| *** peristeri has quit IRC | 00:01 | |
| *** diogogmt has quit IRC | 00:21 | |
| openstackgerrit | Merged openstack/gnocchi: Log retrieve/store data speed in Carbonara based drivers https://review.openstack.org/281674 | 00:46 |
|---|---|---|
| *** thorst has joined #openstack-telemetry | 00:47 | |
| *** thorst has quit IRC | 00:51 | |
| *** thorst has joined #openstack-telemetry | 01:11 | |
| *** diogogmt has joined #openstack-telemetry | 01:19 | |
| *** jroll has quit IRC | 01:22 | |
| *** jroll has joined #openstack-telemetry | 01:24 | |
| openstackgerrit | Kairo Araujo proposed openstack/ceilometer: Initial seed of hacking https://review.openstack.org/282124 | 01:27 |
| *** thorst has quit IRC | 01:36 | |
| *** thorst has joined #openstack-telemetry | 01:36 | |
| *** Liuqing has joined #openstack-telemetry | 01:44 | |
| *** thorst has quit IRC | 01:45 | |
| *** liusheng has joined #openstack-telemetry | 01:48 | |
| *** cheneydc has joined #openstack-telemetry | 01:51 | |
| *** ljxiash has joined #openstack-telemetry | 01:53 | |
| *** chlong_ has joined #openstack-telemetry | 01:55 | |
| *** david-lyle_ has joined #openstack-telemetry | 02:04 | |
| *** david-lyle_ is now known as david-lyle | 02:05 | |
| openstackgerrit | Merged openstack/gnocchi: indexer: replace get_metrics() by list_metrics() https://review.openstack.org/274683 | 02:06 |
| openstackgerrit | Merged openstack/gnocchi: rest: implement groupby in resource/metric aggregation https://review.openstack.org/274684 | 02:06 |
| *** thorst has joined #openstack-telemetry | 02:07 | |
| *** thorst_ has joined #openstack-telemetry | 02:08 | |
| *** thorst has quit IRC | 02:11 | |
| openstackgerrit | liusheng proposed openstack/aodh: Add composite rule alarm evaluator https://review.openstack.org/261226 | 02:17 |
| *** liamji has joined #openstack-telemetry | 02:58 | |
| openstackgerrit | Zi Lian Ji proposed openstack/ceilometer: Enable the Load Balancer v2 for the Ceilometer(Part One) https://review.openstack.org/274518 | 03:01 |
| *** krotscheck_dr is now known as krotscheck_afk | 03:18 | |
| *** ljxiash has quit IRC | 03:26 | |
| *** links has joined #openstack-telemetry | 03:30 | |
| *** liamji has quit IRC | 03:30 | |
| *** david-lyle_ has joined #openstack-telemetry | 03:42 | |
| *** david-lyle has quit IRC | 03:45 | |
| *** david-lyle_ has quit IRC | 04:03 | |
| *** david_lyle__ has joined #openstack-telemetry | 04:03 | |
| *** jwcroppe has quit IRC | 04:07 | |
| *** achatterjee has joined #openstack-telemetry | 04:08 | |
| *** david_lyle__ has quit IRC | 04:08 | |
| *** jwcroppe has joined #openstack-telemetry | 04:08 | |
| *** jwcroppe has quit IRC | 04:13 | |
| *** dikonoor has joined #openstack-telemetry | 04:36 | |
| *** thorst_ has quit IRC | 04:37 | |
| *** thorst has joined #openstack-telemetry | 04:38 | |
| *** thorst has quit IRC | 04:46 | |
| *** Liuqing has quit IRC | 04:47 | |
| *** Liuqing has joined #openstack-telemetry | 04:55 | |
| *** yprokule has joined #openstack-telemetry | 05:00 | |
| *** ljxiash has joined #openstack-telemetry | 05:02 | |
| *** achatterjee has quit IRC | 05:03 | |
| *** liamji has joined #openstack-telemetry | 05:12 | |
| *** dave-mccowan has quit IRC | 05:28 | |
| *** rcernin has joined #openstack-telemetry | 05:38 | |
| *** thorst has joined #openstack-telemetry | 05:44 | |
| *** _nadya_ has quit IRC | 05:46 | |
| *** thorst has quit IRC | 05:50 | |
| *** cheneydc has quit IRC | 05:57 | |
| *** cheneydc has joined #openstack-telemetry | 05:57 | |
| *** ljxiash has quit IRC | 06:00 | |
| *** achatterjee has joined #openstack-telemetry | 06:04 | |
| *** ljxiash has joined #openstack-telemetry | 06:05 | |
| *** liamji has quit IRC | 06:07 | |
| *** liamji has joined #openstack-telemetry | 06:07 | |
| *** Liuqing has quit IRC | 06:22 | |
| *** belmoreira has joined #openstack-telemetry | 06:23 | |
| *** Liuqing has joined #openstack-telemetry | 06:24 | |
| *** dikonoor has quit IRC | 06:27 | |
| *** cheneydc has quit IRC | 06:34 | |
| *** ljxiash has quit IRC | 06:47 | |
| *** thorst has joined #openstack-telemetry | 06:49 | |
| *** ljxiash has joined #openstack-telemetry | 06:51 | |
| *** liamji has quit IRC | 06:52 | |
| *** liamji has joined #openstack-telemetry | 06:52 | |
| *** thorst has quit IRC | 06:56 | |
| *** _nadya_ has joined #openstack-telemetry | 07:09 | |
| *** chlong_ has quit IRC | 07:41 | |
| *** pcaruana has joined #openstack-telemetry | 07:45 | |
| *** safchain has joined #openstack-telemetry | 07:45 | |
| *** _nadya_ has quit IRC | 07:45 | |
| openstackgerrit | Merged openstack/ceilometer: Gnocchi: fix ResourcesDefinitionException for py3 https://review.openstack.org/281190 | 07:48 |
| *** Liuqing has quit IRC | 07:53 | |
| *** thorst has joined #openstack-telemetry | 07:54 | |
| *** thorst has quit IRC | 08:01 | |
| *** boris-42 has joined #openstack-telemetry | 08:04 | |
| *** shardy has joined #openstack-telemetry | 08:10 | |
| *** ildikov has quit IRC | 08:35 | |
| *** lsmola has joined #openstack-telemetry | 08:43 | |
| *** annasort_ has joined #openstack-telemetry | 08:45 | |
| openstackgerrit | Julien Danjou proposed openstack/gnocchi: storage: make sure we delete old measures respecting archive policy https://review.openstack.org/279146 | 08:48 |
| *** annasort has quit IRC | 08:48 | |
| *** annasort_ is now known as annasort | 08:48 | |
| jd__ | sileht: https://review.openstack.org/#/c/279146/ to review and we're ready to go :) | 08:54 |
| * sileht is on it | 08:57 | |
| openstackgerrit | Xia Linjuan proposed openstack/telemetry-specs: Collect neutron security group meters for instance ports https://review.openstack.org/273475 | 08:58 |
| *** thorst has joined #openstack-telemetry | 08:58 | |
| *** thorst has quit IRC | 09:06 | |
| *** _nadya_ has joined #openstack-telemetry | 09:25 | |
| *** yassine has joined #openstack-telemetry | 09:29 | |
| *** efoley has joined #openstack-telemetry | 09:30 | |
| *** efoley_ has joined #openstack-telemetry | 09:30 | |
| *** efoley__ has joined #openstack-telemetry | 09:31 | |
| *** efoley_ has quit IRC | 09:33 | |
| *** efoley has quit IRC | 09:34 | |
| *** efoley_ has joined #openstack-telemetry | 09:34 | |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ceilometer: Fix SDR file parsing for Intel Node Manager https://review.openstack.org/282248 | 09:35 |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ceilometer: Fix SDR file parsing for Intel Node Manager https://review.openstack.org/282248 | 09:40 |
| openstackgerrit | Julien Danjou proposed openstack/gnocchi: storage: make sure we delete old measures respecting archive policy https://review.openstack.org/279146 | 09:41 |
| *** marcin12345 has quit IRC | 09:49 | |
| *** ildikov has joined #openstack-telemetry | 10:01 | |
| *** thorst has joined #openstack-telemetry | 10:04 | |
| *** cdent has joined #openstack-telemetry | 10:11 | |
| *** thorst has quit IRC | 10:12 | |
| *** Liuqing has joined #openstack-telemetry | 10:12 | |
| *** Liuqing has quit IRC | 10:14 | |
| openstackgerrit | Igor Degtiarov proposed openstack/ceilometer: [WIP][POC] bracketer event transformer https://review.openstack.org/266488 | 10:16 |
| *** idegtiarov_ has joined #openstack-telemetry | 10:20 | |
| *** shardy has quit IRC | 10:34 | |
| *** ljxiash has quit IRC | 10:36 | |
| *** liamji has quit IRC | 10:36 | |
| *** ljxiash has joined #openstack-telemetry | 10:37 | |
| *** ljxiash_ has joined #openstack-telemetry | 10:38 | |
| *** cdent has quit IRC | 10:41 | |
| *** ljxiash has quit IRC | 10:41 | |
| *** ljxiash_ has quit IRC | 10:42 | |
| *** ildikov has quit IRC | 10:43 | |
| openstackgerrit | Merged openstack/gnocchi: storage: make sure we delete old measures respecting archive policy https://review.openstack.org/279146 | 10:50 |
| openstackgerrit | zhaozhilong proposed openstack/ceilometer: Alphabetized setup.cfg for cleanliness https://review.openstack.org/282274 | 10:52 |
| openstackgerrit | zhaozhilong proposed openstack/aodh: Alphabetized setup.cfg for cleanliness https://review.openstack.org/282282 | 11:08 |
| *** thorst has joined #openstack-telemetry | 11:09 | |
| *** thorst has quit IRC | 11:16 | |
| *** cdent has joined #openstack-telemetry | 11:25 | |
| *** linkmark has joined #openstack-telemetry | 11:27 | |
| *** shardy has joined #openstack-telemetry | 11:31 | |
| *** efoley_ has quit IRC | 11:32 | |
| *** efoley__ has quit IRC | 11:32 | |
| *** zmasek has joined #openstack-telemetry | 11:34 | |
| *** efoley__ has joined #openstack-telemetry | 11:36 | |
| *** efoley_ has joined #openstack-telemetry | 11:36 | |
| *** dave-mccowan has joined #openstack-telemetry | 11:49 | |
| *** chlong_ has joined #openstack-telemetry | 12:11 | |
| *** thorst has joined #openstack-telemetry | 12:13 | |
| *** cheneydc has joined #openstack-telemetry | 12:17 | |
| *** thorst has quit IRC | 12:21 | |
| *** thorst has joined #openstack-telemetry | 12:23 | |
| *** cheneydc has quit IRC | 12:26 | |
| *** cheneydc has joined #openstack-telemetry | 12:27 | |
| *** shardy has quit IRC | 12:29 | |
| *** efoley__ has quit IRC | 12:30 | |
| *** efoley_ has quit IRC | 12:30 | |
| *** toddnni has quit IRC | 12:31 | |
| *** r-mibu has quit IRC | 12:32 | |
| *** r-mibu has joined #openstack-telemetry | 12:32 | |
| *** ityaptin_ has quit IRC | 12:33 | |
| *** gordc has joined #openstack-telemetry | 12:35 | |
| *** ljxiash has joined #openstack-telemetry | 12:39 | |
| *** cheneydc has quit IRC | 12:39 | |
| *** ljxiash has quit IRC | 12:43 | |
| *** efoley_ has joined #openstack-telemetry | 12:45 | |
| *** efoley__ has joined #openstack-telemetry | 12:45 | |
| idegtiarov_ | gordc, Hi! | 13:04 |
| gordc | idegtiarov_: o/ | 13:04 |
| gordc | whatsup? | 13:04 |
| idegtiarov_ | gordc, we already have POC for spec https://review.openstack.org/#/c/162167/ | 13:04 |
| *** krotscheck_afk is now known as krotscheck | 13:05 | |
| idegtiarov_ | gordc, I'm going finish with it in two three days | 13:06 |
| gordc | idegtiarov_: ok. to be honest, i'm not confident on merging this but we'll see how it goes | 13:06 |
| idegtiarov_ | gordc, I mean go from POC to implemented feature | 13:06 |
| gordc | idegtiarov_: if you want it merged, you should really have it done now... next week the gate will be garbage... the week after is release week | 13:06 |
| gordc | idegtiarov_: i guess it doesn't hurt if you keep working on it. i just can't really promise anything for mitaka. | 13:07 |
| *** ityaptin has joined #openstack-telemetry | 13:08 | |
| *** zmasek has quit IRC | 13:10 | |
| idegtiarov_ | gordc, tbh our colleagues from Intel appear to be interested in that feature | 13:10 |
| gordc | that's fair. i was interested as well. but we are ~1 week away from release and there's nothing to look at yet. | 13:12 |
| gordc | it's very difficult to be confident that this will merge. | 13:12 |
| gordc | idegtiarov_: let's continue working on it and see how it plays out. | 13:13 |
| gordc | we cant possibly consider adding it as a FFE. | 13:13 |
| idegtiarov_ | gordc, ok | 13:13 |
| gordc | s/cant/can/ | 13:13 |
| idegtiarov_ | gordc, thanks | 13:13 |
| openstackgerrit | Merged openstack/python-aodhclient: Return the error message of API request failure https://review.openstack.org/280519 | 13:15 |
| *** davidsha has joined #openstack-telemetry | 13:31 | |
| *** davidsha has left #openstack-telemetry | 13:32 | |
| *** nicodemus_ has joined #openstack-telemetry | 13:34 | |
| gordc | f... why does my virtualbox vm keep hanging. | 13:36 |
| * cdent hands gordc a dead chicken | 13:42 | |
| openstackgerrit | Chaozhe Chen(ccz) proposed openstack/gnocchi: Use '#flake8: noqa' to skip file check https://review.openstack.org/282340 | 13:43 |
| gordc | cdent: you read that story about SF tech bro? | 13:45 |
| cdent | the one who doesn't like the poor folk laying on his streets? | 13:46 |
| cdent | or some other douchebro? | 13:46 |
| gordc | cdent: that dude. | 13:47 |
| gordc | classic SF. | 13:47 |
| *** shardy has joined #openstack-telemetry | 13:47 | |
| gordc | oh darn. his company website is back up. | 13:49 |
| gordc | i don't get it. this dude is selling me an ssh client? | 13:50 |
| jd__ | sileht: looks like Gnocchi is finally out of integrated queue in Zuul | 13:55 |
| sileht | jd__, yeah thanks to gordc | 13:55 |
| jd__ | thanks gordc | 13:55 |
| sileht | jd__, he doesn't known that one of its patch was resolving the issue :) | 13:56 |
| jd__ | he is a self-ignoring genius | 13:56 |
| gordc | lol what did i do? | 13:57 |
| cdent | I've often thought that gordc is some kind of idiot savant | 13:57 |
| cdent | because his usual response to everything is "lol what did I do?" | 13:57 |
| gordc | oh gate stuff? yeah, it's all guessing. everyone is guessing. | 13:57 |
| sileht | that was this one: https://review.openstack.org/#/c/271379/ | 13:58 |
| sileht | I proposed the same code before but with another commit msg and mine got -2 | 13:59 |
| gordc | sileht: yeah, i thought so. not my problem, a birdie told me to do it like that. | 13:59 |
| gordc | sileht: i'm not surprised. i told you. always put 'fix' in title. | 14:00 |
| gordc | no one reads the rest. | 14:00 |
| * sileht should follow gordc advice more often | 14:00 | |
| sileht | lol | 14:00 |
| jd__ | rofl | 14:00 |
| *** diogogmt has quit IRC | 14:01 | |
| *** diogogmt has joined #openstack-telemetry | 14:03 | |
| *** toddnni has joined #openstack-telemetry | 14:04 | |
| *** Ephur has joined #openstack-telemetry | 14:11 | |
| openstackgerrit | Merged openstack/gnocchi: Allows to use cradox with ceph storage https://review.openstack.org/279448 | 14:18 |
| *** peristeri has joined #openstack-telemetry | 14:27 | |
| *** links has quit IRC | 14:27 | |
| openstackgerrit | Claudiu Belu proposed openstack/ceilometer: Hyper-V: replaces in-tree hyper-v utils usage with os_win https://review.openstack.org/280454 | 14:28 |
| *** cdent has quit IRC | 14:33 | |
| *** ddieterly has joined #openstack-telemetry | 14:34 | |
| *** ddieterly has quit IRC | 14:37 | |
| *** diogogmt has quit IRC | 14:37 | |
| *** ddieterly has joined #openstack-telemetry | 14:37 | |
| *** KrishR has joined #openstack-telemetry | 14:38 | |
| *** ddieterly has quit IRC | 14:38 | |
| *** ddieterly has joined #openstack-telemetry | 14:38 | |
| openstackgerrit | Merged openstack/gnocchi: Use '#flake8: noqa' to skip file check https://review.openstack.org/282340 | 14:39 |
| *** gordc has quit IRC | 14:42 | |
| *** openstackgerrit_ has quit IRC | 14:43 | |
| *** gordc has joined #openstack-telemetry | 14:43 | |
| *** jwcroppe has joined #openstack-telemetry | 14:47 | |
| *** diogogmt has joined #openstack-telemetry | 14:49 | |
| *** ddieterl_ has joined #openstack-telemetry | 14:56 | |
| *** ddieterly has quit IRC | 14:58 | |
| *** efoley_ has quit IRC | 14:59 | |
| *** efoley__ is now known as efoley | 14:59 | |
| ityaptin | gordc: Hi! | 15:00 |
| gordc | ityaptin: hey. whatsup? | 15:01 |
| *** openstackgerrit_ has joined #openstack-telemetry | 15:01 | |
| *** linkmark has quit IRC | 15:02 | |
| ityaptin | Have a question about your comment about nova polling patch. I think realtime discussion will be more productive than comments. | 15:03 |
| gordc | ityaptin: sure | 15:03 |
| gordc | so my main concern is right now, it looks like each agent is still calling nova-api. | 15:04 |
| ityaptin | How I understood you came to conclusion that we have a same calls of nova-api with global cache and without it, didn't you? | 15:05 |
| ityaptin | Yes, every agent can call nova-api, when cache is expired | 15:06 |
| ityaptin | Did you see this scheme: https://docs.google.com/drawings/d/1y_ELw6zDdnI1fKxfSOdRv0lPk47tMnexV7Ea4LDLe0M/edit | 15:06 |
| ityaptin | I add a timeline for the operations in the new workflow | 15:07 |
| gordc | ityaptin: nice drawing. | 15:09 |
| gordc | that's how i imagine it working | 15:09 |
| gordc | but if you look at https://review.openstack.org/#/c/267078/18/ceilometer/compute/discovery/instance.py | 15:09 |
| gordc | line 87 | 15:10 |
| gordc | that is called on every single agent | 15:10 |
| gordc | also, your cache expiry is unrelated to polling interval. | 15:11 |
| *** dave-mccowan has quit IRC | 15:11 | |
| gordc | although i guess that last part might not be a concern as you are always polling in current design. | 15:12 |
| *** links has joined #openstack-telemetry | 15:13 | |
| ityaptin | About > that is called on every single agent. we can't acquire a lock at L85 if resource_update_lock is not expired or is not deleted. So, if it exists and fresh self.cache.lock.acquire() returns a False and we go to the collecting vms of hypervisor | 15:14 |
| ityaptin | Yes, how I think we have a two big wins with this patch. 1 - from the spec: We replace many little requests to the nova with 1 big request from the 1 compute agent per cache expiring interval. | 15:16 |
| *** dhellmann has quit IRC | 15:16 | |
| gordc | ityaptin: ah i see. | 15:17 |
| gordc | ityaptin: your cache check is in the lock. | 15:17 |
| gordc | i'd make it more obvious that we're checking cache instead of hiding it in lock. | 15:18 |
| *** dhellmann has joined #openstack-telemetry | 15:19 | |
| gordc | that said, that still leaves us with metadata that is potentially unrelated to measurement no? | 15:19 |
| ityaptin | 2. We can reduce polling interval for the instance measurements for the collecting actual info cpu, mem usages. It allows to make a better autoscaling and etc, because currently we have a tool for fast data getting but can't collect data often | 15:19 |
| jd__ | Gnocchi 2.0.0 tagged and out | 15:19 |
| gordc | ityaptin: i under 2 is the purpose of design. | 15:19 |
| gordc | s/under/understand/ | 15:19 |
| gordc | jd__: nice. | 15:19 |
| ityaptin | jd__: Good news:) | 15:20 |
| sileht | jd__, \o/ | 15:20 |
| jd__ | sileht: feel free to release a gnocchiclient next, using a version number that makes sense (not sure if we broke any compat) | 15:21 |
| sileht | jd__, I think we didn't | 15:21 |
| gordc | so right now agent 1 will poll, get all instance metadata, continue. it will continue to reuse cache for next 10min (depending on expiry). the metadata 10min ago may or may not match the instance you are polling now. | 15:22 |
| ityaptin | gordc: For hitting this points I will to make a invalidation resource_lock if we catch a update notification. | 15:23 |
| gordc | ityaptin: i think the point was we don't want the polling agents to have to understand notifications | 15:24 |
| ityaptin | gordc: Notification agent catch a notification and if we use global_cache it invalidate a resource_update_lock key. So, in the next polling interval we update the cache with a new nova instances | 15:25 |
| *** dave-mccowan has joined #openstack-telemetry | 15:25 | |
| ityaptin | gordc: polling agent will not understand notifications | 15:25 |
| gordc | right but the entire caching mechanism now relies on both polling and notification agent | 15:26 |
| ityaptin | Yes, if user want fast polling. If not - he can use a default configuration or update cache more often than 10 minutes. | 15:27 |
| *** ljxiash_ has joined #openstack-telemetry | 15:27 | |
| ityaptin | gordc: About cache checking under the lock - we can't expire cache for the every instance, because we make a change-since request to the nova-api and it returns not all instances but only changes. So, in the normal workflow we only updates cache. | 15:31 |
| gordc | i see | 15:32 |
| *** ljxiash has joined #openstack-telemetry | 15:33 | |
| *** ljxiash_ has quit IRC | 15:33 | |
| *** cdent has joined #openstack-telemetry | 15:36 | |
| gordc | ityaptin: i'll take a look again. i guess it's less of an issue now that you've pointed out the lock acquiring also checks cache status | 15:36 |
| gordc | ityaptin: i'd prefer it be more obvious but just a preference for now. | 15:36 |
| *** yprokule has quit IRC | 15:38 | |
| ityaptin | gordc: I'll try to make it more clean in the next patchset with your, cdent and liusheng notes. It will be tomorrow I think. I will glad if you review it on Monday. | 15:38 |
| gordc | sure | 15:39 |
| gordc | thanks for explanation | 15:39 |
| ityaptin | no problems :) | 15:39 |
| *** liamji has joined #openstack-telemetry | 15:40 | |
| gordc | ityaptin: one more question. what happens if we have cache and an instance is created? | 15:41 |
| ityaptin | Currently we will get previous number of the instances during the cache is not expired. With a notification using for the cache updating/invalidation - we will update cache after compute.instance.create notification. | 15:42 |
| gordc | :( yeah, that's what i'm not liking. personally i'd rather just avoid notifications and poll as much as defined. if you specify poll every 5s, it should not surprise anyone if your api is hit every 5s | 15:45 |
| gordc | if we depend on notifications, we not have an unwanted link between the two agents, we need to deal with guaranteed delivery, and we have to deal with timing of notification processing and polling intervals | 15:46 |
| gordc | s/not/now/ | 15:46 |
| ityaptin | gordc: Maybe you're right. It's a huge risks. Alternative is to keep list of orphan (created not by nova) hypervisor vms and check new vms which is not in instance cache and not in orphan_vms list with nova-api. | 15:50 |
| ityaptin | I did not researched it yet. | 15:50 |
| gordc | ityaptin: yeah. definitely something to consider. i need to think a bit more too. the notification agent requirement and potential timing sync issues worries. me | 15:52 |
| gordc | biab. | 15:52 |
| *** annasort has quit IRC | 15:53 | |
| *** annasort has joined #openstack-telemetry | 15:53 | |
| ityaptin | gordc: In the context of notification agent and cache - of course we will n't update the cache from notification if notification is older than last cache updating. We can check it, if run get(lock_key, now() - notification_timestamp). | 15:56 |
| *** Liuqing has joined #openstack-telemetry | 15:57 | |
| ityaptin | gordc: Yes, we have a chance what time at the node the notification agent and node with compute agent is not synced? But i thought it is a rule to sync time between nodes on the environment | 15:59 |
| ityaptin | I can finish main functional asap and start researches way to catch every instance update after main functional will be merged or will have only little shortcomings. | 16:04 |
| ityaptin | gordc: How do you think, is it suitable option? | 16:06 |
| *** pcaruana has quit IRC | 16:10 | |
| *** g3ek has quit IRC | 16:14 | |
| *** g3ek has joined #openstack-telemetry | 16:15 | |
| *** liamji has quit IRC | 16:15 | |
| *** ljxiash has quit IRC | 16:17 | |
| gordc | ityaptin: well it's not a matter of syncing system times. once you start relying on queue. you have no guarantee when stuff is processed. it's async | 16:18 |
| *** ljxiash has joined #openstack-telemetry | 16:19 | |
| *** efoley has quit IRC | 16:19 | |
| ityaptin | gordc: Agreed. With this approach we can't update instance cache after instance updated for sure, but it looks like a way to reduce nova load and support fresh instance metadata nevertheless. | 16:26 |
| ityaptin | gordc: Maybe list of orphan instances + updates after new orphan instance found + notification hooks the cache invalidating may be useful in compound? | 16:27 |
| gordc | ityaptin: tbh, this seems to be more and more complex. | 16:29 |
| gordc | i'd much rather we just poll as frequently as defined in pipeline (but minimise it to at most one agent polling it) OR | 16:31 |
| gordc | we just ignore any instance we don't have metadata for. | 16:31 |
| ityaptin | gordc: > "we just ignore any instance we don't have metadata for". How it works now. | 16:31 |
| gordc | ityaptin: i see. so the cache is really less of a cache and more instance discoverer. | 16:34 |
| ityaptin | we can set interval_to_update_resource option equal to the pipeline interval. with it we can make 1 nova request instead of N and have fresh metadata for the every polling. But for the thousands compute nodes it will be hard for the little polling interval (1 minute) for example | 16:34 |
| gordc | i don't mind taht compared to notification appropriate... which should make it explicit though | 16:34 |
| _nadya_ | gordc: we may not update the cache using notification agent, I think it may be ok | 16:35 |
| *** belmoreira has quit IRC | 16:35 | |
| gordc | well in theory, if you have 1000 hosts, the first agent would populate cache for all. the rest of agents would see cache is valid for their poll interval (ie. cache is less than 1minute old) and reuse it | 16:36 |
| ityaptin | Yep | 16:36 |
| gordc | only once cache becomes less than 1 minute old, does another agent poll | 16:36 |
| gordc | it basically makes interval_to_update_resource == smallest poll interval. | 16:37 |
| gordc | i'm ok with just leaving it as is. but let's make interval_to_update_resource more obvious that it's really instance discoverer interval. | 16:38 |
| ityaptin | gordc: Ok | 16:38 |
| *** liamji has joined #openstack-telemetry | 16:38 | |
| gordc | ityaptin: thanks. | 16:39 |
| gordc | let's play around with it more. but i'm less worried now. | 16:40 |
| *** ddieterl_ is now known as ddieterly | 16:50 | |
| *** links has quit IRC | 16:53 | |
| *** rcernin has quit IRC | 16:58 | |
| openstackgerrit | Xia Linjuan proposed openstack/ceilometer: Enable the Load Balancer v2 for notification https://review.openstack.org/281777 | 17:15 |
| openstackgerrit | Xia Linjuan proposed openstack/ceilometer: Enable the Load Balancer v2 for notification https://review.openstack.org/281777 | 17:25 |
| openstackgerrit | Julien Danjou proposed openstack/python-gnocchiclient: Use overtest to run Gnocchi https://review.openstack.org/281955 | 17:27 |
| *** rcernin has joined #openstack-telemetry | 17:30 | |
| *** Ephur has quit IRC | 17:35 | |
| *** ljxiash has quit IRC | 17:35 | |
| openstackgerrit | Zi Lian Ji proposed openstack/ceilometer: Enable the Load Balancer v2 for the Ceilometer(Part One) https://review.openstack.org/274518 | 17:38 |
| *** ljxiash has joined #openstack-telemetry | 17:42 | |
| liamji | gordc: hi | 17:43 |
| *** _nadya_ has quit IRC | 17:52 | |
| gordc | liamji: o/ | 17:54 |
| liamji | Gordon： : ) I have got +1 from Michael Johnson(https://launchpad.net/~johnsom) who is the LBaaS expert in the neutron team and he helps to check whether I use the correct neutron client method. If you have time, could you help to review the code https://review.openstack.org/#/c/274518/13 | 17:55 |
| gordc | liamji: ok. | 17:55 |
| liamji | gordc: thanks for your help | 17:55 |
| gordc | will do it after lunch (and after i fix some gnocchi issue) | 17:55 |
| *** annasort has quit IRC | 17:56 | |
| liamji | gordc: sure. hava a nice day. | 17:56 |
| *** annasort has joined #openstack-telemetry | 17:56 | |
| *** liamji has quit IRC | 18:00 | |
| *** ljxiash has quit IRC | 18:01 | |
| *** _nadya_ has joined #openstack-telemetry | 18:10 | |
| *** Liuqing has quit IRC | 18:12 | |
| *** chlong_ has quit IRC | 18:18 | |
| *** datravis has quit IRC | 18:22 | |
| *** ddieterly has quit IRC | 18:24 | |
| *** ddieterly has joined #openstack-telemetry | 18:24 | |
| *** KrishR has quit IRC | 18:25 | |
| *** eglynn has quit IRC | 18:25 | |
| *** tonyb has quit IRC | 18:25 | |
| *** krotscheck has quit IRC | 18:25 | |
| *** pasquier-s has quit IRC | 18:25 | |
| *** fc__ has quit IRC | 18:25 | |
| *** fc__ has joined #openstack-telemetry | 18:25 | |
| *** krotscheck has joined #openstack-telemetry | 18:25 | |
| *** eglynn has joined #openstack-telemetry | 18:26 | |
| *** pasquier-s has joined #openstack-telemetry | 18:26 | |
| *** gordc has quit IRC | 18:27 | |
| *** raginbajin has quit IRC | 18:27 | |
| *** hparekh has quit IRC | 18:27 | |
| *** raginbaj- has joined #openstack-telemetry | 18:27 | |
| *** hparekh has joined #openstack-telemetry | 18:27 | |
| *** david-lyle has joined #openstack-telemetry | 18:27 | |
| *** raginbaj- is now known as raginbajin | 18:29 | |
| openstackgerrit | gordon chung proposed openstack/gnocchi: fix partitioning https://review.openstack.org/282453 | 18:29 |
| *** gordc has joined #openstack-telemetry | 18:29 | |
| *** tonyb has joined #openstack-telemetry | 18:31 | |
| *** KrishR has joined #openstack-telemetry | 18:51 | |
| *** lsmola has quit IRC | 18:56 | |
| *** _nadya_ has quit IRC | 18:56 | |
| *** rcernin has quit IRC | 18:56 | |
| openstackgerrit | gordon chung proposed openstack/ceilometer: improve notification processing https://review.openstack.org/275741 | 19:02 |
| openstackgerrit | Kairo Araujo proposed openstack/ceilometer: Initial seed of hacking https://review.openstack.org/282124 | 19:02 |
| *** vkmc has joined #openstack-telemetry | 19:03 | |
| gordc | nicodemus_: i messed up partitioning again. very good chance you'll see slow/under-utilised metricd workers | 19:05 |
| nicodemus_ | gordc: patch set 7 is not good? | 19:07 |
| gordc | nicodemus_: no. it won't actually dynamically configure blocks... | 19:08 |
| gordc | https://review.openstack.org/#/c/282453/ | 19:08 |
| nicodemus_ | gordc: one thing I did see, requirements.txt from master should contain "lz4" | 19:08 |
| *** datravis has joined #openstack-telemetry | 19:09 | |
| gordc | i did a bad rebase so it always sets the block_size to lowest value. | 19:09 |
| gordc | nicodemus_: it's actually listed here: https://github.com/openstack/gnocchi/blob/master/setup.cfg#L45 | 19:10 |
| nicodemus_ | gordc: so far I didn't see metricd to fall behind, but then again I'm using a new empty ceph pool. I'll apply the latest patch | 19:10 |
| *** KrishR has quit IRC | 19:11 | |
| gordc | do you have a lot of workers? | 19:11 |
| nicodemus_ | gordc: oh... ok (we add some stuff to setup.cfg, I didn't add that new line) | 19:11 |
| nicodemus_ | gordc: 16 workers | 19:12 |
| gordc | yeah. so right now it'll process at most 16*16 metrics at a time. | 19:12 |
| gordc | it really should be processing 16*(16 to 256) metrics at a time. | 19:13 |
| gordc | you probably won't notice much if have less metrics. | 19:13 |
| gordc | or if you have a larger number of workers like you do. | 19:14 |
| *** ildikov has joined #openstack-telemetry | 19:18 | |
| *** rcernin has joined #openstack-telemetry | 19:23 | |
| *** diogogmt has quit IRC | 19:24 | |
| *** KrishR has joined #openstack-telemetry | 19:25 | |
| *** diogogmt has joined #openstack-telemetry | 19:26 | |
| *** ddieterly has quit IRC | 19:28 | |
| *** ddieterly has joined #openstack-telemetry | 19:33 | |
| *** ddieterl_ has joined #openstack-telemetry | 19:34 | |
| *** ildikov has quit IRC | 19:35 | |
| *** ddieterly has quit IRC | 19:36 | |
| *** diogogmt has quit IRC | 19:40 | |
| *** rcernin has quit IRC | 19:45 | |
| *** diogogmt has joined #openstack-telemetry | 19:47 | |
| *** ddieterl_ has quit IRC | 19:48 | |
| *** ddieterly has joined #openstack-telemetry | 19:48 | |
| *** ddieterl_ has joined #openstack-telemetry | 19:51 | |
| *** vkmc has left #openstack-telemetry | 19:53 | |
| *** ddieterly has quit IRC | 19:53 | |
| *** kairo has quit IRC | 19:58 | |
| *** davidlenwell has quit IRC | 20:01 | |
| *** gordc has quit IRC | 20:02 | |
| *** davidlenwell has joined #openstack-telemetry | 20:06 | |
| *** ddieterl_ has quit IRC | 20:07 | |
| *** ddieterly has joined #openstack-telemetry | 20:07 | |
| *** ddieterl_ has joined #openstack-telemetry | 20:08 | |
| nicodemus_ | hello sileht, do you have a moment? | 20:16 |
| *** yassine has quit IRC | 20:21 | |
| *** eglynn has quit IRC | 20:26 | |
| openstackgerrit | Claudiu Belu proposed openstack/ceilometer: Hyper-V: replaces in-tree hyper-v utils usage with os_win https://review.openstack.org/280454 | 20:38 |
| *** thorst has quit IRC | 20:54 | |
| *** thorst has joined #openstack-telemetry | 20:54 | |
| *** vishwanathj has joined #openstack-telemetry | 20:55 | |
| *** thorst has quit IRC | 20:59 | |
| *** vishwana_ has joined #openstack-telemetry | 21:00 | |
| *** vishwanathj has quit IRC | 21:03 | |
| *** ddieterl_ has quit IRC | 21:05 | |
| *** ddieterly has quit IRC | 21:05 | |
| *** idegtiarov has quit IRC | 21:15 | |
| *** peristeri has quit IRC | 21:20 | |
| *** Ephur has joined #openstack-telemetry | 21:23 | |
| *** ddieterly has joined #openstack-telemetry | 21:25 | |
| *** ddieterl_ has joined #openstack-telemetry | 21:25 | |
| *** shardy has quit IRC | 21:31 | |
| *** vishwana_ has quit IRC | 21:35 | |
| *** vishwanathj has joined #openstack-telemetry | 21:41 | |
| *** dave-mccowan has quit IRC | 21:44 | |
| *** kairo has joined #openstack-telemetry | 21:52 | |
| *** nicodemus_ has quit IRC | 21:52 | |
| openstackgerrit | Kairo Araujo proposed openstack/ceilometer: Initial seed of hacking https://review.openstack.org/282124 | 21:56 |
| *** dave-mccowan has joined #openstack-telemetry | 21:57 | |
| *** safchain has quit IRC | 22:00 | |
| openstackgerrit | Kairo Araujo proposed openstack/ceilometer: Initial seed of hacking https://review.openstack.org/282124 | 22:09 |
| kairo | =/ | 22:09 |
| *** annasort has quit IRC | 22:10 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ceilometer: Updated from global requirements https://review.openstack.org/282550 | 22:12 |
| *** changbl has quit IRC | 22:12 | |
| *** changbl has joined #openstack-telemetry | 22:40 | |
| *** annasort has joined #openstack-telemetry | 22:41 | |
| *** annasort has quit IRC | 22:45 | |
| *** david-lyle has quit IRC | 22:50 | |
| *** david-lyle has joined #openstack-telemetry | 22:51 | |
| *** david-lyle has quit IRC | 22:54 | |
| *** david-lyle has joined #openstack-telemetry | 22:54 | |
| *** ddieterl_ has quit IRC | 23:14 | |
| *** ddieterly has quit IRC | 23:14 | |
| *** cdent has quit IRC | 23:20 | |
| *** david-lyle_ has joined #openstack-telemetry | 23:29 | |
| *** david-lyle has quit IRC | 23:30 | |
| *** pradk has quit IRC | 23:37 | |
| *** KrishR has quit IRC | 23:38 | |
| *** jwcroppe has quit IRC | 23:53 | |
| *** jwcroppe has joined #openstack-telemetry | 23:54 | |
| *** jwcroppe has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!