| *** Drago has quit IRC | 00:54 | |
| *** zhurong has joined #senlin | 01:05 | |
| *** yanyanhu has joined #senlin | 01:24 | |
| *** guoshan has joined #senlin | 01:26 | |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Define new config options for receiver notifying https://review.openstack.org/373861 | 01:52 |
|---|---|---|
| Qiming | yanyanhu, fyi, I'm attempting to fix gate problem by bumping sdk version | 01:54 |
| yanyanhu | Qiming, great | 01:54 |
| Qiming | https://review.openstack.org/374469 | 01:55 |
| yanyanhu | so there will be sdk0.9.6 | 01:55 |
| Qiming | basically, 0.9.6 was bad | 01:55 |
| yanyanhu | oh, I see | 01:55 |
| Qiming | that was causing the gate breaking | 01:55 |
| Qiming | brian has kindly cut a 0.9.7 last night | 01:55 |
| yanyanhu | nice | 01:55 |
| Qiming | but he didn't buy in our idea to change 'generated_at' back to 'timestamp' | 01:56 |
| Qiming | so I reverted that patch to make hime happy | 01:56 |
| Qiming | http://git.openstack.org/cgit/openstack/python-openstacksdk/commit/?id=95ba61263ed5f3ef02025913604cf94c33de0c60 | 01:56 |
| Qiming | actually, I like this patch from miaohb: http://git.openstack.org/cgit/openstack/python-openstacksdk/commit/?id=619032ea1876078f22a1d8dec9509db4a9769f63 | 01:56 |
| Qiming | I tried to argue with him | 01:57 |
| yanyanhu | ok. imo, timestamp is better for 'event' which is kind of timeserial record | 01:57 |
| Qiming | yes, https://review.openstack.org/#/c/372276/ | 01:58 |
| Qiming | but as a compromise, I have proposed a workaround at client side | 01:58 |
| Qiming | https://review.openstack.org/374126 | 01:59 |
| yanyanhu | I see. that can be an alternate. | 02:01 |
| yanyanhu | hi, Qiming, about the event, I'm thinking maybe we can consider to store them into non-sql db, e.g. mongo in future. | 02:04 |
| Qiming | em, yes | 02:05 |
| yanyanhu | since it will grow very fast and could be used as history by user | 02:05 |
| Qiming | that was something we have on newton agenda but wasn't able to deliver | 02:05 |
| openstackgerrit | Tony Breeds proposed openstack/senlin: [DNM] Gate test https://review.openstack.org/374487 | 02:05 |
| yanyanhu | I see | 02:05 |
| Qiming | event can have some backends, db, file, messaging etc | 02:06 |
| yanyanhu | right | 02:06 |
| Qiming | we can pipe them to something remote | 02:06 |
| yanyanhu | yes, that will be the complete solution | 02:06 |
| Qiming | before doing that, I was trying to make versioning work | 02:07 |
| yanyanhu | hope can have time to work on it in next cycle | 02:07 |
| Qiming | then we don't lose control on the format | 02:07 |
| yanyanhu | yep | 02:07 |
| yanyanhu | format compatible is important | 02:07 |
| Qiming | before we have a stable definition of event format, I'm not feeling comfortable to dump events to any external sinks | 02:08 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Define new config options for receiver notifying https://review.openstack.org/373861 | 02:09 |
| yanyanhu | I see, it makes sense | 02:09 |
| Qiming | basically a notification contains the following fields: | 02:10 |
| Qiming | notification fields: | 02:10 |
| Qiming | timestamp: timeutils.utcnow(), with microseconds | 02:10 |
| Qiming | priority: string, e.g. INFO, WARN, ... | 02:10 |
| Qiming | publisher_id: string, e.g. clustering.engine | 02:10 |
| Qiming | event_type: string, e.g. cluster.create.started | 02:10 |
| Qiming | payload: dict, should be from a versioned object | 02:10 |
| Qiming | need a protocol to convert/converge all events and notifications | 02:10 |
| yanyanhu | yes | 02:13 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Define new config options for receiver notifying https://review.openstack.org/373861 | 02:13 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Define new config options for receiver notification https://review.openstack.org/373861 | 02:30 |
| openstackgerrit | RUIJIE YUAN proposed openstack/senlin: Not necessary to handle 'ResourceNotFound' https://review.openstack.org/374505 | 02:40 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Define new config options for receiver notification https://review.openstack.org/373861 | 02:55 |
| openstackgerrit | RUIJIE YUAN proposed openstack/senlin: Lighten the load https://review.openstack.org/374516 | 03:09 |
| *** guoshan has quit IRC | 03:52 | |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add config options for Zaqar queue https://review.openstack.org/374531 | 05:13 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add config options for Zaqar queue https://review.openstack.org/374531 | 05:14 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add config options for Zaqar queue https://review.openstack.org/374531 | 05:44 |
| *** guoshan has joined #senlin | 06:00 | |
| *** AlexeyAbashkin has quit IRC | 06:44 | |
| *** AlexeyAbashkin has joined #senlin | 06:46 | |
| *** lvdongbing has joined #senlin | 06:50 | |
| Qiming | hi, guoshan | 06:52 |
| Qiming | about the batching policy | 06:53 |
| guoshan | yes? | 06:53 |
| Qiming | it is not about policy.json | 06:54 |
| Qiming | it is about a usage scenario | 06:54 |
| Qiming | say you want to create a cluster of 1000 VMs | 06:54 |
| Qiming | you don't want to have senlin issue 1000 requests to nova simultaneously | 06:54 |
| Qiming | you will want 10 batches, for example, each of which contain 100 requests | 06:55 |
| Qiming | this applies to other cluster operations as well | 06:55 |
| Qiming | the idea is to avoid DOS attack on other services | 06:56 |
| Qiming | another aspect of batching policy is about update | 06:56 |
| Qiming | when you are updating a cluster, you may want to keep at least a certain number of nodes active at any time during the update | 06:56 |
| guoshan | oh, batches means lots of nodes operation at the same time？ | 06:57 |
| Qiming | when updating a cluster of 100 VMs, for example, you may want to have at least 80 VMs active at any time, so you have to update the cluster in 5 batches, in each batch, you update only 20 VMs | 06:57 |
| Qiming | after 5 batches, the whole cluster is updated, but at any time, you still have 80 VMs up and running | 06:58 |
| Qiming | batching is about dividing a huge group of operations into several batches | 06:58 |
| guoshan | thx, I got that | 06:59 |
| Qiming | so ... you can see there are primarily two ideas combined | 06:59 |
| Qiming | the first one is about avoiding DOS attack, the second one is about keeping clusters operational | 07:00 |
| Qiming | the second one is of a higher priority in my opinion | 07:00 |
| Qiming | as for the first one, we can check if it fits into the same conceptual design or not | 07:01 |
| Qiming | we can come up a throttling policy if it doesn't fit into the batching design | 07:01 |
| ruijie | Looks like nova do not support update a batch of servers? | 07:02 |
| openstackgerrit | lvdongbing proposed openstack/senlin: Revise schema _validate_version https://review.openstack.org/374571 | 07:13 |
| Qiming | ruijie, seems so | 07:17 |
| Qiming | from their api doc, http://developer.openstack.org/api-ref/compute | 07:17 |
| Qiming | I only saw create in batch | 07:17 |
| ruijie | yep | 07:17 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Cleanse zaqar driver for functions unused https://review.openstack.org/374606 | 08:12 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add new config option for Zaqar queue https://review.openstack.org/374531 | 08:35 |
| openstackgerrit | Lu lei proposed openstack/python-senlinclient: Never use MagicMock https://review.openstack.org/374630 | 09:14 |
| openstackgerrit | lvdongbing proposed openstack/python-senlinclient: Fix help msg of node-check and node-recover https://review.openstack.org/374653 | 09:41 |
| *** lvdongbing has quit IRC | 09:41 | |
| *** zhurong has quit IRC | 10:00 | |
| *** yanyanhu has quit IRC | 10:09 | |
| *** guoshan has quit IRC | 10:36 | |
| *** xiaozhuangqing has joined #senlin | 10:57 | |
| *** guoshan has joined #senlin | 11:19 | |
| openstackgerrit | xiaozhuangqing proposed openstack/senlin: Replace 'MagicMock' with 'Mock' https://review.openstack.org/373607 | 11:35 |
| *** guoshan has quit IRC | 11:41 | |
| *** guoshan has joined #senlin | 11:41 | |
| *** guoshan has quit IRC | 11:46 | |
| *** zigo has quit IRC | 11:54 | |
| *** zigo has joined #senlin | 11:58 | |
| *** zigo is now known as Guest83601 | 11:59 | |
| *** guoshan has joined #senlin | 11:59 | |
| *** Guest83601 has quit IRC | 12:03 | |
| *** zhurong has joined #senlin | 12:09 | |
| *** guoshan has quit IRC | 12:09 | |
| *** guoshan has joined #senlin | 12:10 | |
| *** zigo_ has joined #senlin | 12:12 | |
| *** guoshan_ has joined #senlin | 12:14 | |
| *** guoshan has quit IRC | 12:14 | |
| *** zigo_ has quit IRC | 12:17 | |
| *** zigo_ has joined #senlin | 12:18 | |
| *** guoshan_ has quit IRC | 12:38 | |
| *** guoshan has joined #senlin | 12:39 | |
| openstackgerrit | Merged openstack/python-senlinclient: Replace 'MagicMock' with 'Mock' https://review.openstack.org/373608 | 12:45 |
| *** guoshan has quit IRC | 12:46 | |
| *** guoshan has joined #senlin | 12:46 | |
| *** guoshan has quit IRC | 12:47 | |
| *** zigo_ has quit IRC | 12:48 | |
| *** zigo_ has joined #senlin | 12:51 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-senlinclient: Updated from global requirements https://review.openstack.org/374808 | 13:16 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/senlin: Updated from global requirements https://review.openstack.org/373785 | 13:16 |
| *** guoshan has joined #senlin | 13:18 | |
| openstackgerrit | Merged openstack/senlin: Updated from global requirements https://review.openstack.org/373785 | 14:14 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Tweak CLUSTER_ADD_NODE action for capacity update https://review.openstack.org/373016 | 14:15 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Tweak CLUSTER_DEL_NODES for base capacity https://review.openstack.org/373587 | 14:16 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Fix capacity calculation in scaling actions https://review.openstack.org/373610 | 14:17 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Move is_engine_dead test to common utils https://review.openstack.org/373736 | 14:18 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Remove unuse keystone interface functions https://review.openstack.org/373809 | 14:18 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Clean neutron driver interface for unused calls https://review.openstack.org/374134 | 14:19 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Cleanse nova driver for unused calls https://review.openstack.org/374166 | 14:19 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Correct driver calls in affinity policy https://review.openstack.org/374172 | 14:20 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Cleanse zaqar driver for functions unused https://review.openstack.org/374606 | 14:20 |
| openstackgerrit | Qiming Teng proposed openstack/python-senlinclient: Fix list of event with timestamp field https://review.openstack.org/374126 | 14:22 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Revise TODO.rst https://review.openstack.org/373793 | 14:23 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: API support node replace https://review.openstack.org/369291 | 14:23 |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: modify profile_update https://review.openstack.org/373241 | 14:23 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Fix typos in api-ref/source/policies.inc https://review.openstack.org/373910 | 14:23 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Use new min_size/max_size when eval cluster status https://review.openstack.org/374050 | 14:24 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Define new config options for receiver notification https://review.openstack.org/373861 | 14:24 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Replace 'MagicMock' with 'Mock' https://review.openstack.org/373607 | 14:24 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Add new config option for Zaqar queue https://review.openstack.org/374531 | 14:24 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Revise schema _validate_version https://review.openstack.org/374571 | 14:25 |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Fix user/policy_types.rst error https://review.openstack.org/372116 | 14:25 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Not necessary to handle 'ResourceNotFound' https://review.openstack.org/374505 | 14:25 |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Fix error in tutorial/receivers.rst https://review.openstack.org/373165 | 14:25 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Lighten the load https://review.openstack.org/374516 | 14:25 |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Fix error in registry.py https://review.openstack.org/372886 | 14:26 |
| *** ruijie has quit IRC | 14:26 | |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Fix cluster_check from health manage https://review.openstack.org/369937 | 14:26 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Modify eval status check https://review.openstack.org/374096 | 14:28 |
| openstackgerrit | XueFeng Liu proposed openstack/senlin: Fix FEATURES.rst error https://review.openstack.org/372844 | 14:28 |
| openstackgerrit | Qiming Teng proposed openstack/python-senlinclient: Fix list of event with timestamp field https://review.openstack.org/374126 | 14:37 |
| *** Drago has joined #senlin | 14:40 | |
| *** guoshan has quit IRC | 14:51 | |
| *** zhurong has quit IRC | 14:58 | |
| openstackgerrit | Merged openstack/senlin: Fix error in actions.rst https://review.openstack.org/372938 | 15:12 |
| openstackgerrit | Merged openstack/senlin: Fix error in deletion_v1.rst https://review.openstack.org/373139 | 15:12 |
| *** zigo_ is now known as zigo | 15:14 | |
| openstackgerrit | Merged openstack/senlin: Fix error in template.rst https://review.openstack.org/372939 | 15:33 |
| openstackgerrit | Merged openstack/senlin: Fix a typo in template.rst https://review.openstack.org/373580 | 15:33 |
| openstackgerrit | Merged openstack/senlin: fix typos in TODO.rst https://review.openstack.org/373613 | 15:43 |
| openstackgerrit | Merged openstack/senlin: Fix user/policy_types.rst error https://review.openstack.org/372116 | 15:44 |
| openstackgerrit | Merged openstack/senlin: Fix typos in api-ref/source/policies.inc https://review.openstack.org/373910 | 15:44 |
| openstackgerrit | Merged openstack/senlin: Fix error in registry.py https://review.openstack.org/372886 | 15:44 |
| openstackgerrit | Merged openstack/senlin: Fix FEATURES.rst error https://review.openstack.org/372844 | 15:44 |
| openstackgerrit | Merged openstack/senlin: Fix error in tutorial/receivers.rst https://review.openstack.org/373165 | 15:52 |
| openstackgerrit | Merged openstack/senlin: Replace 'MagicMock' with 'Mock' https://review.openstack.org/373607 | 15:53 |
| *** zigo has quit IRC | 16:41 | |
| *** zigo has joined #senlin | 16:51 | |
| *** zigo is now known as Guest18656 | 16:52 | |
| *** Guest18656 has quit IRC | 16:56 | |
| *** zigo_ has joined #senlin | 16:59 | |
| *** zigo_ has quit IRC | 17:07 | |
| *** zigo_ has joined #senlin | 17:11 | |
| *** zigo_ is now known as zigo | 19:30 | |
| *** Qiming has quit IRC | 21:37 | |
| *** AlexeyAbashkin has quit IRC | 21:37 | |
| *** chenfei has quit IRC | 21:37 | |
| *** zhenguo has quit IRC | 21:37 | |
| *** AlexeyAbashkin has joined #senlin | 21:38 | |
| *** chenfei has joined #senlin | 21:38 | |
| *** zhenguo has joined #senlin | 21:40 | |
| *** zhenguo has quit IRC | 21:40 | |
| *** Qiming has joined #senlin | 21:41 | |
| *** zhenguo has joined #senlin | 21:50 | |
| *** Drago has quit IRC | 22:51 | |
| *** zhurong has joined #senlin | 23:48 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!