Wednesday, 2015-06-03

*** yfujioka_ has joined #openstack-ceilometer00:02
*** ddieterly has joined #openstack-ceilometer00:02
*** yfujioka has quit IRC00:03
*** sputnik13 has quit IRC00:08
*** shakamunyi has joined #openstack-ceilometer00:13
*** barra204 has joined #openstack-ceilometer00:13
*** dikonoor has joined #openstack-ceilometer00:15
*** pauldix has joined #openstack-ceilometer00:23
openstackgerritRohit Jaiswal proposed openstack/ceilometer: Fix catch-all event def to exclude !event
*** rjaiswal has quit IRC00:26
*** fawadkhaliq has quit IRC00:48
*** promulo__ has joined #openstack-ceilometer00:52
*** promulo has quit IRC00:56
*** _cjones_ has quit IRC01:06
*** ddieterly has quit IRC01:08
*** superfly_ has joined #openstack-ceilometer01:12
*** superfly_ is now known as superflyy01:14
*** barra204 has quit IRC01:24
*** shakamunyi has quit IRC01:24
*** superflyy has quit IRC01:24
*** ddieterly has joined #openstack-ceilometer01:30
*** rjaiswal has joined #openstack-ceilometer02:02
*** vivek_nandavanam has joined #openstack-ceilometer02:06
vivek_nandavanamHi, what is the default retention period for Ceilometer?02:08
rjaiswalThink it's disabled by default but can enabled by specifying metering_time_to_live in ceilometer.conf02:21
rjaiswalValue in seconds02:21
*** robsparker has joined #openstack-ceilometer02:23
*** haigang has quit IRC02:26
*** haigang has joined #openstack-ceilometer02:27
*** haigang has quit IRC02:30
vivek_nandavanamSo by default, are the samples kept in the database forever?02:31
vivek_nandavanamLooking at
vivek_nandavanamWhen the database dispatcher is configured as data store, you have the option to set a time_to_live option (ttl) for samples. By default the time to live value for samples is set to -1, which means that they are kept in the database forever.02:31
*** haigang has joined #openstack-ceilometer02:32
vivek_nandavanamtime_to_live is a different option compared to metering_time_to_live02:32
*** haigang has quit IRC02:33
*** haigang has joined #openstack-ceilometer02:34
vivek_nandavanamNevermind I see that time_to_live is deprecated02:35
vivek_nandavanamrjaiswal: Thanks02:35
*** Kennan2 has joined #openstack-ceilometer02:37
*** Kennan has quit IRC02:38
*** haigang has quit IRC02:38
*** haigang has joined #openstack-ceilometer02:46
*** Kennan2 is now known as Kennan02:49
*** pauldix has quit IRC02:55
*** ddaskal has joined #openstack-ceilometer03:00
*** jfluhmann has joined #openstack-ceilometer03:03
*** ddieterly has quit IRC03:10
*** robsparker has quit IRC03:26
*** Longgeek_ has quit IRC03:53
*** Longgeek has joined #openstack-ceilometer03:54
*** alexpilotti has joined #openstack-ceilometer03:54
*** Longgeek_ has joined #openstack-ceilometer03:57
*** ddaskal has quit IRC03:59
*** changbl_ has joined #openstack-ceilometer04:00
*** Longgeek has quit IRC04:00
openstackgerritMerged openstack/ceilometer: Disable rgw pollster when aws module not found
*** Longgeek has joined #openstack-ceilometer04:02
*** yatin has joined #openstack-ceilometer04:03
*** Longgeek_ has quit IRC04:04
*** cmyster_ has joined #openstack-ceilometer04:06
*** cmyster_ has quit IRC04:06
*** Longgeek_ has joined #openstack-ceilometer04:08
*** Longgeek has quit IRC04:10
*** Longgeek_ has quit IRC04:12
*** ddieterly has joined #openstack-ceilometer04:15
*** Longgeek has joined #openstack-ceilometer04:18
*** Longgeek has quit IRC04:18
*** Longgeek has joined #openstack-ceilometer04:19
*** ddieterly has quit IRC04:20
*** Longgeek has quit IRC04:20
*** Longgeek has joined #openstack-ceilometer04:20
*** Longgeek_ has joined #openstack-ceilometer04:23
*** Longgee__ has joined #openstack-ceilometer04:25
*** Longgeek has quit IRC04:25
*** Longgeek_ has quit IRC04:26
*** Longgeek has joined #openstack-ceilometer04:26
*** Longgeek_ has joined #openstack-ceilometer04:28
*** Longge___ has joined #openstack-ceilometer04:29
*** Longgee__ has quit IRC04:30
*** Longgeek has quit IRC04:32
*** Longgeek_ has quit IRC04:32
*** changbl_ has quit IRC04:40
*** jfluhmann has quit IRC04:43
*** pauldix has joined #openstack-ceilometer04:43
*** pauldix has quit IRC04:48
*** _cjones_ has joined #openstack-ceilometer04:48
*** dikonoor has quit IRC04:52
*** ildikov has quit IRC04:54
*** _cjones_ has quit IRC04:55
*** ildikov has joined #openstack-ceilometer05:13
*** ddieterly has joined #openstack-ceilometer05:16
*** ddieterly has quit IRC05:21
*** robsparker has joined #openstack-ceilometer05:25
*** exploreshaifali has joined #openstack-ceilometer05:27
*** robsparker has quit IRC05:29
*** yprokule has joined #openstack-ceilometer05:32
*** ildikov has quit IRC05:38
*** exploreshaifali has quit IRC05:43
*** ildikov has joined #openstack-ceilometer06:04
*** shakamunyi has joined #openstack-ceilometer06:16
*** barra204 has joined #openstack-ceilometer06:17
*** ddieterly has joined #openstack-ceilometer06:17
*** _nadya_ has joined #openstack-ceilometer06:18
*** ddieterly has quit IRC06:22
*** belmoreira has joined #openstack-ceilometer06:26
*** rjaiswal_ has joined #openstack-ceilometer06:27
*** Longgeek has joined #openstack-ceilometer06:35
*** Longge___ has quit IRC06:37
*** rjaiswal_ has quit IRC06:40
*** _nadya_ has quit IRC06:48
*** shardy has joined #openstack-ceilometer06:55
*** sudorandom has quit IRC07:16
*** exploreshaifali has joined #openstack-ceilometer07:17
*** sudorandom has joined #openstack-ceilometer07:18
*** ddieterly has joined #openstack-ceilometer07:18
*** haigang has quit IRC07:19
*** _nadya_ has joined #openstack-ceilometer07:23
*** ddieterly has quit IRC07:23
*** belmoreira has quit IRC07:42
*** belmoreira has joined #openstack-ceilometer07:46
*** exploreshaifali has quit IRC08:08
*** Kennan2 has joined #openstack-ceilometer08:11
*** sudorandom has quit IRC08:11
*** Kennan has quit IRC08:11
*** shardy_ has joined #openstack-ceilometer08:16
*** shardy has quit IRC08:17
*** ddieterly has joined #openstack-ceilometer08:19
*** sudorandom has joined #openstack-ceilometer08:19
*** shardy_ has quit IRC08:22
*** shardy has joined #openstack-ceilometer08:22
*** ddieterly has quit IRC08:23
*** vivek_nandavanam has quit IRC08:23
openstackgerritJulien Danjou proposed openstack/gnocchi: Remove AvoidDictInterface
*** yassine_ has joined #openstack-ceilometer08:27
jd__therve: damn everything got merged fast :)08:27
*** cmyster has quit IRC08:31
*** _nadya_ has quit IRC08:34
*** _nadya_ has joined #openstack-ceilometer08:34
*** _nadya_ has quit IRC08:39
*** safchain has joined #openstack-ceilometer08:57
*** cdent has joined #openstack-ceilometer09:10
kaliyahello, I'm facing is this ceilometer, oslo or both?09:12
cdentkaliya: do you have a dist-packages/ceilometer*.egg-info/entry_points.txt file?09:15
cdent(or something similar)09:15
cdent(the path may be different from what I'm showing)09:15
*** claudiub has joined #openstack-ceilometer09:17
kaliyacdent: now not, I must login to the system. What's there in that file?09:17
cdentthat lists the entry points that stevedore will use to find out things like what dispatchers are available09:18
cdentit's a quite long file, mine looks like this:09:18
cdentat line 49 is "[ceilometer.dispatcher]"09:19
cdentthe error message you're showing is suggesting that that information is not present or not being read for some reason09:19
*** ddieterly has joined #openstack-ceilometer09:19
kaliyawhat is the dispatcher in ceilometer? a core component?09:20
cdentthe dispatcher "dispatches" collected meters to a next stage, usually database storage09:21
kaliyacdent: thanks much indeed, today I check09:22
*** afazekas has joined #openstack-ceilometer09:22
cdentIf it is the case that you don't have an entry_points.txt then something went wrong with your installation and lots of stuff will not work properly09:23
*** ddieterly has quit IRC09:24
kaliyagood starting point, thanks cdent ;) I have still alot to learn about ceilometer09:30
cdentme too! :)09:31
thervejd__, Yeah apparently people don't care :)09:54
*** yfujioka_ has quit IRC10:02
*** _nadya_ has joined #openstack-ceilometer10:06
*** pauldix has joined #openstack-ceilometer10:10
*** exploreshaifali has joined #openstack-ceilometer10:12
*** pauldix has quit IRC10:14
cdentjd__: this has happened twice in a row now on
cdentso I'm guessing there's something not giving backs its postgress connections10:16
cdent(apparently it is use extra letter s day for me)10:17
openstackgerritJulien Danjou proposed openstack/gnocchi: Raise the PostgreSQL connections to 200 + Remove AvoidDictInterface
openstackgerritJulien Danjou proposed openstack/gnocchi: Use pg_ctl for initdb
cdentI'll look at the test code and see if I can find any dangling references10:19
jd__yeah :(10:19
*** ddieterly has joined #openstack-ceilometer10:20
*** ddieterly has quit IRC10:25
*** robsparker has joined #openstack-ceilometer10:25
*** robsparker has quit IRC10:31
openstackgerritMerged openstack/ceilometer: Updated from global requirements
*** cmyster has joined #openstack-ceilometer10:39
openstackgerritMerged openstack/ceilometer: Fixes DiskInfoPollster AttributeError exception
*** _nadya_ has quit IRC10:46
*** _nadya_ has joined #openstack-ceilometer10:48
*** exploreshaifali has quit IRC10:54
*** pauldix has joined #openstack-ceilometer11:05
*** julim has joined #openstack-ceilometer11:06
*** exploreshaifali has joined #openstack-ceilometer11:10
*** cmyster has quit IRC11:14
*** cmyster has joined #openstack-ceilometer11:15
*** cmyster has joined #openstack-ceilometer11:15
*** ddieterly has joined #openstack-ceilometer11:21
*** ddieterly has quit IRC11:26
*** cmyster has quit IRC11:26
*** cmyster has joined #openstack-ceilometer11:27
*** robsparker has joined #openstack-ceilometer11:41
*** ddieterly has joined #openstack-ceilometer11:49
*** ddieterly has quit IRC12:11
*** pauldix has quit IRC12:12
*** _nadya_ has quit IRC12:19
*** robsparker has quit IRC12:34
*** gordc has joined #openstack-ceilometer12:38
openstackgerritgordon chung proposed openstack/ceilometer: WSME version >=0.7 correctly returns a 405
*** ddieterly has joined #openstack-ceilometer12:43
cdentjd__: I think I figured it out, it's basically a problem with the way the gabbi tests were calling setup_app. patch forthcoming12:44
cdent(with my changes I can get the tests to pass with a postgres -N 6)12:45
jd__cdent: feel free to remove the -N 200 option then if not necessary :)12:45
*** yatin has quit IRC12:52
*** primechuck has joined #openstack-ceilometer12:52
*** pauldix has joined #openstack-ceilometer12:52
*** primechuck has quit IRC12:53
*** primechu_ has joined #openstack-ceilometer12:53
*** rwsu has joined #openstack-ceilometer12:54
openstackgerritMerged openstack/gnocchi: Raise the PostgreSQL connections to 200 + Remove AvoidDictInterface
*** _nadya_ has joined #openstack-ceilometer12:57
*** ddieterly has quit IRC12:58
openstackgerritMerged openstack/gnocchi: Use pg_ctl for initdb
*** ddieterly has joined #openstack-ceilometer13:03
openstackgerritChris Dent proposed openstack/gnocchi: Handle indexer connections more cleanly in tests
gordcjd__: cdent: we have any plans on capping wsme<0.7.0 for stable/kilo? i assume we are leaving the breaking changes from 0.7.0 as is?13:14
jd__no se13:14
cdentI am not wise in the ways of these capping things13:15
cdentbut I suppose for safety the right kilo thing might be wsme=0.713:15
gordcjd__: what's 'se'? what i google 'no se' it just assumes i mean 'nose'13:16
cdentel habla español13:17
gordccdent: oh fancy hairpiece on the 'n'13:17
gordccdent: i'm just asking for
cdentI knöw rîght13:17
gordcif we cap requirements, we technically don't need that backport13:17
gordcwhy do you know how to put hats/roofs/hairpieces on so many letters?13:18
cdentI don't know much about capping requirements, but I do know about capping letters13:19
gordclol word.13:19
cdentit's cuz I use one of these new fangled mac things with easy ways to compose characters13:19
cdentI tend to think that capping requirements makes things hard for people who live in the real world13:20
gordci use this new fangled windows thing now and i realised fedora might not be so bad afterall13:20
cdentoh dear13:20
cdentIf we backport that test, then we also have to backport a change in requirements to require >=0.7, otherwise the test will fail13:21
gordccdent: yeah.13:21
cdentwhich feels hairy13:21
gordccdent: i'll check with dhellmann13:21
cdenthe probably at least knows what the normal thing to do is13:21
cdentI think we should never backport anything and require people always upgrade :)13:22
cdentburn world! burn!13:22
gordccdent: that's what i say all the time. but then ildikov says no.13:22
*** robsparker has joined #openstack-ceilometer13:22
dhellmanngordc, cdent : if wsme 0.7 breaks kilo a cap seems in order13:22
cdentSorry BüRN13:22
gordcdhellmann:  :) thanks. will cap it in kilo13:22
cdentthat was easy :)13:23
* ildikov is reading history :)13:23
cdentabout 15 billion years ago...13:24
ildikovcdent: do we support rolling upgrade with no downtime?13:24
openstackgerritChris Dent proposed openstack/gnocchi: Require pecan 0.9, gaining 405 and unicode fixes
ildikovcdent: I guess that equals to no and now you're a big fan of backporting, right? :)13:31
* cdent thinks about how to phrase what he thinks...13:32
cdentI guess: If we are far too willing and able to backport things then the likelihood of us or anyone else ever getting around to implementing clean rolling upgrades decreases13:33
cdentThis is in the same camp of: If we cap requirements all the time then we are being bad citizens to the wide open open source community because we are not encouraging bugs to flow freely and it is bugs which are the driving force of this kind of work.13:33
cdentSo my fundamental position is: on the open source side of things it is very important to always be moving forward. On the commerce side of things (the consultants, the distros, the people who sell stuff), that's where concerns about backporting should really be located.13:34
cdentI recognize that this is not entirely practical nor pragmatic and compromises will always need to be made, but when each of us takes different positions the diversity of opinions leads, eventually, to the best outcomes13:36
cdentor if you want me to be less serious:13:36
cdentmeh, I don't run clouds, I jus write code13:36
ildikovfrom user and vendor perspective you cannot operate on trunk, so some amount of backporting is needed, just like backward compatibility13:39
ildikovthis does not mean that I like it, but we need to accept the need for it13:39
ildikovwhat I'm usually aiming for is balance and trying not to mention gordc's favourite word too many times :)13:41
*** changbl_ has joined #openstack-ceilometer13:43
cdent:) balance is good13:44
cdentI tend to be politically oriented towards making things as hard as possible for vendors13:44
cdentand as interesting/easy as possible for individuals13:45
jasonamyersbasically cdent hates me13:45
cdenttotes dood, you are so the Enëmy13:46
openstackgerritMerged openstack/ceilometer: Move eventlet using commands into own directory
jasonamyersall evil and corporatey13:46
cdentyeah, that's one of the things I noticed about you in Vancouver13:46
cdentI thought: Man, that jason guy is such a suit13:46
openstackgerritMerged openstack/ceilometer: TraitText value restricted to max length 255
ildikovwell, I prefer to earn some salary too, so it's not that black and white IMHO :)13:51
ildikovjasonamyers: no worries, I'm trying to keep cdent on the ground to not make your life complete hell :)13:52
cdentI'm serious though: I think it is _very_ important that we balance the weight of enterprisey concerns with other concerns; of course it is not black and white in _action_ but sometimes it is useful when _discussing_ to speak in black and white so the conversation can move along quickly13:52
cdentthe outcome of the discussions will lead to good stuff (eventually)13:53
cdentideally it will mean that jasonamyers does all the work13:53
jasonamyerscdent: agreed in my case Im not the end customer so Im often not advocating my own position13:54
ildikovcool, I always wanted to opem a bakery... :D13:54
jasonamyersand speaking of corporate apparently we're buying Piston cloud13:55
jasonamyersoff to learn about that13:55
*** claudiub has quit IRC13:55
cdentthat's kind of a bummer, consolidation makes my inner anarchist sad13:56
*** ildikov has quit IRC14:00
gordcjasonamyers: breaking news?14:01
gordccdent: i'm actually wearing a suit today. it's corporate picture day14:01
jasonamyersjust announced it seems14:01
jasonamyersgordc: wins this corporate award them14:02
jasonamyersnever been in a suit14:02
*** rbak has joined #openstack-ceilometer14:02
jasonamyersonly a tux to get married14:02
cdentgordc: you gonna replace your avatar image with your new suit styleee14:02
*** Kennan2 has quit IRC14:02
pradjasonamyers, interesting news14:02
gordccdent: i should. i need to buy some more suits. people will start noticing it's the same shirt, tie, suit.14:03
*** Kennan has joined #openstack-ceilometer14:03
*** ildikov has joined #openstack-ceilometer14:03
gordcit's like when you get carded and you're wearing the exact same clothes as your id...14:03
cdentdo you speak from personal experience on that one?14:05
openstackgerritMerged openstack/gnocchi: Handle indexer connections more cleanly in tests
gordccdent: all the time. i only have a few shirts. and asians look like children so we get carded constantly.lol14:06
gordcjasonamyers: :)14:09
*** idegtiarov has joined #openstack-ceilometer14:10
*** rjaiswal_ has joined #openstack-ceilometer14:12
*** llu has quit IRC14:12
*** llu has joined #openstack-ceilometer14:13
*** dikonoor has joined #openstack-ceilometer14:13
*** triggerz is now known as tobasco14:14
dikonoorgordc: Hi. Implemented your comment for Could you take a look?14:14
openstackgerritJulien Danjou proposed openstack/ceilometer: sqlalchemy: use JSON from sqlalchemy-utils
dikonoorgordc:I have some queries on event rbac. I noticed that I was not able to put the regular rules for events in the ceilometer policy.json file (like you can do for meters). The only rule that works in for admins i e only admins can view events. zqfan pointed out that events don't have user and project information stored with it14:18
dikonoorgordc: With the current event schema , it's not possible to add role-based or user or project based restrictions to events. For eg. If I want to add a restriction that says - only the user who generated an event can view it - that don't work14:20
dikonoorgordc: Similarly , I cannot add a restriction that says that only users from a speific project should be able to view their events (and not users from another project)..14:21
*** david-lyle has quit IRC14:25
*** jruano has joined #openstack-ceilometer14:26
*** rjaiswal__ has joined #openstack-ceilometer14:26
*** rjaiswal_ has quit IRC14:26
gordcdikonoor: yes. that is not possible. i'm not sure we can ensure there's always a tenant or user id for every message14:28
openstackgerritMerged openstack/gnocchi: Require pecan 0.9, gaining 405 and unicode fixes
dikonoorgordc: As auditing is now moved as events, a rbac rule that says something like - "user can see only their audit data" would add a lot of value.14:28
gordcdikonoor: do you know what type of changes would be required?14:29
dikonoorgordc: user/project info for all events may not be present but I believe it will be for all audit records.14:29
*** rjaiswal__ has quit IRC14:30
dikonoorgordc: On the changes, my guess is that the generated events should have user/project info but I haven't looked at where that change has to go in14:30
dikonoorgordc: second change  would be on the policy enforcement for events , which now allows only admins to view it14:31
*** sputnik13 has joined #openstack-ceilometer14:31
gordcdikonoor: if the user/project info are in the traits part of event, can we still apply rbac rules?14:31
*** rwsu has quit IRC14:31
*** _nadya_ has quit IRC14:32
dikonoorgordc: I will have to check that.. So I guess your thought process is that for events where user/project data is available , that could be added in the traits and then rbac could run on that14:33
*** rjaiswal_ has joined #openstack-ceilometer14:33
dikonoorgordc: Let me check and get back. Ceilometer rbac as a whole looks very different from how it's implemented in the other services..14:33
gordcdikonoor: :( i think part of it was to ensure backward compatibility. but if you can point out differences maybe we should move it as close to other implementations as possible to ensure consistency14:35
*** exploreshaifali has quit IRC14:36
*** barra204 has quit IRC14:36
*** shakamunyi has quit IRC14:36
dikonoorgordc: I did not quite understand the need for a (which then calls oslo_policy)..but I guess it's for backward compatibility.. I noticed that apart from the rules (in policy.json) that involve roles , other rules don't work..for eg. stuff like these "owner": "user_id:%(user_id)s",14:37
openstackgerritDoug Hellmann proposed openstack/python-ceilometerclient: Drop use of 'oslo' namespace package
dikonoorgordc: These sort of rules work fine in all other projects( those that use oslo-incubator policy code and those who have moved to oslo_policy)14:38
gordcdikonoor: hmm. maybe we can open a bug for that.14:38
dikonoorgordc: I tried only for meter samples . I was trying to see if only the user who generated a meter samples could view the samples..14:39
dikonoorgordc: ok..I will open one..14:39
*** llu has quit IRC14:41
*** llu has joined #openstack-ceilometer14:41
gordcdikonoor: cool cool. thanks14:41
dikonoorgordc: . I did not completely understand what this segregation rule is needed for and the method get_limited_to.. looks like it's trying to restrict the access to certain users?14:42
*** afazekas has quit IRC14:43
dikonoorgordc: I went through the logic but I feel this can be removed and this support can be provided by routing the request directly to oslo_policy and by setting the right rules in the policy.json14:43
gordcdikonoor: hmm. to be honest, i'm not super familiar with rbac work...14:44
gordcdikonoor: but yeah, if we can move the handling outside of ceilometer and provide the same functionality, i'm all for it.14:44
dikonoorgordc: For all requests , rbac should be controlled by the policy.json ..So let me start by opening couple of bugs.. Will add details there. ..Who would be the right person to contact for ceilometer rbac?14:45
gordcdikonoor: not sure we have a contact for rbac specifically. fabiog implemented it. but if you can, i'm hoping you can fix it. :)14:46
dikonoorgordc:  :)14:47
dikonoorgordc: could you take a look at when you get a chance ? Thanks14:48
*** robsparker has quit IRC14:54
*** rjaiswal_ has quit IRC15:05
*** changbl_ has quit IRC15:10
*** dikonoor has quit IRC15:11
*** changbl_ has joined #openstack-ceilometer15:11
karolynHi gordc, sorry, I've been away. I'm getting back to now. The cpu_util calculation for overcommit on kvm. I saw you had a comment about if it's a new agent15:17
*** ddaskal has joined #openstack-ceilometer15:18
*** rwsu has joined #openstack-ceilometer15:19
karolyngordc: so the code uses the Linux kernel schedstats counters instead of libvirt. So that makes it a new agent, correct? or am I misunderstanding what "agent" means in this context?15:23
*** alexpilotti has quit IRC15:24
gordckarolyn: sorry. was in a meeting.15:31
gordckarolyn: yeah, i guess the question was more whether it needed to function on the host or if it could exists outside host15:31
gordcbut considering you mention kernel schdstats i'm assuming it's on the host? or is it on the instance?15:32
*** karolyn has quit IRC15:34
*** karolyn has joined #openstack-ceilometer15:34
karolynIt's on the host15:34
*** david-lyle has joined #openstack-ceilometer15:39
karolyngordc: Not sure if my previous message went through. I lost my connection... It is on the host.15:40
gordckarolyn: sorry again, didn't get a prompt for your message15:40
gordckarolyn: but that's good to know it's on the host and not instance.15:41
karolyngordc: Do you think it's okay to do this in the cpu_util metric? or would you prefer it be a different metric?15:42
cdentgordc: since DinaBelova is otherwise occupied should we make me the QA liaison (since I'm kind of it by default anyway):
*** david-lyle has quit IRC15:47
*** alexpilotti has joined #openstack-ceilometer15:48
gordckarolyn: i think it's ok to replace it as i think that's what people actually expect from the meter.15:48
gordccdent: sure. i'm ok with it. i think DinaBelova said she'd be busy with internal stuff for part of this cycle still but she can correct me otherwise15:49
gordckarolyn: does the cpu_util value still behave the same when it's not overcommit?15:49
cdentroger that15:49
* cdent updates15:49
*** adurbin__ has joined #openstack-ceilometer15:50
karolynit will always use schdstats...but when it's not overcommitted, it should come up with the same value as libvirt was before15:50
gordckarolyn: ah i see. so i guess we don't need the transformer for libvirt to generate cpu_util meters after this.15:52
gordckarolyn: i think it's ok to replace. we can just mark it as docimpact and we'll note the change.15:52
*** david-lyle has joined #openstack-ceilometer15:53
karolyngordc: okay, I'll rebase and make a few other tweeks and put out another patch15:57
gordckarolyn: awesome. thanks!15:57
*** belmoreira has quit IRC16:07
*** _cjones_ has joined #openstack-ceilometer16:07
*** afazekas has joined #openstack-ceilometer16:12
*** gordc is now known as gordc_afk16:19
*** barra204 has joined #openstack-ceilometer16:21
*** david-lyle has quit IRC16:22
*** ddieterly has quit IRC16:26
*** barra204 has quit IRC16:28
*** yassine_ has quit IRC16:31
*** yatin has joined #openstack-ceilometer16:34
*** dand has joined #openstack-ceilometer16:39
*** safchain has quit IRC16:41
*** ddieterly has joined #openstack-ceilometer16:51
*** dscottsf has joined #openstack-ceilometer16:54
*** dscottsf has quit IRC16:55
*** afazekas has quit IRC16:55
*** Longgeek has quit IRC16:55
*** yprokule has quit IRC16:57
*** belmoreira has joined #openstack-ceilometer16:58
*** dand has quit IRC17:03
*** yatin has quit IRC17:05
*** ildikov has quit IRC17:07
*** afazekas has joined #openstack-ceilometer17:09
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements
*** david-lyle has joined #openstack-ceilometer17:13
*** david-lyle has quit IRC17:13
*** dand has joined #openstack-ceilometer17:14
*** david-lyle has joined #openstack-ceilometer17:15
*** megh has joined #openstack-ceilometer17:27
*** david-lyle has quit IRC17:28
*** megh_ has joined #openstack-ceilometer17:28
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements
*** megh has quit IRC17:32
*** ddieterly has quit IRC17:32
*** gordc_afk is now known as gordc17:39
openstackgerritOpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements
openstackgerritMerged openstack/python-ceilometerclient: [unittest] Increase client and shell modules cover
*** belmoreira has quit IRC18:02
*** afazekas has quit IRC18:07
rjaiswalGordc: updated, please have a look. Thanks18:09
*** ildikov has joined #openstack-ceilometer18:11
gordcrjaiswal: cool cool. will look18:29
*** david-lyle has joined #openstack-ceilometer18:33
openstackgerritMerged openstack/python-ceilometerclient: Drop use of 'oslo' namespace package
*** cdent has quit IRC18:50
*** megh_ has quit IRC18:57
*** barra204 has joined #openstack-ceilometer18:57
*** megh has joined #openstack-ceilometer18:58
*** barra204 has quit IRC19:02
openstackgerritMerged openstack/python-ceilometerclient: Catch missing 404 exceptions in client
*** shardy_ has joined #openstack-ceilometer19:06
*** shardy has quit IRC19:08
*** cdent has joined #openstack-ceilometer19:09
*** shardy_ has quit IRC19:12
*** shardy has joined #openstack-ceilometer19:12
*** megh has quit IRC19:15
*** rbak_ has joined #openstack-ceilometer19:15
openstackgerritMerged openstack/gnocchi: Switch to PyMySQL and enable MySQL on py34
*** rbak has quit IRC19:18
openstackgerritMerged openstack/gnocchi: indexer: enable MySQL schema migration test
*** megh has joined #openstack-ceilometer19:46
*** megh has quit IRC19:51
pradgordc, hey19:53
gordcprad: so about exchanges...19:53
gordci don't recall what fabiog's idea was.19:54
gordci saw he reopened his notification_meter_pipeline... which i still don't think we need19:54
pradgordc, yea i thought he wanted to put something at the pipeline level.. vaguely recall19:54
gordcprad: ok. yeah i think that's it19:54
gordci think the idea was that he wanted a pipeline to control notification meters vs polling meters19:55
jasonamyershe wants a way to drain to kafka19:55
pradgordc, yea something like that19:55
pradgordc, but it isnt very clear to me how that would work19:55
gordcso that can be done as of today. we pass in the pipeline file to each of the agents... each agent has it's own copy regardless of polling/notification19:56
gordcit just so happens that everyone is lazy and passes the same pipeline file to both polling and notification agents19:56
gordcjasonamyers: you can correct me if i'm wrong but i'm pretty sure the drain to kafka stuff is possible today?19:57
pradi think drain to kafka was a different discussion? dint think that was related to exchange control stuff we discussed at the summit19:58
*** barra204 has joined #openstack-ceilometer19:58
*** _nadya_ has joined #openstack-ceilometer20:00
gordcprad: i'll reread the spec.20:00
gordcbut i think we should leave exchange control out of your declarative meters work for now.20:01
pradgordc, thx.. i'll update the spec to make it a bit more clear where i'll be adding the new base class etc.. but i need some clarity on how we go about with exchange control. that part of the spec is not flushed out20:01
pradgordc, well i'm cool with that, but the exchange control opts have to go somewhere when the files are nuked?20:02
gordcprad: true true20:02
jasonamyersgordc: it is he wants to avoid hitting the queue a second time I think20:02
jasonamyersI wasnt towing the corporate line for that as it seemed unneccessary20:03
jasonamyersto add that20:03
*** barra204 has quit IRC20:03
jasonamyersthus his frustration with me20:04
gordcprad: imo i think we should just add an option. you list the exchanges you want to listen to, by default we include all exchanges20:04
*** megh has joined #openstack-ceilometer20:05
gordcjasonamyers: i see... i think he was against having our polling agents requeue to be processed by notification agent.20:05
pradgordc, and the base notification agent class will process that?20:05
jasonamyersgordc: yeah that as well20:05
gordcprad ^20:05
pradhmm ok.. i'll play around with it and see..20:06
pradi'll mark the spec WIP till i flush out the exchange control piece20:06
pradgordc, feel free to review anyway20:06
gordcjasonamyers: i'm undecided on that... i think it'd move a lot of the processing load off polling agents (and by association off compute nodes)20:06
gordcthat said, i don't know if there's much load from polling agents (aside from hitting apis)20:07
gordcbut probably something to investigate... it'd also make our polling agents 'declarative' as well20:07
*** alexpilotti has quit IRC20:17
pradgordc, when you get a chance glance through it again
*** ddaskal has quit IRC20:25
gordcprad: sure. i might need to look at it later tonight. going to go coach some baseball in a bit and have to finish up some other stuff20:26
pradgordc, sure np .. enjoy your baseball20:27
*** _nadya_ has quit IRC20:34
*** idegtiarov has quit IRC20:48
*** gordc has quit IRC20:58
*** swamireddy has quit IRC21:03
*** cdent_ has joined #openstack-ceilometer21:05
*** cdent has quit IRC21:06
*** cdent_ is now known as cdent21:06
*** swamireddy has joined #openstack-ceilometer21:24
*** jruano has quit IRC21:34
*** dontalton has joined #openstack-ceilometer21:39
*** alexpilotti has joined #openstack-ceilometer21:45
*** adurbin__ has quit IRC21:57
*** megh has quit IRC22:00
*** rbak_ has quit IRC22:06
*** primechu_ has quit IRC22:15
*** megh has joined #openstack-ceilometer22:15
*** primechuck has joined #openstack-ceilometer22:20
*** cdent has quit IRC22:40
*** jruano has joined #openstack-ceilometer22:46
*** barra204 has joined #openstack-ceilometer22:48
*** megh has quit IRC22:48
*** htruta_ has joined #openstack-ceilometer22:49
*** dand has quit IRC22:49
*** barra204 has quit IRC22:53
*** megh has joined #openstack-ceilometer22:54
*** pauldix has quit IRC22:55
*** jruano has quit IRC22:59
*** primechuck has quit IRC23:05
*** sputnik13 has quit IRC23:27
*** sputnik13 has joined #openstack-ceilometer23:28
*** sputnik13 has quit IRC23:31
*** ddieterly has joined #openstack-ceilometer23:33
*** sputnik13 has joined #openstack-ceilometer23:33
*** dontalton has quit IRC23:37
*** rwsu has quit IRC23:41
*** titilambert has quit IRC23:52
*** titilambert has joined #openstack-ceilometer23:52
*** yfujioka has joined #openstack-ceilometer23:52
*** sputnik13 has quit IRC23:58

Generated by 2.14.0 by Marius Gedminas - find it at!