Wednesday, 2014-10-01

*** sbfox has joined #openstack-ceilometer01:02
*** r-daneel has quit IRC01:04
*** sbfox has quit IRC01:08
*** sbfox has joined #openstack-ceilometer01:18
*** sbfox has quit IRC01:24
*** changbl has joined #openstack-ceilometer01:40
*** _nadya_ has joined #openstack-ceilometer01:46
*** sbfox has joined #openstack-ceilometer01:48
*** nosnos has joined #openstack-ceilometer01:57
*** hhuang has joined #openstack-ceilometer02:03
*** sbfox has quit IRC02:13
*** fnaval has quit IRC02:32
*** vissborg has quit IRC02:41
*** harlowja is now known as harlowja_away02:42
*** vissborg has joined #openstack-ceilometer02:44
*** X019 has quit IRC02:45
*** _nadya_ has quit IRC02:46
*** fnaval has joined #openstack-ceilometer02:49
*** sbfox has joined #openstack-ceilometer02:49
*** sbfox has quit IRC03:09
*** cmyster_away is now known as cmyster03:16
*** ccrouch has quit IRC03:18
*** ccrouch has joined #openstack-ceilometer03:18
*** sbfox has joined #openstack-ceilometer03:20
*** nosnos has quit IRC03:21
*** nosnos has joined #openstack-ceilometer03:22
*** ccrouch has quit IRC03:23
*** ccrouch has joined #openstack-ceilometer03:25
*** nosnos has quit IRC03:26
*** deepthi has joined #openstack-ceilometer03:51
*** sdake_ has joined #openstack-ceilometer03:55
*** ccrouch has quit IRC04:02
*** ccrouch has joined #openstack-ceilometer04:02
*** ccrouch has quit IRC04:02
*** ccrouch has joined #openstack-ceilometer04:02
*** ccrouch has quit IRC04:07
*** sdake_ has quit IRC04:14
*** nosnos has joined #openstack-ceilometer04:20
*** deepthi has quit IRC04:45
*** hhuang has quit IRC04:47
*** ildikov has quit IRC04:49
openstackgerritMonty Taylor proposed a change to openstack/ceilometer: Install retrying before tooz
*** deepthi has joined #openstack-ceilometer05:04
*** amalagon has quit IRC05:10
*** Longgeek has joined #openstack-ceilometer05:11
*** yatin has joined #openstack-ceilometer05:30
*** yatin has quit IRC05:39
*** amalagon has joined #openstack-ceilometer05:46
*** k4n0 has joined #openstack-ceilometer05:47
*** ildikov has joined #openstack-ceilometer05:47
*** X019 has joined #openstack-ceilometer05:57
*** yatin has joined #openstack-ceilometer06:04
openstackgerritOpenStack Proposal Bot proposed a change to openstack/ceilometer: Imported Translations from Transifex
*** IvanBerezovskiy has joined #openstack-ceilometer06:08
*** nsaje has quit IRC06:15
*** zul has quit IRC06:31
*** Ala has joined #openstack-ceilometer06:39
*** ifarkas has joined #openstack-ceilometer06:44
*** zul has joined #openstack-ceilometer06:52
openstackgerritOpenStack Proposal Bot proposed a change to openstack/ceilometer: Updated from global requirements
*** _nadya_ has joined #openstack-ceilometer07:28
*** safchain has joined #openstack-ceilometer07:32
*** sbfox has quit IRC07:45
*** sbfox has joined #openstack-ceilometer07:47
*** lsmola has joined #openstack-ceilometer08:04
*** admin0 has joined #openstack-ceilometer08:14
*** admin0 has left #openstack-ceilometer08:20
*** eglynn has joined #openstack-ceilometer08:21
*** _nadya_ has quit IRC08:21
*** yassine has joined #openstack-ceilometer08:30
openstackgerritA change was merged to openstack/python-ceilometerclient: switch to oslo.utils
*** NellyK has joined #openstack-ceilometer08:53
*** Daviey has quit IRC08:54
*** cdent has joined #openstack-ceilometer08:56
*** NellyK has quit IRC08:58
openstackgerritJoe Hakim Rahme proposed a change to openstack/ceilometer: Run unit tests against MySQL
*** Daviey has joined #openstack-ceilometer09:03
*** ajayaa has joined #openstack-ceilometer09:07
ajayaaDinaBelova, Hi.09:08
DinaBelovaajayaa, good morning :)09:08
ajayaagm. :)09:08
ajayaaHave you come across a situation where a components does not specify resource_id in notification.09:08
DinaBelovaor you probably have a day now :)09:08
ajayaaIt's afternoon.09:09
DinaBelovaajayaa, hm, personally nope09:09
DinaBelovabut I remember it were neutron (?) problems with it09:09
ajayaaIn mdb we don't have a unique identifier for a table in.09:09
ajayaaThe project_name+table_name could be used.09:10
ajayaaIn mdb we use combination of project_id and table name to identify a table.09:10
DinaBelovaajayaa, I guess ityaptin may help you :)09:11
DinaBelovaityaptin, may you take a look on how are we solving it right now?09:11
ajayaaI humbly ask other cores to also give suggestions.09:14
ajayaaeglynn, ^^09:14
silehtajayaa, resource_id is a mandatory field of a sample09:15
eglynnajayaa: the resource_id doesn't have to be a UUID, it just needs to be unique09:15
ajayaaeglynn, Thanks09:15
ajayaasileht, thanks09:15
eglynnajayaa: so if there isn't a "natural" pre-existing identifier in for the magnetoDB resources, then you'd have to fabricate something from the other attributes09:16
ajayaaI will try to combine two fields to create a resource_id.09:16
ajayaaOne other question I have is, Is aggregation of rows handled by ceilometer in any way?09:16
eglynnajayaa: cool, that sounds reasonable (... we've done similar things before, for example with instance NICs IIRC)09:16
eglynnajayaa: aggregation of rows? ... do you mean aggregation of samples?09:17
eglynnajayaa: currently it's only done on-demand via the statistics API09:17
eglynnajayaa: i.e. there is no ongoing eager roll-up of samples09:17
eglynnajayaa: however, adding such continuous downsampling logic is one of the goals of the gnocchi project09:18
ajayaaWe have a use case like, If a user reads x and y number of rows respectively in two requests, we should combine them to one.09:18
ajayaaeglynn, So I guess ceilometer is not the place for such metrics.09:19
eglynnajayaa: I don't fully understand the context there09:19
ajayaaCan I add something to the statistics api to support my use case?09:19
ajayaaBy statistics api, I assume I have to add something to ceilometer.09:20
eglynnajayaa: well, the statistics API is an existing complete part of the ceilometer API09:20
eglynnajayaa: but as I said I don't fully understand your usecase yet09:21
eglynnajayaa: note that we can use the arithmetic transformer to combine two primary samples into a derived sample09:21
ajayaaand would it be done when I call statistics api?09:21
eglynnajayaa: the statistics API allows you to aggregate over a *single* metric09:22
eglynnajayaa: e.g. gimme the average or standard deviation or max for each 5 min period for this meter over this time duration09:22
ajayaaeglynn, I will go through statistics api and ask you more questions.09:23
eglynnajayaa: in your usecase, does magnetoDB emit *separate* samples for X & Y that you want to *combine* into a sample for meter Z?09:23
eglynnajayaa: .... does the magnetoDB notification handler ...09:23
ajayaaNo. A simple use case is, A user reads 10 rows from a table in time t1 and the same user reads 10 more rows from the same table in time t2.09:24
ajayaaI was thinking those two could be combined to a single row in db.09:24
eglynnajayaa: so these times t1 & t2 are emitted by magnetoDB as a separate notifications?09:25
ajayaaeglynn, Yes09:25
eglynnajayaa: i.e. every single read operation is mapped to an individual sample?09:25
eglynnajayaa: that sounds like quite a heavyweight approach, assuming magnetoDB is intended to scale to very large API throughput09:26
ajayaaeglynn, What kind of problems would it create? I thought we would be updating one field in db for each notification.09:27
eglynnajayaa: volume of messages on the AMQP bus, volume of data written to the metering store09:28
eglynnajayaa: how about using a periodic task to emit a notification containing average times per caller in the preceeding X minutes?09:29
eglynnajayaa: the statistics API will allow you to do the aggregation over the raw per-DB-request timings if you take that approach09:30
ajayaaeglynn, That is a nice approach also. We would need to do the bookkeeping in magnetodb.09:30
eglynnajayaa: however I would be concerned about the volume of sample generated in that case09:30
ajayaaeglynn, What I am thinking right now is, let' catch the basic notifications from magnetodb right now and if we face scalability problems later we would look for other ways.09:32
eglynnajayaa: ok, but do bear that aspect in mind09:32
ajayaaI think if zeromq is supported in oslo.messaging we can mitigate the scalability issue of amqp.09:32
silehteglynn, jd__ about heat/ceilometer/autoscaling, we are all OK with endpoint /v1/resource/instance-group and /v1/entity-aggregate ?09:34
eglynnajayaa: zeromq support in the community has been dwindling09:35
ajayaaeglynn, Sadly yes.09:36
silehtajayaa, in oslo.messaging the more promise 'brokerless' driver will be the new AMQP 1.0 protocol driver09:37
eglynnsileht, jd__: I thought at the midcycle we'd decided against the entity-aggregate and separate entity approaches for the AS group09:37
silehtajayaa, the current experimental code doesn't yet provide the brokerless feature, but this is planned by the developers that write this driver09:37
eglynnsileht, jd__: ... and instead decided to support cross entity aggregation in gnocchi based on a resource attribute like server_group09:38
eglynnsileht, jd__: ... e.g. GET /v1/resource/{}/entities/cpu_util&aggregation=6009:38
eglynnsee for some notes taken at the time (read from "only allow cross-entity aggregation ...")09:39
silehteglynn, oh I have missed, I was stuck on the 'Dimensionality for entity measurements' mail09:39
silehteglynn, thx09:40
silehteglynn, I will start hacking around that, so09:41
silehtjd__, ^09:41
eglynnsileht: a-ha, yeah, so IIRC that mail thread happened a week before midcycle, and then continued F2F and morphed into the proposal in the etherpad09:41
eglynnsileht: cool, thanks!09:41
*** nellysmitt has joined #openstack-ceilometer09:42
*** eoutin has joined #openstack-ceilometer09:44
idegtiaroveglynn: Hi! In patch ( I proposed additional client for Mongo ReplicaSet db, do we need this possibility or I should leave only standard one?09:48
eglynnidegtiarov: so my main concern was whether the AutoReconnect-raising semantics were consistent across the 2 mongo client types09:49
eglynnidegtiarov: ... if that is the case, then fine, I can see the advantage in using the replicaset version to avoid having to enumerate all the mongo hosts in the configured URL09:50
idegtiaroveglynn: In my testing raising of exception AutoReconnect was the same in both cases, but now I have doubts that it is not a real problem to configure url with some hosts09:52
idegtiaroveglynn: So the question is how many hosts could be in mongodb in replica set?09:53
eglynnidegtiarov: isn't the replicaset approach more flexible when a new mongo replica is added?09:53
eglynnidegtiarov: i.e. the replicaset client will "discover" the new replica?09:54
idegtiaroveglynn: yes, it will find it09:55
eglynnidegtiarov: ... whereas the normal mongo client would need to have the connection URL config explicitly updated?09:55
eglynnidegtiarov: ... which would also require a service restart to take effect?09:55
idegtiaroveglynn: yes you are right09:56
eglynn(i.e. ceilometer service restart, of the collector & service)09:56
eglynnyeah, so in that case, I think there is some value in using the replicaset client09:56
eglynnidegtiarov: ^^^ also I think the replicaset client will in the future give more flexibility around configuring read preferences09:57
idegtiaroveglynn: I agree09:57
eglynnidegtiarov: e.g. allowing reads from secondaries, or insisting all reads need to go to the primary09:57
*** Longgeek has quit IRC10:01
idegtiaroveglynn:  actually it tried to configure replica client with read_preference eq to PRIMARY_PREFERRED, but I still have an Exception when primary mongo was closed.10:02
*** Longgeek has joined #openstack-ceilometer10:03
*** Longgeek has quit IRC10:04
idegtiaroveglynn:  And another question is do you satisfied with the my and Dina answers on your questions for this patch
*** Longgeek has joined #openstack-ceilometer10:04
idegtiaroveglynn: It still be signed to juno rc110:05
*** eglynn has quit IRC10:06
jd__sileht: cool ide10:11
eglynn-officeafkidegtiarov: juno-rc1 was released yesterday10:12
*** eglynn-officeafk is now known as eglynn-office10:12
*** Longgeek_ has joined #openstack-ceilometer10:13
eglynn-officeidegtiarov: and that bug doesn't show up as targetted on
DinaBelovaeglynn-office, I guess idegtiarov meant that "we need to set some nice milestone for it ":)10:13
DinaBelovanot the released already :)10:13
eglynn-officeidegtiarov: a-ha, it because is marked as duplicate10:13
eglynn-officeeglynn-office: this is the real bug
eglynn-officeidegtiarov: ^^^10:14
eglynn-officeidegtiarov: I'll retarget to kilo-1, makes sense?10:14
DinaBelovaeglynn-office, idegtiarov it works for me10:15
idegtiaroveglynn-office:  thank!10:15
eglynn-officeidegtiarov: you'll need to change the Closes-Bug tag in as it refers now to the duplicate bug10:16
*** Longgeek_ has quit IRC10:16
*** Longgeek has quit IRC10:16
idegtiaroveglynn-office: ok, I'll do that10:16
*** Longgeek has joined #openstack-ceilometer10:16
*** Longgeek_ has joined #openstack-ceilometer10:22
*** Longgeek_ has quit IRC10:23
*** Longgeek_ has joined #openstack-ceilometer10:24
*** Longgeek has quit IRC10:26
*** nsaje has joined #openstack-ceilometer10:27
*** Longgeek_ has quit IRC10:28
*** Longgeek has joined #openstack-ceilometer10:28
*** Longgeek has quit IRC10:30
*** Longgeek has joined #openstack-ceilometer10:31
*** Longgeek_ has joined #openstack-ceilometer10:33
jd__sileht: I've added you to gnocchi-core (Cc eglynn-office)10:33
jd__if you guys can do some review that'd help10:33
silehtjd__, ack10:33
*** Longgeek has quit IRC10:36
eglynn-officejd__: cool10:45
*** X019 has quit IRC10:46
*** Longgeek_ has quit IRC10:56
*** _nadya_ has joined #openstack-ceilometer10:58
*** X019 has joined #openstack-ceilometer10:58
openstackgerritA change was merged to stackforge/gnocchi: Add config generator support
openstackgerritA change was merged to stackforge/gnocchi: sqlalchemy: remove with_for_update()
openstackgerritA change was merged to stackforge/gnocchi: tests: fix race condition on archive policies
openstackgerritA change was merged to stackforge/gnocchi: indexer: remove entities from Entity
openstackgerritA change was merged to stackforge/gnocchi: rest: add support for create/get archive policies
*** asalkeld has left #openstack-ceilometer11:15
*** asalkeld has joined #openstack-ceilometer11:17
*** Longgeek has joined #openstack-ceilometer11:20
*** Longgeek has quit IRC11:21
*** Longgeek has joined #openstack-ceilometer11:22
openstackgerritJulien Danjou proposed a change to stackforge/gnocchi: rest: enable Keystone auth_token middleware by default
*** Ala has quit IRC11:26
*** asalkeld has left #openstack-ceilometer11:27
*** Longgeek has quit IRC11:28
*** Longgeek has joined #openstack-ceilometer11:28
*** Ala has joined #openstack-ceilometer11:36
*** ajayaa has quit IRC11:37
openstackgerritA change was merged to stackforge/gnocchi: Update oslo-incubator
*** vkmc has joined #openstack-ceilometer11:42
openstackgerritA change was merged to stackforge/gnocchi: rest: enable Keystone auth_token middleware by default
*** _nadya_ has quit IRC11:44
vkmcjd__, eglynn-office hi! I was wondering if you were willing to mentor in next OPW round :)11:46
vkmcor, in general, if someone want to mentor an OPW applicant in next OPW round11:47
vkmcthere is an applicant interested in Ceilometer11:47
jd__why not11:47
vkmcfor those who are not aware of the program, here is some information :)
vkmchey jd__! :)11:48
DinaBelovavkmc, hehe, nice possibility for nice women :)11:49
vkmcDinaBelova, indeed! :)11:50
silehtjd__, how can we disable the middleware now ? ;)11:50
jd__sileht: via the option, why?11:50
DinaBelovavkmc, someone from the core team is supposed to be a mentor?11:51
* DinaBelova is not the specialist11:51
silehtjd__, I don't know how to create a empty list in the config file11:51
DinaBelovabut that's interesting quesiton :)11:51
jd__sileht: middlewares= ?11:51
silehtjd__, if I put just 'middlewares = ', it try the load the middleware ''11:51
silehtjd__, oslo.cfg returns ['']11:51
jd__crappy lib11:52
eglynn-officevkmc: hey! yeah I'm still mentoring amalagon from the current OPW round, so hadn't yet thought about the next round11:52
jd__sileht: so let's fix that in Gnocchi and oslo.config, which one do you want to fix? :)11:52
vkmcDinaBelova, not neccesarily :)11:52
eglynn-officevkmc: ... I'll put my thinking cap and see if I can come up with some project ideas11:52
vkmceglynn-office, that's cool!11:52
vkmceglynn-office, is the idea in the ideas page still available?
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: [MongoDB] Fix bug with 'bad' chars in metadatas keys
eglynn-officevkmc: that's Ana's in-progress project in the current round ... looks like the ideas page hasn't been refreshed since the last application period11:54
silehtjd__, I will fix gnocchi first to be able to works :)11:55
vkmcDinaBelova, OPW don't demand much from the mentor.. the idea is to help women to get involved with open source communities so it's good if the mentor knows a lot about the code base but they don't have to be experts11:55
jd__sileht: ok11:55
vkmcDinaBelova, if you are interested, eglynn-office and jd__ can give you more details on how is to mentor someone in OPW :)11:55
*** Longgeek has quit IRC11:55
vkmceglynn-office, good to know... yeah I have been reaching mentors to do so, we are getting lots of applicants this round :o11:56
jd__sileht: I'm writing a patch for oslo.config in the mean time11:57
eglynn-officeDinaBelova: just forwarded you an introductory mail from annegentle setting out expectations for mentors etc.11:57
vkmcjd__, eglynn-office, DinaBelova, nellysmitt is the OPW applicant I was talking about11:59
DinaBelovaeglynn-office, vkmc thanks!12:00
* DinaBelova reading the mail12:00
jd__sileht: thumbs up12:00
DinaBelovanellysmitt, hello :) btw, what's your TZ? :)12:02
nellysmitthi :) UTC + 4 (Moscow time) now12:03
DinaBelovanellysmitt, perfect :D12:03
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Add a gnocchi dispatcher
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Don't try to load empty middleware
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Add server_group to instance resource
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Fix typo in gnocchi wsgi script
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Keep gnocchi/ empty
DinaBelovaeglynn-office, jd__, vkmc, nellysmitt - actually I'm interested in this activity12:04
nellysmittbut our government mess it up, they gonna permamently change it *again* this fall ... so it will be  -1 hr (closer to EU time)12:04
vkmcDinaBelova, that's so great to hear!12:05
DinaBelovanellysmitt, your "government" phrase looks like my Russian one :D12:05
DinaBelovaThey are going to freeze the 'winter' time forever this fall12:06
eglynn-officeDinaBelova: cool :)12:06
DinaBelovanellysmitt, indeed12:06
DinaBelovaeglynn-office, are you ok with it?12:07
*** sdake_ has joined #openstack-ceilometer12:07
* DinaBelova reading the contract12:07
*** sdake_ has quit IRC12:07
*** sdake_ has joined #openstack-ceilometer12:07
*** isviridov is now known as isviridov_na12:07
eglynn-officeDinaBelova: yeah of course, it would be great to have another OPW intern working on ceilometer12:07
* DinaBelova is trying to find any referencies to sell the house/car to the Gnome Foundation12:07
DinaBelovaeglynn-office, cool-cool12:08
eglynn-officeDinaBelova: (IIRC that would be our 3rd one after Terri and Ana)12:08
DinaBelovaeglynn-office, it looks so, yeah12:08
DinaBelovaeglynn-office, I was not looking through the list of possible tasks here12:08
DinaBelovaeglynn-office, is the link published by vkmc valid?12:09
eglynn-officeDinaBelova: yeah, I agree the contract is way OTT ... but at least the GNOME foundation is being upfront about it this time12:09
DinaBelovaeglynn-office :)12:09
eglynn-officeDinaBelova: link valid? ... this one, do you mean ?12:10
DinaBelovaa-ha, me reading it now :)12:10
DinaBelovalooks nice task btw12:10
eglynn-officeDinaBelova: that's amalagon's project in the current round12:10
eglynn-officeDinaBelova: ... the proposal predated gnocchi, so the actual work was rebased on that12:11
DinaBelovayeah, I see that - that's why I'm asking why it has strange status on this wiki page (Outdated: Waiting for mentor confirmation)12:11
DinaBelovaand if there are any other ideas on surface :)12:12
*** yatin has quit IRC12:12
DinaBelovaeglynn-office ^^, sorry, missing your IRC name everywhere it's possible :)12:13
vkmcheh I added that to distinguish current/last round project ideas12:13
DinaBelovavkmc, a-ha12:14
eglynn-officeDinaBelova: here's that edit
DinaBelovaeglynn-office, what do you think about some ideas being collected for the next round?12:14
DinaBelovaa-ha, I got it12:14
openstackgerritA change was merged to openstack/ceilometer: Run unit tests against MySQL
*** admin0 has joined #openstack-ceilometer12:15
eglynn-officeDinaBelova: I'm gonna remove that idea from the page since it relates to the last round ... but yes, good to collect fresh ideas there12:15
DinaBelovaeglynn-office, nellysmitt, vkmc - for now I have no clear ideas for the next round :) I guess eglynn-office will help me here :D12:15
nellysmittok :) thanks anyway12:16
DinaBelovanellysmitt, np :)12:16
vkmcmaybe nellysmitt could start working on her first contribution12:16
vkmcwhile you look for a project for her12:16
DinaBelovavkmc, eglynn-office - will it be okay to say my final decision tomorrow? If I'll have enough time for sure to beocme a mentor?12:17
vkmcDinaBelova, sure thing12:17
nellysmittof course12:17
vkmcDinaBelova, take your time and feel free to reach previous round mentors to ask them about time commitments and so :) flaper87 and jpich volunteered to mentor new mentors12:18
*** rbowen has joined #openstack-ceilometer12:18
DinaBelovavkmc, a-ha, thanks!12:20
*** hhuang has joined #openstack-ceilometer12:20
vkmcDinaBelova, and thanks for thinking about it :)12:21
eglynn-officeDinaBelova: cool, no problem :)12:21
vkmclooks like this could work as a first contribution12:26
vkmcthe author has abandoned the change... I wonder if he is still working on that12:26
*** sdake_ has quit IRC12:29
*** admin0 has quit IRC12:29
eglynn-officeDinaBelova: ^^^ IIRC one of your colleagues had a script to automate the assertEqual param re-ordering in the ceilometer units?12:30
eglynn-officevkmc: yeah, that would be good community service all right (though not necessarily very educational about ceilometer)12:31
eglynn-officevkmc: ... gotta start somewhere though I guess12:31
vkmceglynn-office, hmm I'm afraid so yeah12:31
vkmcyeah that's the idea12:31
vkmcwhat about this one?
vkmcis a docs bug, but it requires learning a bit about ceilometer's tests12:32
eglynn-officevkmc: yeap, a doc patch is always a good confidence-builder12:32
ildikovvkmc: eglynn-office: that patch simply expired because of the -1 from Jenkins, I don't know how much effort would that be to rebase it, maybe it could worth a try12:32
*** _nadya_ has joined #openstack-ceilometer12:34
ildikovvkmc: if the docco bug will be the one nellysmitt can ask me regarding to sphinx (docco build) issues, if I'm around, I have some experience with it12:34
vkmcildikov, sounds great! thanks!12:35
ildikovvkmc: nellysmitt: of course other questions are welcomed also, I will try to help, if I can :)12:35
vkmcnellysmitt, would you like to tackle this?
vkmc:D awesome12:35
*** julim has joined #openstack-ceilometer12:36
eglynn-officeildikov: cool, thank you!12:37
nellysmitthmm, I think I can try :)12:39
ildikovvkmc: nellysmitt: the ceilo client bug can be useful to get experience with git/gerrit commands like rebase, and also running the unit tests with tox, etc without too much coding effort12:39
ildikovvkmc: nellysmitt: I will try to check if the patch can be rebased easily enough to the master branch before my next meeting12:40
nellysmittildikov: ok thanks :)12:41
vkmcildikov, thanks for that, I think it would be a nice practice with git/gerrit too12:41
*** hhuang has quit IRC12:42
*** Longgeek has joined #openstack-ceilometer12:42
*** Longgeek has quit IRC12:42
*** Longgeek has joined #openstack-ceilometer12:43
nellysmittcould you give me a short plan what to do/install first?12:44
vkmcnellysmitt, here is a guide on how to install your development environment
vkmcand this one... using DevStack
eglynn-officeDinaBelova: we could brainstorm some project ideas here
ildikovnellysmitt: you will need tox and then you can run both unit tests and doc build with tox12:46
eglynn-officeDinaBelova: ... I've gotten the ball rolling with an extract from an email I sent last time round on a thread discussing potential OPW ideas12:46
eglynn-officeDinaBelova: (... we ended up rejecting that idea, and went with the period-spanning stats instead)12:46
*** Longgeek_ has joined #openstack-ceilometer12:46
ildikovvkmc: nellysmitt: I checked the client patch, it has many constraints, but looks straightforward to fix it, at least in the filed I checked12:46
*** Longgeek_ has quit IRC12:47
DinaBelovaeglynn-office, a-ha, thanks for the etherpad creation!12:47
ildikovvkmc: nellysmitt: in one part of the conflict you can find the current state of the code and the other part is the change, they are easily identifyable blocks, so ugly work, but does not seem to be difficult12:48
*** Longgeek_ has joined #openstack-ceilometer12:48
nellysmittok I'll start doing it :) whom can I ask if I have any questions?12:48
ildikovvkmc: nellysmitt: so maybe you can check and practice how to solve merge conflicts in gerrit, if you're interested and then solve the issue of the original bug, which is also an easy job, so in general it looks a good practice regarding how git/gerrit works, if it is needed12:49
*** Longgeek has quit IRC12:49
eglynn-officenellysmitt: thanks! ... just pop up on this channel if you've questions12:50
nellysmittyeah, that's a good idea ... never worked with gerrit before (sure worked with git)12:50
ildikovnellysmitt: I will drop off soon, but as eglynn-office mentioned, just drop your questions here, someone, who's around should answer12:50
nellysmittok thanks a lot guys :)12:51
vkmcthanks again ildikov, it's a great task to start contributing to the stack12:51
ildikovnellysmitt: ok, then I suggest to check out what the bug report is about and then check that ceiloclient patch and if you think you can deal with it, then I will assign that bug to you and help if you need12:52
vkmcnellysmitt, if you need help about the OPW process itself, just let me know12:52
nellysmitt:) thx all, sure I will, need to start first :)12:53
ildikovnellysmitt: first just set up your env based on the docs vkmc linked above and then run the unit tests and a docco build against master as it should work after solving possible missing dependencies issues12:55
nellysmittildikov: will do :)12:56
*** thomasem has joined #openstack-ceilometer12:56
ildikovnellysmitt: I need to drop off now, as jd__ would say: happy hacking! :) and also welcome :)12:57
nellysmitthihi :D12:58
*** gordc has joined #openstack-ceilometer12:59
*** ildikov has quit IRC13:04
*** ildikov has joined #openstack-ceilometer13:05
*** ildikov is now known as ildikov_afk13:05
*** admin0 has joined #openstack-ceilometer13:14
*** nosnos has quit IRC13:28
*** nosnos has joined #openstack-ceilometer13:29
*** nosnos has quit IRC13:33
*** nellysmitt has quit IRC13:36
*** jaypipes has joined #openstack-ceilometer13:42
*** nellysmitt has joined #openstack-ceilometer13:47
* cdent must go to bank13:48
*** hhuang has joined #openstack-ceilometer13:52
*** deepthi has quit IRC13:53
gordcanyone have any problems with me branching events into it's own database (similar to what we did for alarms)?13:53
eglynn-officegordc: nope, that seems like a reasonable pattern to follow13:54
gordceglynn-office: cool cool. just starting to look at events so i figured that's step one.13:55
*** ildikov has joined #openstack-ceilometer14:01
*** ildikov_afk has quit IRC14:02
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Ensures only one resource is created into the backend
silehtjd__, I guess you will find this bug funny:
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Add a gnocchi dispatcher
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Keep gnocchi/ empty
jd__sileht: uh? is this supposed to pass?14:11
silehtjd__, it doesn't pass :(14:11
*** fnaval has quit IRC14:12
jd__sileht: yeah that's not surprising14:12
jd__sileht: the tests use only one database which is never cleaned up14:12
*** fnaval has joined #openstack-ceilometer14:12
jd__so you can't do this kind of test14:12
jd__I wouldn't mind having an empty db for each test, but so far I managed to write the test in some smart way and it never has been a problem :)14:13
silehtjd__, I have written the API endpoint for  /v1/resource/instance/
silehtjd__, and when I have use list_resources to get all resources with, the list is never what I expect14:15
silehtjd__, I will generate a server_group name different for each tests, it should avoir the problem14:16
openstackgerritA change was merged to stackforge/gnocchi: Don't try to load empty middleware
*** fnaval has quit IRC14:17
openstackgerritPradeep Kilambi proposed a change to openstack/ceilometer: Fix neutron client to catch 404 exceptions
silehtjd__, my test pass now \o/14:17
*** openstackgerrit has quit IRC14:18
*** openstackgerrit has joined #openstack-ceilometer14:18
jd__yeah that's what you should do indeed14:18
silehtjd__, in the ethernet I have seen {server_group=foobar} as filter, this means we plan to use multiple filters like: {server_group=foobar,display_name=barhost} ?14:21
jd__ethernet haha :)14:21
jd__let me check the interpad14:21
*** nellysmitt has quit IRC14:23
jd__sileht: likely14:24
gordcjd__: you mean 'let me *edit* the etherpad'14:24
jd__I'm just not sure about the URL format14:24
jd__gordc: haah14:24
*** ildikov has quit IRC14:25
silehtjd__, for now, I have implemented: /v1/resource/instance/
silehtjd__, but I can easly do /v1/resource/instance/{,display_name=toto}/entities/cpu_util&aggregation=6014:25
jd__that sounds attracting14:26
silehtjd__, or better /v1/resource/instance/{server_group:foobar,display_name:toto}/entities/cpu_util&aggregation=6014:26
jd__: is better than = ?14:26
silehtjd__, directly a dict14:26
jd__I guess that can be unicode so likely it's gonna be OK14:27
jd__I don't see limitation14:27
jd__and I assume we won't use = or : in the key names :)14:27
jd__sileht: though I think = is more clear for a filter, WDYT?14:28
jd__sileht: you could be fancy and replace the , by & to indicate it's an AND so we can add support for | later ? ;)14:28
silehtjd__, yes, and it's easy to convert that to dict too14:28
*** rbak has joined #openstack-ceilometer14:28
silehtjd__, hum, I afraid that pecan won't like a & at this place14:29
silehtthis remembers me the metadata filtering of the APIv2 of ceilometer14:30
jd__oh right14:30
jd__sileht: then we use ∧ ? :p14:31
jd__'LOGICAL AND' (U+2227)14:31
silehtjd__, that one should workds :)14:31
silehtjd__, I go for /v1/resource/instance/server_group=foobar∧display_name=toto/entities/cpu_util&aggregation=6014:35
eglynn-officesileht: looks good to me14:37
eglynn-officesileht: ... is the idea to extend later to be equivalent to the ceilo v2 api complex query semantics? (... add logical or, not and bracketing)14:37
*** dhellmann has joined #openstack-ceilometer14:38
*** fnaval has joined #openstack-ceilometer14:38
*** dhellmann has quit IRC14:38
silehteglynn-office, it should be simplier and safer that ceilo v2 api, because everything have a type now _o/ :)14:38
eglynn-officesileht: yeap, no more free-form metadata fun14:39
*** pnascimento has joined #openstack-ceilometer14:40
*** dhellmann has joined #openstack-ceilometer14:41
*** _nadya_ has quit IRC14:43
*** nellysmitt has joined #openstack-ceilometer14:45
pnascimentohello stackers :)14:46
*** openstackgerrit has quit IRC14:47
*** openstackgerrit has joined #openstack-ceilometer14:48
gordcpnascimento: hey, whatsup?14:49
*** alexpilotti has joined #openstack-ceilometer14:50
pnascimentoI have some questions, if anyone can help me :)14:52
pnascimento1st question: In this new update of Ceilometer API (v2) the response of /meters is OldSample or Sample?  Because this appears on the documentation: "This class is deprecated in favor of Sample." but in the response to the "ceilometer sample-list" I receive list of OldSamples, expected behaviour?14:53
pnascimento2nd: I am developing a system on Java that consumes Ceilometer API, this consumption takes place in two steps: 1) get token which works successfully with Keystone API, 2) get samples on Ceilometer, which returns a 400 Bad Request with "{"error_message": {"debuginfo": null, "faultcode": "Client", "faultstring": "Missing argument: \"samples\""}}" and on the documentation there is no "sample" argument of request body14:55
pnascimentoIf anyone can help, I will be very thankful:)14:56
*** dhellman_ has joined #openstack-ceilometer14:57
*** dhellmann_ has quit IRC14:57
gordcpnascimento: i'd think the first part would probably be a oversight -- probably should be considered a bug but i'm not sure from a backward-compatibility standpoint15:00
gordcpnascimento: is your system using the client or calling the api directly?15:01
*** dhellmann_ has joined #openstack-ceilometer15:01
pnascimentogordc: a bug? how so? the expected output should be Sample, right?15:02
*** openstackgerrit has quit IRC15:02
pnascimentogordc: on my Java application I'm calling the API15:02
*** openstackgerrit has joined #openstack-ceilometer15:03
pnascimentogordc: the interaction with Keystone to get the token, works perfectly...15:03
gordcpnascimento: yeah. i'd think you should get Sample and not OldSample but it might be that way to preserve backward-compat... (although i think only real diff is change in attribute names.)15:04
*** isviridov_na is now known as isviridov_meetin15:04
*** dhellmann_ has quit IRC15:05
*** dhellmann_ has joined #openstack-ceilometer15:06
pnascimentogordc: this backward compatibility is by configuration or forced, if it is a configuration where can I define that ?15:06
gordcpnascimento: i'm not sure there's a way... i would probably open a bug saying sample-list should call Sample and not OldSample.15:08
*** nellysmitt has quit IRC15:08
gordcwhat's the api call you are using when you get the error message? my only guess is you are not passing in a meter name when retrieving samples.15:08
pnascimentogordc: this is a test Sample =>
eglynn-officepnascimento: yeah, the meters endpoint returns OldSample whereas the samples endpoint returns [New]Sample15:12
eglynn-officepnascimento: so if you really needed the Sample as opposed to OldSample representation15:12
silehtjd__, I continue to have the ipc issue, when I run many times the tests suite on tooz15:13
jd__sileht: what's the traceback?15:13
eglynn-officepnascimento: ... you could replace /v2/meters/cpu for example with /v2/samples?q.field=meter&q.op=eq&q.value=cpu15:14
silehtjd__, this is because sysv_ipc.Semaphore().remove() is never called15:14
jd__sileht: ah right :/15:14
eglynn-officepnascimento: but as gordc says, it should be pretty much the same info except labelled differently15:14
pnascimentoeglynn-office: confirm something, please, those development changes aren't updated on documentation, I am right?15:15
silehtjd__, the tooz IPCLock have a destroy method but it never called15:15
eglynn-officepnascimento: "development changes"?15:15
jd__sileht: yeah it was not part of the tooz protocol15:15
jd__we should fix Gnocchi I guess15:15
silehtjd__, I guess it should be called when we are sure we are the latest client that call release()15:16
pnascimentoeglynn-office: sorry about the english, I meant "code" or "implementation"...15:16
eglynn-officepnascimento: is generated from the code so it correctly identifies OldSample as the representation returned from /v2/meters15:16
silehtjd__, yes but in gnocchi the lock name depends on the entity id ...15:16
eglynn-officepnascimento: see
jd__sileht: or can we fix it in tooz by tracking get_lock() calls and refcount?15:17
silehtjd__, having somthing to list all locks would be cool15:17
jd__like override __del__ on lock or something15:17
jd__CRAZY SHIT15:17
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: [MongoDB] Fix bug with 'bad' chars in metadatas keys
*** mitz_ has quit IRC15:18
pnascimentoeglynn-office: I did see that :) but I could not find documentation about /v2/samples :(15:18
*** dhellman_ has quit IRC15:20
pnascimentoeglynn-office: I tried http://localhost:8777/v2/samples?q.field=meter&q.op=eq&q.value=cenas and it worked perfectly as you said it :) Thank you very much :)15:20
pnascimentoeglynn-office: and I have to say, that with one answer from you, I think I have resolved my two problems :D ty15:21
*** yassine has quit IRC15:21
amalagonhey jd__ : just wanted to make sure, there's currently no way to get the archive_policy just from knowing the entity_id alone?15:22
eglynn-officepnascimento: np! ... yeah, fair point about the docco15:23
*** mitz_ has joined #openstack-ceilometer15:23
eglynn-officepnascimento: ... we do cover complex queries on the samples endpoint
eglynn-officepnascimento: ... but not simple queries15:23
*** Ala has quit IRC15:24
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: Implement redesigned separator in names of columns in HBase
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer: [HBase] Add migration script for new row separate design
*** sbfox has quit IRC15:26
*** dhellmann_ has quit IRC15:26
*** dhellman_ has joined #openstack-ceilometer15:27
*** dhellman_ has quit IRC15:28
*** dhellmann_ has joined #openstack-ceilometer15:28
pnascimentoeglynn-office: I should have read all the documentation, I missed this /v2/query/samples on Complex Queries :( but at this stage I do not pretend do {Simple, Complex} Querying, I just need to "list" the samples of a given meter, but you already helped at that... once again, thank you :) (thankfully, I waited for someone to answer me and did no created a  useless bug :P )15:28
jd__amalagon: humm no but we should add that15:30
jd__amalagon: adding to my TODO15:30
eglynn-officepnascimento: well if you want to be completeist about it, you could file a bug on the absence of docco on the /v2/samples endpoint :)15:31
amalagonjd__: cool, thanks!15:32
*** dhellmann has quit IRC15:34
*** dhellmann_ is now known as dhellmann15:34
*** jasondotstar has joined #openstack-ceilometer15:34
*** yassine has joined #openstack-ceilometer15:36
*** hhuang has quit IRC15:38
*** eoutin has quit IRC15:40
pnascimentoeglynn-office: on this endpoint /v2/samples can I mix the query on querystring with a Complex Query object on body request? or can I have it all in the querystring? like "orderby" or "limit" ?15:41
eglynn-officejd__, amalagon: ... would indexer.get_resource('entity', entity_id) not give the relevant archive_policy in the entity returned?15:42
*** mikedillion has joined #openstack-ceilometer15:42
jd__eglynn-office: ah you're right15:44
jd__amalagon: ^15:44
amalagonohh, let me try that15:44
eglynn-officepnascimento: /v2/samples is a different endpoint to /v2/query/samples ... it supports limit, but not orderby or complex constraints (combined with logic OR, NOT etc.)15:44
jd__amalagon: GET /v1/resource/<entity-id> should return it15:44
*** X019 has quit IRC15:46
*** X019 has joined #openstack-ceilometer15:46
eglynn-officeamalagon: yeah, use what jd__ suggests ^^^ if you're accessing it from outside the API service15:47
eglynn-officeamalagon: ... but, if you want to access the archive policy from within a custom aggregator, more direct to go straight to the indexer15:47
pnascimentoeglynn-office: I have found out that by trying :P15:48
amalagoneglynn-office, jd__ : cool, thank you!15:50
*** _nadya_ has joined #openstack-ceilometer15:51
*** nellysmitt has joined #openstack-ceilometer15:54
*** ifarkas has quit IRC16:04
*** ildikov has joined #openstack-ceilometer16:05
*** k4n0 has quit IRC16:05
*** thomasem has quit IRC16:11
openstackgerritMehdi Abaakouk proposed a change to stackforge/gnocchi: Really cleanup ipc lock created by tooz
*** aviau has quit IRC16:17
*** zul has quit IRC16:17
*** harlowja_away has quit IRC16:17
*** openstackgerrit has quit IRC16:18
*** openstackgerrit has joined #openstack-ceilometer16:18
*** zul has joined #openstack-ceilometer16:18
*** aviau has joined #openstack-ceilometer16:19
*** admin0 has quit IRC16:19
*** hhuang has joined #openstack-ceilometer16:21
*** amalagon has quit IRC16:21
*** thomasem has joined #openstack-ceilometer16:23
*** thomasem has quit IRC16:30
*** pnascimento has quit IRC16:36
*** hhuang has quit IRC16:37
*** nikhil_k is now known as nikhil_k|afk16:38
*** _nadya_ has quit IRC16:39
*** Longgeek_ has quit IRC16:47
*** thomasem has joined #openstack-ceilometer16:49
*** safchain has quit IRC16:50
*** thomasem_ has joined #openstack-ceilometer16:57
*** thomasem has quit IRC16:57
*** thomasem_ has quit IRC16:58
*** sbfox has joined #openstack-ceilometer17:00
*** thomasem has joined #openstack-ceilometer17:02
*** amalagon has joined #openstack-ceilometer17:04
*** harlowja has joined #openstack-ceilometer17:15
*** sbfox has quit IRC17:16
*** sbfox1 has joined #openstack-ceilometer17:16
*** sbfox1 has quit IRC17:17
*** sbfox has joined #openstack-ceilometer17:17
*** isviridov_meetin is now known as isviridov_away17:17
*** sbfox1 has joined #openstack-ceilometer17:20
*** sbfox has quit IRC17:21
*** lsmola has quit IRC17:31
*** _nadya_ has joined #openstack-ceilometer17:42
*** sbfox1 has quit IRC17:44
*** nellysmitt has quit IRC18:01
*** sbfox has joined #openstack-ceilometer18:14
*** zhiyan has quit IRC18:16
*** zhiyan has joined #openstack-ceilometer18:17
*** nellysmitt has joined #openstack-ceilometer18:18
*** sbfox has quit IRC19:03
*** nsaje has quit IRC19:12
*** admin0 has joined #openstack-ceilometer19:18
*** admin0 has quit IRC19:18
*** alexpilotti has quit IRC19:21
*** _nadya_ has quit IRC19:26
*** thomasem has quit IRC19:53
*** thomasem has joined #openstack-ceilometer19:55
*** nellysmitt has quit IRC19:56
*** nellysmitt has joined #openstack-ceilometer19:56
*** thomasem has quit IRC19:59
*** nellysmitt has quit IRC20:01
*** X019 has quit IRC20:06
*** X019 has joined #openstack-ceilometer20:07
*** sbfox has joined #openstack-ceilometer20:34
*** _robot has quit IRC20:46
*** _robot has joined #openstack-ceilometer20:46
*** mikedillion has quit IRC20:51
*** rbowen is now known as rcb_afk20:51
* cdent calls it20:56
*** cdent has quit IRC20:56
*** jasondotstar has quit IRC21:02
openstackgerritPradeep Kilambi proposed a change to openstack/ceilometer: Fix neutron client to catch 404 exceptions
openstackgerritgordon chung proposed a change to openstack/ceilometer: isolate event storage models
*** sbfox has quit IRC21:39
*** sbfox has joined #openstack-ceilometer21:42
*** sbfox1 has joined #openstack-ceilometer21:45
*** sbfox has quit IRC21:47
openstackgerritgordon chung proposed a change to openstack/ceilometer: isolate event storage models
*** rbak has quit IRC21:52
*** mikedillion has joined #openstack-ceilometer22:10
*** flwang1 has quit IRC22:13
*** gordc has quit IRC22:17
*** _nadya_ has joined #openstack-ceilometer22:27
*** _nadya_ has quit IRC22:31
*** sbfox1 has quit IRC22:33
*** asalkeld has joined #openstack-ceilometer22:35
*** renatoarmani has joined #openstack-ceilometer22:52
*** yassine has quit IRC22:57
*** vkmc has quit IRC23:02
*** mikedillion has quit IRC23:09
*** renatoarmani has quit IRC23:25
*** mikedillion has joined #openstack-ceilometer23:55
*** sbfox has joined #openstack-ceilometer23:55
*** sbfox has quit IRC23:59

Generated by 2.14.0 by Marius Gedminas - find it at!