*** zhurong has quit IRC | 00:03 | |
*** adrian_otto has joined #heat | 01:12 | |
*** bana_k has joined #heat | 01:21 | |
*** zh has joined #heat | 01:33 | |
*** adrian_otto1 has joined #heat | 01:33 | |
*** adrian_otto has quit IRC | 01:34 | |
*** vishwanathj has joined #heat | 01:43 | |
*** vishwanathj has quit IRC | 01:43 | |
*** yanyanhu has joined #heat | 01:44 | |
*** elynn has joined #heat | 01:54 | |
*** adrian_otto1 has quit IRC | 01:59 | |
*** jamielennox is now known as jamielennox|away | 02:10 | |
*** steveg_afk has quit IRC | 02:28 | |
*** jamielennox|away is now known as jamielennox | 02:30 | |
openstackgerrit | Rico Lin proposed openstack/heat: Handle stop rpc service at engine listener service https://review.openstack.org/353909 | 02:30 |
---|---|---|
*** bana_k has quit IRC | 02:40 | |
*** bana_k has joined #heat | 02:53 | |
openstackgerrit | Merged openstack/heat: Catch NotSupported when cancelling a nested stack https://review.openstack.org/352989 | 02:55 |
*** zhurong has joined #heat | 02:59 | |
*** bana_k has quit IRC | 03:08 | |
*** ramishra has quit IRC | 03:49 | |
*** ramishra has joined #heat | 03:51 | |
*** sabeen has joined #heat | 03:58 | |
*** sabeen1 has joined #heat | 04:00 | |
openstackgerrit | Merged openstack/heat: Add support for a resource level external_id https://review.openstack.org/135492 | 04:00 |
*** sabeen has quit IRC | 04:02 | |
*** chlong has joined #heat | 04:21 | |
*** vishwanathj has joined #heat | 04:25 | |
*** elynn has quit IRC | 04:27 | |
*** bana_k has joined #heat | 04:30 | |
*** Drago1 has joined #heat | 04:41 | |
*** Drago1 has quit IRC | 04:41 | |
*** Drago1 has joined #heat | 04:42 | |
*** zhurong has quit IRC | 04:43 | |
*** zhurong has joined #heat | 04:44 | |
*** Drago1 has quit IRC | 04:45 | |
*** bana_k has quit IRC | 04:46 | |
*** jaosorior has joined #heat | 04:47 | |
*** elynn has joined #heat | 04:50 | |
*** elynn has quit IRC | 04:55 | |
*** vishwanathj has quit IRC | 04:55 | |
*** elynn has joined #heat | 04:56 | |
*** crose has quit IRC | 05:15 | |
*** rook has quit IRC | 05:15 | |
*** vishwanathj has joined #heat | 05:15 | |
*** vishwanathj is now known as vishwanathj_zzz | 05:16 | |
*** rook has joined #heat | 05:17 | |
*** rook is now known as Guest14039 | 05:18 | |
*** crose has joined #heat | 05:18 | |
*** rcernin has joined #heat | 05:25 | |
openstackgerrit | shizhihui proposed openstack-dev/heat-cfnclient: Delete openstack/common in tox https://review.openstack.org/355302 | 05:35 |
*** elynn has quit IRC | 05:55 | |
*** elynn has joined #heat | 05:59 | |
*** elynn has quit IRC | 06:04 | |
openstackgerrit | huangtianhua proposed openstack/heat: Provides 'if' intrinsic function https://review.openstack.org/283479 | 06:12 |
openstackgerrit | huangtianhua proposed openstack/heat: Support condition for output https://review.openstack.org/347182 | 06:12 |
openstackgerrit | huangtianhua proposed openstack/heat: Add a Macro class to heat.engine.function https://review.openstack.org/322248 | 06:12 |
openstackgerrit | huangtianhua proposed openstack/heat: Add validating of section 'conditions' https://review.openstack.org/345975 | 06:12 |
*** zhurong has quit IRC | 06:15 | |
*** zhurong has joined #heat | 06:16 | |
openstackgerrit | huangtianhua proposed openstack/heat: Add validating of section 'conditions' https://review.openstack.org/345975 | 06:19 |
openstackgerrit | huangtianhua proposed openstack/heat: Support condition for resource https://review.openstack.org/347123 | 06:26 |
*** david-lyle has joined #heat | 06:36 | |
*** david-lyle_ has quit IRC | 06:39 | |
openstackgerrit | huangtianhua proposed openstack/heat: Support condition for output https://review.openstack.org/347182 | 06:40 |
*** rwsu has joined #heat | 06:44 | |
*** sabeen1 has quit IRC | 06:44 | |
openstackgerrit | huangtianhua proposed openstack/heat: Provides 'if' intrinsic function https://review.openstack.org/283479 | 06:44 |
openstackgerrit | huangtianhua proposed openstack/heat: Add a Macro class to heat.engine.function https://review.openstack.org/322248 | 06:44 |
openstackgerrit | Rico Lin proposed openstack/heat: Handle stop rpc service at engine listener service https://review.openstack.org/353909 | 06:47 |
*** skraynev has joined #heat | 06:56 | |
*** chlong has quit IRC | 07:08 | |
*** jpich has joined #heat | 07:21 | |
*** skraynev has quit IRC | 07:24 | |
openstackgerrit | shizhihui proposed openstack-dev/heat-cfnclient: Add Python 3.5 classifier and venv https://review.openstack.org/355302 | 07:25 |
*** skraynev has joined #heat | 07:25 | |
*** chlong has joined #heat | 07:26 | |
*** akuznetsov has joined #heat | 07:29 | |
*** duvarenkov has joined #heat | 07:38 | |
*** svkr_ has joined #heat | 07:45 | |
*** dimtruck is now known as zz_dimtruck | 07:59 | |
*** zzzeek has quit IRC | 08:00 | |
*** sorantis has joined #heat | 08:00 | |
*** sdake has joined #heat | 08:01 | |
*** zzzeek has joined #heat | 08:02 | |
*** svkr_ has quit IRC | 08:04 | |
*** chlong has quit IRC | 08:07 | |
*** sorantis has quit IRC | 08:12 | |
*** sdake has quit IRC | 08:15 | |
*** svkr_ has joined #heat | 08:16 | |
*** sorantis has joined #heat | 08:17 | |
*** ccamacho has joined #heat | 08:26 | |
*** ccamacho has quit IRC | 08:26 | |
*** ccamacho has joined #heat | 08:26 | |
openstackgerrit | huangtianhua proposed openstack/heat: Support condition for resource https://review.openstack.org/347123 | 08:29 |
openstackgerrit | huangtianhua proposed openstack/heat: Provides 'if' intrinsic function https://review.openstack.org/283479 | 08:33 |
openstackgerrit | huangtianhua proposed openstack/heat: Support condition for output https://review.openstack.org/347182 | 08:33 |
openstackgerrit | huangtianhua proposed openstack/heat: Add a Macro class to heat.engine.function https://review.openstack.org/322248 | 08:33 |
*** sdake has joined #heat | 08:35 | |
*** bana_k has joined #heat | 08:37 | |
*** sdake has quit IRC | 08:43 | |
*** prazumovsky has joined #heat | 09:14 | |
*** elynn has joined #heat | 09:20 | |
openstackgerrit | Rico Lin proposed openstack/heat: Support exponential backoff retry on resource sync https://review.openstack.org/328613 | 09:23 |
*** elynn has quit IRC | 09:27 | |
*** elynn has joined #heat | 09:28 | |
*** sshnaidm has quit IRC | 09:29 | |
*** jaosorior is now known as jaosorior_brb | 09:38 | |
*** bana_k has quit IRC | 09:41 | |
*** sdake has joined #heat | 09:46 | |
*** sorantis has quit IRC | 09:58 | |
*** elynn has quit IRC | 10:16 | |
*** zz_dimtruck is now known as dimtruck | 10:18 | |
*** yanyanhu has quit IRC | 10:22 | |
openstackgerrit | Merged openstack/heat: Add validating of section 'conditions' https://review.openstack.org/345975 | 10:27 |
*** dimtruck is now known as zz_dimtruck | 10:28 | |
*** jaosorior_brb is now known as jaosorior | 10:31 | |
*** KanagarajM has joined #heat | 10:53 | |
*** evilt0ne has joined #heat | 10:58 | |
*** zhurong has quit IRC | 11:05 | |
*** skraynev is now known as skraynev__ | 11:05 | |
*** sorantis has joined #heat | 11:18 | |
*** jaosorior has quit IRC | 11:34 | |
*** jaosorior has joined #heat | 11:35 | |
*** dprince has joined #heat | 11:52 | |
*** sdake has quit IRC | 12:02 | |
*** sorantis_ has joined #heat | 12:06 | |
*** sdake has joined #heat | 12:06 | |
*** sorantis has quit IRC | 12:08 | |
*** sorantis_ is now known as sorantis | 12:08 | |
openstackgerrit | Peter Razumovsky proposed openstack/heat: Keystone::Endpoint observe reality implementation https://review.openstack.org/300973 | 12:10 |
openstackgerrit | shizhihui proposed openstack/python-heatclient: Updeate the home-page in setup.cfg https://review.openstack.org/355435 | 12:13 |
*** zz_dimtruck is now known as dimtruck | 12:18 | |
*** sshnaidm|afk has joined #heat | 12:26 | |
*** dimtruck is now known as zz_dimtruck | 12:28 | |
*** steveg_afk has joined #heat | 12:30 | |
*** pradk has joined #heat | 12:33 | |
openstackgerrit | Peter Razumovsky proposed openstack/heat: Nova::Server observe reality implementation https://review.openstack.org/244066 | 12:35 |
*** dyasny has joined #heat | 12:37 | |
*** sdake has quit IRC | 12:38 | |
*** elynn has joined #heat | 12:40 | |
*** zh has quit IRC | 12:41 | |
*** elynn has quit IRC | 12:45 | |
*** elynn has joined #heat | 12:46 | |
openstackgerrit | Peter Razumovsky proposed openstack/heat: Neutron::Net observe reality implementation https://review.openstack.org/255287 | 12:49 |
*** elynn_ has joined #heat | 12:54 | |
*** elynn has quit IRC | 12:55 | |
openstackgerrit | Peter Razumovsky proposed openstack/heat: Neutron::FloatingIP observe reality implement https://review.openstack.org/256264 | 13:03 |
*** zz_dimtruck is now known as dimtruck | 13:04 | |
*** ccamacho has quit IRC | 13:05 | |
*** alee has quit IRC | 13:08 | |
*** elynn_ has quit IRC | 13:13 | |
*** jcoufal has joined #heat | 13:13 | |
openstackgerrit | Peter Razumovsky proposed openstack/heat: Neutron::FloatingIP observe reality implement https://review.openstack.org/256264 | 13:24 |
openstackgerrit | Peter Razumovsky proposed openstack/heat: Neutron::Subnet observe reality implementation https://review.openstack.org/255753 | 13:37 |
*** sorantis has quit IRC | 13:42 | |
openstackgerrit | Peter Razumovsky proposed openstack/heat: Neutron::Router observe reality implementation https://review.openstack.org/255776 | 13:43 |
*** KanagarajM has quit IRC | 13:47 | |
*** sorantis has joined #heat | 13:50 | |
*** jasond has joined #heat | 13:50 | |
*** chlong has joined #heat | 13:52 | |
*** dyasny has quit IRC | 13:54 | |
*** ayoung has joined #heat | 13:55 | |
*** dprince has quit IRC | 13:55 | |
openstackgerrit | Peter Razumovsky proposed openstack/heat: Neutron::Net observe reality implementation https://review.openstack.org/255287 | 13:55 |
openstackgerrit | Alexander Ignatyev proposed openstack/heat: Support new osprofiler API https://review.openstack.org/344514 | 13:55 |
*** prazumovsky has quit IRC | 13:56 | |
*** sorantis has quit IRC | 13:56 | |
*** sorantis has joined #heat | 13:59 | |
*** adrian_otto has joined #heat | 14:01 | |
*** dyasny has joined #heat | 14:03 | |
*** sorantis has quit IRC | 14:05 | |
*** sorantis has joined #heat | 14:06 | |
*** sdake has joined #heat | 14:10 | |
*** adrian_otto has quit IRC | 14:13 | |
*** adrian_otto has joined #heat | 14:15 | |
*** adrian_otto has quit IRC | 14:20 | |
*** fragatina has joined #heat | 14:20 | |
*** fragatina has quit IRC | 14:21 | |
*** sorantis has quit IRC | 14:21 | |
*** fragatina has joined #heat | 14:21 | |
*** jtomasek is now known as jtomasek|biab | 14:23 | |
*** KanagarajM has joined #heat | 14:24 | |
*** spotz_zzz is now known as spotz | 14:26 | |
*** spotz is now known as spotz_zzz | 14:26 | |
*** dprince has joined #heat | 14:27 | |
*** spotz_zzz is now known as spotz | 14:27 | |
*** akuznetsov has quit IRC | 14:28 | |
*** dimtruck is now known as zz_dimtruck | 14:29 | |
*** spedione|AWAY is now known as spedione | 14:31 | |
*** vijendar has joined #heat | 14:32 | |
*** jasond has quit IRC | 14:33 | |
*** jasond has joined #heat | 14:34 | |
*** zz_dimtruck is now known as dimtruck | 14:34 | |
*** zaneb has joined #heat | 14:36 | |
*** zhurong has joined #heat | 14:37 | |
*** zhurong has quit IRC | 14:39 | |
*** vishwanathj_zzz is now known as vishwanathj | 14:40 | |
*** akuznetsov has joined #heat | 14:43 | |
*** charlesr has joined #heat | 14:47 | |
*** sdake has quit IRC | 14:48 | |
*** elynn has joined #heat | 14:51 | |
*** sabeen has joined #heat | 14:59 | |
*** dyasny has quit IRC | 15:00 | |
*** spzala has joined #heat | 15:01 | |
*** akuznetsov has quit IRC | 15:04 | |
*** dyasny has joined #heat | 15:05 | |
*** rcernin has quit IRC | 15:06 | |
*** dprince has quit IRC | 15:12 | |
*** sabeen1 has joined #heat | 15:15 | |
*** elynn has quit IRC | 15:15 | |
*** sabeen has quit IRC | 15:17 | |
*** elynn has joined #heat | 15:18 | |
*** jtomasek|biab is now known as jtomasek | 15:20 | |
*** svkr_ has quit IRC | 15:20 | |
*** jasond has quit IRC | 15:23 | |
*** elynn has quit IRC | 15:25 | |
*** Drago has joined #heat | 15:31 | |
*** Drago1 has joined #heat | 15:32 | |
*** Drago1 has quit IRC | 15:32 | |
*** charlesr has quit IRC | 15:32 | |
*** Drago1 has joined #heat | 15:33 | |
*** Drago has quit IRC | 15:35 | |
*** Drago1 has quit IRC | 15:36 | |
*** Drago has joined #heat | 15:36 | |
*** harlowja_at_home has joined #heat | 15:40 | |
*** bana_k has joined #heat | 15:54 | |
*** spotz is now known as spotz_zzz | 15:55 | |
*** jpich has quit IRC | 16:00 | |
*** adrian_otto has joined #heat | 16:02 | |
*** KanagarajM has quit IRC | 16:07 | |
*** cwolferh has joined #heat | 16:09 | |
*** jtomasek is now known as jtomasek|afk | 16:10 | |
*** dyasny has quit IRC | 16:21 | |
*** dprince has joined #heat | 16:24 | |
*** spedione is now known as spedione|AWAY | 16:25 | |
*** harlowja_at_home has quit IRC | 16:28 | |
*** Drago has quit IRC | 16:31 | |
*** bana_k has quit IRC | 16:31 | |
*** bwallis has quit IRC | 16:36 | |
*** daneyon_ has joined #heat | 16:38 | |
*** daneyon has quit IRC | 16:40 | |
*** dyasny has joined #heat | 16:40 | |
*** Drago has joined #heat | 16:45 | |
*** dyasny has quit IRC | 16:45 | |
*** spedione|AWAY is now known as spedione | 16:49 | |
*** dyasny has joined #heat | 16:50 | |
*** adrian_otto has quit IRC | 16:54 | |
*** adrian_otto has joined #heat | 16:57 | |
*** sdake has joined #heat | 17:00 | |
*** bwallis has joined #heat | 17:02 | |
*** daneyon has joined #heat | 17:03 | |
*** jaosorior has quit IRC | 17:06 | |
*** daneyon_ has quit IRC | 17:06 | |
*** daneyon_ has joined #heat | 17:07 | |
*** harlowja has joined #heat | 17:08 | |
*** dprince has quit IRC | 17:09 | |
*** bana_k has joined #heat | 17:09 | |
*** daneyon has quit IRC | 17:09 | |
*** wirehead_ has quit IRC | 17:15 | |
*** sabeen1 has quit IRC | 17:15 | |
*** wirehead_ has joined #heat | 17:18 | |
*** vijendar has quit IRC | 17:19 | |
*** daneyon has joined #heat | 17:22 | |
*** ayoung has quit IRC | 17:25 | |
*** vijendar has joined #heat | 17:25 | |
*** daneyon_ has quit IRC | 17:26 | |
*** adrian_otto1 has joined #heat | 17:31 | |
*** adrian_otto has quit IRC | 17:35 | |
*** iurygregory has quit IRC | 17:36 | |
*** dprince has joined #heat | 17:36 | |
*** rcernin has joined #heat | 17:37 | |
*** GonZo2000 has joined #heat | 17:38 | |
*** sabeen has joined #heat | 17:39 | |
*** ayoung has joined #heat | 17:41 | |
*** vijendar has quit IRC | 17:42 | |
*** GonZo2K has joined #heat | 17:44 | |
*** GonZo2000 has quit IRC | 17:44 | |
*** iurygregory has joined #heat | 17:53 | |
*** vijendar has joined #heat | 17:56 | |
openstackgerrit | Zane Bitter proposed openstack/heat: Do str_replace in a single pass https://review.openstack.org/355583 | 18:01 |
*** GonZo2K has quit IRC | 18:10 | |
*** GonZo2K has joined #heat | 18:10 | |
*** GonZo2K has quit IRC | 18:10 | |
*** GonZo2K has joined #heat | 18:10 | |
*** GonZo2K has quit IRC | 18:12 | |
*** GonZo2K has joined #heat | 18:12 | |
*** GonZo2K has quit IRC | 18:13 | |
*** dyasny has quit IRC | 18:21 | |
*** neophy has joined #heat | 18:33 | |
*** dyasny has joined #heat | 18:36 | |
*** dyasny has quit IRC | 18:36 | |
*** ayoung has quit IRC | 18:37 | |
*** dyasny has joined #heat | 18:37 | |
*** rcernin has quit IRC | 18:42 | |
*** spzala has quit IRC | 18:48 | |
*** spzala has joined #heat | 18:48 | |
*** spzala has quit IRC | 18:51 | |
*** spzala has joined #heat | 18:52 | |
*** pt_15 has joined #heat | 18:57 | |
*** dyasny has quit IRC | 19:00 | |
*** vijendar has quit IRC | 19:01 | |
*** vijendar has joined #heat | 19:07 | |
*** Drago has quit IRC | 19:23 | |
*** dyasny has joined #heat | 19:30 | |
*** adrian_otto1 has quit IRC | 19:31 | |
*** rajpatel has joined #heat | 19:38 | |
*** blomquisg_gone is now known as blomquisg | 19:40 | |
*** coreyob has quit IRC | 19:40 | |
*** spotz_zzz is now known as spotz | 19:41 | |
*** ayoung has joined #heat | 19:48 | |
*** fragatina has quit IRC | 19:53 | |
*** daneyon_ has joined #heat | 19:55 | |
*** daneyon has quit IRC | 19:58 | |
*** jasond has joined #heat | 20:00 | |
openstackgerrit | Julian proposed openstack/heat-specs: New Resource Type OS::Cinder::Quota https://review.openstack.org/355627 | 20:02 |
openstackgerrit | Julian proposed openstack/heat-specs: New Resource Type OS::Cinder::Quota https://review.openstack.org/355627 | 20:03 |
*** coreyob has joined #heat | 20:05 | |
*** neophy has quit IRC | 20:14 | |
*** dyasny has quit IRC | 20:17 | |
*** Drago has joined #heat | 20:24 | |
*** sdake has quit IRC | 20:32 | |
*** fragatina has joined #heat | 20:41 | |
*** fragatina has quit IRC | 20:42 | |
*** fragatina has joined #heat | 20:43 | |
*** adrian_otto has joined #heat | 20:44 | |
stevebaker | morning | 20:49 |
openstackgerrit | Zane Bitter proposed openstack/heat: Do str_replace in a single pass https://review.openstack.org/355583 | 20:54 |
*** sdake has joined #heat | 20:57 | |
*** vijendar has quit IRC | 20:58 | |
*** dprince has quit IRC | 20:59 | |
stevebaker | zaneb: hey, this may well be ready for review https://review.openstack.org/#/c/278231 | 21:06 |
*** spedione is now known as chris_hultin | 21:06 | |
*** spzala has quit IRC | 21:07 | |
*** spzala has joined #heat | 21:07 | |
zaneb | stevebaker: no thoughts on my comment? | 21:08 |
stevebaker | oh yeah, that | 21:08 |
*** sdake has quit IRC | 21:08 | |
zaneb | er, this one specifically https://review.openstack.org/#/c/278231/16/heat/tests/engine/service/test_stack_resources.py | 21:09 |
stevebaker | what you really want from a deleted stack is the event list, not the resource list | 21:09 |
zaneb | hmmm... fair | 21:10 |
zaneb | so our position is that it doesn't matter, and deleted stacks will only have events and not resources? (and specifically, there's no way to filter the events down to a particular resource) | 21:12 |
zaneb | I think I can buy that | 21:12 |
stevebaker | that sounds reasonable to me. client side filtering only | 21:12 |
*** adrian_otto has quit IRC | 21:12 | |
*** spzala has quit IRC | 21:13 | |
*** GonZo2K has joined #heat | 21:13 | |
*** GonZo2K has quit IRC | 21:13 | |
*** GonZo2K has joined #heat | 21:13 | |
*** sdake has joined #heat | 21:17 | |
stevebaker | I wonder if zzzeek is about? I think zaneb and I both have misunderstandings about the sqlalchemy identity map which would be good to clarify | 21:20 |
zzzeek | stevebaker: hi | 21:20 |
stevebaker | zzzeek: ye | 21:20 |
stevebaker | zzzeek: heyt | 21:20 |
zaneb | lol | 21:20 |
stevebaker | alskdh | 21:20 |
zaneb | is it Friday yet? | 21:21 |
stevebaker | I have a cold | 21:21 |
stevebaker | zzzeek: so some context, I'm building off the enginefacade change https://review.openstack.org/#/c/330800 to try some caching strategies which reduce the number of sql queries in a request | 21:22 |
zzzeek | stevebaker: OK is there code here trying to do that yet | 21:23 |
stevebaker | zzzeek: One case is where I know individual stack records will be fetched, so I do a query to get them all, I hold a reference to the results so that the identity map stays hot, then I expect no individual fetches to happen - but I'm not seeing that | 21:24 |
stevebaker | zzzeek: https://review.openstack.org/#/c/317221 | 21:24 |
zzzeek | stevebaker: ok by "fetch" the only "fetch" you won't see is a get() of those identities | 21:25 |
zzzeek | stevebaker: it will not cache a collection of those objects, or if you said query(MyObject).filter(<whatever>), etc. | 21:25 |
stevebaker | zzzeek: thats right, I want my get() calls to not result in sql being logged | 21:26 |
zzzeek | stevebaker: so, if you have a Session, and load objects into a persistent collection, then you use that same session for a get() later, and also the object isn't expired, you wont see SQL | 21:26 |
zzzeek | so you need to check 1. same session 2. definitely a get and 3. object is not expired | 21:27 |
*** adrian_otto has joined #heat | 21:27 | |
zzzeek | also you can look in the identity map directly session.identity_map, | 21:27 |
stevebaker | zzzeek: and I'm wondering if the identity map is scoped to the session, which is scoped to the @db_context.reader annotated function - which makes it not useful for this situation | 21:27 |
*** chris_hultin is now known as spedione|AWAY | 21:27 | |
zzzeek | stevebaker: the identity map is local to a session. there's no other scope that would be appropriate | 21:28 |
zzzeek | however, @db_context.reader is using a single Session that came up from the top-level @db_context decorator in the request | 21:29 |
zzzeek | unless you are breaking out @writer and @reader in different databases | 21:29 |
*** rajpatel has quit IRC | 21:29 | |
zaneb | istr I checked and they were all using the same session | 21:30 |
zzzeek | zaneb: OK so you'd want to peek in session.identity_map as you pdb through and see it's there | 21:31 |
zaneb | by 'checked' I mean I read the code ;) | 21:31 |
*** Zer0Byte__ has joined #heat | 21:32 | |
stevebaker | zzzeek: I don't think there is a "top level" @db_context reader, just a series of calls to @db_context.reader functions. Should I be doing something to ensure there is a default session on the context? | 21:32 |
stevebaker | zzzeek: like annotating our service methods? | 21:32 |
stevebaker | for example, list_stack_resources https://review.openstack.org/#/c/317221/15/heat/engine/service.py | 21:33 |
zzzeek | stevebaker: OK, so, if there are sibling calls to @db_context.reader and there's no db_context enclosing it, then it would be making a new session / transaction in each call, becasue there's not an existing one. though that contradicts what zaneb just said | 21:33 |
zzzeek | oh | 21:33 |
zzzeek | "read the code" :) yeah you can't see it by reading :) | 21:33 |
zaneb | I probably remembered wrong then | 21:33 |
*** adrian_otto has quit IRC | 21:34 | |
zaneb | zzzeek: I dispute that. all the computer does is read the code ;) | 21:34 |
stevebaker | zaneb: I think https://review.openstack.org/#/c/330800 changes that | 21:34 |
zzzeek | stevebaker: do you ultimately want there to be just one database transaction for this whole operation? | 21:34 |
stevebaker | zzzeek: for read-only calls yeah. otherwise it would depend | 21:35 |
*** jcoufal has quit IRC | 21:35 | |
zzzeek | stevebaker: so you'd want to have some top level "@db_context" that frames where the transaction is. now this is kind of a thing I've seen, where we all have our RPC-style object interfaces that don't know anthing about the DB as the "top level" | 21:37 |
zzzeek | im assumign thats the case here? | 21:37 |
stevebaker | zzzeek: exactly, list_stack_resources is an RPC interface | 21:38 |
zaneb | we ought to be setting up a session in service.py (when we get an RPC message and create the context). then we want to start a transaction when we call the db code. IIUC. | 21:38 |
stevebaker | maybe we've discovered the use for this https://review.openstack.org/#/c/330800/9/heat/db/api.py | 21:38 |
*** sdake has quit IRC | 21:39 | |
zzzeek | zaneb: yeah, that's kind of what you'd need to do | 21:39 |
zzzeek | so, enginefacade is tricky, but this is one of the patterns it hoped to "force". e.g., lets know where our transactions are demarcated | 21:40 |
zzzeek | and lets not over-demarcate | 21:40 |
zaneb | http://git.openstack.org/cgit/openstack/heat/tree/heat/common/context.py#n141 | 21:41 |
zzzeek | zaneb: in your "let's do a request!" it would just be doing "with enginefacade.writer():" before it calls down further. | 21:41 |
zzzeek | zaneb: assuming the API is RESTful you could discern reader/writer by GET/PUT/POST/etc | 21:41 |
stevebaker | zaneb: that goes away https://review.openstack.org/#/c/330800/9/heat/common/context.py | 21:42 |
zzzeek | zaneb: or jsut use writer() in all cases | 21:42 |
zaneb | stevebaker: ah, that'll be the trouble then | 21:43 |
stevebaker | zzzeek: what is the tradeoff of having a default reader vs a default writer | 21:45 |
zzzeek | stevebaker: it is only meant to apply a hint in the use case that an application is routing "read only" SQL to a read slave DB | 21:45 |
zzzeek | i dont think anyone's doing that, however when i first did enginefacade poeple were proposing reader/writer separation on the mailing list | 21:46 |
zzzeek | it's really not very critical except at like...you know Web Scale :) | 21:47 |
stevebaker | isn't that how our galera clusters are set up? | 21:47 |
zzzeek | galeras are multi master. so there's no "read slave" concept | 21:47 |
zzzeek | reader/writer can stil be used to route write transactions vs. readers to different galera nodes | 21:47 |
stevebaker | I thought we configured to write to one master to work around that "deadlock" pattern | 21:48 |
zzzeek | stevebaker: i guess in theory, enginefacade would need some kind of "wrap this block in sibling support" type of feature | 21:48 |
zzzeek | stevebaker: yeah we do. but all the "readers" go to the same DB also | 21:48 |
*** adrian_otto has joined #heat | 21:51 | |
stevebaker | zzzeek: so it sounds like decorating here isn't actually the right solution for us https://review.openstack.org/#/c/330800/12/heat/db/sqlalchemy/api.py instead we should decorate our entry points - which will either be our RPC object interface or the starting point of the async jobs they spawn? | 21:52 |
zzzeek | stevebaker: IMO you should decorate regardless, but if you did have a global "make sure the context has the Session" then yes, I suppose you wouldn't need all the decorators on each method | 21:53 |
*** sdake has joined #heat | 21:54 | |
stevebaker | zzzeek: I'm fine with leaving them there too | 21:54 |
zzzeek | stevebaker: if you are changing transaction scope to be at the top of all your requests, that's fine but you'd want to make sure there's no weird side effects of that. nova etc. hvae some spots where they retry transactions and such | 21:54 |
stevebaker | ok | 21:55 |
*** harlowja has quit IRC | 21:55 | |
zaneb | I don't think we want to change transaction scope | 22:04 |
*** GonZoPT has joined #heat | 22:05 | |
*** GonZo2K has quit IRC | 22:05 | |
zzzeek | zaneb: so, the other way to "cache" things is just to cache them anywhere. to get them into the Session that you're querying against you can get them in there using merge() | 22:06 |
zzzeek | or even add(). | 22:06 |
zzzeek | merge(obj, dont_load=True) is more the "I've pulled this object from a cache" function | 22:07 |
zaneb | please no ;) | 22:07 |
stevebaker | zzzeek, zaneb: I'm suggesting we cache at our versioned object layer, here is a case which can't use the identity map because its caching filter() calls | 22:09 |
stevebaker | https://review.openstack.org/#/c/317220/14 | 22:09 |
zzzeek | stevebaker: coarse-grained caching is usually better, yup | 22:09 |
stevebaker | zzzeek: I looked into implementing a sqlalchemy query cache but the dogpile example hurt my brain and I just couldn't see the advantage | 22:10 |
*** fragatin_ has joined #heat | 22:10 | |
zzzeek | stevebaker: yeah that is a thing I worked up at a gig some years ago where we really needed to cache everythign everywhere for a website | 22:10 |
zzzeek | and it was like big object graphs | 22:10 |
zzzeek | openstack doesn't seem to do much with deep graphs | 22:10 |
zzzeek | I think a coarse grained cache thing is easier, dogpile has that with just the regalar function cache decorators | 22:11 |
zzzeek | keystone shoudl be using it via oslo.cache | 22:11 |
stevebaker | zzzeek: we wouldn't use dogpile, I just didn't see a benefit of caching one layer lower than what anybody in heat is familiar with ;) | 22:11 |
zzzeek | stevebaker: sure | 22:11 |
*** fragatina has quit IRC | 22:13 | |
*** fragatin_ has quit IRC | 22:14 | |
*** fragatina has joined #heat | 22:15 | |
*** jistr has quit IRC | 22:18 | |
*** jistr has joined #heat | 22:19 | |
*** sdake has quit IRC | 22:20 | |
*** GonZoPT has quit IRC | 22:22 | |
*** GonZo2000 has joined #heat | 22:24 | |
*** jasond has quit IRC | 22:24 | |
*** sabeen has quit IRC | 22:32 | |
*** evilt0ne has quit IRC | 22:32 | |
*** sdake has joined #heat | 22:35 | |
*** Drago has quit IRC | 22:42 | |
*** GonZo2000 has quit IRC | 22:48 | |
*** GonZo2000 has joined #heat | 22:50 | |
*** elynn has joined #heat | 22:55 | |
*** elynn_ has joined #heat | 22:56 | |
*** sdake has quit IRC | 22:57 | |
*** elynn has quit IRC | 23:00 | |
*** Drago has joined #heat | 23:10 | |
*** Drago has quit IRC | 23:10 | |
*** Drago has joined #heat | 23:10 | |
*** elynn_ has quit IRC | 23:12 | |
*** GonZo2000 has quit IRC | 23:22 | |
*** tochi has joined #heat | 23:29 | |
*** GonZo2000 has joined #heat | 23:31 | |
*** bwallis has quit IRC | 23:33 | |
*** harlowja has joined #heat | 23:36 | |
*** zhurong has joined #heat | 23:49 | |
*** bwallis has joined #heat | 23:55 | |
openstackgerrit | Merged openstack/heat: Add version keyword argument to RAX glanceclient https://review.openstack.org/343054 | 23:55 |
*** GonZo2000 has quit IRC | 23:57 | |
*** zhurong has quit IRC | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!