16:00:14 <dims__> #startmeeting oslo
16:00:14 <dims__> courtesy ping for GheRivero, amotoki, amrith, bknudson, bnemec, dansmith, dhellmann, dougwig, e0ne, flaper87, garyk, harlowja, haypo,
16:00:14 <dims__> courtesy ping for ihrachyshka, jd__, jecarey, johnsom, jungleboyj, kgiusti, kragniz, lifeless, lintan, ozamiatin, redrobot, rpodolyaka,
16:00:15 <openstack> Meeting started Mon Oct 12 16:00:14 2015 UTC and is due to finish in 60 minutes.  The chair is dims__. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:16 <dims__> courtesy ping for sergmelikyan, sreshetnyak, sileht, sreshetnyak, stevemar, therve, thinrichs, toabctl, viktors, zhiyan, zzzeek
16:00:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:19 <rpodolyaka1> o/
16:00:19 <openstack> The meeting name has been set to 'oslo'
16:00:19 <redrobot> o/
16:00:21 <ozamiatin> o/
16:00:23 <kgiusti> o/
16:00:25 <harlowja_at_home> sup
16:00:27 <kragniz> o/
16:00:32 <amrith> ./
16:00:34 <kzaitsev_mb> 0/
16:00:35 * bknudson on fleek.
16:00:35 <toabctl> hi
16:00:40 <amrith> hello!
16:00:42 <ihrachys> o/
16:00:46 <johnsom> o/
16:00:50 <e0ne> hi
16:00:52 <haypo> \o/
16:00:56 <abhishekk> o/
16:00:59 * haypo has two arms
16:01:17 <dims__> :)
16:01:19 <dhellmann> o/
16:01:30 <dims__> welcome everyone, let's get rolling
16:01:31 <dims__> #topic Red flags for/from liaisons
16:01:44 <johnsom> Nothing to report
16:01:45 <amrith> I have a question in that area.
16:01:55 <jd__> o/
16:02:00 <amrith> I saw an email about moving mask_password() from utils back to o-o.
16:02:16 <sileht> o/
16:02:18 <bknudson> keystone unit tests are failing locally due to requests!=2.8.0... still looking into it
16:02:21 <amrith> could someone tell me what that's all about (or maybe for the purposes of this meeting whether that email subject line was misleading).
16:02:28 <bknudson> also, need a requirements bump for oslo.cache
16:02:39 <dims__> amrith: not going to happen
16:02:45 <amrith> trove uses the method so I wanted to know about it. thanks dims__
16:02:51 <amrith> nothing else from me.
16:02:54 <bknudson> not sure if the requests!=2.8.0 is going to affect the gate but it's not working locally
16:03:06 <jungleboyj> o/
16:03:10 <jungleboyj> Nothing from Cinder.
16:03:25 <dims__> amrith: there was a concern about backports and a suggestion o-o may be easier
16:03:46 <dims__> thanks jungleboyj
16:03:54 <dhellmann> bknudson: the oslo.cache requirements update is approved
16:03:57 <dims__> bknudson: please file a bump if not already done
16:04:04 <dims__> jungleboyj: thanks!
16:04:07 <amrith> dims__, thanks. I'm not sure I follow that rationale but I guess we don't have to hold up this meeting for that (the list of things I don't understand is very long).
16:04:16 <jungleboyj> dims__: You are welcome twice.
16:04:17 <bknudson> hmm, now keystone unit tests are failing with requests.exceptions.InvalidSchema ... odd. I'll keep looking.
16:04:19 <dims__> amrith: let's do it in the open discussion
16:04:21 <dhellmann> bknudson: flaper87 knows about the requests issue, you might coordinate with him
16:04:53 <dims__> bknudson: there's long email threads, most recently with updates from zigo
16:05:07 <bknudson> oh, I haven't caught up on the ml.
16:05:09 <dims__> switching topics
16:05:49 <dims__> bknudson: short story requests uses a vendorized version of urllib3 which may or may not be an actual urllib3 release causing issues
16:06:01 <dims__> #topic Releases for Mitaka
16:06:01 <dims__> #link https://review.openstack.org/#/c/233326/
16:06:01 <dims__> Releases pushed out today after verifying tests this morning https://travis-ci.org/dims/
16:06:28 <dims__> so, who ever is interested in bumps in versions, please file openstack/requirements reviews
16:06:33 <harlowja_at_home> woah, lots of releases, nice
16:06:59 <dhellmann> dims__: if those releases are done we should merge the patch
16:07:00 <dims__> harlowja_at_home: y, ran travis based tests a few times since fri
16:07:07 <dims__> dhellmann: yes, about to
16:07:10 <dhellmann> dims__: ++
16:07:23 <harlowja_at_home> dims__,  cools
16:08:07 <dims__> should be a ton of emails on openstack-announce :)
16:08:51 <dims__> we caught a bunch of issues last week with updates to neutron, heat, cinder, nova etc for testresources/testscenarios
16:09:02 <dims__> so oslo.db is at 3.0.0
16:09:40 <dims__> there was issues between nova and o.vo which was addressed already as well
16:10:07 <haypo> note: oslo.report 0.6 changes the signal from SIGUSR1 to SIGUSR2 by default
16:10:08 <dims__> and a neutron issue as it was using implementation details in oslo.messaging i believe
16:10:20 <haypo> oslo.reports*
16:10:24 <dims__> sorry oslo.policy
16:10:35 <dims__> haypo: thanks for the highlight
16:11:02 <haypo> i know that dhellmann was not a big fan of this change :-D
16:11:02 <dims__> so let's keep a look out on zuul today and please let me know if you see breaks
16:11:40 <dims__> haypo: i still have a TODO to fix the update constraints stuff to let us use oslo.db[xyz] in requirements files
16:12:15 <dims__> k switching topics
16:12:18 <haypo> dims__: ah, lifeless change on requirements was merged?
16:12:32 <dims__> haypo: which one?
16:12:43 <haypo> dims__: to export dependencies in oslo.db
16:12:59 <bknudson> signals don't provide enough info... we've got a network use it.
16:13:00 <haypo> i don't know how this stuff is called, something like oslo.db[mysql] to require PyMySQL
16:13:06 <dims__> haypo: i believe so
16:13:10 <haypo> oh ok
16:13:41 <dims__> haypo: it works in oslo.db itself. other projects can't use that syntax for now (see my TODO remark)
16:13:51 <haypo> ok
16:14:24 <dims__> bknudson: "signals"?
16:14:44 <bknudson> SIGUSR1, SIGUSR2 etc.
16:14:53 <bknudson> it's an outdated concept
16:15:08 <dims__> right
16:15:11 <haypo> bknudson: oslo.reports is only needed when a process is dying or stuck
16:15:23 <haypo> bknudson: so for example when it doesn't communicate to the network anymore :)
16:15:52 <bknudson> but luckily it responds to signals?
16:16:06 <dims__> bknudson: yes, it does
16:16:07 <haypo> bknudson: that's the bet of oslo.reports
16:16:41 <dims__> switching topics
16:16:46 <dims__> #topic openstack spec reviews - oslo-incubator changes to add wrapper classes for request-id
16:16:46 <dims__> #link https://blueprints.launchpad.net/oslo-incubator/+spec/wrapper-classes-to-return-request-id
16:16:58 <abhishekk> hi
16:17:07 <dims__> question: do we let things in till tokyo? into oslo-incubator?
16:17:20 <dims__> question: do we need a library for things abhishekk needs?
16:17:24 <harlowja_at_home> 'These new classes will be added in common package of oslo-incubator (apiclient/base.py).' :(
16:17:27 <harlowja_at_home> oh noes, lol
16:17:52 <bknudson> as long as oslo-incubator is switched to being obviously private (_ prefix) I'm ok with it.
16:17:59 <dims__> i had asked abhishekk to check what they will do in projects that don't use the apiclient/* stuff
16:18:21 <bknudson> it's going to create a lot of work if we can't have a stable interface.
16:18:28 <abhishekk> i am checking on it
16:18:39 <bknudson> although, stable interfaces create a lot of work, too
16:18:51 <dims__> dhellmann: thoughts? as this came from cross-project spec / TC
16:21:14 <dims__> abhishekk: will approve that spec once the meeting is over. my feeling is that we can let the changes in for now. we'll have to deal with _ prefix etc anyway whether we have this change or not
16:21:17 * harlowja_at_home thinks doug is somewhere else, ha
16:21:27 <abhishekk> thank you dims__
16:21:43 <haypo> i would prefer to stop modifying oslo-incubator
16:21:48 <haypo> and instead start to work on a real library
16:22:08 <haypo> this spec looks like a good motivation to create it
16:22:23 <harlowja_at_home> +1
16:22:24 <dims__> abhishekk: what do you think?
16:22:48 <haypo> would it be possible to make baby step, like starting with a subset of the apiclient API?
16:22:53 <bknudson> is that library openstack-sdk?
16:23:11 <abhishekk> dims__: i am not aware about making new liabrary
16:23:14 <haypo> bknudson: what?
16:23:35 <haypo> bknudson: i proposed to create a new library, ex: oslo.apiclient
16:23:41 <dims__> abhishekk: would you be able to help with pulling out code you need into a library?
16:23:53 <bknudson> there's a group working on an openstack sdk library to replace the python-*client
16:24:07 <abhishekk> dims__: I will check it out
16:24:15 <haypo> bknudson: i'm not talking about the openstack sdk thing, it's unrelated IMHO
16:24:33 <dims__> abhishekk: check this out - https://wiki.openstack.org/wiki/Oslo/CreatingANewLibrary
16:24:48 <haypo> abhishekk: i can try to help you to do that
16:24:55 <harlowja_at_home> me too
16:24:56 <abhishekk> bknudson: for openstacksdk I have filed anothr blueprint
16:25:08 <dims__> haypo: we've previously told people that they should just switch to something else, probably in openstack-sdk
16:25:13 <abhishekk> dims__, hypo, joshua thank  you
16:25:18 <harlowja_at_home> np
16:25:37 <dims__> nice segue to the next topic
16:25:40 <abhishekk> https://blueprints.launchpad.net/python-openstacksdk/+spec/return-request-id-to-caller
16:25:40 <dims__> #topic summit request for sessions
16:25:40 <dims__> #link https://etherpad.openstack.org/p/mitaka-oslo-summit-planning
16:25:43 <bknudson> if the funcationality is in openstacksdk, we can use it in python-keystoneclient
16:25:56 <harlowja_at_home> oh man, tokyo here we come
16:25:58 <dims__> abhishekk: please add oslo.apiclient to the mitaka etherpad for consideration
16:26:10 <abhishekk> dims__: sure
16:26:10 <harlowja_at_home> do we know who is all going to tokyo btw?
16:26:30 <harlowja_at_home> or maybe better question, who isn't going to tokyo
16:26:57 <haypo> dims__: oh ok (for openstack-sdk)
16:27:25 <dims__> i'll need help later today to finalize the list of topics, anyone free later today?
16:27:31 <kgiusti> o/ - no travel budget :(
16:27:37 <dims__> kgiusti: ouch :(
16:27:50 <harlowja_at_home> dims__, sure let me know, we can make up some creative session names, hahaha
16:27:57 <dims__> harlowja_at_home: ack thanks!
16:28:12 <dims__> #topic Open discussion
16:28:12 <dims__> Any stuck reviews? or specs?
16:28:28 <dims__> kgiusti: proton still holding up against devstack+tempest?
16:28:40 <kgiusti> not exactly stuck, but very embarrassing: https://review.openstack.org/#/c/233094/
16:29:02 <dims__> haha :) +2
16:29:03 <harlowja_at_home> since i thought it was cool
16:29:05 <harlowja_at_home> http://tempsend.com/E38B61A9FA/6D80/oslo.svg
16:29:10 <harlowja_at_home> print that out and put on your wall folks
16:29:10 <harlowja_at_home> lol
16:29:18 <dims__> harlowja_at_home: nice :)
16:29:29 <bknudson> looks like a big bang
16:29:32 <harlowja_at_home> :)
16:29:35 <bknudson> or traces from the supercollider
16:29:38 <harlowja_at_home> lol
16:29:42 <haypo> kgiusti: i didn't understand https://review.openstack.org/#/c/233094/ but i see gate-tempest-dsvm-full-zmq FAILURE
16:29:54 <harlowja_at_home> the oslobigbang, ha
16:29:57 <haypo> kgiusti: why do you expect the gate to fail?
16:30:17 <harlowja_at_home> bknudson,  https://review.openstack.org/#/c/232729/
16:30:18 <ozamiatin> haypo: it was failing before
16:30:25 <harlowja_at_home> http://tempsend.com/44D56E1EE3/A632/infra.svg is a bigger bang, lol
16:30:31 <ozamiatin> haypo: it is non-voting for that reason
16:30:40 <kgiusti> haypo: I don't expect it to fail - that change only affects tox on amqp1
16:30:49 <jungleboyj> harlowja_at_home: Oh my ... So many oslo things.
16:30:50 <haypo> hum wait, the patch is for AMQP but the failing gate is ZMP
16:30:53 <harlowja_at_home> lol
16:30:53 <haypo> it's unrelated, right?
16:31:02 <harlowja_at_home> (http://tempsend.com/1F392B692C/1FFD/os.svg ) every project
16:31:04 <kgiusti> haypo: dumps protocol trace msgs over "the wire"
16:31:23 <haypo> kgiusti: anyway, i trust you, it's now approved :=
16:31:24 <haypo> :)
16:31:30 <kgiusti> haypo: I had expected it to fail gate - it was consistently failing before
16:31:37 <ozamiatin> haypo: yeah
16:31:38 <kgiusti> haypo: thanks1
16:31:55 <kgiusti> haypo: hmmm - you trust me, eh?
16:32:04 <haypo> kgiusti: i was confused by "the gate". there are a lot of gates
16:32:12 <haypo> kgiusti: you worked on AMQP more than me :)
16:32:14 <kgiusti> haypo: I've got a deal on some land in florida....
16:32:26 <kgiusti> haypo: :)
16:32:33 <haypo> kgiusti: i only trust you for AMQP things :)
16:32:39 <harlowja_at_home> jungleboyj, http://tempsend.com/44D56E1EE3/A632/infra.svg is much bigger bang, lol
16:32:55 <haypo> if you are bored, another pending review for python3 : https://review.openstack.org/#/c/233093/
16:33:30 <haypo> dims__: again, i'm not sure that an oslo session is required for python3, i still it in the etherpad
16:33:55 <harlowja_at_home> is python4 out yet?
16:33:57 <harlowja_at_home> :)
16:34:05 <jungleboyj> harlowja_at_home: Oy vey!
16:34:15 <haypo> harlowja_at_home: lol
16:34:20 <haypo> harlowja_at_home: almost
16:34:24 <harlowja_at_home> :)
16:34:57 <amrith> dims__, I have one for open discussion
16:35:09 <dims__> amrith: what's up?
16:35:20 <amrith> next steps on oslo.messaging encryption
16:35:23 <dims__> haypo: probably not
16:35:37 <dims__> amrith: did you pick an approach yet?
16:35:42 <amrith> https://bugs.launchpad.net/oslo.messaging/+bug/1504525
16:35:42 <openstack> Launchpad bug 1504525 in oslo.messaging "oslo messaging should provide a mechansim to generate a completely encrypted channel" [Undecided,New] - Assigned to Amrith (amrith)
16:35:57 <harlowja_at_home> haypo, i'd be interseted in hearing peoples thoughts about what to do now that python3 support is nearly everywhere, would any project actually start only using python3 (and therefore people can actually use python3 features/functionality)
16:35:57 <amrith> dims__, I'm not sure which is better
16:36:17 <dims__> amrith: can you please enumerate the 2 approaches?
16:36:23 <amrith> that would seem like something that people more experience in o.m should provide some color.
16:36:27 <amrith> so the two approaches are these ...
16:36:29 <harlowja_at_home> haypo, aka, drop 2.7 completely for some project?
16:36:32 <bknudson> python3 support is nearly everywhere?
16:36:36 <harlowja_at_home> lol
16:36:50 <harlowja_at_home> nearly, lol :-/
16:37:01 <harlowja_at_home> for some defintion of near
16:37:01 <amrith> a) given that o.m already has a concept of a driver for each transport, make encryption common for all transports; i.e. implement encryption above the driver.
16:37:02 <harlowja_at_home> lol
16:37:14 <amrith> b) provide implementations of encryption per-driver
16:37:19 <haypo> bknudson: 6 applications works on Python 3, http://blogs.rdoproject.org/7807/python-3-status-openstack-liberty
16:38:07 <dims__> sileht: kgiusti: ozamiatin: thoughts please ^^^
16:38:53 <bknudson> haypo: the other applications are the ones everybody needs
16:39:08 <kgiusti> current drivers support "over the wire" encryption (ssl/gssapi/etc).  This would involve end-2-end encryption.
16:39:21 <kgiusti> I think that' sbetter done above the drivers
16:39:42 <kgiusti> the tricky part would be the key exchange - that's above my paycheck, sadly.
16:39:49 <amrith> kgiusti, when you say current drivers support "over the wire" ... how so?
16:40:06 <amrith> that would be a matter of how one configured the underlying transport.
16:40:10 <haypo> harlowja_at_home: the next milestone is to run functional tests on python 3
16:40:19 <haypo> harlowja_at_home: without this, it's hard to trust python 3 support :-/
16:40:20 <bknudson> kgiusti: we had the kite project for a while for key exchange.
16:40:22 <amrith> the best I've found, for example to enable ssl on rabbit, is to do https://www.rabbitmq.com/ssl.html
16:40:28 <kgiusti> amrith: I think rabbitmq supports SSL encryption
16:40:36 <harlowja_at_home> haypo, good point
16:40:40 <haypo> harlowja_at_home: the first step would be to add an option to install some apps and clients on py3
16:40:52 <kgiusti> amrith: proton (amqp1) support SSL and SASL-based encryption
16:40:56 <bknudson> https://github.com/OpenStack/kite
16:41:19 <kgiusti> amrith: but this is only on the wire to the broker.  The secure tunnel ends at the broker
16:41:39 <kgiusti> amrith: of course, both directions (from and to broker) are encrypted
16:41:47 <bknudson> I guess it's still around: https://review.openstack.org/#/q/project:openstack/kite,n,z
16:42:00 <amrith> kgiusti, what's proton?
16:42:09 <kgiusti> amrith: but messages on the broker are not, unless the sending endpoint has encrypted the payload.
16:42:36 <kgiusti> amrith: sorry - that's the amqp1 driver - I really shouldn't be using the term 'broker'
16:42:52 <kgiusti> amrith: no wait - not using 'proton' (freudian slip)
16:43:27 <amrith> kgiusti, sounds like a discussion we can take offline (to #openstack.oslo)
16:43:27 <kgiusti> amrith: what I really need to do is document how to setup SSL and/or SASL using amqp1 driver...
16:43:38 <kgiusti> amrith: +1
16:44:15 <kgiusti> bknudson: +1 that's what I was trying to remember !
16:45:30 <amrith> there's an oslo conf setting, if that helps ;)
16:47:15 <amrith> dims__, kgiusti let's take this to #openstack.oslo for later
16:47:20 <dims__> ok, so let's wrap up folks?
16:47:31 <dims__> thanks everyone
16:47:38 <dims__> #endmeeting