| *** adrian_otto has joined #heat | 00:09 | |
| *** rajpatel has joined #heat | 00:10 | |
| *** adrian_otto has quit IRC | 00:23 | |
| *** rajpatel1 has joined #heat | 00:34 | |
| *** rajpatel has quit IRC | 00:35 | |
| *** rajpatel1 has quit IRC | 01:01 | |
| *** david-lyle_ has joined #heat | 01:04 | |
| *** david-lyle has quit IRC | 01:08 | |
| *** david-lyle_ is now known as david-lyle | 01:10 | |
| *** yanyanhu has joined #heat | 01:12 | |
| *** vishwanathj has joined #heat | 01:20 | |
| *** xek_ has joined #heat | 01:28 | |
| *** jistr has quit IRC | 01:29 | |
| *** abramley has quit IRC | 01:29 | |
| *** Tristitia has quit IRC | 01:29 | |
| *** xek has quit IRC | 01:29 | |
| *** jtomasek has quit IRC | 01:29 | |
| *** jtomasek has joined #heat | 01:34 | |
| *** abramley has joined #heat | 01:35 | |
| *** jistr has joined #heat | 01:35 | |
| *** Tristitia has joined #heat | 01:36 | |
| *** code-R has joined #heat | 01:36 | |
| *** code-R has quit IRC | 01:40 | |
| *** zhihui has joined #heat | 01:54 | |
| *** vijendar has joined #heat | 01:59 | |
| *** dimtruck is now known as zz_dimtruck | 01:59 | |
| *** vijendar has quit IRC | 02:04 | |
| *** bana_k has joined #heat | 02:17 | |
| *** zz_dimtruck is now known as dimtruck | 02:35 | |
| *** dimtruck is now known as zz_dimtruck | 02:45 | |
| *** sdake has joined #heat | 02:46 | |
| *** bana_k has quit IRC | 02:46 | |
| *** bana_k has joined #heat | 02:57 | |
| *** code-R has joined #heat | 02:58 | |
| *** david-lyle has quit IRC | 03:09 | |
| *** code-R has quit IRC | 03:14 | |
| *** code-R has joined #heat | 03:22 | |
| *** gberginc has joined #heat | 03:25 | |
| *** gberginc has quit IRC | 03:26 | |
| *** code-R_ has joined #heat | 03:30 | |
| *** bana_k has quit IRC | 03:32 | |
| *** code-R has quit IRC | 03:32 | |
| *** sdake has quit IRC | 03:41 | |
| *** adrian_otto has joined #heat | 03:45 | |
| *** adrian_otto has quit IRC | 03:47 | |
| *** code-R_ has quit IRC | 03:48 | |
| *** zz_dimtruck is now known as dimtruck | 03:49 | |
| *** adrian_otto has joined #heat | 03:52 | |
| *** dimtruck is now known as zz_dimtruck | 03:59 | |
| *** vijendar has joined #heat | 04:05 | |
| *** flwang1 has quit IRC | 04:08 | |
| *** vijendar has quit IRC | 04:10 | |
| *** adrian_otto1 has joined #heat | 04:13 | |
| *** adrian_otto has quit IRC | 04:16 | |
| *** code-R has joined #heat | 04:18 | |
| *** code-R has quit IRC | 04:27 | |
| *** gberginc has joined #heat | 04:30 | |
| *** adrian_otto1 has quit IRC | 04:33 | |
| *** adrian_otto has joined #heat | 04:35 | |
| *** adrian_otto has quit IRC | 04:36 | |
| *** adrian_otto has joined #heat | 04:36 | |
| *** adrian_otto has quit IRC | 04:44 | |
| openstackgerrit | Merged openstack/heat: Using oslo_config's min attribute in heat https://review.openstack.org/359064 | 04:45 |
|---|---|---|
| *** adrian_otto has joined #heat | 04:46 | |
| openstackgerrit | Merged openstack/heat: Correct invalid git URL in LogABug config https://review.openstack.org/382968 | 04:46 |
| openstackgerrit | Merged openstack/heat: Keystone role assignment observe reality method https://review.openstack.org/303428 | 04:46 |
| *** adrian_otto has quit IRC | 05:01 | |
| *** adrian_otto has joined #heat | 05:04 | |
| *** jaosorior has joined #heat | 05:07 | |
| *** jaosorior has quit IRC | 05:10 | |
| *** bana_k has joined #heat | 05:11 | |
| *** jaosorior has joined #heat | 05:11 | |
| *** exploreshaifali has joined #heat | 05:31 | |
| *** ishant|away has quit IRC | 05:32 | |
| *** kirankv has quit IRC | 05:35 | |
| *** zz_dimtruck is now known as dimtruck | 05:50 | |
| *** dimtruck is now known as zz_dimtruck | 05:59 | |
| *** vijendar has joined #heat | 06:06 | |
| *** gberginc has quit IRC | 06:12 | |
| *** vijendar has quit IRC | 06:12 | |
| *** adrian_otto has quit IRC | 06:14 | |
| *** gberginc has joined #heat | 06:16 | |
| *** yprokule has joined #heat | 06:20 | |
| *** adrian_otto has joined #heat | 06:23 | |
| *** bana_k has quit IRC | 06:26 | |
| *** code-R has joined #heat | 06:28 | |
| *** exploreshaifali has quit IRC | 06:29 | |
| *** prazumovsky has joined #heat | 06:32 | |
| *** code-R has quit IRC | 06:33 | |
| *** jprovazn has joined #heat | 06:33 | |
| openstackgerrit | Merged openstack/python-heatclient: handle empty sections in environment files https://review.openstack.org/383785 | 06:35 |
| *** exploreshaifali has joined #heat | 06:37 | |
| *** rcernin has joined #heat | 06:48 | |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/heat: Add http_proxy_to_wsgi middleware to Heat CFN endpoint https://review.openstack.org/384314 | 06:48 |
| *** exploreshaifali has quit IRC | 06:52 | |
| *** code-R has joined #heat | 06:53 | |
| *** code-R has quit IRC | 06:58 | |
| *** gberginc has quit IRC | 07:02 | |
| *** fabian4 has joined #heat | 07:08 | |
| fabian4 | hi, i have a quesstion, util.py[WARNING]: Running scripts-user (<module 'cloudinit.config.cc_scripts_user' from '/usr/lib/python2.7/site-packages/cloudinit/config/cc_scripts_user.pyc'>) failed ? pleae give me a help , my centos7 install cloudbase-0.7.5 | 07:09 |
| *** gberginc has joined #heat | 07:22 | |
| openstackgerrit | Merged openstack/heat: Add networking-sfc flow-classifier resource plug-in https://review.openstack.org/277988 | 07:23 |
| *** bana_k has joined #heat | 07:25 | |
| *** exploreshaifali has joined #heat | 07:25 | |
| *** dbecker has joined #heat | 07:27 | |
| *** prazumovsky has quit IRC | 07:28 | |
| *** prazumovsky has joined #heat | 07:33 | |
| *** jpich has joined #heat | 07:38 | |
| *** ccamacho has joined #heat | 07:38 | |
| *** ccamacho is now known as ccamacho|mtg | 07:39 | |
| *** prazumovsky has quit IRC | 07:45 | |
| *** flwang1 has joined #heat | 07:49 | |
| openstackgerrit | Deepak proposed openstack/heat: Cleanup Newton Release Notes https://review.openstack.org/384331 | 07:50 |
| *** zz_dimtruck is now known as dimtruck | 07:50 | |
| *** bana_k has quit IRC | 07:59 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** dimtruck is now known as zz_dimtruck | 08:00 | |
| *** zzzeek has joined #heat | 08:01 | |
| *** sorantis has joined #heat | 08:05 | |
| *** vijendar has joined #heat | 08:08 | |
| *** gberginc has quit IRC | 08:10 | |
| *** vijendar has quit IRC | 08:12 | |
| *** exploreshaifali has quit IRC | 08:14 | |
| *** vijendar has joined #heat | 08:18 | |
| *** adrian_otto has quit IRC | 08:20 | |
| *** ccamacho|mtg is now known as ccamacho | 08:21 | |
| *** vijendar has quit IRC | 08:22 | |
| *** adrian_otto has joined #heat | 08:23 | |
| *** yassine has joined #heat | 08:25 | |
| *** yassine is now known as Guest20409 | 08:25 | |
| *** Guest20409 is now known as yassou | 08:25 | |
| *** exploreshaifali has joined #heat | 08:27 | |
| openstackgerrit | Thomas Herve proposed openstack/heat: Use RPC to retrieve nested stack output https://review.openstack.org/383839 | 08:33 |
| *** trananhkma has joined #heat | 08:43 | |
| *** sdake has joined #heat | 08:43 | |
| *** vishwana_ has joined #heat | 08:46 | |
| *** dsneddon_1 has quit IRC | 08:46 | |
| *** dsneddon_1 has joined #heat | 08:49 | |
| *** evilt0ne_ has joined #heat | 08:49 | |
| *** pleia2_ has joined #heat | 08:50 | |
| *** dgonzalez_ has joined #heat | 08:53 | |
| *** coreyob_ has joined #heat | 08:53 | |
| *** mfisch` has joined #heat | 08:53 | |
| *** andreaf_ has joined #heat | 08:53 | |
| *** Qiming_ has joined #heat | 08:54 | |
| *** vishwanathj has quit IRC | 08:54 | |
| *** andreaf has quit IRC | 08:54 | |
| *** skraynev has quit IRC | 08:54 | |
| *** mfisch has quit IRC | 08:54 | |
| *** pleia2 has quit IRC | 08:54 | |
| *** coreyob has quit IRC | 08:54 | |
| *** Qiming has quit IRC | 08:54 | |
| *** hoonetorg has quit IRC | 08:54 | |
| *** gmoro has quit IRC | 08:54 | |
| *** evilt0ne has quit IRC | 08:54 | |
| *** dgonzalez has quit IRC | 08:54 | |
| *** dgonzalez_ is now known as dgonzalez | 08:54 | |
| *** andreaf_ is now known as andreaf | 08:54 | |
| *** Qiming_ is now known as Qiming | 08:54 | |
| *** skraynev has joined #heat | 09:01 | |
| *** gmoro has joined #heat | 09:01 | |
| *** hoonetorg has joined #heat | 09:01 | |
| *** kirankv has joined #heat | 09:21 | |
| *** ishant|away has joined #heat | 09:21 | |
| *** code-R has joined #heat | 09:21 | |
| *** gberginc has joined #heat | 09:23 | |
| *** Guest10079 is now known as zigo | 09:23 | |
| *** code-R_ has joined #heat | 09:24 | |
| openstackgerrit | Peter Razumovsky proposed openstack/heat: Keystone::Endpoint observe reality implementation https://review.openstack.org/300973 | 09:27 |
| *** code-R has quit IRC | 09:27 | |
| *** prazumovsky has joined #heat | 09:27 | |
| *** eliqiao has quit IRC | 09:28 | |
| *** ishant|away has quit IRC | 09:29 | |
| *** eliqiao has joined #heat | 09:29 | |
| *** eliqiao has quit IRC | 09:29 | |
| prazumovsky | Hi! Can someone check new property tags for server: https://review.openstack.org/#/c/374797/ I have some doubts about client versioning, so will be appreciate if KanagarajM take a look on it:) | 09:30 |
| *** kirankv has quit IRC | 09:30 | |
| *** eliqiao has joined #heat | 09:32 | |
| *** gberginc has quit IRC | 09:32 | |
| *** eliqiao has quit IRC | 09:32 | |
| *** eliqiao has joined #heat | 09:33 | |
| *** derekh has joined #heat | 09:36 | |
| *** caowei has quit IRC | 09:36 | |
| *** shardy has joined #heat | 09:37 | |
| *** caowei has joined #heat | 09:37 | |
| *** jaosorior has quit IRC | 09:38 | |
| *** shardy has quit IRC | 09:38 | |
| *** jaosorior has joined #heat | 09:38 | |
| *** shardy has joined #heat | 09:39 | |
| openstackgerrit | Deepak proposed openstack/heat: Fixed the too long characters https://review.openstack.org/384388 | 09:44 |
| *** evilt0ne_ has quit IRC | 09:45 | |
| *** evilt0ne has joined #heat | 09:48 | |
| *** evilt0ne has joined #heat | 09:48 | |
| *** zz_dimtruck is now known as dimtruck | 09:51 | |
| *** yassou has quit IRC | 09:54 | |
| therve | shardy, Hi! Time to chat about my favorite subject, memory usage? :) | 09:54 |
| *** dimtruck is now known as zz_dimtruck | 10:01 | |
| *** yanyanhu has quit IRC | 10:01 | |
| *** jistr is now known as jistr|mtg | 10:03 | |
| *** prazumovsky has quit IRC | 10:03 | |
| *** exploreshaifali has quit IRC | 10:09 | |
| *** exploreshaifali has joined #heat | 10:12 | |
| *** vijendar has joined #heat | 10:19 | |
| *** vijendar has quit IRC | 10:23 | |
| *** Guest2483 is now known as rook | 10:31 | |
| *** caowei has quit IRC | 10:38 | |
| *** adrian_otto has quit IRC | 10:46 | |
| *** code-R_ has quit IRC | 10:55 | |
| *** kirankv has joined #heat | 11:11 | |
| *** ishant|away has joined #heat | 11:11 | |
| *** yassine has joined #heat | 11:13 | |
| *** yassine is now known as Guest75406 | 11:13 | |
| *** Alexey_Abashkin_ has joined #heat | 11:24 | |
| *** exploreshaifali has quit IRC | 11:25 | |
| *** Alexey_Abashkin has quit IRC | 11:27 | |
| *** exploreshaifali has joined #heat | 11:28 | |
| *** jistr|mtg is now known as jistr | 11:34 | |
| *** ccamacho is now known as ccamacho|lunch | 11:37 | |
| *** sdake has quit IRC | 11:41 | |
| *** zaneb has joined #heat | 11:48 | |
| *** dprince has joined #heat | 11:49 | |
| *** zz_dimtruck is now known as dimtruck | 11:52 | |
| *** code-R has joined #heat | 11:53 | |
| *** code-R_ has joined #heat | 11:56 | |
| *** code-R has quit IRC | 11:58 | |
| *** dimtruck is now known as zz_dimtruck | 12:02 | |
| openstackgerrit | Mohankumar proposed openstack/heat: Add networking-sfc port-chain resource plug-in https://review.openstack.org/277940 | 12:04 |
| *** exploreshaifali has quit IRC | 12:07 | |
| *** evilt0ne has quit IRC | 12:09 | |
| *** shardy is now known as shardy_mtg | 12:12 | |
| *** exploreshaifali has joined #heat | 12:12 | |
| *** shardy_mtg is now known as shardy | 12:12 | |
| shardy | therve: Hey, sorry in meetings most of today - happy to chat about memory usage tho :) | 12:13 |
| *** ccamacho|lunch is now known as ccamacho | 12:15 | |
| *** exploreshaifali has quit IRC | 12:16 | |
| *** vijendar has joined #heat | 12:20 | |
| *** vijendar has quit IRC | 12:24 | |
| *** EricGonczer_ has joined #heat | 12:29 | |
| *** blomquisg is now known as blomquisg_gone | 12:36 | |
| *** shardy is now known as shardy_mtg | 12:37 | |
| *** evilt0ne has joined #heat | 12:39 | |
| *** evilt0ne has quit IRC | 12:39 | |
| *** evilt0ne has joined #heat | 12:39 | |
| therve | shardy_mtg, https://review.openstack.org/#/c/383839/ is the latest of the attempts I made. One of the biggest improvements | 12:44 |
| therve | It's unclear exactly why it matters so much, but it does | 12:44 |
| shardy_mtg | therve: aha! Very interesting, thanks I'll test it later, but lgtm | 12:46 |
| therve | shardy_mtg, I tried http://paste.openstack.org/show/585172/ change in tht, which I was hoping to be effectively the same as the heat change | 12:47 |
| therve | But it didn't work, and I don't really understand why | 12:47 |
| therve | It looks like data available during output is not in resource creation | 12:47 |
| *** shardy_mtg is now known as shardy | 12:49 | |
| shardy | therve: interesting, I was trying to do similar until I hit bug #1629899 (now fixed and not related to your issue AFAICS) | 12:50 |
| openstack | bug 1629899 in heat "repeat for_each can't reference attributes" [Medium,Fix released] https://launchpad.net/bugs/1629899 - Assigned to Rabi Mishra (rabi) | 12:50 |
| therve | shardy, Hum, that might be it though | 12:50 |
| * therve tries again | 12:50 | |
| shardy | therve: one issue is output validation is less strict than resource property validation atm due to bug #1599114 | 12:51 |
| openstack | bug 1599114 in heat "Outputs aren't correctly validated" [Medium,In progress] https://launchpad.net/bugs/1599114 - Assigned to Oleksii Chuprykov (ochuprykov) | 12:51 |
| shardy | therve: so it's possible there's other issues like bug #1629899 which mean stuff breaks when moved into OS::Heat::Value | 12:51 |
| openstack | bug 1629899 in heat "repeat for_each can't reference attributes" [Medium,Fix released] https://launchpad.net/bugs/1629899 - Assigned to Rabi Mishra (rabi) | 12:51 |
| therve | shardy, It's worrying if we end up breaking current output usage with more validation | 12:51 |
| shardy | therve: Yeah, well it probably just means more validation bugs where stuff breaks due to get_attr returning None | 12:52 |
| therve | Sigh | 12:52 |
| shardy | it'd be good to figure out a more robust approach to that in general than all the special-cases for validation in the functions | 12:52 |
| therve | Yeah, that's on my list. We somewhat have a plan, need to see how it does in the face of the codebase | 12:53 |
| shardy | therve: so, is the issue that every output is resolving all the yaql for the entire tree of all nested stacks? | 12:53 |
| shardy | and moving to RPC breaks that chain? | 12:53 |
| therve | shardy, Possibly. Although, I tried to move just the yaql evaluation to RPC, and that didn't improve things. | 12:54 |
| therve | So there is something else, maybe just the sheer size of data passed to yaql | 12:54 |
| shardy | therve: Ok, well good that we're making progress narrowing down the issue tho, thanks for digging into it :) | 12:54 |
| therve | Yeah it's quite the haystack | 12:55 |
| therve | So yeah my tht patch still fails with "resources.CephStorageServiceChain: Property error: resources.RoleData.properties.value: No method "where" for receiver None matches supplied arguments" | 12:56 |
| shardy | therve: Ok, so that's due to yaql getting None during validation | 12:56 |
| shardy | so we'll need the query to handle that with a default somewhere | 12:56 |
| shardy | (or at least I assume that's the issue as I hit the same myself) | 12:57 |
| therve | OK, I'll add some debug | 12:59 |
| *** iurygregory has joined #heat | 13:00 | |
| shardy | therve: FYI I do partial updates by heat stack-update -x <nested stack ID> fixed-template.yaml, which shortens the debug cycle a bit | 13:01 |
| therve | shardy, Ah good idea, thanks | 13:02 |
| *** shardy is now known as shardy_mtg | 13:03 | |
| *** jprovazn has quit IRC | 13:03 | |
| *** EricGonc_ has joined #heat | 13:08 | |
| *** huangtianhua has joined #heat | 13:08 | |
| therve | Ah, I think I got it | 13:08 |
| therve | shardy_mtg, It gets None from {get_attr: [ServiceChain, role_data]} | 13:08 |
| therve | Because ServiceChain specifies concurrent, we don't generate the dependencies | 13:09 |
| *** EricGonczer_ has quit IRC | 13:09 | |
| shardy_mtg | therve: aha | 13:09 |
| therve | Thus... we don't wait for the resources, and data is not there | 13:09 |
| *** shardy_mtg is now known as shardy | 13:09 | |
| shardy | therve: isn't that a bug in service chain? | 13:09 |
| shardy | e.g even in concurrent mode we should wait for all nested resources to be complete? | 13:10 |
| therve | Possibly, I don't know enough about it to tell | 13:10 |
| shardy | we just don't serialize their creation | 13:10 |
| shardy | therve: FWIW I was somewhat involved with service chain stuff and IIRC the intention was only to serialize (or not) the creation of resources in the nested stack | 13:11 |
| shardy | it just adds a depends_on between each resource in the chain | 13:11 |
| therve | OK. So the chain itself should depend on all the resources | 13:11 |
| shardy | therve: yeah, that's my understanding | 13:11 |
| therve | OK, will try that :) | 13:11 |
| *** tiantian has quit IRC | 13:11 | |
| therve | Hum no, what I'm saying doesn't make sense | 13:15 |
| therve | It's done in a nested stack, so it should wait for all resources inside it | 13:15 |
| *** jdandrea has joined #heat | 13:15 | |
| shardy | therve: ah, I thought you were saying the parent resource didn't wait for complete | 13:17 |
| therve | shardy, Well there is no explicit depends_on, but there shouldn't be | 13:18 |
| *** vijendar has joined #heat | 13:24 | |
| *** jprovazn has joined #heat | 13:24 | |
| *** vijendar has quit IRC | 13:33 | |
| zaneb | therve: I don't think it's a ResourceChain problem, likely just a standard returns-none-during-validation problem | 13:35 |
| therve | zaneb, Yeah you're probably right. I don't understand where though. | 13:36 |
| *** EricGonc_ has quit IRC | 13:38 | |
| *** EricGonczer_ has joined #heat | 13:41 | |
| therve | zaneb, Can you enlighten me a bit on that subject? | 13:43 |
| therve | It doesn't seem we filter None everywhere in functions result, should we? | 13:43 |
| *** blomquisg_gone is now known as blomquisg | 13:44 | |
| therve | zaneb, Things like https://review.openstack.org/#/c/349459/ ? | 13:44 |
| zaneb | my memory is fuzzy here | 13:45 |
| zaneb | but I think it's something along the lines that we try to validate resource definitions prior to creating any resources | 13:45 |
| zaneb | and in the process we try resolving the functions (which we probably shouldn't) | 13:46 |
| zaneb | so if any of their inputs resolve to None then they need some sane way of handling it | 13:46 |
| therve | Yeah that seems to be the case | 13:46 |
| zaneb | resulting in lots of ugly special cases :( | 13:47 |
| therve | Right, and we don't handle that systematically | 13:47 |
| zaneb | and bugs in all the places where we forget to put special cases :( | 13:47 |
| zaneb | yep | 13:47 |
| therve | We just stumble upon those... | 13:47 |
| therve | Urg that's not a way to start the week :) | 13:48 |
| zaneb | lol | 13:48 |
| therve | zaneb, https://review.openstack.org/#/c/383839/ is a better way. WDYT? | 13:48 |
| therve | It makes some great improvements for tripleo, not sure for the general case though | 13:49 |
| zaneb | haven't read the patch yet, but the idea seems sound | 13:49 |
| *** sdake has joined #heat | 13:49 | |
| zaneb | therve: I'm curious what the theory is behind http://paste.openstack.org/show/585172/ ? I wouldn't have expected that to make a big difference | 13:50 |
| *** code-R_ has quit IRC | 13:50 | |
| therve | zaneb, So, the theory is that it moves output calculation into the nested stack | 13:50 |
| therve | So that output becomes "static" | 13:51 |
| therve | Thus not requiring the heat patch delegating output to RPC | 13:51 |
| zaneb | ah, ok, so it only has to dig down one level | 13:51 |
| zaneb | good thinking | 13:51 |
| therve | Right. I'm trying to make it work though, because of those None values popping out | 13:51 |
| therve | So I haven't confirmed that it actually works | 13:52 |
| zaneb | what would really have made a big difference is if you could ditch that whole stack, and replace the whole child stack with an OS::Heat::Value | 13:52 |
| *** zz_dimtruck is now known as dimtruck | 13:53 | |
| zaneb | not sure how many of those stacks have zero resources though (obviously those are the only ones it works for) | 13:53 |
| therve | Wouldn't that create the data in the top level stack though? | 13:53 |
| *** sdake has quit IRC | 13:54 | |
| *** sdake has joined #heat | 13:55 | |
| therve | Oh wow, that is deeply broken | 13:56 |
| therve | I can filter out None data when it's top level | 13:56 |
| therve | But not when it's nested into something else | 13:56 |
| zaneb | there are templates like http://git.openstack.org/cgit/openstack/tripleo-heat-templates/tree/puppet/services/ceilometer-base.yaml that have no resources and exist purely for laundering parameters into outputs. that's the case that OS::Heat::Value can potentially replace | 13:57 |
| *** code-R has joined #heat | 13:57 | |
| therve | zaneb, Yes, it can. I mean that by delegating to a nested stack, you delegate the calculation, thus the memory can be released | 13:57 |
| therve | Ditching the stack would have the opposite effect of what I want to achieve in tripleo | 13:58 |
| zaneb | therve: I'm much, much less worried about hanging on to memory after doing a calculation than I am about loading all of the nested stacks simultaneously | 13:59 |
| therve | zaneb, That's what I thought too, but that's not my testing showed | 13:59 |
| *** jasond has joined #heat | 13:59 | |
| therve | zaneb, With the output over RPC patch, memory is much improved, but I still load the nested stacks | 14:00 |
| therve | The issue is really with the stack.outputs[op].get_value() call | 14:00 |
| therve | Which I guess call function.resolve down the line | 14:00 |
| zaneb | therve: that's pretty much what I would expect, because the maximum number of stacks you have to load *simultaneously* is only the depth of the tree (vs. all of them), and they're spread among multiple engines | 14:01 |
| zaneb | so depth=4, spread across 4 engines, means 1 stack loaded at a time | 14:02 |
| *** dimtruck is now known as zz_dimtruck | 14:03 | |
| zaneb | conversely, all stacks loaded in-memory in a single engine is like 350 stacks loaded at a time | 14:03 |
| therve | Hum, possibly | 14:03 |
| therve | tripleo stack is not very deep though | 14:03 |
| zaneb | yeah, it's only 3 or 4 deep I think | 14:03 |
| therve | But that's a good point, I'd need to check that | 14:04 |
| zaneb | the problem is that you get the outputs by calling self._nested(), which loads the stack in-memory and *keeps it there* | 14:04 |
| zaneb | currently, I mean | 14:05 |
| therve | Yeah, but we still call self.nested() | 14:05 |
| therve | But maybe only the top level one | 14:05 |
| zaneb | and then that stack calls self.nested() in its resources, so they get loaded in memory and kept there too | 14:05 |
| therve | Right, that's an interesting thing to check | 14:06 |
| zaneb | so by the time you're done, you will have basically every stack (at least all of the ones we do get_attr on) loaded at once | 14:06 |
| zaneb | that's why we use so much memory | 14:06 |
| *** EricGonc_ has joined #heat | 14:07 | |
| therve | Yep | 14:07 |
| therve | Usage of self.nested is much bigger than looking at StackResource let me think, too | 14:07 |
| zaneb | so the RPC thing is a good fix, not so much because it spreads the workload (although that's good) as because the memory gets freed after the RPC call returns | 14:07 |
| *** EricGonczer_ has quit IRC | 14:07 | |
| therve | Right exactly | 14:08 |
| *** sorantis has quit IRC | 14:09 | |
| *** ayoung has joined #heat | 14:13 | |
| *** spotz_zzz is now known as spotz | 14:13 | |
| *** zz_dimtruck is now known as dimtruck | 14:13 | |
| *** bnemec has joined #heat | 14:15 | |
| therve | shardy, So interestingly, tripleo completely relies on output validation to be what it is to work at all :) | 14:24 |
| *** rajpatel has joined #heat | 14:26 | |
| shardy | therve: I'm not that surpised tbh, I expected to need some fixes when the heat output validation patches land, but I've not had time to test with them | 14:26 |
| shardy | that's the problem with the current inconsistent validation behavior :( | 14:26 |
| shardy | therve: patches welcome ;) | 14:27 |
| *** Qiming has quit IRC | 14:27 | |
| therve | shardy, I wonder if we should that patch goes in TBH | 14:27 |
| therve | I don't know how we're going to make things work with it | 14:28 |
| shardy | therve: Yeah it definitely runs the risk of breaking currently working things :( | 14:28 |
| shardy | perhaps we can combine enabling output validation with a HOT version bump for ocata? | 14:28 |
| shardy | kind of ugly, but at least then it'd be an opt-in interface change | 14:29 |
| shardy | or an API flag enabling a stricter validation mode I guess | 14:29 |
| therve | shardy, There is a really simple thing when using yaql. If you do data: {something: {get_attr...}}. data is not None, because it's a dict, but you get None in the key | 14:29 |
| therve | So I don't know how we can work around that | 14:29 |
| shardy | therve: IIRC I worked around it by doing the get_attr in a list then filtering for non-None values | 14:30 |
| therve | shardy, Right, you need to handle it in the expression | 14:30 |
| shardy | but yeah, it's kind of a tricky case | 14:30 |
| shardy | yup | 14:30 |
| therve | Urg :) | 14:30 |
| shardy | it'd be much nicer if that wasn't needed tho | 14:30 |
| therve | Yeah, it feels validation is more a burden than a help here. | 14:32 |
| *** Qiming has joined #heat | 14:33 | |
| *** rajpatel1 has joined #heat | 14:33 | |
| *** chris_hultin|AWA is now known as chris_hultin | 14:34 | |
| *** rajpatel has quit IRC | 14:34 | |
| *** Drago has joined #heat | 14:35 | |
| *** Drago has quit IRC | 14:35 | |
| *** annegentle has joined #heat | 14:35 | |
| *** Drago has joined #heat | 14:35 | |
| *** tanvirt has joined #heat | 14:37 | |
| *** openstack has joined #heat | 18:26 | |
| *** openstack has joined #heat | 19:08 | |
| *** sabeen has joined #heat | 19:10 | |
| *** dprince has joined #heat | 19:15 | |
| *** flwang1 has quit IRC | 19:21 | |
| *** spotz is now known as spotz_zzz | 19:21 | |
| *** yprokule has quit IRC | 19:27 | |
| *** jdandrea has joined #heat | 19:29 | |
| *** shardy has joined #heat | 19:33 | |
| *** shardy has quit IRC | 19:33 | |
| *** Zer0Byte__ has joined #heat | 19:34 | |
| *** bana_k has quit IRC | 19:45 | |
| *** Zer0Byte__ has quit IRC | 19:50 | |
| stevebaker | morning | 19:50 |
| *** adrian_otto has joined #heat | 19:59 | |
| *** spzala has joined #heat | 20:00 | |
| *** bana_k has joined #heat | 20:04 | |
| *** randallburt has quit IRC | 20:05 | |
| *** dimtruck is now known as zz_dimtruck | 20:07 | |
| *** dprince has quit IRC | 20:16 | |
| *** spotz_zzz is now known as spotz | 20:19 | |
| *** flwang1 has joined #heat | 20:23 | |
| *** annegentle has quit IRC | 20:26 | |
| *** annegentle has joined #heat | 20:27 | |
| *** blomquisg is now known as blomquisg_gone | 20:27 | |
| *** rajpatel has quit IRC | 20:28 | |
| *** rajpatel has joined #heat | 20:30 | |
| *** annegentle has quit IRC | 20:30 | |
| *** vijendar has quit IRC | 20:33 | |
| *** adrian_otto has quit IRC | 20:37 | |
| *** sorantis has joined #heat | 20:38 | |
| *** adrian_otto has joined #heat | 20:44 | |
| *** adrian_otto has quit IRC | 20:47 | |
| *** Zer0Byte__ has joined #heat | 20:48 | |
| *** adrian_otto has joined #heat | 20:48 | |
| *** jprovazn has quit IRC | 21:03 | |
| *** blomquisg_gone is now known as blomquisg | 21:12 | |
| *** hoonetorg has quit IRC | 21:15 | |
| *** adrian_otto has quit IRC | 21:15 | |
| *** chris_hultin is now known as chris_hultin|AWA | 21:24 | |
| *** hoonetorg has joined #heat | 21:28 | |
| openstackgerrit | Zane Bitter proposed openstack/heat: Avoid loading nested stacks in memory where possible https://review.openstack.org/384718 | 21:32 |
| *** hoonetorg has quit IRC | 21:36 | |
| *** evilt0ne has quit IRC | 21:38 | |
| *** zz_dimtruck is now known as dimtruck | 21:38 | |
| *** tlashchova has quit IRC | 21:41 | |
| *** hoonetorg has joined #heat | 21:53 | |
| *** dfflanders has joined #heat | 21:54 | |
| *** jasond has quit IRC | 22:02 | |
| *** jdandrea has quit IRC | 22:10 | |
| *** sabeen has quit IRC | 22:15 | |
| *** jamielennox is now known as jamielennox|away | 22:32 | |
| *** spzala has quit IRC | 22:33 | |
| *** spzala has joined #heat | 22:36 | |
| *** rajpatel has quit IRC | 22:37 | |
| *** rajpatel has joined #heat | 22:38 | |
| *** spzala has quit IRC | 22:39 | |
| *** sdake has quit IRC | 22:44 | |
| *** jamielennox|away is now known as jamielennox | 22:48 | |
| *** gus_ is now known as gus | 22:52 | |
| *** EricGonc_ has quit IRC | 22:58 | |
| *** vijendar has joined #heat | 22:58 | |
| *** EricGonczer_ has joined #heat | 23:01 | |
| *** sabeen has joined #heat | 23:01 | |
| *** vijendar has quit IRC | 23:04 | |
| *** vijendar has joined #heat | 23:08 | |
| *** sdake has joined #heat | 23:12 | |
| *** sdake_ has joined #heat | 23:17 | |
| *** adrian_otto has joined #heat | 23:19 | |
| *** hoonetorg has quit IRC | 23:20 | |
| *** sdake has quit IRC | 23:21 | |
| *** sorantis has quit IRC | 23:31 | |
| *** hoonetorg has joined #heat | 23:35 | |
| *** asalkeld has joined #heat | 23:38 | |
| *** rajpatel has quit IRC | 23:42 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!