Thursday, 2015-06-04

« Wednesday, 2015-06-03 Index Friday, 2015-06-05 »
*** alexpilotti has quit IRC00:03
openstackgerritBen Nemec proposed openstack/oslo.messaging: Remove leftover oslo.config reference  https://review.openstack.org/18821400:04
*** sputnik13 has joined #openstack-oslo00:05
*** sputnik13 has quit IRC00:07
openstackgerritJoshua Harlow proposed openstack/taskflow: Refactor machine builder + runner into single unit  https://review.openstack.org/18821700:07
*** sputnik13 has joined #openstack-oslo00:09
*** sputnik13 has quit IRC00:10
*** jaypipes has quit IRC00:27
*** stpierre has quit IRC00:29
openstackgerritJoshua Harlow proposed openstack/futurist: Remove dependency on oslo.utils (replace with small util code)  https://review.openstack.org/18780400:34
*** vilobhmm1 has quit IRC00:40
openstackgerritJames Polley proposed openstack-dev/pbr: WIP: more extensive testing of requirement parsing  https://review.openstack.org/18782300:43
openstackgerritJames Polley proposed openstack-dev/pbr: Fix everything by requiring setuptools>=16.  https://review.openstack.org/18784600:43
*** gordc has joined #openstack-oslo00:43
*** bknudson has joined #openstack-oslo00:43
*** takedakn has joined #openstack-oslo00:53
*** takedakn has quit IRC00:54
*** takedakn has joined #openstack-oslo00:54
*** kgiusti has quit IRC00:59
*** zzzeek has quit IRC01:04
*** takedakn has quit IRC01:06
*** takedakn has joined #openstack-oslo01:06
*** rjaiswal has joined #openstack-oslo01:12
rjaiswalDoes oslo.config has a way of reloading a configuration file?01:13
*** openstack has joined #openstack-oslo01:22
dims_search that page for the method01:22
*** Kennan has joined #openstack-oslo01:22
rjaiswalthanks dims01:23
*** openstack has joined #openstack-oslo01:37
*** yamahata has quit IRC01:46
openstackgerritJoshua Harlow proposed openstack/futurist: Remove dependency on oslo.utils (replace with small util code)  https://review.openstack.org/18780401:47
openstackgerritJoshua Harlow proposed openstack/futurist: Rename futures.py -> _futures.py and expose at futurist module  https://review.openstack.org/18738201:47
*** boris-42 has quit IRC01:48
*** openstack has quit IRC01:52
*** openstack has joined #openstack-oslo01:54
rjaiswaldims: is it possible to just reload a single configuration file using reload_config_files?01:54
openstackgerritJames Polley proposed openstack-dev/pbr: WIP: more extensive testing of requirement parsing  https://review.openstack.org/18782301:58
openstackgerritJames Polley proposed openstack-dev/pbr: Fix everything by requiring setuptools>=16.  https://review.openstack.org/18784601:58
openstackgerritMerged openstack/oslo.messaging: Remove leftover oslo.config reference  https://review.openstack.org/18821402:01
*** _amrith_ is now known as amrith02:02
tchaypolifeless: ^^^ i still expect that to fail in the 3 places - as far as i can tell setuptools 16 doesn’t actually fix things. But I’m now happy that we’re doing a proper test  (with the in-tree pbr version) and I’ll be able to tell if any fix I come up with works02:05
tchayposo now all I need to do is find a fix..02:05
lifelesstchaypo: so, it worked for me :)02:15
lifelesstchaypo: let me know when theres a failed run to poke at02:16
rjaiswaldims_: don't think reload_co fig_files can be used to reload a specific config file?02:19
*** gcb has quit IRC02:35
*** ashleighfarnham has joined #openstack-oslo02:36
*** mriedem_away has quit IRC02:39
*** bknudson has quit IRC02:39
openstackgerritJoshua Harlow proposed openstack/taskflow: Use a lru cache to limit the size of the internal file cache  https://review.openstack.org/17610402:40
*** ashleighfarnham has quit IRC02:41
*** sputnik13 has joined #openstack-oslo02:47
*** stevemar has joined #openstack-oslo02:58
*** yamahata has joined #openstack-oslo03:00
*** sputnik13 has quit IRC03:01
*** sputnik13 has joined #openstack-oslo03:02
*** dims_ has quit IRC03:07
openstackgerritJoshua Harlow proposed openstack/tooz: Add watch file driver support  https://review.openstack.org/18824603:08
lifelesstchaypo: indeed, it failed only on python2603:09
*** sputnik13 has quit IRC03:15
*** takedakn has quit IRC03:28
*** sputnik13 has joined #openstack-oslo03:35
*** sputnik13 has quit IRC03:42
tchaypoho hum.03:45
tchaypoon the python-2.6 output03:46
tchaypoSuccessfully installed pbr-0.003:46
tchaypothat’s not the version number I expected to see03:47
*** sputnik13 has joined #openstack-oslo03:48
*** sputnik13 has quit IRC03:49
*** gcb has joined #openstack-oslo03:49
*** sputnik13 has joined #openstack-oslo03:50
*** tobe has joined #openstack-oslo03:51
*** EmilienM|off has quit IRC03:55
openstackgerritJoshua Harlow proposed openstack/taskflow: Use a lru cache to limit the size of the internal file cache  https://review.openstack.org/17610404:00
tchaypoin any case, it’s not a very good test yet. I wouldn’t say that it “didn’t fail”, just that it “didn’t report an Invalid Environment Marker"04:00
*** EmilienM has joined #openstack-oslo04:01
*** gordc has quit IRC04:03
lifelesstchaypo: well true; so three issues; install the pbr we have (for the installed-pbr cases); figure out what sup with 2.6; make the case actually check the desired thing is installed04:05
*** sdake has joined #openstack-oslo04:05
lifelesstchaypo: the 0.0 thing indicates a git tree with no tags04:06
lifelesstchaypo: not sure why pbr's tree would be like that04:06
lifelessmust run; ciao (available on hangouts if you need to brain)04:06
*** dims_ has joined #openstack-oslo04:07
*** dims_ has quit IRC04:13
*** sputnik13 has quit IRC04:14
*** krotscheck has quit IRC04:19
tchaypoand if possible, figure out why it breaks on my system differently to how it breaks under jenkins..04:19
*** mordred has quit IRC04:21
*** sputnik13 has joined #openstack-oslo04:23
*** mordred has joined #openstack-oslo04:25
*** krotscheck has joined #openstack-oslo04:25
*** sdake has quit IRC04:34
*** rushiagr_away is now known as rushiagr04:42
*** rpodolyaka1 has joined #openstack-oslo04:47
*** rpodolyaka1 has quit IRC04:50
openstackgerritJoshua Harlow proposed openstack/taskflow: Use the sqlalchemy-utils json type instead of our own  https://review.openstack.org/18825804:56
openstackgerritJoshua Harlow proposed openstack/taskflow: Use the sqlalchemy-utils json type instead of our own  https://review.openstack.org/18825804:57
*** rpodolyaka1 has joined #openstack-oslo04:59
*** stevemar has quit IRC05:04
*** tsekiyama has quit IRC05:10
openstackgerritJames Polley proposed openstack-dev/pbr: WIP: more extensive testing of requirement parsing  https://review.openstack.org/18782305:14
openstackgerritJames Polley proposed openstack-dev/pbr: Fix everything by requiring setuptools>=16.  https://review.openstack.org/18784605:14
*** sputnik13 has quit IRC05:17
*** tobe has quit IRC05:22
openstackgerritJoshua Harlow proposed openstack/futurist: Put features into the README.rst  https://review.openstack.org/18772705:23
*** tobe has joined #openstack-oslo05:40
*** rpodolyaka1 has quit IRC05:49
openstackgerritMerged openstack/oslo.utils: Replace parse_strtime with parse_isotime in older/newer  https://review.openstack.org/18800906:07
openstackgerritOpenStack Proposal Bot proposed openstack/oslo.messaging: Imported Translations from Transifex  https://review.openstack.org/18827006:08
openstackgerritMerged openstack/oslo.utils: timeutils: fix newer/older comparison with TZ aware datetime  https://review.openstack.org/18801006:17
*** tobe has quit IRC06:18
*** flaper87 has quit IRC06:33
*** flaper87 has joined #openstack-oslo06:33
*** exploreshaifali has joined #openstack-oslo06:42
*** rpodolyaka1 has joined #openstack-oslo06:51
*** rpodolyaka1 has quit IRC06:52
*** amotoki has joined #openstack-oslo06:52
*** ndipanov has quit IRC06:53
*** rpodolyaka1 has joined #openstack-oslo06:54
*** tobe has joined #openstack-oslo06:57
*** amotoki has quit IRC07:03
jd__dhellmann: it's documented in the docstring IIRC, I told them on the review they were doing it wrong, they ignored it it seems07:19
openstackgerritJames Polley proposed openstack-dev/pbr: Fix everything by requiring setuptools>=16.  https://review.openstack.org/18784607:20
*** Kiall has quit IRC07:24
*** Kiall has joined #openstack-oslo07:25
*** amotoki has joined #openstack-oslo07:27
*** exploreshaifali has quit IRC07:31
*** dims_ has joined #openstack-oslo07:44
*** dims_ has quit IRC07:49
*** rpodolyaka1 has quit IRC07:51
*** ihrachyshka has joined #openstack-oslo07:54
*** ndipanov has joined #openstack-oslo08:07
openstackgerritMerged openstack/taskflow: Replace lock_utils lock(s) with fasteners package  https://review.openstack.org/18524908:11
flaper87sileht: the py-amqp patch landed08:13
flaper87sileht: anything else you need from me on the kombu side?08:13
*** yamahata has quit IRC08:14
silehtflaper87, I don't think so08:14
*** yassine has joined #openstack-oslo08:16
silehtflaper87, thanks !08:17
flaper87sileht: kk, lemme know08:18
*** amotoki has quit IRC08:20
*** amotoki has joined #openstack-oslo08:20
*** rpodolyaka1 has joined #openstack-oslo08:31
openstackgerritMerged openstack/tooz: Use MySQL default port when not set explicitly  https://review.openstack.org/18798208:35
*** tobe has quit IRC08:39
*** tobe has joined #openstack-oslo08:39
*** gcb has quit IRC08:41
*** ozamiatin has joined #openstack-oslo08:51
*** salv-orlando has quit IRC09:08
*** amotoki has quit IRC09:08
*** cdent has joined #openstack-oslo09:21
*** rpodolyaka1 has quit IRC09:23
*** rpodolyaka1 has joined #openstack-oslo09:27
*** tobe has quit IRC09:33
*** amotoki has joined #openstack-oslo09:38
*** amotoki has quit IRC09:42
*** e0ne has joined #openstack-oslo09:44
*** jaosorior has joined #openstack-oslo09:45
*** dims_ has joined #openstack-oslo09:45
*** yassine has quit IRC09:49
*** dims_ has quit IRC09:51
*** tobe has joined #openstack-oslo09:53
*** dims_ has joined #openstack-oslo10:03
*** rpodolyaka1 has quit IRC10:03
*** salv-orlando has joined #openstack-oslo10:09
*** salv-orlando has quit IRC10:17
*** Kennan2 has joined #openstack-oslo10:20
*** Kennan has quit IRC10:20
*** salv-orlando has joined #openstack-oslo10:20
*** ozamiatin has quit IRC10:21
*** salv-orlando has quit IRC10:21
*** Kennan2 is now known as Kennan10:21
*** salv-orlando has joined #openstack-oslo10:23
*** boris-42 has joined #openstack-oslo10:33
dims_@eezhova : ping11:09
*** yassine_ has joined #openstack-oslo11:09
eezhovadims_, pong11:09
dims_eezhova: what else needs to be done on oslo.service? so we can try a test release to see if the CI systems and pypi work?11:09
eezhovadims_, we have to fix its API and remove usage of global config https://github.com/openstack/oslo-specs/blob/master/specs/liberty/graduate-oslo-service.rst#work-items11:11
eezhovadims_, does a "test release" mean that a library can be not yet ready for adoption at that point of time?11:13
*** exploreshaifali has joined #openstack-oslo11:15
*** shardy_ has joined #openstack-oslo11:21
*** shardy has quit IRC11:22
*** rpodolyaka1 has joined #openstack-oslo11:24
*** shardy_ has quit IRC11:26
*** shardy has joined #openstack-oslo11:27
*** e0ne is now known as e0ne_11:39
*** e0ne_ has quit IRC11:49
*** ozamiatin has joined #openstack-oslo11:51
*** ihrachyshka_ has joined #openstack-oslo12:02
dims_eezhova: right12:02
*** ihrachyshka has quit IRC12:03
dims_may be we can do it in a week or so eezhova12:04
eezhovadims_, then it's definitely worth a try12:04
eezhovadims_, okay12:04
*** tobe has quit IRC12:06
*** e0ne has joined #openstack-oslo12:13
*** mtanino has joined #openstack-oslo12:16
*** ihrachyshka has joined #openstack-oslo12:20
*** ihrachyshka_ has quit IRC12:23
openstackgerritDavanum Srinivas (dims) proposed openstack/oslo-incubator: Add new projects to dashboard  https://review.openstack.org/18760312:24
*** stpierre has joined #openstack-oslo12:26
dims_team, updated the dashboards with new projects etc https://wiki.openstack.org/wiki/Oslo#Review_Links12:31
dims_these were the changes - https://review.openstack.org/#/c/187603/12:31
*** gordc has joined #openstack-oslo12:32
*** kgiusti has joined #openstack-oslo12:35
*** rushiagr is now known as rushiagr_away12:36
*** stevemar has joined #openstack-oslo12:41
*** zzzeek has joined #openstack-oslo12:42
*** bknudson has joined #openstack-oslo12:43
*** rpodolyaka1 has quit IRC12:51
openstackgerritCedric Brandily proposed openstack/oslo.db: Allow to fail instead of skip in DbFixture  https://review.openstack.org/18771712:58
*** e0ne is now known as e0ne_13:01
*** e0ne_ is now known as e0ne13:04
*** mtanino has quit IRC13:05
*** mriedem has joined #openstack-oslo13:15
*** rpodolyaka1 has joined #openstack-oslo13:20
*** rpodolyaka1 has quit IRC13:20
*** rushiagr_away is now known as rushiagr13:28
*** sputnik13 has joined #openstack-oslo13:29
dims_folks, looks like we need eyes on zzzeek's enginefacade work in oslo.db - https://review.openstack.org/#/c/138215/13:31
*** amrith is now known as _amrith_13:35
dims_zzzeek: here's a devstack change that runs oslo.db master + your review - https://review.openstack.org/#/c/188395/13:40
dims_zzzeek: if you need changes in nova, thrown in a nova review and add another Depends-On for that new review in  https://review.openstack.org/#/c/188395/13:40
dims_zzzeek: ok?13:40
dims_zzzeek: you can type 188395 in http://status.openstack.org/zuul/ to watch the jobs13:41
openstackgerritMarian Horban proposed openstack/oslo-incubator: Graceful shutdown WSGI server  https://review.openstack.org/18653313:41
zzzeekdims_: OK so the depends-on thing only applies to the “devstack” project ?13:42
*** rpodolyaka1 has joined #openstack-oslo13:43
*** rpodolyaka1 has quit IRC13:48
dims_zzzeek: you can add a depends-on in any project on any other project, however it may not work in some circumstances13:49
*** rpodolyaka1 has joined #openstack-oslo13:50
zzzeekdims_: thanks13:50
*** rpodolyaka1 has quit IRC13:50
dims_zzzeek: since i know i can run devstack with oslo.db master using LIBS_FROM_GIT (see stackrc) i started there, and then i said, i need this other oslo.db review so i threw in a Depends-On13:50
dims_by default that will use nova master, but you will probably need changes in nova. so then make those changes in another review against nova and add the depends-on in the devstack review13:51
dims_makes sense?13:51
dims_zzzeek: ^^13:52
zzzeekdims_: yes13:52
dims_cool. stepping out for a bit13:53
*** sigmavirus24_awa is now known as sigmavirus2413:55
*** e0ne is now known as e0ne_13:59
zzzeekok13:59
mdboothzzzeek:     def _transaction_scope(self, context):13:59
mdbooth        if self._replace:13:59
mdbooth            restore_factory = _context_manager._root._root_factory13:59
mdbooth            _context_manager._root._root_factory = self._root._root_factory13:59
mdbooth_context_manager is global13:59
zzzeekyes13:59
zzzeeki just commented on that13:59
mdboothYou did?14:00
zzzeekso, ill make that tlocal and we’re done ?14:00
zzzeekyes14:00
zzzeekin the gerrit14:00
zzzeekdoes cells work this way, that the use of a different DB is local to an initial method ?14:00
zzzeeklike   def do_something(self):  with now_we_will_use_cells:  do_something_else()14:01
mdboothWell, I think we're already constraining a context to a single thread of execution, right?14:01
mdboothSo as long as the switch is confined to the current thread I can't see how it could interfere with anything else under any circumstances14:01
zzzeekmdbooth: ill make that change (somehow)14:02
zzzeekmight have to make _root a tlocal.14:02
zzzeekits been awhile14:02
mdboothzzzeek: I was also mulling passing in a specific config root14:02
mdboothzzzeek: Yeah, me too14:02
mdboothLet me find the code which explains what I'm talking about14:03
zzzeekmdbooth: OK like “with enginefacade.replace(my_other_factory).using(context): “ ?14:03
mdboothflying by my notoriously poor memory here...14:04
*** e0ne_ has quit IRC14:04
zzzeekmdbooth: acutally “with alt_mgr1.replace.using(context):” shuld be able to just put something on the context14:04
mdboothSo right now oslo.db essentially hardcodes the location of config in nova.conf14:04
dhellmannjd__: sorry, I lost the thread you were replying to in my scrollback. What docstring/API is being used wrong? :-/14:05
mdboothin [db] perhaps, or [default]14:05
mdboothWhatever14:05
mdboothenginefacade always pulls it from the same location in the current config14:05
jd__dhellmann: isotime() IIRC14:05
mdbooththe cells stuff defines a parallel config in [api]14:05
openstackgerritOpenStack Proposal Bot proposed openstack/taskflow: Updated from global requirements  https://review.openstack.org/18840914:06
*** e0ne has joined #openstack-oslo14:06
dhellmanndims_, eezhova : if you release oslo.service, someone is going to start using it. You probably want to fix up the API first.14:06
mdboothIt would be trivial for the library to be passed an alternate config key14:06
mdboothAnd default it to the current14:06
mdboothiirc (not a given)14:06
mdboothLook in nova/db/sqlalchemy/api.py14:07
mdboothdef _create_facade(conf_group):14:07
zzzeekmdbooth: OK as far as config file that gets intoi oslo.config14:08
zzzeekmdbooth: we hvae a CONF and we look at CONF.database14:08
mdboothThat's it14:08
mdboothif that could be CONF.<foo>14:08
mdboothdefaulting to database14:08
zzzeekmdbooth: so….do we look at CONF.api or just a different “CONF” and still look at .database14:08
zzzeekmdbooth: OK immmm not sure how oslo.config works in this regard14:08
mdboothIf we could just pass it in, that would be great14:08
zzzeekmdbooth: these config keys all have to be set up14:08
mdboothYes14:08
zzzeekmdbooth: its a very inconvenient system14:08
mdboothIndeed14:08
zzzeekmdbooth: but also this is kind of a differnt thing compared to just getting “we want a differnt config in this context” to work for starters14:09
zzzeekmdbooth: its an addition to that14:09
mdboothIncidentally, I hit the problem with the 'load config on demand' approach which you've copied to preserve the previous behaviour14:09
*** sdake has joined #openstack-oslo14:09
mdboothYou can't override it in a unit test, because it hasn't been loaded yet14:09
dhellmannjd__: ok, we need to figure out the backwards-compatibility story for REST APIs that use that method to produce date strings. IIRC, the original issue was the fact that we had to move some of your code in oslo.serialization off of master into a feature branch.14:10
dhellmannmdbooth: why do applications want to control the location of values in the config file?14:10
mdboothzzzeek: Yes, this is not a fundamental issue. The global context was a genuine problem, whereas this would just be tidier.14:10
zzzeekmdbooth: OK so yes the CONF thing we just need to say conf.register_opts(options.database_opts, ‘api’) somewhere14:10
*** sdake_ has joined #openstack-oslo14:11
zzzeekmdbooth: and not on every request :)14:11
mdboothzzzeek: It's normally done at module level, so it happens on import14:11
zzzeekmdbooth: fine14:11
mdboothdhellmann: It's still hardcoded, but in this case the application wants to use multiple configurations14:12
mdboothSo the location should be hardcoded by the application, not the library14:12
dhellmannmdbooth: ah, ok, if you want to connect to multiple databases that makes more sense14:12
*** sputnik13 has quit IRC14:12
eezhovadhellmann, I have already started working on making necessary changes. Can I add you as the reviewer so you can probably take a look? There already is a couple of patches on review.14:13
dhellmanneezhova: I'll take a quick look now14:14
mdboothzzzeek: Ideally this would mean that oslo_db would expose a library call to register config variables under a specific key14:14
mdboothWhich would then be called by the application14:14
*** sdake has quit IRC14:15
mdboothSo in nova we'd have, at module level, dbconf = oslo_db.register_conf('database')14:15
mdboothapiconf = oslo_db.register_conf('api')14:15
mdboothor whatever14:15
dhellmannmdbooth: why wouldn't the engine just do that when it's called?14:15
mdboothdhellmann: That's what it does right now14:15
dhellmannso it just needs the name of the group to use, right?14:15
mdboothThe only issue I've hit so far with that is with unit tests14:16
*** mtanino has joined #openstack-oslo14:16
mdboothConfig is normally loaded at import14:16
mdboothWhich means that a unit test can tweak bits14:16
mdboothBut if you load config when the function runs, you can't set its values beforehand14:16
dhellmannwe have a fixture for that14:16
mdboothThere may be other reasons14:16
mdboothIn that case it may be moot :)14:17
*** rpodolyaka1 has joined #openstack-oslo14:17
mdboothHowever, it does seem to be the pattern14:17
dhellmannhrm, that fixture isn't in the docs, hang on14:17
dhellmannoops, looking in the wrong place, here it is: http://docs.openstack.org/developer/oslo.config/fixture.html14:18
mdboothThe other issue is determinism14:18
mdboothIf you load the config at import time, you can be sure that if you query a value it has been loaded14:18
mdboothotherwise you can be caught tearing your hair out trying to work out why your values contain None when the config value is clearly set14:19
mdboothAnd it's just an ordering thing14:19
* mdbooth points to patches of missing hair :)14:19
*** dulek has joined #openstack-oslo14:19
openstackgerritCedric Brandily proposed openstack/oslo.db: Allow to fail instead of skip in DbFixture  https://review.openstack.org/18771714:20
*** mtanino_ has joined #openstack-oslo14:21
dhellmannmdbooth: the libraries try to register options right before they use them, and rely on the app to have read the config file. That lets us populate a configobj in a test setup method.14:21
mdboothzzzeek dhellmann: All of the above is besides the point, btw. The principal would be to avoid having different code for the 'default' and 'api' db initialisations just because one used the default config location and the other didn't.14:21
mdboothAny means of achieving that is fine, and the above is orthogonal to that.14:21
*** jecarey has joined #openstack-oslo14:21
dhellmannmdbooth: right, that's why I was suggesting it makes more sense to pass the "profile name" or whatever into the engine, and let the engine worry about registering options14:21
*** mtanino has quit IRC14:23
mdboothdhellmann: I expect you've given this way more thought than me :)14:23
dhellmannmdbooth: we have done this once or twice before, yes :-)14:23
*** sdake has joined #openstack-oslo14:25
*** sputnik13 has joined #openstack-oslo14:27
*** sdake_ has quit IRC14:28
dhellmanneezhova: https://review.openstack.org/#/c/186360/ is phase 1, right? later changes will move the config object to be an argument in the API?14:30
eezhovadhellmann, yes that is right14:31
dhellmanneezhova: ok, good, just double-checking :-)14:31
*** dulek has quit IRC14:32
*** dulek_ has joined #openstack-oslo14:32
*** dulek_ has quit IRC14:32
*** _amrith_ is now known as amrith14:32
*** dulek_ has joined #openstack-oslo14:32
dhellmanneezhova: ok, I think the options moves are the inverse of what we want. We would like to have a single entry point that returns all of the options for oslo.service so that projects using oslo.service can just add "oslo.service" to the list of namespaces in their config generator configuration file.14:35
dhellmanneezhova: we aren't always consistent with how we implement that in the libraries, from what I see14:35
dhellmanneezhova: hmm, although maybe we don't expect apps to always use all of the service features, like the backdoor?14:36
dhellmannin which case having more than one entry point might makes sense...14:36
dhellmanndims_: what do you think? ^^14:37
eezhovadhellmann, In https://review.openstack.org/#/c/187607/1 I have made a single entrypoint for getting a list of options available in the library14:38
dhellmanneezhova: ah! ok, in that commit you could remove the other list_opts functions, since that's what those were for in the incubated code14:38
*** mtanino_ has quit IRC14:39
eezhovadhellmann, okay, I wanted to remove them in a follow-up patch, but it's really better to do it in https://review.openstack.org/#/c/187607/114:40
* dhellmann shrugs14:40
*** exploreshaifali has quit IRC14:41
dhellmanneezhova: +2 on both14:41
*** mtanino has joined #openstack-oslo14:43
dims_dhellmann: +1 to multiple entrypoints14:46
dims_eezhova: either way is ok i guess14:47
eezhovadims_, dhellmann: I've updated https://review.openstack.org/#/c/187607 so that it also removes multiple entrypoints. But either way is ok for me too, I can revert it if needed.14:49
dims_eezhova: let's leave it alone :)14:50
dims_i mean leave your review as is14:50
eezhovadims_, ok, got it:) just curious: if we have multiple entrypoints then we have to list them one by one in setup.cfg?14:51
dims_eezhova: yes14:51
*** rpodolyaka1 has quit IRC14:53
*** sputnik13 has quit IRC14:55
*** sdake has quit IRC15:00
*** ozamiatin has quit IRC15:03
*** alexpilotti has joined #openstack-oslo15:11
dhellmanndims_: the only benefit I can see to having >1 ep is if a project uses the service library, but doesn't use some parts of it (like the backdoor, or ssl)15:15
dhellmannI don't know what the usage patterns look like, so I don't know if we have that case or not15:15
*** amrith is now known as _amrith_15:19
dims_dhellmann: right, eezhova do you want to check quickly?15:27
dhellmanndims_, eezhova : good idea15:28
eezhovadims_, dhellmann: I can say for sure that not so many projects use sslutils. I've found only sahara to do so for now15:29
eezhovadims_, dhellmann: will check for backdoor15:30
openstackgerritMerged openstack/oslo.service: Move the option definitions into a private file  https://review.openstack.org/18636015:30
dhellmanneezhova: ok, it sounds like maybe separate entry points, with some good documentation to remind consuming projects to update their config generator file, is the right approach15:31
dims_dhellmann: i had a dashboard update when you get a chance - https://review.openstack.org/#/c/187603/15:31
*** dulek_ has quit IRC15:32
* dims_ stepping out for a bit. will be off and on this afternoon15:34
dhellmanndims_: +215:34
eezhovadhellmann, dims_ : okay, I got it15:37
eezhovadhellmann, dims_ : it seems to me that eventlet_backdoor is used only by service.py15:37
dhellmanneezhova: is it always turned on?15:38
viktorsdims_: hi! Can you please look at https://review.openstack.org/#/c/187963/  ?15:39
viktorsit should be an easy one15:40
dims_viktors: heading out. will do it in a couple of hours15:41
eezhovadhellmann, it's used only here by the Launcher class (https://github.com/openstack/oslo.service/blob/master/oslo_service/service.py#L100)15:41
eezhovadhellmann, and this line executes if we launch something with ServiceLauncher https://github.com/openstack/oslo.service/blob/master/oslo_service/service.py#L50115:41
*** viktors is now known as viktors|afk15:43
*** dims_ has quit IRC15:43
openstackgerritMichael Bayer proposed openstack/oslo.db: Implement new oslo.db.sqlalchemy.enginefacade module  https://review.openstack.org/13821515:43
bknudsonnotifications question related to using rabbit ... can the keystore have a password?15:47
bknudsonI don't see a config option for the keystore password15:49
openstackgerritJeremy Stanley proposed openstack-dev/pbr: More explicit data_files install location docs  https://review.openstack.org/18845115:54
*** sdake has joined #openstack-oslo16:02
*** ihrachyshka has quit IRC16:04
openstackgerritJoshua Harlow proposed openstack/tooz: Ensure locks can not be created outside of the root file driver directory  https://review.openstack.org/18846716:08
*** jaypipes has joined #openstack-oslo16:08
*** rpodolyaka1 has joined #openstack-oslo16:09
*** shardy_ has joined #openstack-oslo16:14
openstackgerritMerged openstack-dev/pbr: Move CapturedSubprocess fixture to base  https://review.openstack.org/18782216:14
*** shardy has quit IRC16:15
*** ihrachyshka has joined #openstack-oslo16:19
*** shardy_ has quit IRC16:19
*** shardy has joined #openstack-oslo16:20
*** yassine_ has quit IRC16:28
cdentI have a question on the proper way, when using oslo_db, to drop all the tables given that I have a session.EngineFacade. I can see that there is a drop_all_objects, but as far as I know I have no BackendImpl?16:31
*** yamahata has joined #openstack-oslo16:36
*** dims_ has joined #openstack-oslo16:39
*** amotoki has joined #openstack-oslo16:39
*** amotoki has quit IRC16:45
*** sdake_ has joined #openstack-oslo16:48
*** sdake has quit IRC16:52
openstackgerritDavanum Srinivas (dims) proposed openstack/oslo-incubator: Remove versionutils  https://review.openstack.org/18242816:55
dims_dhellmann: pushed a tiny fix to requirements.txt in that patch ^^^16:55
openstackgerritMerged openstack/taskflow: Updated from global requirements  https://review.openstack.org/18840916:55
*** ashleighfarnham has joined #openstack-oslo16:59
*** nikhil_k is now known as nikhil-afk17:01
*** vilobhmm has joined #openstack-oslo17:03
*** salv-orlando has quit IRC17:12
*** rpodolyaka1 has quit IRC17:14
*** nikhil-afk is now known as nikhil_k17:17
*** Kennan2 has joined #openstack-oslo17:18
*** Kennan has quit IRC17:20
*** e0ne has quit IRC17:22
*** Viswanath has joined #openstack-oslo17:25
*** sputnik13 has joined #openstack-oslo17:27
*** ihrachyshka has quit IRC17:27
*** Viswanath has quit IRC17:30
*** ndipanov has quit IRC17:30
dhellmanneezhova: ok, so it sounds like the backdoor thing is tied to the launcher, and those options should all be returned in one entry point. The ssl things are optional, and should be part of a separate entry point.17:31
dhellmannbknudson: keystore?17:31
bknudsondhellmann: the keystore is used for ssl. you keep your certs in there17:31
dhellmanndims_: ack, thanks17:31
bknudsonwhen you're using java17:32
dhellmannoh, ok, I don't know anything about that :-/17:32
bknudsony, no problem. I'll keep looking into it17:32
bknudsonmaybe it's a server-side issue17:32
*** Glass_saga has quit IRC17:34
*** salv-orlando has joined #openstack-oslo17:40
*** dims_ has quit IRC17:40
dhellmannharlowja, jd__, sileht, bnemec, flaper87, lifeless  : another namespace package update is ready to merge https://review.openstack.org/#/c/181959/317:48
harlowjawoot17:48
*** _amrith_ is now known as amrith17:51
*** yamahata has quit IRC17:55
*** yamahata has joined #openstack-oslo17:55
*** prad has quit IRC17:57
*** rushiagr is now known as rushiagr_away17:57
*** prad has joined #openstack-oslo17:59
*** Viswanath has joined #openstack-oslo18:01
*** barra204 has joined #openstack-oslo18:01
*** Viswanath has quit IRC18:04
*** barra204 has quit IRC18:07
*** dims_ has joined #openstack-oslo18:08
lifelesscool18:09
ekarlsoHi guys, I wanna use o.messaging to send app specific notifications allmost like notificiations format but without the {"oslo.message": {}} wrapping data ?=18:11
ekarlsolike I can define the data structure myself18:12
dims_ekarlso: i don't see a way to do it. sileht may have a better view, but it's too late in his TZ now18:19
mriedemjamespage: hey a question on https://bugs.launchpad.net/oslo.messaging/+bug/1436788 - i think we're seeing the same thing here,18:21
openstackLaunchpad bug 1436788 in oslo.messaging "rabbitmq heartbeat failures don't reset connections" [Medium,Fix released] - Assigned to Mehdi Abaakouk (sileht)18:21
mriedemjamespage: when you say "until the system tcp timeout is reached." - did you have tcp_retries2 changed at all or just left as the default (~15 minutes)?18:21
*** timsim has joined #openstack-oslo18:23
*** amotoki has joined #openstack-oslo18:28
*** cdent has quit IRC18:29
*** cdent has joined #openstack-oslo18:33
*** amotoki has quit IRC18:33
*** openstackstatus has joined #openstack-oslo18:40
*** ChanServ sets mode: +v openstackstatus18:40
-openstackstatus- NOTICE: Gerrit has been restarted to clear an issue with its event stream. Any change events between 17:25 and 18:38 UTC should be rechecked or have their approvals reapplied to initiate testing.18:42
*** ajo has quit IRC18:47
*** tsbot_ has quit IRC18:51
*** rpodolyaka1 has joined #openstack-oslo19:03
*** exploreshaifali has joined #openstack-oslo19:14
*** salv-orlando has quit IRC19:15
*** Viswanath has joined #openstack-oslo19:18
*** Viswanath has quit IRC19:21
*** salv-orlando has joined #openstack-oslo19:23
*** cdent has quit IRC19:39
mriedemdhellmann: have you seen this? https://review.openstack.org/#/c/185287/19:47
*** exploreshaifali has quit IRC19:50
*** e0ne has joined #openstack-oslo19:55
*** Viswanath has joined #openstack-oslo19:55
*** kgiusti has left #openstack-oslo19:57
dhellmanndims_, lifeless, harlowja, bnemec : the namespace package removal in oslo.vmware is ready: https://review.openstack.org/#/c/181987/320:00
dhellmannmriedem: looking20:00
dhellmannmriedem: yes, that's the right thing for them to do be doing20:00
*** Viswanath has quit IRC20:00
mriedemdhellmann: ok20:00
dhellmannmriedem: we can turn that back on after pip has a better dependency resolver20:01
mriedemi wasn't sure if they were just putting their heads in the sand over some bad reqs20:01
dhellmannwell, we all are, but we need to be for now :-/20:01
mriedemok, +2 on the backport then https://review.openstack.org/#/c/188330/ - thanks20:02
dhellmannmriedem: approved the backport20:03
mriedemcool, thanks20:03
dhellmannmriedem: thanks for checking :-)20:03
openstackgerritDoug Hellmann proposed openstack/oslo.i18n: Remove oslo namespace package  https://review.openstack.org/18197120:06
lifelessdhellmann: mriedem: so arguably it should never be on20:06
lifelessas that code path can never *fix* it20:07
*** Viswanath has joined #openstack-oslo20:07
openstackgerritDoug Hellmann proposed openstack/oslo.db: Remove oslo namespace package  https://review.openstack.org/18196720:07
openstackgerritDoug Hellmann proposed openstack/oslo.serialization: Remove oslo namespace package  https://review.openstack.org/18198020:08
openstackgerritDoug Hellmann proposed openstack/oslo.utils: Remove oslo namespace package  https://review.openstack.org/18198420:09
*** Viswanath has quit IRC20:10
openstackgerritDoug Hellmann proposed openstack/oslo.messaging: Remove oslo namespace package  https://review.openstack.org/18197420:11
openstackgerritOpenStack Proposal Bot proposed openstack/oslotest: Updated from global requirements  https://review.openstack.org/18856920:13
*** Viswanath has joined #openstack-oslo20:16
*** amotoki has joined #openstack-oslo20:17
*** stevemar has quit IRC20:18
*** Viswanath has quit IRC20:20
dhellmannlifeless: yeah, the default is false20:21
*** amotoki has quit IRC20:22
*** sputnik13 has quit IRC20:22
*** sputnik13 has joined #openstack-oslo20:23
openstackgerritDoug Hellmann proposed openstack/oslo.i18n: Add requirements for pre-release test scripts  https://review.openstack.org/18857320:25
openstackgerritDoug Hellmann proposed openstack/oslo.i18n: Add requirements for pre-release test scripts  https://review.openstack.org/18857320:26
harlowjazzzeek qq, i'm trying to convert some text columns (that store json) into using the native json type (that sqlalchemy-utils has); and am trying to make a migraiton to do this, http://paste.openstack.org/show/263815/20:31
harlowjathat though craps out with20:31
harlowja'ProgrammingError: (ProgrammingError) column "meta" cannot be cast automatically to type json'20:31
harlowja'HINT:  Specify a USING expression to perform the conversion.'20:31
harlowjasupposedly u have to do http://baudehlo.com/2014/04/28/postgresql-converting-text-columns-to-json/ (or similar)20:32
harlowjahow does one make that happen via alembic (and only for postgres)20:32
zzzeekharlowja: i dont know that “using” is an option we have for alter_column I probably told folks to render the ALTER as a SQL string20:32
zzzeekharlowja: obviously you say if op.bind.name == ‘postgresql’ for the simple verison of that :)20:32
harlowjak20:32
harlowjajd__ u might be interested in that to, think u were trying this also recently20:33
harlowja*in https://review.openstack.org/#/c/185658/20:33
harlowjapretty sure u have to migrate the data vs just change the model, due to above error20:34
*** tsekiyama has joined #openstack-oslo20:43
*** tsbot_ has joined #openstack-oslo20:44
*** Viswanath has joined #openstack-oslo20:44
*** Viswanath has quit IRC20:47
*** rpodolyaka1 has quit IRC20:54
*** sputnik13 has quit IRC21:01
*** sputnik13 has joined #openstack-oslo21:01
*** sputnik13 has quit IRC21:03
*** e0ne has quit IRC21:05
*** sputnik13 has joined #openstack-oslo21:07
*** Kennan has joined #openstack-oslo21:18
*** Kennan2 has quit IRC21:19
*** salv-orlando has quit IRC21:35
*** salv-orlando has joined #openstack-oslo21:41
krotscheckbknudson: So, I reread through your comments about the keystone middleware two days ago, and I'm having a hard time parsing exactly what kinds of precedents you were suggesting. Do you have a moment to help me walk through my options?21:47
bknudsonkrotscheck: sure21:47
krotscheckbknudson: So, one of the things you suggested was creating middleware A, that figures out the configuration, which is then read by middleware B, is that correct?21:48
bknudsonkrotscheck: y, this might be what swift does with auth_token.21:48
krotscheckbknudson: I dislike that approach, it feels brittle.21:49
krotscheckGiven that it creates an implicit dependency between two middlewares, but that dependency is not really enforced.21:49
bknudsonmiddleware A would be specific to the application whereas middleware B is generic21:50
bknudsony, it does create a dependency.21:50
bknudsonthat would only be enforced by B failing to init or something21:50
*** mriedem has quit IRC21:51
krotscheckOh, I see.21:51
krotscheckSo we could provide something like a generic 'read my oslo config' middleware, which someone has the option of using or using with something else.21:51
krotscheckI mean replacing with something else.21:52
openstackgerritJoshua Harlow proposed openstack/taskflow: Use the sqlalchemy-utils json type instead of our own  https://review.openstack.org/18825821:52
krotscheckAnd then modify the cors middleware to check any options added to the request context that would inform its logic.21:52
harlowjajd__ u might need something like in 18825821:52
bknudsonkrotscheck: that's pretty nead.21:54
bknudsonneat21:54
bknudsonwe could use the same thing in auth_token middleware21:54
krotscheckbknudson: Indeed.21:55
krotscheckbknudson: Though, all that would do is grab the global config instance and put it somewhere else.21:56
bknudsonthe default would21:56
bknudsonbut the deployer or application could override use their own custom middleware21:57
bknudsonwhich uses a different config instance21:57
bknudsonI don't know... have to experiment.21:57
bknudsonmaybe we can find the swift middleware if we look hard enough21:58
bknudsonhttp://docs.openstack.org/developer/swift/overview_auth.html#configuring-swift-to-use-keystone21:59
bknudsonthis is saying to put the config for keystonemiddleware in paste.ini ... which we discourage21:59
bknudsoncache = swift.cache -- they have their special swift cache22:00
bknudson"cache is set to swift.cache. This means that the middleware will get the Swift memcache from the request environment."22:00
bknudsonsomehow swift sets swift.cache in the request environment22:00
bknudsonI was thinking that other applications could put the "Swift" CONF in the request environment the same way.22:01
bknudsonwouldn't be swift but whatever the application is22:01
*** stevemar has joined #openstack-oslo22:05
*** amotoki has joined #openstack-oslo22:06
*** jecarey has quit IRC22:09
*** amotoki has quit IRC22:11
*** sigmavirus24 is now known as sigmavirus24_awa22:13
rjaiswalIs it possible to re-load a specific config file on-the-fly using oslo.config?22:17
dhellmannrjaiswal: oslo.config views all of the config files as a single data store, so it doesn't have a way to load just one of them again. Why not reload them all?22:20
*** amrith is now known as _amrith_22:22
*** bknudson has quit IRC22:28
rjaiswaldehellmann: good answer, thanks22:29
krotscheckbknudson: The more I'm reading up about this, the more I'm coming to realize that using oslo _at all_ may be a bad idea.22:47
krotscheckOh, he left.22:48
krotscheckDoh22:48
krotscheckOh well, time to call it a day anyway.22:48
*** zzzeek has quit IRC22:50
*** gordc has quit IRC22:51
*** prad has quit IRC22:55
openstackgerritJames Polley proposed openstack-dev/pbr: WIP: more extensive testing of requirement parsing  https://review.openstack.org/18782322:57
openstackgerritJames Polley proposed openstack-dev/pbr: Fix everything by requiring setuptools>=16.  https://review.openstack.org/18784622:57
harlowjakrotscheck having __init__ and function args, and objects and using docstrings on those may be the right way to go for all these things, if someone wants to reload it, they throw away the object and make a new one?22:59
*** gordc has joined #openstack-oslo23:15
openstackgerritMerged openstack/oslo.concurrency: Remove oslo namespace package  https://review.openstack.org/18195923:19
openstackgerritMerged openstack/oslo-incubator: Add new projects to dashboard  https://review.openstack.org/18760323:26
*** enikanorov_ has quit IRC23:29
*** enikanorov__ has joined #openstack-oslo23:29
openstackgerritJoshua Harlow proposed openstack/taskflow: Remove all 'lock_utils' now that fasteners provides equivalents  https://review.openstack.org/18862923:30
*** _amrith_ is now known as amrith23:34
*** vilobhmm has quit IRC23:38
*** sdake_ is now known as sdake23:44
*** salv-orlando has quit IRC23:53
*** ashleighfarnham has quit IRC23:56
« Wednesday, 2015-06-03 Index Friday, 2015-06-05 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!