Monday, 2019-09-30

*** Chaiko has joined #openstack-oslo02:08
ChaikoNoen som anbefaler en god start for å lære python?02:09
*** ccamel has quit IRC02:26
*** camelCaser has joined #openstack-oslo02:27
*** zaneb has quit IRC02:36
*** Chaiko has quit IRC02:46
*** toabctl has joined #openstack-oslo04:07
*** Luzi has joined #openstack-oslo05:01
*** jaosorior has joined #openstack-oslo05:03
*** pcaruana has joined #openstack-oslo05:24
*** pcaruana has quit IRC05:30
*** iurygregory has quit IRC06:04
*** pcaruana has joined #openstack-oslo06:41
*** iurygregory has joined #openstack-oslo07:01
*** rcernin has quit IRC07:14
*** tesseract has joined #openstack-oslo07:17
*** e0ne has quit IRC07:19
*** iurygregory has quit IRC07:34
*** tesseract has quit IRC07:43
*** tesseract has joined #openstack-oslo07:43
*** ralonsoh has joined #openstack-oslo08:00
*** iurygregory has joined #openstack-oslo08:11
gsantomaggiogood morning @bnemec08:23
gsantomaggioHope the answer is clear08:23
gsantomaggiohttps://review.opendev.org/#/c/68360008:23
gsantomaggioor better...  the comment is clear :)08:24
gsantomaggioand also, Kombu allows different Json Serialization types:08:29
gsantomaggiohttps://kombu.readthedocs.io/en/stable/userguide/producers.html#serialization08:29
*** e0ne has joined #openstack-oslo08:31
*** tosky has joined #openstack-oslo09:11
openstackgerritMerged openstack/oslo.messaging master: Align message serialization between drivers  https://review.opendev.org/68360009:26
*** sum12 has quit IRC09:39
*** sum12 has joined #openstack-oslo09:52
*** tosky has quit IRC10:17
*** sean-k-mooney has quit IRC10:18
*** tosky has joined #openstack-oslo10:18
*** sean-k-mooney has joined #openstack-oslo10:22
*** rcernin has joined #openstack-oslo11:28
*** ansmith has quit IRC11:29
*** ianychoi has quit IRC11:39
*** ianychoi has joined #openstack-oslo11:40
*** goldyfruit has quit IRC12:18
*** ansmith has joined #openstack-oslo12:31
*** rcernin has quit IRC12:50
*** zaneb has joined #openstack-oslo13:04
*** goldyfruit has joined #openstack-oslo13:18
*** raildo has joined #openstack-oslo13:20
*** Luzi has quit IRC13:37
bnemecgsantomaggio: Thanks, I left a reply too. I think we're going to need a followup to that patch since it already merged.13:43
bnemecI don't like that the rabbit driver is more strict than the test driver. :-/13:43
gsantomaggio@bnemec FakeDriver is used only for the tests, is that right?13:46
bnemecgsantomaggio: Yes. Which means other projects are relying on it to prevent them from merging something that will break the other drivers.13:47
bnemecWith that patch, it would be possible for a project to start passing oslo.messaging a message that the rabbit driver can't serialize, but it might pass their tests because the fake driver is more lenient.13:47
gsantomaggiothe fact now, they are using the same json class13:51
gsantomaggiothis is what Kombu uses by default https://usercontent.irccloud-cdn.com/file/VQShqeCt/Screenshot%20from%202019-09-30%2015-51-53.png13:52
gsantomaggio@bnemec  ^^^13:52
gsantomaggiothat should be same used here:13:53
gsantomaggiohttps://review.opendev.org/#/c/683600/2/oslo_messaging/_drivers/impl_fake.py13:53
gsantomaggio( right?)13:53
gsantomaggioyes, it is the same13:53
bnemecgsantomaggio: But that patch changed it from json.dumps to jsonutils.dumps. The new function can serialize more object types than json.dumps.13:54
gsantomaggio@bnemec  ops you are right13:54
gsantomaggiosorry13:54
bnemecnp13:55
bnemecThis situation is why I didn't +2 the patch in the first place. The fake driver should be as strict as the strictest real driver and now it's not.13:55
bnemecWe can either revert and make it strict again, or we can make the rabbit driver less strict to match.13:56
bnemecI don't currently have a strong opinion on which we do. I wouldn't be surprised if it turned out that the reason the rabbit driver uses json.dumps is simply that it pre-dates the existence of jsonutils.13:57
*** redrobot has joined #openstack-oslo13:59
gsantomaggio@bnemec  https://github.com/openstack/oslo.serialization/blob/master/oslo_serialization/jsonutils.py#L20014:00
gsantomaggioOk so, oslo_serialization.dumps is just a wrapper around14:00
gsantomaggiojson.dumps14:00
gsantomaggio> the rabbit driver uses json.dumps14:03
gsantomaggioThat's not totally correct14:03
gsantomaggio > the rabbit driver uses json.dumps14:03
gsantomaggiois inside Kombu14:03
dansmithbnemec: Sorry for just seeing it now, but this looks concerning to me: https://review.opendev.org/#/c/683600/214:03
gsantomaggioit is not under olso control14:03
dansmithbnemec: er, not seeing it *until* now14:03
bnemecdansmith: I agree.14:04
bnemecgsantomaggio: Right, but you said we could override what kombu uses, didn't you?14:04
dansmithoh I see we're discussing it now :)14:04
bnemecIf not, then we need to revert that change.14:04
bnemecYep, seeing that had merged was my Monday morning welcome back. :-)14:05
gsantomaggio@bnemec  teorically yes, but the parameter is not exposed14:05
gsantomaggioso in kombu you can change it, in olso you can't14:05
dansmithbnemec: I didn't see what the strictness was choking on.. usually strict is good for something like this14:05
gsantomaggiosee for example:14:07
gsantomaggiohttps://docs.openstack.org/mitaka/config-reference/compute/rpc.html14:07
gsantomaggiothere are a few Kombu paramenters, the JSON serializer is not exported14:07
*** smcginni1 has joined #openstack-oslo14:08
bnemecdansmith: Apparently nova was concerned about this: https://github.com/openstack/nova/blob/09bf71407f4c0d1ddbef89c489ec87c3bca0b7b2/nova/compute/rpcapi.py#L82014:08
dansmithbnemec: right, but.. oh I see, datetimes14:09
dansmithand jsonutils.dumps() converts those to something portable?14:09
*** adriant has quit IRC14:09
bnemecgsantomaggio: I don't want it to be configurable by the user, I just want us to be able to tell kombu to use jsonutils.14:09
bnemecdansmith: Yeah, jsonutils has handlers for more types than the regular json module.14:09
bnemecAlthough it's a little weird to me that round-tripping it like that actually works.14:10
dansmithokay, as long as it's really handling them and not just quietly passing them or doing something dumb14:10
dansmithbnemec: yeah, that seems weird, so I wonder if it's stripping timezones or something14:10
gsantomaggiook14:11
gsantomaggiojust moemnt14:11
*** smcginni1 is now known as smcginnis_14:11
bnemecjsonutils uses to_primitive to serialize: https://github.com/openstack/oslo.serialization/blob/master/oslo_serialization/jsonutils.py#L6014:11
bnemecI guess it does this with datetimes: https://github.com/openstack/oslo.serialization/blob/master/oslo_serialization/jsonutils.py#L12014:12
dansmithwhich is TZ-unaware14:12
dansmithwhich means it's converting back to a TZ-unaware datetime on the other side14:13
gsantomaggioit should be enough to change https://github.com/openstack/oslo.messaging/blob/master/oslo_messaging/_drivers/impl_rabbit.py#L122114:13
gsantomaggioby adding the serialization14:13
dansmithhttps://github.com/openstack/oslo.utils/blob/master/oslo_utils/timeutils.py#L4014:13
bnemecdansmith: Do you see a problem with using jsonutils for this?14:16
bnemecEspecially if projects are already using it to pre-serialize before passing messages to us.14:16
dansmithbnemec: well, like I said in my review, converting to it now means that we'll generate a different on-the-wire format which older services will choke on14:16
dansmithif datetimes are the reason, and we just silently generate a string of a certain format, the other side may do something weird,14:17
dansmithso I dunno14:17
dansmithlike if I'm running a train conductor which is sending messages to a stein compute14:17
*** pcaruana has quit IRC14:18
dansmithI guess theoretically we shouldn't have that situation because we'd be failing at the sender right now, but it pushes the failure to the receiver which will be a lot harder to debug14:18
dansmithI like to be ultra conservative in stuff like this, so I lean towards no change, but...14:18
bnemecThat's true, this way projects have to be explicit about what they're doing.14:20
dansmithyup, especially given this will silently drop the TZ, so if they hand a GMT datetime to the rpc layer, it will come out as a local one on the other end (well, undefined really)14:21
dansmithI definitely understand the desire to just have that work though14:21
bnemecOkay, I'm going to propose a revert.14:23
bnemecThat patch alone is definitely wrong, and it's not clear to me that we should be changing this behavior.14:24
bnemecI'll send an email to the list explaining the reasoning and we can discuss it with the wider audience there.14:24
dansmithsounds good to me14:24
gsantomaggio@bnemec teorically we can do something like:14:25
gsantomaggio```14:25
gsantomaggio  self._producer.publish(14:25
gsantomaggio                msg,14:25
gsantomaggio                mandatory=transport_options.at_least_once if14:25
gsantomaggio                transport_options else False,14:25
gsantomaggio                exchange=exchange,14:25
gsantomaggio                routing_key=routing_key,14:25
gsantomaggio                expiration=timeout,14:25
gsantomaggio                compression=self.kombu_compression,14:25
gsantomaggio                serializer="jsonutils.JSONECndoder")14:25
gsantomaggio```14:25
gsantomaggio * `jsonutils.JSONEncoder`14:25
gsantomaggioor something like that14:25
gsantomaggiobut it does not work :(14:27
openstackgerritBen Nemec proposed openstack/oslo.messaging master: Revert "Align message serialization between drivers"  https://review.opendev.org/68572414:30
bnemecgsantomaggio: That's okay, it's not clear to me that we want to change this anyway. I'm going to send an email to the list explaining why I proposed a revert.14:31
gsantomaggio@bnemec  ok14:33
*** Akira72 has joined #openstack-oslo14:45
*** smcginnis_ is now known as smcginnis14:46
*** Akira72 is now known as yosshy14:47
*** ansmith_ has joined #openstack-oslo14:50
*** goldyfruit_ has joined #openstack-oslo14:51
*** ansmith has quit IRC14:51
*** goldyfruit has quit IRC14:53
moguimarbnemec: meeting?15:01
bnemec#startmeeting oslo15:02
bnemecCourtesy ping for bnemec, jungleboyj, moguimar, hberaud, kgiusti, redrobot, stephenfin, johnsom, gsantomaggio15:02
bnemec#link https://wiki.openstack.org/wiki/Meetings/Oslo#Agenda_for_Next_Meeting15:02
openstackMeeting started Mon Sep 30 15:02:30 2019 UTC and is due to finish in 60 minutes.  The chair is bnemec. Information about MeetBot at http://wiki.debian.org/MeetBot.15:02
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.15:02
*** openstack changes topic to " (Meeting topic: oslo)"15:02
openstackThe meeting name has been set to 'oslo'15:02
jungleboyj:-)15:02
moguimaro/15:02
bnemecFighting with server parts that don't want to fit. :-/15:03
gsantomaggioo/15:03
stephenfino/15:04
jungleboyj:-(  I am doing that with a computer for Church.15:04
jungleboyjNeed some weird power connector for the SATA drive from the motherboard.15:04
*** iurygregory has quit IRC15:05
bnemecThat's annoying.15:05
bnemec#topic Red flags for/from liaisons15:06
*** openstack changes topic to "Red flags for/from liaisons (Meeting topic: oslo)"15:06
bnemecOther than hardware issues, any problems to report?15:06
jungleboyjYeah, having to order some weird part from China.15:06
jungleboyjNo Red Flags from Cinder.15:06
jungleboyjOur RC is out.  Hopefully all is well.15:06
moguimarNo flags from Barbican.15:06
bnemecWe're still not doing releases for Oslo right now, so I don't expect any new issues to occur.15:06
bnemecI think that's all of our present liaisons, so...15:08
bnemec#topic Releases15:08
*** openstack changes topic to "Releases (Meeting topic: oslo)"15:08
bnemecAs I noted, we won't do any master releases until Train ships, just in case a feature release would be needed on the stable branch.15:09
bnemecThat would probably only happen if we needed to bump a requirement or something.15:09
bnemecAs a result, not much else to say on this topic.15:09
bnemec#topic Action items from last meeting15:09
*** openstack changes topic to "Action items from last meeting (Meeting topic: oslo)"15:09
bnemec"bnemec to send email about courtesy ping list for Ussuri"15:10
bnemecDone15:10
*** kgiusti has joined #openstack-oslo15:10
bnemec"Anyone who wishes to continue receiving courtesy pings needs to add themselves to the new list in https://wiki.openstack.org/wiki/Meetings/Oslo#Agenda_Template"15:10
bnemecI see a few people have added their names to the new list, so I'm going to say this one is good too.15:10
bnemec"Anyone interested in being Oslo PTL for V cycle contact bnemec ASAP"15:11
bnemecI had one other person contact me about this, not for V but for a possible future cycle.15:11
bnemecSo I have a couple of people I'll be including on PTL stuff for this cycle.15:12
bnemecThanks for stepping up. :-)15:12
bnemecAnd that's it for action items.15:12
bnemec#topic oslo.messaging fake driver revert15:12
*** openstack changes topic to "oslo.messaging fake driver revert (Meeting topic: oslo)"15:12
bnemecI sent an email to the list about this too, but I've proposed a revert of the JSON serialization patch that merged over the weekend.15:13
bnemecIt makes the fake driver less strict about what it will accept than the rabbit driver, and that's not good.15:13
bnemecFor more details, see the IRC scrollback or my email.15:14
bnemec#link http://lists.openstack.org/pipermail/openstack-discuss/2019-September/009812.html15:14
bnemecI don't have a lot else to say about that and I think I'd prefer to keep future discussion mostly on the list/revert review since this is a cross-project thing.15:15
bnemec#topic  oslo.messaging devstack plugins15:15
*** openstack changes topic to "oslo.messaging devstack plugins (Meeting topic: oslo)"15:15
bnemeckgiusti: ansmith_: This one is mostly for you.15:15
kgiustikk15:16
bnemecI submitted the change to branch the messaging devstack plugins last week, and I'm pretty sure there were no meaningful changes to either of them.15:16
bnemecDo we want to keep branching them?15:16
kgiustiI'm conflicted - I just got back from vaca and need to think about this a bit15:17
bnemeckgiusti: Okay, no worries. We don't really need an answer until next cycle so you've got time. :-)15:18
kgiustistill catching up on email ;)15:18
kgiustibnemec: +115:18
bnemecOkay, just wanted to get that one out there, so we can move on.15:19
bnemec#topic  Ping list for Ussuri15:19
*** openstack changes topic to "Ping list for Ussuri (Meeting topic: oslo)"15:19
bnemecAnother reminder that in the near future I'll be clearing the ping list and switching to the new Ussuri one on the agenda.15:19
bnemec#link https://wiki.openstack.org/wiki/Meetings/Oslo#Agenda_Template15:19
bnemecIf you want to continue receiving courtesy pings at the start of the Oslo meeting you'll need to indicate that by adding yourself to the new list.15:20
bnemecAnd that's pretty much it for that topic too.15:21
bnemec#topic  Oslo project update topics15:21
*** openstack changes topic to "Oslo project update topics (Meeting topic: oslo)"15:21
bnemecWe have a project update slot at the Shanghai summit, so I've started collecting topics for the presentation.15:21
bnemecI'm tracking them on the oslo etherpad.15:21
bnemec#link https://etherpad.openstack.org/p/oslo-shanghai-topics15:21
bnemecIt's the first topic on the list.15:22
bnemecIf you have anything you think should be mentioned that isn't already there, please feel free to add it.15:22
bnemecFor the moment just a one line summary is enough. I'll ask for more details as we get closer if I need them.15:22
bnemecAs a reminder, this talk is basically about what we did, what we're currently working on, and what we're planning to work on in the future.15:23
bnemecSo anything that fits those categories is fair game.15:23
bnemec#action Add items to the project update topic list in https://etherpad.openstack.org/p/oslo-shanghai-topics as appropriate15:24
bnemecThat's it for agenda topics, and I'm still skipping the wayward review until train ships, so...15:25
bnemec#topic Open discussion15:25
*** openstack changes topic to "Open discussion (Meeting topic: oslo)"15:25
bnemecAnything else before we call it?15:26
bnemecAlright, thanks for joining everyone!15:29
bnemec#endmeeting15:29
*** openstack changes topic to "OpenStack Common Libraries | https://wiki.openstack.org/wiki/Oslo"15:29
openstackMeeting ended Mon Sep 30 15:29:08 2019 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:29
openstackMinutes:        http://eavesdrop.openstack.org/meetings/oslo/2019/oslo.2019-09-30-15.02.html15:29
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/oslo/2019/oslo.2019-09-30-15.02.txt15:29
openstackLog:            http://eavesdrop.openstack.org/meetings/oslo/2019/oslo.2019-09-30-15.02.log.html15:29
*** yosshy has left #openstack-oslo15:29
*** pcaruana has joined #openstack-oslo15:46
*** ansmith_ has quit IRC15:46
*** tesseract has quit IRC15:55
*** ansmith_ has joined #openstack-oslo16:16
*** e0ne has quit IRC16:27
openstackgerritMerged openstack/mox3 master: Update the constraints url  https://review.opendev.org/68556516:39
*** jaosorior has quit IRC16:46
*** raildo_ has joined #openstack-oslo16:50
*** raildo has quit IRC16:53
*** raildo_ has quit IRC17:07
*** raildo has joined #openstack-oslo17:10
*** e0ne has joined #openstack-oslo17:28
*** e0ne has quit IRC17:36
openstackgerritMerged openstack/stevedore master: Update the constraints url  https://review.opendev.org/68555017:41
openstackgerritMerged openstack/oslo.rootwrap master: Update the constraints url  https://review.opendev.org/68554717:43
openstackgerritMerged openstack/futurist master: Update the constraints url  https://review.opendev.org/68556917:49
openstackgerritMerged openstack/automaton master: Update the constraints url  https://review.opendev.org/68556617:58
openstackgerritMerged openstack/oslo.policy master: Update the constraints url  https://review.opendev.org/68554818:05
openstackgerritMerged openstack/oslo.context master: Update the constraints url  https://review.opendev.org/68555318:05
openstackgerritMerged openstack/oslo.privsep master: Update the constraints url  https://review.opendev.org/68493718:09
*** e0ne has joined #openstack-oslo18:19
*** gmann is now known as gmann_afk18:19
openstackgerritMerged openstack/oslo.utils master: Update the constraints url  https://review.opendev.org/68555418:19
openstackgerritMerged openstack/oslo.utils master: Support "qemu-img info" virtual size in QEMU 4.1 and later  https://review.opendev.org/68224718:19
openstackgerritMerged openstack/oslo.service master: Update the constraints url  https://review.opendev.org/68493118:26
openstackgerritMerged openstack/oslo.reports master: Update the constraints url  https://review.opendev.org/68493318:32
openstackgerritMerged openstack/oslo.concurrency master: Update the constraints url  https://review.opendev.org/68554918:33
*** jbadiapa has joined #openstack-oslo19:17
*** e0ne has quit IRC19:41
*** e0ne has joined #openstack-oslo20:14
*** e0ne has quit IRC20:19
*** jbadiapa has quit IRC20:27
*** kgiusti has left #openstack-oslo20:35
*** jbadiapa has joined #openstack-oslo20:46
*** ralonsoh has quit IRC20:49
*** pcaruana has quit IRC20:50
*** raildo has quit IRC20:54
*** irclogbot_1 has quit IRC20:54
*** irclogbot_3 has joined #openstack-oslo20:56
*** irclogbot_3 has quit IRC21:01
*** irclogbot_2 has joined #openstack-oslo21:02
*** irclogbot_2 has quit IRC21:07
*** irclogbot_1 has joined #openstack-oslo21:08
*** irclogbot_1 has quit IRC21:13
*** irclogbot_1 has joined #openstack-oslo21:14
*** e0ne has joined #openstack-oslo21:18
*** irclogbot_1 has quit IRC21:19
*** irclogbot_0 has joined #openstack-oslo21:20
*** e0ne has quit IRC21:24
*** irclogbot_0 has quit IRC21:25
*** irclogbot_1 has joined #openstack-oslo21:26
*** irclogbot_1 has quit IRC21:33
*** irclogbot_2 has joined #openstack-oslo21:36
*** irclogbot_2 has quit IRC21:41
*** irclogbot_0 has joined #openstack-oslo21:42
*** irclogbot_0 has quit IRC21:47
*** irclogbot_0 has joined #openstack-oslo21:48
*** irclogbot_0 has quit IRC21:53
*** irclogbot_0 has joined #openstack-oslo21:54
*** goldyfruit_ has quit IRC21:56
*** irclogbot_0 has quit IRC21:59
*** irclogbot_2 has joined #openstack-oslo22:00
*** irclogbot_2 has quit IRC22:05
*** irclogbot_2 has joined #openstack-oslo22:06
*** adriant has joined #openstack-oslo22:14
*** irclogbot_2 has quit IRC22:15
*** irclogbot_0 has joined #openstack-oslo22:16
*** goldyfruit_ has joined #openstack-oslo22:23
*** tosky has quit IRC22:24
*** rcernin has joined #openstack-oslo22:24
*** harlowja has quit IRC22:36
*** jbadiapa has quit IRC22:53
*** ansmith_ has quit IRC23:10
*** ansmith has joined #openstack-oslo23:12
*** goldyfruit_ has quit IRC23:21

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!