/file/default.json | 09:35 |
| eglynn | DinaBelova: yep, Monasca uses this IIUC? | 09:35 |
| DinaBelova | eglynn, well, I do not actually remember what tool did they use | 09:35 |
| DinaBelova | eglynn, but suddenly I understood that's nice thing fof both opentsdb and influxdb | 09:36 |
| eglynn | sudipto: so between #1 & #2 above, seems to me that the resource-tracker is not truly suited as an "extension point" | 09:36 |
| eglynn | DinaBelova: coolness | 09:36 |
| sudipto | eglynn: Ok. The resource tracker spec that i shared had this topic: Potential to use these mechanisms to provide/consume metrics to/from external services | 09:36 |
| eglynn | sudipto: I think that extensibility notion expressed in https://etherpad.openstack.org/p/IcehouseNovaExtensibleSchedulerMetrics is out of date | 09:37 |
| sudipto | eglynn: That made me think that if i want to make a nova filter using a ceilometer gathered data - it could be done via this logic. | 09:37 |
| sudipto | eglynn: Oh ok. | 09:37 |
| *** ajc___ has quit IRC | 09:38 | |
| eglynn | sudipto: I'll confirm my take on that with the nova guys | 09:38 |
| sudipto | eglynn: Ok that'd be great. Do you want me to be a part of the conversation in some sort of way? If so, do let me know...Otherwise i can revert to you on IRC (or any other suitable means you may suggest) | 09:39 |
| *** f13o has quit IRC | 09:47 | |
| DinaBelova | eglynn, may I ask you to take a look on the https://review.openstack.org/#/c/126395/4/specs/kilo/elasticsearch-event-db.rst,cm ? Actually I mean the discussion about how does this spec will coordinate with gnocchi in future, and it looks like overall idea of how events idea will (or won't) live in gnocchi is kind of vagues to me. I belived that we'll store events in gnocchi (probably via indexer concept, ia resources changes history or | 09:54 |
| DinaBelova | whatever), but after some conversaiton with Gordon I'm not so sure | 09:54 |
| eglynn | DinaBelova: I'll have a look | 09:55 |
| DinaBelova | eglynn, thank you sir! | 09:55 |
| idegtiarov | eglynn: DinaBelova: Thank Dina for rising that question, I'm also interested in illumination Events in gnocchi approach. | 09:58 |
| eglynn | sudipto: so I confirmed my understanding of the nova RT->scheduler interaction with a colleague who's more embedded in nova | 10:14 |
| eglynn | sudipto: the upshot is that the nova RT is not really suitable as an external extension point, as I thought | 10:15 |
| eglynn | sudipto: now there is work afoot to stabilize the RT->scheduler interaction | 10:15 |
| openstackgerrit | Lena Novokshonova proposed openstack/ceilometer: Improve tools/make_test_data.sh correctness https://review.openstack.org/141061 | 10:15 |
| eglynn | sudipto: i.e. ensure it's versioned, objectified and all that goodness | 10:15 |
| eglynn | sudipto: ... and potentially to eventually move it out of the nova DB as a next step | 10:15 |
| eglynn | sudipto: but that's all very forward looking, with a time horizon extending from kilo into lemming | 10:15 |
| eglynn | sudipto: so not something to be basing a more current stragey on | 10:16 |
| *** liusheng has quit IRC | 10:21 | |
| *** liusheng has joined #openstack-ceilometer | 10:21 | |
| *** ajc_ has joined #openstack-ceilometer | 10:23 | |
| _elena_ | ildikov, it may be interesting for you https://review.openstack.org/#/c/142139/ | 10:26 |
| liusheng | Hi DinaBelova, could you please take a look at the pagination spec https://review.openstack.org/#/c/105343/, we(ZhiQiang Fan and me) plan to upload patches for this currently. thanks:) | 10:32 |
| DinaBelova | liusheng ok, will take a look | 10:33 |
| liusheng | DinaBelova: thanks a lot :) | 10:33 |
| *** f13o has joined #openstack-ceilometer | 10:34 | |
| DinaBelova | liusheng, +2ed it + approved | 10:35 |
| ildikov | _elena_: sure, I saw the mail about, I will check soon, thanks | 10:35 |
| liusheng | DinaBelova: thanks | 10:36 |
| *** yatin has quit IRC | 10:36 | |
| *** yatin has joined #openstack-ceilometer | 10:37 | |
| *** claudiub has joined #openstack-ceilometer | 10:38 | |
| *** alexpilotti has joined #openstack-ceilometer | 10:41 | |
| idegtiarov | liusheng: Hi! please take a look at my answer for you patch https://review.openstack.org/#/c/141781 | 10:43 |
| idegtiarov | liusheng: suppose I can clarify my proposition | 10:45 |
| *** ryanpetrello has joined #openstack-ceilometer | 10:46 | |
| sudipto | eglynn: Sorry i had stepped out. So it sounds like - if we want to do this, it would have to be tied to Nova - short term? | 10:46 |
| idegtiarov | liusheng: *can meant could | 10:46 |
| ildikov | _elena_: eglynn: can I ask for some clarification on this duplicated samples topic? | 10:46 |
| ildikov | _elena_: eglynn: I've checked the logs, but a quick chat would help for my holiday brain to catch up correctly :) | 10:47 |
| eglynn | sudipto: well, the thought was the nova resource-tracker is not suited to being extended by services outside nova (such as ceilo), for the reasons given above | 10:47 |
| ildikov | _elena_: eglynn: so my question is that the docs should be about to note that sample duplication can happen or do we have a "workaround" for the pipeline config? | 10:48 |
| eglynn | ildikov: can you remind me of which duplicated samples topic you mean? (mentioned in llu's docco patch, or?) | 10:48 |
| ildikov | eglynn: no, the transformer pollster topic, that had a bug report, which was discussed on the last week's meeting IIRC, but I had to leave earlier, so I missed that discussion | 10:49 |
| *** ryanpetrello has quit IRC | 10:51 | |
| eglynn | ildikov: do you mean the discussion after the meeting on https://bugs.launchpad.net/ceilometer/+bug/1401566 ? | 10:51 |
| ildikov | eglynn: yes, that's the one | 10:51 |
| liusheng | idegtiarov: thanks, looking | 10:51 |
| sudipto | eglynn: Hmm, so for hovering around on the same point, but so it means that now, if i have to implement a nova filter for "network" or "memory" bandwidth: 1. It probably is not possible short term. 2. It can be done via the virt drivers like libvirt but not via something like ceilometer. Is that the right understanding? | 10:52 |
| eglynn | ildikov: we decided that the bug report was invalid, as the pipeline.yaml essentially requests that duplication | 10:52 |
| sudipto | eglynn: Bad english. I meant - sorry for hovering around the same point :) | 10:52 |
| ildikov | eglynn: so the note in the docs should be something like 'be aware of meters that are contained by multiple pipelines will be dupplicated'? | 10:53 |
| ildikov | eglynn: I guess it means the all meters and disk and network pipelines in the default config | 10:54 |
| eglynn | sudipto: OK, so the first proposal was to extend the nova RT from *outside* nova (e.g. ceilometer) ... that doesn't look like a good idea | 10:54 |
| eglynn | sudipto: OTOH it's sane to work *within* nova to (a) report extra data up to the scheduler from the resource-tracker, and (b) filter on this data in different ways on the nova-scheduler side | 10:54 |
| ildikov | eglynn: there was smth about the Xen version in the logs, but if I saw right it only polls the rate data, so it shouldn't trigger the disk pipeline, but anyway | 10:55 |
| sudipto | eglynn: Thanks that explains :) | 10:56 |
| eglynn | ildikov: the docco note should say that if the current hypervisor inspector allows the disk and network rates to be gathered directly ... | 10:58 |
| eglynn | ildikov: ... then the disk_source:disk_sink and network_source:network_sink pipelines are not required | 10:58 |
| eglynn | ildikov: and to include them in the pipeline.yaml will lead to duplication | 10:58 |
| _elena_ | ildikov, i did as advised ityaptin, but maybe there is an option to do something differently as a better | 10:58 |
| eglynn | ildikov: or alternatively, the disk_source:disk_sink and network_source:network_sink pipelines in the pipeline.yaml can be thought of as a work-around for libvirt | 10:59 |
| ildikov | eglynn: currently vmware and xen only allows to poll the rate data, I don't know HyperV, but for Libvirt the rate meters are calculated, but in general makes sense | 11:01 |
| eglynn | ildikov: yep, the only hypervisors that allow the rate be gathered directly are vmware & xenapi | 11:01 |
| eglynn | ildikov: IIRC the reason the disk_source:disk_sink and network_source:network_sink pipelines are in the out-of-the-box pipeline.yaml is because the work-around for libvirt pedated the "native" support in vmware & xenapi | 11:02 |
| ildikov | eglynn: but they don't have the raw data, so there shouldn't be any duplication, so we should I guess add a general note that meters that are included in multiple pipelines because of any reason will be duplicated | 11:03 |
| ildikov | eglynn: sure, that's fine, I didn't want to remove it | 11:03 |
| ildikov | eglynn: I just said, that for those two hypervisors we don't have for instance disk.read and disk.write, so the disk pipeline should not be triggered at all | 11:04 |
| ildikov | eglynn: but let's take a step back, as now we are talking about a specific case, which is not really an issue IIUC :) | 11:05 |
| ildikov | eglynn: so basically we would like to note that all the meters that are added to the sources will be polled and if they are added multiple times that will lead to duplication | 11:06 |
| ildikov | eglynn: ... is this correct? | 11:06 |
| eglynn | ildikov: yes, my comments are predicated on the assumption that the same hypervisor allows both disk.write.bytes *and* disk.write.bytes.rate to gathered directly | 11:06 |
| eglynn | ildikov: IIRC, that's what ityaptin stated on IRC after the meeting last week | 11:07 |
| * eglynn checks logs | 11:07 | |
| ildikov | eglynn: as I saw we currently don't have a hypervisor like this | 11:07 |
| * eglynn pastes ... | 11:07 | |
| ildikov | eglynn: the bug is more generic, that is why I suggested to take a step away from the hypervisors | 11:07 |
| eglynn | <ityaptin> eglynn: It's work with defaul pipeline, for meters disk.(read|write).(request|bytes).rates if virt instpector implements method inpect_disk_rates. Xenapi for example | 11:07 |
| eglynn | <ityaptin> eglynn: With disk_sink we create samples 'disk.request.write.rate' in transformer and poll PerDeviceReadBytesPollster by meter_sink. | 11:07 |
| eglynn | ildikov: OK, perhaps I misread the above statement ^^^ if we don't actually have a hypervisor that collects both | 11:08 |
| ildikov | eglynn: looking at the inspector code under xenapi, I only see rates there, but maybe I miss smth there | 11:09 |
| openstackgerrit | Merged openstack/ceilometer-specs: Support pagination for ceilometer APIs https://review.openstack.org/105343 | 11:09 |
| ildikov | eglynn: and in the measurements docco only the *.rate meters are marked for xen and vmware | 11:10 |
| eglynn | ildikov: looking at the vmwame/xenapi code myself just now, I think you are indeed correct | 11:10 |
| eglynn | ildikov: so I'm entirely sure now what ityaptin mean by the above | 11:11 |
| ildikov | eglynn: but the point about the docco here is to warn people if they touch the pipeline.yaml, then they can cause duplication | 11:11 |
| ildikov | eglynn: and what is that you're sure about now? :) | 11:11 |
| eglynn | ildikov: well, if they explicitly duplicate in the pipeline.yaml, then that causes duplication in the samples | 11:13 |
| eglynn | ildikov: ... do we need to explicitly warn about that? | 11:13 |
| eglynn | ildikov: ... not sure that we do, seem kinda self-evident | 11:13 |
| eglynn | ildikov: ... i.e. if you tell ceilometer to do something twice, it'll do the thing twice :) | 11:14 |
| eglynn | ildikov: ... so the confusion was related to the idea that hypervisor was gathering *both* the cumulative and rate meters for disks or network | 11:14 |
| ildikov | eglynn: you can never be sure that it's self evident for a user ;) | 11:14 |
| ildikov | eglynn: as I checked it does not | 11:15 |
| ildikov | eglynn: you checked it too now right? | 11:15 |
| eglynn | ildikov: ... checked that we don't have a hypervisor gathering *both* the cumulative and rate meters for disks or network? | 11:16 |
| eglynn | ildikov: ... yes, I did, and you're right we don't | 11:16 |
| ildikov | eglynn: yes | 11:16 |
| ildikov | eglynn: coolness :) | 11:16 |
| ildikov | eglynn: anyway, a note will not hurt in the admin guide, I just wanted to figure out what exactly we could add this note about | 11:16 |
| eglynn | cool | 11:16 |
| ildikov | eglynn: I mean a note that if you do it twice, it will happen twice | 11:17 |
| eglynn | ildikov: yep, understood, thanks! | 11:17 |
| ildikov | eglynn: thanks for clarification | 11:17 |
| ildikov | _elena_: I will add a comment to your spec based on this chat ^^ and sum it up there | 11:19 |
| ildikov | eglynn: one additional dumb question regarding to storing samples | 11:23 |
| ildikov | eglynn: will the samples be stored in the defined db anyway, or it depends on the defined publisher? | 11:24 |
| eglynn | ildikov: depends on the configured dispatcher | 11:26 |
| eglynn | ildikov: i.e. could be dumped in a file instead of, or as well as, the DB | 11:27 |
| eglynn | ildikov: ... or pushed to remote HTTP API via tongli's new dispatcher | 11:27 |
| ildikov | eglynn: yeap sure, I need to document that new one too... :) | 11:28 |
| ildikov | eglynn: so if we have publishers set in a pipeline, it does not matter | 11:28 |
| eglynn | ildikov: yep, no publishing => no storing | 11:29 |
| eglynn | ildikov: I misread your question ... thought you meant, if we have *no* publishers set in a pipeline | 11:31 |
| ildikov | eglynn: no, so my dumb question is that for instance what happens if we have a udp publisher set in a pipeline | 11:32 |
| ildikov | eglynn: ? | 11:32 |
| eglynn | ildikov: the publisher controls the agent->collector conduit | 11:32 |
| eglynn | ildikov: so we need a collector also listening on UDP | 11:32 |
| eglynn | ildikov: then it dispatches to the DB | 11:32 |
| eglynn | ildikov: so publisher and dispatcher kick in at different points in the chain | 11:33 |
| openstackgerrit | Julien Danjou proposed stackforge/gnocchi: Remove sphinx from Python 3 requirements https://review.openstack.org/142421 | 11:33 |
| *** eglynn is now known as eglynn-afk | 11:34 | |
| ildikov | eglynn: yeap, I know that part, just wasn't sure about the exact chain, but I think I'm ok now | 11:34 |
| *** ajc_ has quit IRC | 11:37 | |
| *** ajc_ has joined #openstack-ceilometer | 11:38 | |
| _elena_ | ildikov, thanks! | 11:38 |
| ildikov | _elena_: np, let me know if you have any questions | 11:40 |
| ildikov | eglynn-afk: thanks again :) | 11:40 |
| _elena_ | ildikov, i tried to correct note as you advised https://review.openstack.org/#/c/142139/4 | 11:46 |
| *** cdent has joined #openstack-ceilometer | 11:56 | |
| ildikov | _elena_: thanks | 11:57 |
| *** ajc_ has quit IRC | 11:58 | |
| *** ajc_ has joined #openstack-ceilometer | 11:59 | |
| *** sudipto has quit IRC | 11:59 | |
| *** ddieterly has joined #openstack-ceilometer | 12:08 | |
| *** Sanskar has joined #openstack-ceilometer | 12:11 | |
| cdent | Oh this is a fine kettle of fish: https://bugs.launchpad.net/ceilometer/+bug/1403280 | 12:20 |
| cdent | DinaBelova: you seen that ^ | 12:20 |
| cdent | the relevant change that busted it: https://github.com/openstack/ceilometer/commit/934a90d4c61ff66f99e5a46480047dec355b5d61 | 12:21 |
| *** yatin has quit IRC | 12:22 | |
| *** ajc_ has quit IRC | 12:23 | |
| *** ryanpetrello has joined #openstack-ceilometer | 12:35 | |
| *** yatin has joined #openstack-ceilometer | 12:39 | |
| *** ryanpetrello has quit IRC | 12:39 | |
| *** eglynn-afk is now known as eglynn | 12:40 | |
| eglynn | cdent: hmmm, revert 934a90d4 before we tag kilo-1? | 12:43 |
| cdent | Assuming the bug is accurate I would guess so, yeah. | 12:44 |
| DinaBelova | cdent, nope | 12:44 |
| cdent | Unless it is something we could fix quickly | 12:44 |
| DinaBelova | eglynn, cdent yeah, it looks like we need to fix this before cutting the kilo-1 | 12:45 |
| eglynn | cdent: I think we'd need some awarness of whether pollsters were expecting a resources list or not | 12:46 |
| eglynn | cdent: doesn't sound like something that should be rushed in before a milestone at the 11th hour | 12:46 |
| eglynn | cdent: (without futher polluting the pollster abstraction) | 12:46 |
| * cdent nods | 12:46 | |
| DinaBelova | eglynn, normal fix -yeah... just revert is quick, but it'll leave all of us with these endless logs | 12:47 |
| cdent | I guess we have a testing gap somewhere | 12:47 |
| eglynn | cdent: how about I revery for now, punt the original bug to kilo-2? | 12:47 |
| cdent | ✔ | 12:47 |
| * DinaBelova nods | 12:47 | |
| *** cdent has quit IRC | 12:58 | |
| openstackgerrit | Eoghan Glynn proposed openstack/ceilometer: Revert "Skip to poll and publish when no resources found" https://review.openstack.org/142435 | 12:59 |
| *** ildikov has quit IRC | 13:01 | |
| eglynn | DinaBelova, cdent: ^^^ | 13:01 |
| DinaBelova | eglynn, +2ed | 13:02 |
| DinaBelova | eglynn, I do not see original bug btw | 13:02 |
| eglynn | DinaBelova: from the orginal review, I don't think there was one :( | 13:02 |
| DinaBelova | eglynn, I suppose we need some | 13:03 |
| eglynn | DinaBelova: thanks! | 13:03 |
| DinaBelova | eglynn, np | 13:03 |
| *** ccrouch has quit IRC | 13:03 | |
| eglynn | DinaBelova: yep, a fresh bug for kilo-2 ... filing now | 13:03 |
| DinaBelova | eglynn, thanks! | 13:03 |
| *** ccrouch has joined #openstack-ceilometer | 13:04 | |
| *** ildikov has joined #openstack-ceilometer | 13:13 | |
| *** Sanskar has quit IRC | 13:15 | |
| *** yatin has quit IRC | 13:22 | |
| eglynn | DinaBelova: FYI comments inline on https://review.openstack.org/#/c/126395/4/specs/kilo/elasticsearch-event-db.rst,cm | 13:47 |
| DinaBelova | eglynn, thanks! will go through them right now | 13:47 |
| *** cdent has joined #openstack-ceilometer | 13:49 | |
| *** jaypipes has joined #openstack-ceilometer | 13:52 | |
| eglynn | DinaBelova: sorry, we collided on the assignment for https://bugs.launchpad.net/ceilometer/+bug/1403505 ... I've reverted it to your original assignment | 14:00 |
| DinaBelova | eglynn, np :) | 14:00 |
| eglynn | ... launchpad not great on refershing the bug view to show other changes | 14:00 |
| DinaBelova | as I've commented, I just though that small code chnage like this will be the nice training for _elena_, that's why I've assigned this to her | 14:00 |
| DinaBelova | eglynn, yeah :D | 14:00 |
| eglynn | DinaBelova: yep, makes perfect sense | 14:01 |
| DinaBelova | eglynn, probably if don't mind, let's give this to her? | 14:01 |
| DinaBelova | oh | 14:01 |
| DinaBelova | already | 14:01 |
| DinaBelova | ЖВ | 14:01 |
| DinaBelova | :D | 14:01 |
| eglynn | DinaBelova: yep, already reverted :) | 14:02 |
| *** thomasem has joined #openstack-ceilometer | 14:03 | |
| *** exploreshaifali has joined #openstack-ceilometer | 14:03 | |
| *** thomasem has quit IRC | 14:03 | |
| DinaBelova | eglynn, hehe, yeah, LP is kind of ridiculous sometimes | 14:03 |
| *** thomasem has joined #openstack-ceilometer | 14:05 | |
| eglynn | DinaBelova: shhh, don't let lifeless hear you say that :) | 14:06 |
| DinaBelova | :D | 14:06 |
| *** julim has joined #openstack-ceilometer | 14:07 | |
| ildikov | sorry guys, I approved that little revert puppy | 14:08 |
| ildikov | ... I saw the quick chat here, so my last info was the aggreement on reverting that change | 14:09 |
| *** eglynn is now known as eglynn-lunch | 14:11 | |
| DinaBelova | ildikov, yes, for sure | 14:13 |
| DinaBelova | we need to revert it now | 14:13 |
| DinaBelova | eglynn-lunch and I were about how to fix it normally and who'll be doing this | 14:14 |
| ildikov | DinaBelova: a-ha, ok, cool :) | 14:14 |
| ildikov | DinaBelova: sure it needs to be fixed normally, but these kind of bugs somehow always appears close to a milestone... | 14:15 |
| ildikov | DinaBelova: Murphy's Law is around us constantly :) | 14:15 |
| DinaBelova | ildikov, well, it's not actually a bug, that's inconvinience | 14:15 |
| DinaBelova | almost | 14:15 |
| DinaBelova | I mean all these dusty logs, etc. | 14:15 |
| ildikov | DinaBelova: I meant the bug that the fix caused | 14:15 |
| DinaBelova | but the old fix for this inconvinience created the *real* bug :) | 14:15 |
| DinaBelova | a-ha | 14:16 |
| DinaBelova | yeah, sorry, did not get your point firsttime | 14:16 |
| DinaBelova | yeah, indeed | 14:16 |
| ildikov | DinaBelova: np, my brain is already in Christmas mode, so annoyed poor eglynn-lunch with some dumb questions today already :) | 14:17 |
| *** ala_ has joined #openstack-ceilometer | 14:17 | |
| *** ddieterly has quit IRC | 14:18 | |
| *** ddieterly has joined #openstack-ceilometer | 14:19 | |
| *** Ala has quit IRC | 14:19 | |
| *** ala_ has quit IRC | 14:20 | |
| *** Ala has joined #openstack-ceilometer | 14:20 | |
| *** fnaval has joined #openstack-ceilometer | 14:21 | |
| DinaBelova | ildikov, my mind is in this mode as well :) | 14:22 |
| ildikov | DinaBelova: :) | 14:23 |
| *** ddieterly has quit IRC | 14:23 | |
| *** eglynn-lunch is now known as eglynn | 14:33 | |
| *** gordc has joined #openstack-ceilometer | 14:41 | |
| *** pcaruana has quit IRC | 14:51 | |
| *** ddieterly has joined #openstack-ceilometer | 14:52 | |
| *** claudiub has quit IRC | 14:57 | |
| *** IvanBerezovskiy1 has quit IRC | 14:58 | |
| *** IvanBerezovskiy has joined #openstack-ceilometer | 14:59 | |
| *** rbak has joined #openstack-ceilometer | 15:10 | |
| *** fnaval has quit IRC | 15:18 | |
| *** fnaval has joined #openstack-ceilometer | 15:18 | |
| *** fnaval has quit IRC | 15:22 | |
| *** exploreshaifali has quit IRC | 15:24 | |
| *** Ala has quit IRC | 15:36 | |
| *** nealph has joined #openstack-ceilometer | 15:36 | |
| *** LinstatSDR has joined #openstack-ceilometer | 15:47 | |
| *** Sanskar has joined #openstack-ceilometer | 15:56 | |
| Sanskar | ! | 15:56 |
| *** prad has joined #openstack-ceilometer | 16:07 | |
| *** Sanskar has quit IRC | 16:20 | |
| *** Sanskar has joined #openstack-ceilometer | 16:21 | |
| *** gordc has quit IRC | 16:23 | |
| *** claudiub has joined #openstack-ceilometer | 16:24 | |
| *** Sanskar has quit IRC | 16:26 | |
| *** Sanskar has joined #openstack-ceilometer | 16:27 | |
| *** Sanskar has quit IRC | 16:39 | |
| *** packet has joined #openstack-ceilometer | 16:56 | |
| *** exploreshaifali has joined #openstack-ceilometer | 16:57 | |
| *** sbfox has joined #openstack-ceilometer | 17:00 | |
| *** harlowja_away is now known as harlowja | 17:01 | |
| *** safchain has quit IRC | 17:04 | |
| *** changbl has quit IRC | 17:05 | |
| *** amalagon has joined #openstack-ceilometer | 17:05 | |
| *** LinstatSDR has quit IRC | 17:17 | |
| *** IvanBerezovskiy has left #openstack-ceilometer | 17:17 | |
| *** LinstatSDR has joined #openstack-ceilometer | 17:18 | |
| *** Sanskar has joined #openstack-ceilometer | 17:27 | |
| *** fnaval has joined #openstack-ceilometer | 17:29 | |
| *** _cjones_ has joined #openstack-ceilometer | 17:34 | |
| *** _cjones_ has quit IRC | 17:34 | |
| *** _cjones_ has joined #openstack-ceilometer | 17:34 | |
| *** _nadya_ has joined #openstack-ceilometer | 17:36 | |
| cdent | eglynn: I recall I promised you a way to play with ceilo + gabbi today. That's still on my list, but I'm not quite there yet. Gonna go for a swim and then get back at it. | 17:36 |
| openstackgerrit | Julien Danjou proposed stackforge/gnocchi: rest: check that the Content-Type is JSON https://review.openstack.org/142517 | 17:42 |
| *** _nadya_ has quit IRC | 17:50 | |
| eglynn | cdent: cool, thanks | 17:50 |
| eglynn | cdent: ... I hope you've got a good wetsuit, or else a solid layer of subcutaneous blubber ;) | 17:50 |
| *** ryanpetrello has joined #openstack-ceilometer | 17:56 | |
| *** exploreshaifali has quit IRC | 17:59 | |
| *** claudiub has quit IRC | 18:06 | |
| *** ryanpetrello has quit IRC | 18:31 | |
| *** ryanpetrello has joined #openstack-ceilometer | 18:50 | |
| cdent | eglynn: went swimming indoors, but I do have a good wetsuit and unfortunately am growing quite the fat layer (thus the desire for a swim) | 18:58 |
| *** nsaje has quit IRC | 19:20 | |
| *** fwdit_ has quit IRC | 19:24 | |
| *** nellysmitt has quit IRC | 19:26 | |
| *** alexpilotti has quit IRC | 19:36 | |
| *** flwang has quit IRC | 19:37 | |
| *** gordc has joined #openstack-ceilometer | 19:40 | |
| *** changbl has joined #openstack-ceilometer | 19:51 | |
| *** _nadya_ has joined #openstack-ceilometer | 19:53 | |
| *** fwdit has joined #openstack-ceilometer | 19:59 | |
| *** ryanpetrello has quit IRC | 20:03 | |
| *** Sanskar has quit IRC | 20:04 | |
| *** nellysmitt has joined #openstack-ceilometer | 20:05 | |
| *** LinstatSDR has quit IRC | 20:10 | |
| *** sbfox has quit IRC | 20:13 | |
| *** flwang1 has joined #openstack-ceilometer | 20:18 | |
| openstackgerrit | Merged stackforge/gnocchi: Remove sphinx from Python 3 requirements https://review.openstack.org/142421 | 20:26 |
| *** drjones has joined #openstack-ceilometer | 20:29 | |
| *** _cjones_ has quit IRC | 20:29 | |
| *** LinstatSDR has joined #openstack-ceilometer | 20:34 | |
| *** _nadya_ has quit IRC | 20:39 | |
| *** gordc has quit IRC | 20:43 | |
| *** r-daneel has joined #openstack-ceilometer | 20:49 | |
| *** drjones has quit IRC | 20:51 | |
| openstackgerrit | Ildiko Vancsa proposed openstack/ceilometer: Remove Sphinx from py33 requirements https://review.openstack.org/142574 | 20:54 |
| *** nellysmitt has quit IRC | 21:11 | |
| *** nealph_ has joined #openstack-ceilometer | 21:13 | |
| *** julim has quit IRC | 21:38 | |
| *** nealph_ has quit IRC | 21:41 | |
| *** nealph_ has joined #openstack-ceilometer | 21:41 | |
| *** flwang1 has quit IRC | 21:42 | |
| *** _cjones_ has joined #openstack-ceilometer | 21:51 | |
| *** _cjones_ has quit IRC | 21:53 | |
| *** _cjones_ has joined #openstack-ceilometer | 21:53 | |
| *** nealph_ has quit IRC | 21:56 | |
| *** promulo__ has joined #openstack-ceilometer | 21:57 | |
| *** promulo has quit IRC | 21:57 | |
| *** changbl has quit IRC | 21:59 | |
| *** srsakhamuri has joined #openstack-ceilometer | 22:07 | |
| *** promulo has joined #openstack-ceilometer | 22:15 | |
| *** promulo__ has quit IRC | 22:16 | |
| *** flwang1 has joined #openstack-ceilometer | 22:22 | |
| openstackgerrit | Phil Neal proposed openstack/ceilometer: Add configuration tables via sqlalchemy migrate scripts https://review.openstack.org/142592 | 22:23 |
| *** prad has quit IRC | 22:26 | |
| *** jaypipes has quit IRC | 22:26 | |
| *** ddieterly has quit IRC | 22:28 | |
| *** LinstatSDR has quit IRC | 22:28 | |
| *** ddieterly has joined #openstack-ceilometer | 22:29 | |
| openstackgerrit | Chris Dent proposed openstack/ceilometer: WIP Hack gabbi into ceilometer in a very basic way. https://review.openstack.org/142594 | 22:30 |
| *** ddieterly has quit IRC | 22:33 | |
| openstackgerrit | Chris Dent proposed openstack/ceilometer: WIP Hack gabbi into ceilometer in a very basic way https://review.openstack.org/142594 | 22:43 |
| *** cdent has quit IRC | 22:44 | |
| *** nealph_ has joined #openstack-ceilometer | 22:53 | |
| *** ryanpetrello has joined #openstack-ceilometer | 22:54 | |
| *** thomasem has quit IRC | 22:55 | |
| *** sbfox has joined #openstack-ceilometer | 22:56 | |
| *** ryanpetrello has quit IRC | 23:03 | |
| *** gordc has joined #openstack-ceilometer | 23:09 | |
| *** nellysmitt has joined #openstack-ceilometer | 23:12 | |
| *** nealph_ has quit IRC | 23:15 | |
| *** nellysmitt has quit IRC | 23:16 | |
| *** salty has joined #openstack-ceilometer | 23:43 | |
| *** salty is now known as Guest1238 | 23:44 | |
| *** Guest1238 is now known as cptsalty | 23:47 | |
| *** sbfox has quit IRC | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!