Thursday, 2014-05-29

*** rbowen has joined #openstack-ceilometer00:00
*** raymondr has quit IRC00:01
*** julim has quit IRC00:09
*** fnaval has joined #openstack-ceilometer00:12
*** isviridov has quit IRC00:20
*** matsuhashi has joined #openstack-ceilometer00:22
*** isviridov has joined #openstack-ceilometer00:22
*** jdonalds has quit IRC00:29
*** r0j4z0 has joined #openstack-ceilometer00:36
*** rbowen has quit IRC00:48
*** _cjones_ has quit IRC00:48
*** _cjones_ has joined #openstack-ceilometer00:48
*** nati_ueno has quit IRC00:50
*** _cjones_ has quit IRC00:53
*** liusheng has quit IRC00:58
*** raymondr has joined #openstack-ceilometer01:03
*** fnaval has quit IRC01:05
*** fnaval has joined #openstack-ceilometer01:05
*** raymondr has quit IRC01:08
*** shakamunyi has quit IRC01:16
*** nati_ueno has joined #openstack-ceilometer01:21
*** liusheng has joined #openstack-ceilometer01:28
liushengmorning all :)01:28
*** nati_ueno has quit IRC01:32
*** nati_ueno has joined #openstack-ceilometer01:42
*** nati_ueno has quit IRC01:42
*** nosnos has joined #openstack-ceilometer01:45
*** Ruetobas has quit IRC02:06
*** Ruetobas has joined #openstack-ceilometer02:12
*** jdonalds has joined #openstack-ceilometer02:14
*** Ruetobas has quit IRC02:16
*** Ruetobas has joined #openstack-ceilometer02:23
*** raymondr_ has joined #openstack-ceilometer02:36
*** jdonalds has quit IRC02:39
*** liusheng has quit IRC02:48
*** matsuhashi has quit IRC03:00
*** matsuhashi has joined #openstack-ceilometer03:04
openstackgerritZhiQiang Fan proposed a change to openstack/ceilometer: Ignore the generated file ceilometer.conf.sample
*** nosnos has quit IRC03:33
*** matsuhashi has quit IRC03:34
*** matsuhashi has joined #openstack-ceilometer03:35
*** matsuhashi has quit IRC03:39
*** shakamunyi has joined #openstack-ceilometer03:49
*** matsuhashi has joined #openstack-ceilometer03:56
*** liusheng has joined #openstack-ceilometer03:57
*** shakamunyi has quit IRC04:05
*** renlt has joined #openstack-ceilometer04:18
*** psharma has joined #openstack-ceilometer04:18
*** raymondr_ has quit IRC04:27
*** jdonalds has joined #openstack-ceilometer04:31
openstackgerritZhiQiang Fan proposed a change to openstack/ceilometer: Enable filter alarms by their type
*** raymondr has joined #openstack-ceilometer04:39
*** jdonalds has quit IRC04:40
*** raymondr has quit IRC04:42
*** nosnos has joined #openstack-ceilometer04:45
*** jdonalds has joined #openstack-ceilometer04:53
*** xiej has joined #openstack-ceilometer04:56
*** xiej has quit IRC04:58
*** jdonalds has quit IRC04:59
*** jdonalds has joined #openstack-ceilometer05:02
*** shakamunyi has joined #openstack-ceilometer05:03
openstackgerritA change was merged to openstack/ceilometer: Update the copyright date in doc
*** jdonalds has quit IRC05:06
*** shakamunyi has quit IRC05:14
*** fnaval_ has joined #openstack-ceilometer05:34
*** fnaval has quit IRC05:36
*** shakamunyi has joined #openstack-ceilometer05:40
*** raymondr has joined #openstack-ceilometer05:43
openstackgerritA change was merged to openstack/python-ceilometerclient: fixed several pep8 issues
*** raymondr has quit IRC05:47
*** shakamunyi has quit IRC05:50
openstackgerritOpenStack Proposal Bot proposed a change to openstack/ceilometer: Imported Translations from Transifex
*** fnaval_ has quit IRC06:07
*** ildikov has quit IRC06:13
*** _nadya_ has joined #openstack-ceilometer06:20
*** fnaval has joined #openstack-ceilometer06:23
*** _nadya_ has quit IRC06:25
*** inc0 has joined #openstack-ceilometer06:27
*** ildikov has joined #openstack-ceilometer06:40
*** fnaval has quit IRC06:42
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Fix doc gate job false success
*** shakamunyi has joined #openstack-ceilometer06:47
*** shakamunyi has quit IRC06:59
*** fnaval has joined #openstack-ceilometer07:00
*** _nadya_ has joined #openstack-ceilometer07:01
*** jdonalds has joined #openstack-ceilometer07:03
*** fnaval has quit IRC07:05
*** _nadya_ has quit IRC07:18
*** shakamunyi has joined #openstack-ceilometer07:26
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Fix doc gate job false success
*** shakamunyi has quit IRC07:36
*** r0j4z0 has quit IRC07:37
*** zqfan has quit IRC07:41
*** raymondr has joined #openstack-ceilometer07:44
*** raymondr has quit IRC07:48
openstackgerritliusheng proposed a change to openstack/ceilometer: Improve the timestamp validation of ceilometer API
*** zhikunliu has joined #openstack-ceilometer07:55
*** matsuhashi has quit IRC07:58
*** matsuhashi has joined #openstack-ceilometer07:59
*** matsuhas_ has joined #openstack-ceilometer08:02
*** zqfan has joined #openstack-ceilometer08:03
*** shakamunyi has joined #openstack-ceilometer08:03
*** matsuhashi has quit IRC08:04
*** cdent has joined #openstack-ceilometer08:12
*** shakamunyi has quit IRC08:14
*** cdent has quit IRC08:16
*** eglynn has joined #openstack-ceilometer08:23
*** cdent has joined #openstack-ceilometer08:23
*** r0j4z0 has joined #openstack-ceilometer08:40
*** shakamunyi has joined #openstack-ceilometer08:41
*** raymondr has joined #openstack-ceilometer08:43
*** raymondr has quit IRC08:47
*** shakamunyi has quit IRC08:52
*** matsuhas_ has quit IRC08:53
*** matsuhashi has joined #openstack-ceilometer08:58
openstackgerritA change was merged to openstack/ceilometer: Ignore the generated file ceilometer.conf.sample
*** inc0_ has joined #openstack-ceilometer09:00
*** inc0 has quit IRC09:02
*** zqfan has quit IRC09:07
*** zqfan has joined #openstack-ceilometer09:08
*** renlt has quit IRC09:12
*** fnaval has joined #openstack-ceilometer09:18
*** shakamunyi has joined #openstack-ceilometer09:19
*** pnascimento has joined #openstack-ceilometer09:21
DinaBelovajd__, o/ Should be the swift container created automatically by gnocchi - or this shuold be done manually?09:24
*** zhikunliu has quit IRC09:30
*** shakamunyi has quit IRC09:31
*** matsuhashi has quit IRC09:33
*** matsuhashi has joined #openstack-ceilometer09:33
*** matsuhashi has quit IRC09:35
*** matsuhashi has joined #openstack-ceilometer09:36
DinaBelovajd__ - I simply have strange situation trying to create entity in the gnocchi and I have 404 not found for the Swift container :D the funny fact is I see the id of the container in the request to swift, that makes me really suspicious)09:36
*** fnaval has quit IRC09:38
openstackgerritChristian Berendt proposed a change to openstack/ceilometer: use logger instead of LOG as name for the Logger object attribute
DinaBelovaamalagon - did you try this thing?09:43
*** shakamunyi has joined #openstack-ceilometer09:58
*** idegtiarov has joined #openstack-ceilometer10:00
*** nosnos has quit IRC10:02
*** nosnos has joined #openstack-ceilometer10:05
*** shakamunyi has quit IRC10:05
*** isviridov has quit IRC10:10
*** isviridov has joined #openstack-ceilometer10:18
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer-specs: Hbase-events-feature specs
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer-specs: Hbase-events-feature specs
* cdent gets lunch and runs errands10:40
*** matsuhashi has quit IRC10:41
*** nosnos has quit IRC10:41
jd__DinaBelova: do you have more information on your issue?10:41
jd__DinaBelova: the container is created by in __init__ IIRC10:42
DinaBelovajd__ -
DinaBelovajd__ - I used this script to ping the API
jd__DinaBelova: hum I think it's a Swift issue, like it looks like your admin account does not exist10:45
jd__the user exists in tempauth, but the swift account is not created10:45
jd__something like that10:45
*** nosnos has joined #openstack-ceilometer10:45
jd__checks that it works with the 'swift' client CLI first :)10:45
DinaBelovaokay, will try to look on this.10:45
DinaBelovayes, it works)10:45
jd__you can create a container with the same parameter you provided to Gnocchi?10:46
DinaBelovawell, I tried listing10:46
DinaBelovaas I'm not quite experienced in swift :d10:46
jd__check if it does the same requests, and also use TEMPAUTH_10:46
jd__ok :)10:46
*** nosnos has quit IRC10:46
*** nosnos has joined #openstack-ceilometer10:47
*** nosnos has quit IRC10:48
*** cdent_ has joined #openstack-ceilometer10:50
*** cdent has quit IRC10:50
*** cdent_ is now known as cdent10:50
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer-specs: Hbase-events-feature specs
DinaBelovajd__ - I've restarted devstack, this error is gone)11:02
DinaBelovaI have new one :D11:02
DinaBelovajd__ - btw - it there a db-sync for gnocchi or not yet?11:03
openstackgerritliusheng proposed a change to openstack/ceilometer: Improve the timestamp validation of ceilometer API
jd__no db sync yet11:11
jd__what's your error DinaBelova ?11:11
*** IvanBerezovskiy has joined #openstack-ceilometer11:11
DinaBelovano tables in my sql db))11:12
DinaBelovathat's why I asked about the db-sync)11:12
jd__haha yeah that's something missing I guess11:12
jd__there's a upgrade() method in the driver11:12
jd__but I didn't write an entry point for that I guess11:12
jd__he you got your first patch11:12
DinaBelovajd__, yes, sure :D11:14
DinaBelovawill do :D11:14
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer-specs: Hbase-events-feature specs
*** pnascimento has quit IRC11:14
*** prad_ has joined #openstack-ceilometer11:17
*** liusheng has quit IRC11:17
*** prad__ has joined #openstack-ceilometer11:22
*** prad_ has quit IRC11:22
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer-specs: Hbase-events-feature specs
amalagon /away11:29
amalagonsorry, ignore the above11:30
*** prad__ has quit IRC11:38
openstackgerritMitsuru Kanabuchi proposed a change to openstack/ceilometer: Add the function of deleting alarm history
*** zqfan_ has joined #openstack-ceilometer11:46
DinaBelovajd__ -
Alexei_987stupid question: def make_timestamp_range(start, end, start_timestamp_op=None, end_timestamp_op=None) Why do we need timestamp_op ??11:50
jd__DinaBelova: cool I'll check that, don't worry your patch will fail, I still have CI issues to fix11:50
DinaBelovaokay, got it11:50
DinaBelovaI've tried it on my testing env - after its running I have sqlite db and entity recorded11:51
eglynnAlexei_987: ... possibly to distinguish between <= and <?11:54
Alexei_987eglynn: it doesn't make sense11:54
Alexei_987eglynn: just like comparing float == float11:54
DinaBelovaeglynn, o/ I guess I have bad-bad English)) I'm about is/are question in the :)11:54
Alexei_987date in milliseconds would never be equal11:55
Alexei_987so it just adds additional code complexity without any reason11:55
eglynnAlexei_987: ... never be equal, or unlikely to be equal?11:56
eglynnAlexei_987: ... do we actually have millisecond granularity on all timestamps?11:56
openstackgerritMitsuru Kanabuchi proposed a change to openstack/ceilometer: Split bunch of samples into multiple messages
Alexei_987eglynn: in most backends we do11:57
Alexei_987eglynn: basically when you query an API there is no way to know that you have sample with date exactly equal11:57
Alexei_987so IMO no need to support <= operator11:57
Alexei_987eglynn: < would be enough for everybody11:58
*** dmsimard has left #openstack-ceilometer11:58
*** prad_ has joined #openstack-ceilometer12:00
amalagonhey eglynn thanks for the comments :)12:02
eglynnAlexei_987: ... do we support such a distinction still in the API?12:02
eglynnAlexei_987: ... so even if it's unlikely to make a difference in practice whether the user chooses < or <=12:02
eglynnAlexei_987: ... seems wrong to ignore their preference, or?12:03
Alexei_987eglynn: actually I care more about driver level12:03
Alexei_987eglynn: if user wants <= we can simply subtract 1 second/millisecond from his query and use <12:03
eglynnAlexei_987: ... the point is that we *allow* the API caller currently to use <=12:04
Alexei_987eglynn: yes :)12:05
eglynnAlexei_987: ... how to force then to use < and subtract 1 milli as you suggest?12:05
Alexei_987eglynn: they can use <= in the API but in driver we can use < with -1 second12:06
eglynnAlexei_987: ... and is the code simplification sufficient to justify leaking that into the API?12:06
Alexei_987eglynn: ok nevermind that :)12:06
Alexei_987eglynn: I just don't like this part of the code12:06
Alexei_987eglynn: but i'm ok with it being around12:07
eglynnAlexei_987: I hear ya ... I just don't think it's important enough to worry about12:07
*** prad_ has quit IRC12:14
*** prad_ has joined #openstack-ceilometer12:15
r0j4z0i think i found a bug on ceilometer12:16
r0j4z0when requesting trait information12:16
r0j4z0of an event12:16
eglynnr0j4z0: can you file a report in launchpad for that bug?12:20
r0j4z0im trying to do it now im firstimer...12:21
DinaBelovar0j4z0, yep - please just create account first )))12:21
DinaBelovathen click 'Report a bug' in the right column on the
DinaBelovaSummary is short line like 'my-bug-about-this'12:22
r0j4z0ok :D i try12:22
DinaBelovaand then it'll be a field for the extra information - you may place there traces, symptoms, etc.12:22
r0j4z0im not sure if is a bug12:24
r0j4z0can you try it ? is on icehouse12:24
r0j4z0when i try to get a integer trait12:24
r0j4z0i get this error: WARNING (http:173) Request returned failure status.12:24
r0j4z0HTTPBadRequest (HTTP 400) ERROR Invalid input for field/attribute value. Value: '20'. Wrong type. Expected '<type 'unicode'>', got '<type 'long'>'12:24
r0j4z0but if i get a string trait it works fine12:25
openstackgerritIgor Degtiarov proposed a change to openstack/ceilometer-specs: Hbase-events-feature specs
*** fnaval has joined #openstack-ceilometer12:30
*** cdent has quit IRC12:39
*** eglynn has quit IRC12:39
*** zqfan_ has quit IRC12:40
amalagono/ DinaBelova12:40
amalagonthe Mismatch error strikes again in Jenkins12:41
*** gordc has joined #openstack-ceilometer12:46
*** raymondr has joined #openstack-ceilometer12:51
*** cdent has joined #openstack-ceilometer12:51
*** raymondr has quit IRC12:52
*** _nadya_ has joined #openstack-ceilometer12:58
DinaBelovaamalagon, it looks like issues Julien has mentioned earlier)12:59
*** fnaval_ has joined #openstack-ceilometer13:01
*** rbowen has joined #openstack-ceilometer13:03
*** fnaval has quit IRC13:03
*** eglynn has joined #openstack-ceilometer13:06
*** thomasem has joined #openstack-ceilometer13:06
*** cmart has joined #openstack-ceilometer13:08
*** jaypipes has quit IRC13:10
*** jdob has joined #openstack-ceilometer13:11
*** ddieterly has joined #openstack-ceilometer13:13
*** eglynn is now known as eglynn|afk13:14
*** eglynn|afk is now known as eglynn_afk13:14
*** eglynn_afk is now known as eglynn-away13:14
*** prad_ has quit IRC13:15
*** jaypipes has joined #openstack-ceilometer13:16
*** shakamunyi has joined #openstack-ceilometer13:16
*** shakayumi has joined #openstack-ceilometer13:19
*** julim has joined #openstack-ceilometer13:21
*** shakamunyi has quit IRC13:21
cdentcmart you around?13:24
cmartcdent: hi! yes, I'm here13:25
cdentcan you remind me of that query that was giving you grief?13:25
cdentyeah, I remember that, but what was the string you were passing that led you there?13:26
cmart"state=string::insufficient data"13:26
cmartthe error was that the pattern doesn´t allow whitespaces..13:26
cmartI got confused by the ^ right at the beginning of the last group13:27
* cdent nods13:27
cmartI solved it by removing the ¬v and the whitespace right next to the ^ symbol13:28
cmart(I also removed the restrictions related to commas and "-", because we could send strings with that kind of characters13:28
cmartI guess that once I upload my patch, we can discuss it appropriately :)13:29
*** fnaval_ has quit IRC13:30
*** eglynn-away is now known as eglynn13:32
cdentsplit_by_op seems over specified to me13:32
cdentit ought to be doable as just one much simpler line13:32
* cdent tries something13:32
*** shakayumi has quit IRC13:34
openstackgerritOpenStack Proposal Bot proposed a change to openstack/python-ceilometerclient: Updated from global requirements
*** shakayumi has joined #openstack-ceilometer13:48
*** prad_ has joined #openstack-ceilometer13:49
*** prad_ has quit IRC13:53
*** fnaval has joined #openstack-ceilometer13:55
*** raymondr has joined #openstack-ceilometer13:58
*** prad_ has joined #openstack-ceilometer14:00
*** prad__ has joined #openstack-ceilometer14:03
*** prad_ has quit IRC14:03
*** prad__ is now known as prad_14:04
*** ildikov_ has joined #openstack-ceilometer14:05
cdentcmart I think I have a patch which covers your needs as well as addressing my concerns with the regex14:06
*** ildikov has quit IRC14:06
* cdent runs many tests14:07
*** inc0_ has quit IRC14:09
*** psharma has quit IRC14:14
*** yfujioka has joined #openstack-ceilometer14:19
*** ityaptin has joined #openstack-ceilometer14:19
idegtiaroveglynn: Eoghan it was some strange happened with my answer for your comment, I have added it separately below in comments.14:20
DinaBelovaidegtiarov - just appeared :-\14:20
DinaBelovafunny Gerrit)14:21
DinaBelovaeglynn, anyway - I'm writing two specs to the qa-specs now14:21
DinaBelovafor the ceilometer14:21
DinaBelovabased on two existing BPs14:21
DinaBelovaevents feature will be the part of scenario testing14:21
DinaBelovalike alarms14:21
idegtiarovthanks Dina, while I try to write that it will be a part of spec you are going to prepare! You have already write about it! :))14:23
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Fix doc gate job false success
eglynnDinaBelova: gremlins in the gerrit machine stealing comments? ;)14:24
DinaBelovaidegtiarov, ;)14:24
DinaBelovaeglynn, hehe14:24
openstackgerritIldiko Vancsa proposed a change to openstack/ceilometer: Fix doc gate job false success
DinaBelovamy little secret)14:24
*** jdonalds has quit IRC14:26
*** ildikov_ has quit IRC14:26
*** ildikov has joined #openstack-ceilometer14:26
cdentanyone: is there a way to indicate in a commit that the current patch obsoletes a pending but not merged patch?14:28
cdenti've  got a changed to ceilometerclient that makes this one redundant:
cdentcmart did you make a bug for your issue?14:30
cmartI have a bug related to that14:30
cmartbut I could create a bug for that, sure14:31
nsajecdent: I'd simply -1 the redundant one and post an explanation in the comment14:31
cdentcmart unless it is necessary to have a bug for the record, I can just submit the patch for review14:32
cdentnsaje: thanks14:32
eglynncdent: normally you could -1 the patch and reference the new patch in the comment14:32
eglynncdent: ... however if the pre-existing patch already has a core +214:32
cdentyeah, in this case it does...14:33
eglynncdent: ... why not just rebase your patch on top of that?14:33
eglynncdent: ... assuming your patch does something similar AND more14:33
cdentit changes all the same code, but in a rather different way14:33
eglynncdent: ... a-ha so splitting out the "AND more" bit into a separate patch rebased on the original patch isn't feasible?14:34
cdentI _think_ rebasing it against the existing will obscure the point14:34
eglynncdent: ... is the original wrong, or just incomplete?14:34
cdentand the +2 patch removes many of the "security" protections in the original code14:35
cdentso if I were king I'd -1 the existing thing...14:35
nsajecdent: you don't need to be a king to -1 a patch :-D14:35
cdentnot even for one day?14:36
eglynncdent: ... k, in that case -1 the original patch and just clearly state it's deficiencies14:36
eglynncdent: ... and add a link to your more complete / more correct patch14:36
nsajecdent: you can be king for the next 20ms!14:37
*** nati_ueno has joined #openstack-ceilometer14:37
cdenthmmm, actually I think I will rebase to the existing, because parts of it are better14:38
eglynncdent: ... OK definitely better to rebase if that's the case14:38
*** yfujioka has quit IRC14:39
cdentwhat's the procedure there, my rebase foo is weak14:39
eglynncdent: ... "nicer" from an etiquette PoV as well :)14:39
eglynncdent: procedure from a git PoV?14:40
cdenti know how to rebase in general14:40
cdentbut rebasing to the thing sitting on gerrit14:40
eglynncdent: well create a fresh branch first14:40
eglynncdent: then ... $ git fetch refs/changes/74/92074/1 && git format-patch -1 --stdout FETCH_HEAD > original14:40
eglynncdent: $ git am original14:41
eglynncdent: is your patch already up on gerrit?14:41
cdentno not yet14:41
eglynncdent: ... if not: $ git format-patch -1 -o my_patch.patch # <=== on the old branch14:42
eglynncdent: ... switch back to new branch14:42
eglynncdent: $ git am my_patch.patch --reject14:42
eglynncdent: $ find . -name "*.rej" # <=== manually fix up these rejected hunks14:42
eglynncdent: $ git add $(git status | grep modified | awk -F: '{print $2}')14:44
eglynncdent: $ git am --resolved14:44
eglynncdent: prolly a zillion other ways of doing it in git14:44
nsajethat is some serious work, shouldn't that be on a wiki somewhere?14:44
cdentthat's what I was thinking14:44
cdentI'm pretty sure it ought to be possible to use the rebase command itself to do much of that work (after fetching from the remote)14:45
eglynncdent: ... as I said, many ways of skinning a cat in git14:46
cdentso easy to love and hate git14:46
eglynncdent: ... the above is not meant to be prescriptive14:47
* cdent nods14:47
eglynncdent: ... just a quick'n'dirty off-the-top-of-my-head way of getting you to where you want to be with that patch14:47
*** fabiog has joined #openstack-ceilometer14:51
*** r0j4z0 has quit IRC14:52
*** fabiog has quit IRC14:53
*** idegtiarov has quit IRC14:53
openstackgerritPradeep Kilambi proposed a change to openstack/ceilometer-specs: Spec for metering LBaaS
*** jmckind has joined #openstack-ceilometer15:11
openstackgerritgordon chung proposed a change to openstack/ceilometer: improve expirer performance for sql backend
*** gordc-mobile has joined #openstack-ceilometer15:22
*** jergerber has joined #openstack-ceilometer15:23
*** nati_ueno has quit IRC15:27
*** rbowen has quit IRC15:31
*** rbowen has joined #openstack-ceilometer15:32
*** cmart has quit IRC15:34
*** promulo has quit IRC15:43
*** shakayumi is now known as shakamunyi15:59
*** Ruetobas has quit IRC16:01
*** llu-laptop has joined #openstack-ceilometer16:02
*** fabiog has joined #openstack-ceilometer16:02
eglynnaviau: hey16:02
eglynnaviau: so you mentioned
*** Ruetobas has joined #openstack-ceilometer16:03
eglynnaviau: ... you think that specifically *excludes* monitoring from the telemetry program?16:03
eglynnaviau: ... that wasn't really the intention16:03
*** gordc-mobile has quit IRC16:03
aviauNo, but I would like to see it included.16:03
ildikovaviau: it is still an open question that should it be included or it should be a separate service/project16:04
ildikovaviau: it was also mentioned in ATL that maybe it could be started under the Ceilometer umbrella and then we will see16:05
* eglynn doesn't want to get rat-holed in a long discussion about the mission statement ...16:05
llu-laptopildikov: but it looks like hp-mon and ceilometer have a lot in common16:05
eglynn... but it was intended not to close off future possibilities16:05
ildikoveglynn: sorry, agreed16:05
eglynnllu-laptop: yeap hp-mon looks like it's got some potential all right16:06
ildikovllu-laptop: I would separate billing related stuff and monitoring from each other16:06
* eglynn is especially interested in hearing about how hp-mon will be proven in production in the HP cloud over the next 6 months or so16:07
fabiogildikov: I agree, even if the data can be somehow shared. So have a common storage strategy can be beneficial for the two projects16:07
nealphI think the hp-mon folks are of the opinion that the two should co-exist (what ildikov ^^^ noted).16:07
llu-laptopildikov: agreed, billing and monitoring should be separated, but architecturally, many common things16:08
*** Ruetobas has quit IRC16:08
eglynnaviau: ... so it seems like someone from outside HP getting involved with hp-mon would be useful16:08
eglynnaviau: ... in the sense of community building etc.16:08
nealpheglynn:they're open to that, absolutely. I think that's the idea of moving to stackforge.16:09
eglynnnealph: yeap, exactly ... co-exist and maybe over time begin to share some elements16:09
ildikovfor me it is a bit difficult to imagine how the separation will work, if we do both in the exact same project, but we will se what we can share and how it could work16:09
eglynnildikov: for example they use kafka as a highly performant/scalable message bus, IIUC16:10
nealphildikov:well, we'll probably have to make it work internally...we're the guinea pigs? :)16:10
aviauOne thing is certain. I think there has to be a separated service. Do you agree on this?16:10
eglynnildikov: ... it would rock if kafka could be layered under the oslo.messaging layer16:10
fabiogaviau +116:11
ildikoveglynn: that's ok, but it is far from doing monitoring inside Ceilometer16:11
eglynnaviau: "separated service" ... what exactly do you mean by that?16:11
ildikovnealph: hmm, I like guineapigs, but I do not like to be one usually ;)16:12
*** nati_ueno has joined #openstack-ceilometer16:12
*** fnaval has quit IRC16:12
aviauI think MONaaS needs a separate API and endpoint.16:12
eglynnildikov: "if we do both in the exact same project" ==> there's the ceilometer project and then there's the telemetry program16:12
eglynnaviau: so multiple APIs can be exposed by a single service, and multiple services can colloborate to provide a single API16:13
ildikoveglynn: a-ha, ok, then let's say that under the Telemetry umbrella and we use only those things in common, like kafka, that really makes sense16:13
*** Ruetobas has joined #openstack-ceilometer16:14
fabiogeglynn: are you advocating to have a telemetry program, a telemetry-common project and a ceilometer and monaas projects?16:14
eglynnildikov: ... well it could be kafka, it could be we also leverage an influxDB integration that they do as part of the startegy to have a "pure" opensource version16:14
eglynnildikov: (... i.e a version minus the vertica DB)16:15
eglynnfabiog: well, we already have a telemetry program16:15
ildikoveglynn: sure, that looks doable, but I wouldn't say it is necessary from the beginning16:15
*** Titilambert has joined #openstack-ceilometer16:16
eglynnfabiog: ... I'm just saying there are possibilities there for two projects to collaborate16:16
eglynnTitilambert: hello16:16
Titilambertare you in meeting ?16:16
ildikoveglynn: I haven't checked the DB part of the HP solution yet, so it's an open question for me16:16
eglynnTitilambert: are we in #openstack-meeting do you mean?16:17
Titilambertis the ceilometer weekly meeting ?16:17
eglynnTitilambert: nope that meeting ending ~16 mins ago16:17
aviauWe continue the meeting here.16:17
eglynnTitilambert: that was on 1500-1600UTC16:17
_nadya_ildikov: db part is interesting. I see only vertica on picture16:18
Titilamberteglynn: Do you think integrate HP solution as monitoring as a service ?16:18
*** vrovachev has quit IRC16:19
eglynnTitilambert: ... we're just on the learning phase, figuring out what it is, and riffing on possibilities16:19
ildikoveglynn: so monitoring and billing are different things, but if one DB solution can efficiently handle both, then I'm not against, we will see16:19
aviauildikov, +116:19
cdentthe time dimension seems much different to me16:19
gordcildikov: i don't think we're looking at one db solution. we're already planning to split meter/alarm/event data into their own db.16:19
*** IvanBerezovskiy has left #openstack-ceilometer16:20
eglynnTitilambert: ... as we agreed at summit, our focus for the Juno cycle is all about solidifying the ceilometer core16:20
eglynnTitilambert: ... so this is more future-oriented discussion16:20
Titilamberteglynn: ok16:20
ildikovgordc: yeap sure, then let's call it shared DB solution for any of these areas, if poss16:20
eglynnTitilambert: ... with an eye to the K summit and beyond16:20
Titilamberteglynn: I see16:21
eglynn_nadya_: AFAIK it requires vertica right now16:21
ildikovgordc: I just reflected to what eglynn wrote, I'm a great fan of separation of responsibilities, like what you mentioned about the "own dbs"16:21
eglynn_nadya_: ... and there's a community edition of vertica with a limitation on total data size16:21
eglynn_nadya_: ... but it may well make sense for them to also have a "pure opensource" metric DB as an option16:22
aviauI dont think you guys are taking the right approach to implementing MONaaS. I think we should start from scratch and evaluate the integration of HP components as we go.16:22
gordcildikov: yeah, i think separating data to different targets makes more sense... (you can always target the data to point to the same db if you want).16:23
ildikov_nadya_: maybe that will be improved also, I did not have time to deeply check the doc and collect the info all together that I saw or heard nowadays :(16:23
gordcaviau: who are 'you guys'? the HP design? Stacktach?16:23
eglynnaviau: before starting from scratch, IMO it makes sense to evaluate the prior art16:23
ildikovgordc: yeap, exactly, I think we're on the same page16:24
gordcildikov: :) cool cool.16:24
openstackgerritChris Dent proposed a change to openstack/python-ceilometerclient: Refactor split_by_op for greater encapsulation
ildikovgordc: I'm just trying to not to be a really bad cop here only a simple bad one ;)16:25
*** nati_ueno has quit IRC16:26
*** isviridov is now known as isviridov|away16:26
TitilambertI guess nobody wants to reinvent Nagios, right ?16:26
gordcildikov: i'll take the really bad cop role.16:26
eglynnaviau: ... as stated a few times already, our focus for the Juno cycle is all about solidifying the ceilometer core16:27
ildikovgordc: cool, tnx :)16:28
aviaueglynn, Right now the blueprint targets 'next' and not Juno.16:28
openstackgerritChris Dent proposed a change to openstack/python-ceilometerclient: Refactor split_by_op for greater encapsulation
openstackgerritChris Dent proposed a change to openstack/python-ceilometerclient: Enable special characters in query keys and values
ildikovaviau: do you mean that we should start to implement MONaaS from scratch independently from the HP one and then we will see?16:29
ildikovaviau: ... or I misunderstood what you wrote above about the not right approach?16:30
*** llu-laptop has quit IRC16:31
eglynnaviau: ... so my recommendation would be to get familiar with hp-mon and figure out where the gaps are (as opposed to assuming a start from a complete green-field)16:32
aviauildikov, I would start "from kind of scratch" and use hp's solution during development.16:34
_nadya_aviau: what's your plan? are you interested in integration in ceilometer or you consider variants? whats the priority?16:34
Titilamberteglynn: you're right, is just the hp solution is not a small project16:35
aviau_nadya_,  To integrate with Ceilometer is not my priority. It would be to cover the use cases defined in the proposition.16:36
aviauWether it is integrated or not.16:36
aviauIn the future, integration will become important.16:36
eglynnTitilambert: sure, it's not small ... question is whether starting independently from scratch makes sense16:37
ildikovaviau: you can start it as a completely separate thing, if you would like to or join to the HP one if that is poss16:37
*** promulo has joined #openstack-ceilometer16:38
Titilamberteglynn: you're right! I think if we start from scratch, we have to know and USE HP solution's ideas and maybe some part of the code16:39
ildikovaviau: ... and then we can see how this under the Telemetry umbrella thing can work out, if you have something that's working16:39
eglynnaviau: ... if you want to benefit from the "multiplier effect" of open source development, it makes sense I think to consider the existing prior art16:40
nealphaviau:I'll echo again...I think the hp-mon folks would welcome contributions from all corners.16:40
ildikovnealph: +116:41
eglynnnealph: ... that's also the impression I got from conversations at summit16:41
aviaueglynn, This is correct. I will work on collaborating with what has already been done.16:44
Titilamberteglynn: oki, so first step: enter into HPcloud-mon, second step: check use cases with HPcloud-mon, then new meeting ?16:44
*** _nadya_ has quit IRC16:44
eglynnaviau, Titilambert: cool, sounds like a plan16:44
eglynnaviau, Titilambert: discuss again when, maybe in a week or two?16:45
ildikovTitilambert: new meeting about use cases, or?16:45
aviauI would like to have some time to dive in HP's solutions. Maybe two weeks?16:46
Titilambertildikov: about our research about HPcloudmon16:46
Titilambertildikov: about our researches about HPcloudmon16:46
ildikovTitilambert: hmm, ok, it can't hurt :)16:47
Titilamberteglynn: about the meeting, I don't know, I don't have time during the next weeks16:47
Titilambertildikov: why ?16:48
Titilambertildikov: oups, i read "it can hurt" :D16:48
eglynnaviau, Titilambert: cool, let's aim to talk again in circa two weeks (... modulo Titilambert's time constraints)16:48
ildikovTitilambert: :)16:48
eglynnright-o gentlepersons, getting late for me ...16:49
eglynn... I'm gonna have to drop off IRC very soon16:49
eglynn... have a good evening all, and thanks for the discussion16:49
Titilamberteglynn: maybe we have to get HP people here16:49
aviauHave a good day :)16:50
eglynnTitilambert: ... yeap I'll reach out to them about that16:50
aviauAs far as I know, roland-hochmuth was supposed to be present.16:50
Titilamberteglynn: cool16:50
ildikovall right guys, then see you soon around :)16:51
fabiogguys, can anyone review please?16:51
fabiogis to update the client to KS v3 API16:52
*** zqfan_ has joined #openstack-ceilometer16:53
*** dhandy has joined #openstack-ceilometer16:54
ildikovnealph: still around?16:55
ildikovI will try to dig into the new Gnocchi design soon and participate from the pipeline side in the integration16:58
ildikovwill you have some time later on to think about the dynamic config a bit?16:58
*** dhandy has quit IRC16:59
nealphabsolutely...have just been pushing that agenda "in-house" and will have some more time to devote next week.16:59
ildikovcool, sounds good17:00
*** shardy is now known as shardy_afk17:01
nealphaviau:Titilambert:ping me or fabio...we work closely with those hp-mon guys and I know they're glad to kvetch on any topic.17:01
nealphany topic you might have questions on, that is.17:02
nealphooo, "kvetch" not what I meant to imply. "chat" is the better word.17:02
ildikovnealph: ping then, when you will have some time, I will also ping if I have some info that we can start with17:02
*** eglynn has quit IRC17:03
openstackgerritChris Dent proposed a change to openstack/python-ceilometerclient: Refactor split_by_op for greater encapsulation
*** DrBacchus has joined #openstack-ceilometer17:05
*** _cjones_ has joined #openstack-ceilometer17:06
nealphildikov:perfect...will do.17:08
ildikovnealph: cool, tnx17:08
*** rbowen has quit IRC17:10
ildikovnealph: we should start with some specs anyway, so I will try to collect as many info as I can about Gnocchi, hopefully we can figure the dynamic config design in parallel with that implementation and also we can have some brainstorming about the basics as a starting point17:10
*** fnaval has joined #openstack-ceilometer17:10
ildikovnealph: ok, I will not annoy you more for today, I think it's the best time to say that my office day is ended :)17:10
nealphildikov: "you don't gotta go home, but you can't stay here." :)17:11
ildikovnealph: LOL :)17:13
ildikovnealph: BTW, what's your TZ?17:13
*** underyx is now known as Underyx|off17:14
*** _nadya_ has joined #openstack-ceilometer17:14
*** DrBacchus is now known as rbowen17:17
nealphildikov: I'm in US Mountain. almost lunch. :D17:21
ildikovnealph: ah, ok :D17:21
ildikovnealph: I will keep it in mind :)17:21
ildikovnealph: ... and also bon apetite :)17:22
*** jergerber has quit IRC17:23
ildikovnealph: it's 19:30 here, I should drop the notebook out of the window to be sure that I will not work more today ;)17:23
*** _nadya_ has quit IRC17:25
* cdent retires17:30
*** cdent has quit IRC17:30
*** ildikov has quit IRC17:35
*** nati_ueno has joined #openstack-ceilometer17:37
*** ityaptin has quit IRC17:38
*** fnaval has quit IRC17:42
*** cmart has joined #openstack-ceilometer17:55
*** _nadya_ has joined #openstack-ceilometer17:55
*** openstackgerrit has quit IRC17:57
*** fabiog has quit IRC18:17
*** raymondr has quit IRC18:20
*** _nadya_ has quit IRC18:20
*** _nadya_ has joined #openstack-ceilometer18:20
*** raymondr_ has joined #openstack-ceilometer18:20
*** ildikov has joined #openstack-ceilometer18:35
*** dmsimard has joined #openstack-ceilometer18:53
dmsimardAny clue if it's possible to have the Swift ceilometer middleware talk to the Ceilometer API instead of the Bus/Database backend ?18:55
*** zqfan_ has quit IRC19:01
*** openstackgerrit has joined #openstack-ceilometer19:07
*** fnaval has joined #openstack-ceilometer19:08
*** openstackgerrit has quit IRC19:13
*** nati_ueno has quit IRC19:16
dmsimardIf this yet does not exist, would it make sense to implement something like this ? Otherwise the swift-proxy servers need access to the message bus. The swift-proxy are likely to be public-facing while the bus is likely isolated.19:17
*** nati_ueno has joined #openstack-ceilometer19:28
DinaBelovaildikov, eglynn, gordc - please take a look on the - it's j1, probably we may keep more attention to it) Quite old change and really nice feature)19:28
gordcdmsimard: have you tried using udp publisher instead? i assume it's possible since it leverages the pipeline.19:29
gordcDinaBelova: will take a look after meeting.19:29
DinaBelovagordc, thanks a lot!19:30
dmsimardgordc: That's what I am looking at as I type this, if the file or udp publishers can help me.19:32
gordcdmsimard: cool cool. the udp functionality might be broken but give it a try.19:33
dmsimardgordc: Trying to figure out how to use them right now. I know they're there but can't find how to use them in the docs :(19:34
*** openstackgerrit has joined #openstack-ceilometer19:34
gordcdmsimard: on a call... you'll need to edit pipeline.yaml and change sink from rpc to udp19:35
*** nati_ueno has quit IRC19:36
*** nati_ueno has joined #openstack-ceilometer19:36
openstackgerritChristian Berendt proposed a change to openstack/ceilometer-specs: fixed typos found by RETF rules
*** _nadya_ has quit IRC19:48
*** nati_ueno has quit IRC19:54
*** nati_ueno has joined #openstack-ceilometer19:57
openstackgerritChris Dent proposed a change to openstack/python-ceilometerclient: Refactor split_by_op for greater encapsulation
dmsimardgordc: It doesn't seem like I can get the collector to listen on the UDP port.19:59
*** fnaval has quit IRC20:00
*** nati_ueno has quit IRC20:01
*** nati_ueno has joined #openstack-ceilometer20:03
*** fnaval has joined #openstack-ceilometer20:05
openstackgerritChristian Berendt proposed a change to openstack/ceilometer-specs: Fixed typos found by RETF rules
gordcdmsimard: working against master/20:11
dmsimardI can do either icehouse or havana.20:12
dmsimardThe UDP settings are in both available in the settings20:13
*** nati_ueno has quit IRC20:15
openstackgerritChris Dent proposed a change to openstack/python-ceilometerclient: Refactor split_by_op for greater encapsulation
-openstackstatus- NOTICE: Gerrit is going offline to correct an issue with a recent project rename. ETA 20:45 UTC.20:19
*** ChanServ changes topic to "Gerrit is going offline to correct an issue with a recent project rename. ETA 20:45 UTC."20:19
gordcdmsimard: hmm. i don't recall any issues with udp in icehouse.20:22
gordci should think you should be able to just add udp://host:port as a publisher and it would work... if not, maybe open a bug.20:22
*** jergerber has joined #openstack-ceilometer20:22
*** openstackgerrit has quit IRC20:25
dmsimardgordc: perhaps the collector end of my config is wrong. I do see in the logs that it loads the udp config (ex: DEBUG ceilometer.openstack.common.service [-] collector.udp_port = 4952 )20:28
dmsimardgordc: But there is nothing listening on UDP 495220:28
*** eglynn has joined #openstack-ceilometer20:31
*** ChanServ changes topic to "OpenStack Telemetry (Ceilometer) |"20:38
-openstackstatus- NOTICE: Gerrit is back online20:38
prad_eglynn, gordc, if you can take a look at should have all the changes discussed and ready to go20:49
*** shakamunyi has quit IRC20:51
*** jmckind has quit IRC20:52
*** shakamunyi has joined #openstack-ceilometer20:52
*** jdob has quit IRC20:54
*** thomasem has quit IRC20:58
cmartAlexei_987, are you there?21:01
Alexei_987cmart: Hi yes21:02
cmartHello.. you are alexei kornienko, right? The one that comment the patch I just submitted, right?21:02
Alexei_987cmart: Yes :)21:03
Alexei_987cmart: change itself is not bad21:03
eglynnprad_: comments inline21:03
eglynnprad_: thanks!21:03
Alexei_987cmart: just need to check with cores/PTL that it's acceptable21:03
Alexei_987eglynn: could you please check this one when you'll have time:
cmartAlexei_987: +1 to that. In my defense, I "joined" ceilometer after the summit :S21:04
Alexei_987eglynn: do we plan to add new features to v2 api?21:04
eglynnAlexei_987: well v2 is on a "soft deprecation path" in my mind21:04
*** promulo has quit IRC21:05
eglynnAlexei_987: ... which means that while it hasn't been *officially* deprecated21:05
eglynnAlexei_987: ... we should concentrate on our innovation on v321:05
Alexei_987eglynn: since we have a good progress in improving v2 performance maybe it will be reasonable to extend it? until we'll have a production ready v3?21:06
Alexei_987so we could release 2.1 or something like that21:07
cmartAlexei_987: .. and my implementation it's hitting performance right in the chin :(21:07
eglynnAlexei_987: ... yep, *soft* deprecation in my view means we can extend the API on a case-by-case basis21:08
eglynnAlexei_987: looking at now21:08
Alexei_987cmart: it's no so bad as you say21:08
Alexei_987cmart: I can only say for sqlalchemy + mongodb but this 2 looks ok21:08
Alexei_987cmart: except for minor pep8 issues21:09
Alexei_987cmart: in memory filtering is not really cool but since you already have a filtered query it should not iterate too many results21:10
eglynncmart: so no postgresql special-casing \o/ :(21:10
eglynncmart: ... I jest :)21:10
cmartAlexei_987: Glad to hear that :) (or not)21:10
Alexei_987cmart: the only thing I could add is that would should now allow to use meter filter *only*21:10
eglynncmart: ... no no postgresql special-casing == goodness :)21:10
Alexei_987cmart: it might be reasonable to fail query and require some additional filters21:11
Alexei_987to avoid read-all case21:11
Alexei_987for sqla21:12
cmartI think that the filtering options are waay too many.. and the schema ,IMHRookieO, doesn't help..21:12
eglynncmart: ... looks sane, -1 purely for lack of test coverage21:13
cmarteglynn: I couldn't find unit tests on that.. Only scenario tests...21:14
Alexei_987these are the ones you need21:14
*** shakamunyi has quit IRC21:14
Alexei_9871 test covers all backends21:14
eglynncmart: scenario tests are good enough for sqlalchemy and mongo (runs against sqlite and real mongodb)21:15
Alexei_987cmart: left you a small note on sqla21:15
cmarteglynn: test_post_invalid_alarm_query at file seems to be a good case to replicate and change21:16
cmartAlexei_987: which tess?21:16
cmartAlexei_987:eglynn: Any tips to make the tests run "faster"? Config options, extra params, any??21:18
Alexei_987cmart: move /tmp on tmpfs21:18
Alexei_987cmart: helps a lot :)21:18
eglynncmart: you can also run tests selectively21:20
cmartAlexei_987: (linux rookie): Do you mean move a /tmp folder that is being created inside ceilometer folder to a new created folder called tmpfs?21:20
cmarteglynn: I usually run them by activating a virtual env and then using testtools21:20
eglynncmart: ... $ tox -epy27 -- package.subpackage.module # <=== for example21:20
Alexei_987tmpfs                   /tmp                    tmpfs   size=4G       0 021:21
eglynn(^^^ runs only that individual module's worth of tests)21:21
Alexei_987you can put such record in your /etc/fstab21:21
Alexei_987and your system wide /tmp folder will stay in RAM21:21
Alexei_987ceilometer tests make heavy use of this folder so it helps a lot21:22
cmartAlexei_987: got it!21:22
Alexei_9874G is the size for /tmp21:22
Alexei_987you can put your own value here21:23
*** julim has quit IRC21:24
cmartAlexie_987: eglynn: thanks for the help!!!21:25
* Alexei_987 thinking about changing nick to Alex_987 :)21:26
eglynnAlexei_987: nah, wouldn't suit you ;)21:28
eglynnAlexei_987: ... the 'ei' adds cultural background :)21:28
dmsimardgordc: I can get the file publisher to work fine. Got to see how to have it read by the collector, now.21:28
Alexei_987eglynn: actually my friends often call me alex21:28
*** cmart has quit IRC21:29
eglynnAlexei_987: ... a-ha, I thought Alex was more a British-ism21:29
*** thomasem has joined #openstack-ceilometer21:29
*** thomasem has quit IRC21:30
*** nati_ueno has joined #openstack-ceilometer21:31
*** thomasem has joined #openstack-ceilometer21:31
Alexei_987eglynn: just in case you have good review mood now could you please also check this one:
*** openstackgerrit has joined #openstack-ceilometer21:34
gordcdmsimard: the file publisher isn't meant to connect to collector. you'd need to work your own magic there.21:34
eglynnAlexei_987: ... sure, looking now21:34
dmsimardgordc: Yeah, okay, I can deal with that - but how would you insert the data into ceilometer ?21:34
gordcdmsimard: when you say 'ceilometer' you mean the database?21:35
dmsimardgordc: Yeah, I'm using a mongodb backend.21:36
gordcdmsimard: using the file publisher? you'd need to write some script to parse and insert it.... i hadn't really thought about it.21:38
dmsimardgordc: Hmm, could these be posted as is to the metering API endpoint ? Like with
gordcdmsimard: seems like it could work.21:39
gordcnot sure how well that'd scale.21:39
*** promulo has joined #openstack-ceilometer21:41
*** Underyx|off is now known as underyx21:44
*** openstackgerrit has quit IRC21:47
*** raymondr_ has quit IRC21:56
*** jergerber has quit IRC22:02
*** shakamunyi has joined #openstack-ceilometer22:11
*** r0j4z0 has joined #openstack-ceilometer22:13
*** gordc has left #openstack-ceilometer22:21
*** shakamunyi has quit IRC22:22
*** nati_ueno has quit IRC22:22
*** shakayumi has joined #openstack-ceilometer22:22
*** yjiang5 is now known as yjiang5_away22:23
*** isviridov|away is now known as isviridov22:24
*** fnaval has quit IRC22:37
*** prad_ has quit IRC22:54
*** nati_ueno has joined #openstack-ceilometer22:56
*** raymondr_ has joined #openstack-ceilometer23:06
*** fnaval has joined #openstack-ceilometer23:13
*** ddieterly has quit IRC23:27
*** fnaval has quit IRC23:42
*** ddieterly has joined #openstack-ceilometer23:51
*** zhikunliu has joined #openstack-ceilometer23:59

Generated by 2.14.0 by Marius Gedminas - find it at!