Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-07-17
18:03:47 mriedem like sahid or vladikr?
18:03:54 mriedem i don't know how this was actually working before
18:04:21 openstackgerrit Matt Riedemann proposed openstack/nova master: Convert HostState.limits objects to primitives https://review.openstack.org/484439
18:05:56 mriedem was conductor serializing that response before for us so the NUMATopologyLimits was getting serialized by conductor?
18:07:47 oomichi alex_xu: most patches are approved on https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/api-no-more-extensions-pike
18:08:09 oomichi alex_xu: need to update the remaining for conflicts
18:31:24 openstackgerrit Matt Riedemann proposed openstack/nova master: Log compute node uuid when the record is created https://review.openstack.org/484447
18:52:46 mriedem dansmith: sfinucan: nvm i think i know how that worked before - jsonutils.to_primitive() wouldn't error out, but wouldn't convert the NUMATopologyLimits either, so it would go through NovaObjectSerializer to do the serialize/deserialize work on the object
18:53:08 mriedem the new change in to_primitive to raise ValueError is what blew it all up because it would try to serialize before hitting the NovaObjectSerializer
18:53:22 dansmith sorry I missed the above ping,
18:53:25 dansmith but I assume I can ignore now
18:53:53 mriedem validation of my understanding, and feelings, is always appreciated, but yeah you can ignore probably
18:54:26 dansmith jsonutils.to_primitive used to fail with "recursion error" or something like that when it hit a non-serializeable object
18:56:15 dansmith mriedem: what change introduced this?
18:56:30 dansmith oh I see, in oslo
18:58:29 mriedem yeah that's what the oslo.serialization change was trying to prevent, the infinite recursion
18:58:37 mriedem by just failing fast
18:59:29 dansmith mriedem: so the thing I'm concerned about is: are we returning something that is only using jsonutils for serialization?
18:59:45 dansmith our object serializer doesn't go hunting in an iterable for any instance of an object to serialize
19:00:55 dansmith doesn't go arbitrarily deep, I mean
19:01:14 mriedem hmm yeah because that's what i was looking at with https://github.com/openstack/nova/blob/master/nova/objects/base.py#L244
19:01:24 mriedem so it goes 1 level?
19:01:32 mriedem well, actually,
19:01:55 mriedem so this would be a like {'limits': {'numa_topology': object}}
19:01:59 mriedem so that should hit it ^
19:02:24 dansmith where is this exactly?
19:02:29 dansmith is this in a returned value?
19:02:32 mriedem yeah
19:02:54 mriedem https://review.openstack.org/#/c/484439/2/nova/scheduler/manager.py@140
19:03:33 mriedem creates a list of these dicts, and each dict will have a 'limits' key that maps to a dict, and that thing has the numa_topology_limits=NUMATopologyLimits object
19:03:34 mriedem map
19:05:02 openstackgerrit Matt Riedemann proposed openstack/nova master: Optimize creating servers with tags https://review.openstack.org/484170
19:05:18 dansmith I guess it will go pretty deep
19:05:24 dansmith had to re-read that a couple times
19:05:43 mriedem yeah so i think that's how it wasn't blowing up before
19:05:52 mriedem well, and was actually serializing the thing
19:06:07 dansmith yeah
19:06:10 mriedem the fact jsonutils.to_primitive ignored it was to our advantage
19:06:25 dansmith yeah
19:06:58 mriedem that change is also being reverted https://review.openstack.org/#/c/484396/
19:07:18 mriedem b/c i pointed out that releasing that under a patch release versoin w/o any prior warning <2 weeks to FF is likely to break other projects
19:07:19 dansmith heh
19:07:25 dansmith yeah I guess so
19:07:53 mriedem we only had prior warning b/c the requirements changes gate on nova unit/functional tests
19:25:51 mriedem oomichi: i'm +2 again on https://review.openstack.org/#/c/479774/ after your comments
19:39:33 Ctesias hi. I'm digging around in nova trying to figure out why my nova_override.conf file isn't overriding nova defaults and I'm curious where `import nova.conf` and `CONF = nova.conf.CONF` are coming from because I don't see those being defined anywhere
19:41:28 mriedem Ctesias: for nova-compute the options are loaded up here https://github.com/openstack/nova/blob/master/nova/cmd/compute.py#L42
19:41:45 mriedem similar pattern for the other service entry points
19:41:59 mriedem https://github.com/openstack/nova/blob/master/nova/config.py#L49
19:43:15 Ctesias Thanks for the pointer.
19:43:31 Ctesias Is putting an override_nova.conf file in the config directory the right way to get overrides into the nova config?
19:44:10 mriedem i've never used nova_override.conf
19:44:17 mriedem so if it works, it's some magic i'm not aware of
19:45:14 Ctesias well... It doesn't work :(
19:45:27 Ctesias Any tips on how I would go about overriding defaults in nova.conf the right way?
19:45:56 mriedem can't you pass overrides on the --config-file cli option
19:46:02 mriedem i'm not sure though
19:46:38 mriedem there could be something in https://docs.openstack.org/oslo.config/latest/
19:47:40 sdague Ctesias: can you explain what you are trying to do with overrides besides modifying the nova.conf directly? Maybe we're just not understanding the question
19:49:49 Ctesias I was under the impression modifying nova.conf directly was no bueno. But I'm trying to enable a barbican with Nova, so I need to set the api_class setting in config file under [keymgr]
19:50:21 Ctesias I created nova_override.conf in the same directory where nova.conf was and put [keymgr] and api_class = <the barbican manager> in it
19:50:24 Ctesias restarted nova
19:50:30 Ctesias but nova had the defaults when it came back
19:51:33 mriedem just modify nova.conf
19:52:00 mriedem nova.conf can be totally blank and it won't override any defaults - the defaults are all in code
19:52:08 mriedem so if you want to override any defaults, it happens in nova.conf
19:52:25 mriedem nova_override.conf is probably some package-specific trick
19:52:33 Ctesias .....Oh.
19:52:46 Ctesias Well awesome. Thanks for your help, really
19:52:50 mriedem yw
19:52:51 Ctesias I've been pulling my hair out for days
20:00:12 openstackgerrit Matt Riedemann proposed openstack/nova master: Support tag instances when boot(4/4) https://review.openstack.org/469800
20:00:55 mriedem Kevin_Zheng: ^ fixed up the functional test failures; it was pretty simple but not obvious until i injected some debug code and ran the tests, and then also had to tweak the tests a bit
20:03:30 mriedem sdague: can you take a gander at ^ ? it's pretty straight-forward and well tested
20:05:49 openstackgerrit Matt Riedemann proposed openstack/nova master: Optimize creating servers with tags https://review.openstack.org/484170
20:06:38 sdague mriedem: I'll add it to the morning list
20:06:56 mriedem thanks - alex_xu might get it overnight then, he was reviewing it before too
20:08:24 oomichi mriedem: thanks, +2
20:10:10 Ctesias Hi again. I'm getting the same behavior with the changes directly in my nova.conf file. I've tried pointing at the directory with --config-dir and directly at the nova.conf file with --config-file with no change in behavior
20:10:20 Ctesias the api_class value is still the default
20:11:19 mriedem Ctesias: which service?
20:11:33 Ctesias nova service
20:11:45 mriedem which of the nova services, there are several
20:11:50 mriedem nova-osapi_compute?
20:11:57 mriedem nova-compute, etc
20:12:21 Ctesias nova-compute
20:13:59 mriedem when you restart nova-compute, do you have debug logging enabled so you can see it dump which config files are being used to and it will dump the options
20:16:46 Ctesias the file is pretty long, there something I can grep for to get the file it's using?
20:17:57 mriedem yeah
20:17:58 mriedem "command line args"
20:18:02 mriedem "config files"
20:18:07 mriedem like here http://logs.openstack.org/47/484447/1/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/b2fbd0c/logs/screen-n-cpu.txt.gz#_Jul_17_18_47_11_345788
20:18:22 mriedem and then it dumps the options it's using
20:19:05 mriedem Ctesias: you're using the wrong group
20:19:09 mriedem it's not [keymgr]
20:19:13 mriedem it's [key_manager]
20:19:21 mriedem http://logs.openstack.org/47/484447/1/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/b2fbd0c/logs/screen-n-cpu.txt.gz#_Jul_17_18_47_11_670125
20:19:28 Ctesias yeah, that was a typo earlier. I've got [key_manager] in the file
20:21:05 Ctesias If I put a breakpoint in nova/keymgr/conf_key_mgr.py I can go to the line after it grabs the nova configuration and see it's the same
20:21:10 Ctesias (Pdb) CONF['key_manager']['api_class'] 'nova.keymgr.conf_key_mgr.ConfKeyManager'

Earlier   Later