*** saneax has quit IRC | 00:13 | |
*** tmckay has joined #openstack-sahara | 00:19 | |
openstackgerrit | Ken Chen proposed openstack/sahara: Add Zookeeper and Sentry in CDH540 scenario tests https://review.openstack.org/201887 | 00:58 |
---|---|---|
*** Longgeek has joined #openstack-sahara | 01:00 | |
*** melaks has left #openstack-sahara | 01:12 | |
*** openstack has joined #openstack-sahara | 01:23 | |
*** raildo1 has joined #openstack-sahara | 01:24 | |
*** htruta_ has joined #openstack-sahara | 01:34 | |
*** melaks has joined #openstack-sahara | 02:06 | |
openstackgerrit | Ken Chen proposed openstack/sahara-specs: Add HDFS HA support in Cloudera plugin https://review.openstack.org/196929 | 02:15 |
openstackgerrit | Ken Chen proposed openstack/sahara: Enable HDFS HA in Cloudera plugin https://review.openstack.org/199464 | 02:20 |
*** Longgeek has joined #openstack-sahara | 02:30 | |
*** Longgeek has quit IRC | 02:35 | |
*** melaks has quit IRC | 02:44 | |
*** Poornima has joined #openstack-sahara | 02:55 | |
*** htruta_ has quit IRC | 03:04 | |
*** saneax has joined #openstack-sahara | 03:33 | |
*** Nikolay_St has quit IRC | 03:41 | |
*** zhiyan has quit IRC | 03:45 | |
*** h00327910_ has quit IRC | 03:46 | |
*** coolsvap|away is now known as coolsvap | 03:47 | |
*** zhiyan has joined #openstack-sahara | 03:54 | |
*** hdd has joined #openstack-sahara | 04:06 | |
*** h00327910_ has joined #openstack-sahara | 04:10 | |
*** hdd has quit IRC | 04:12 | |
*** Longgeek has joined #openstack-sahara | 04:17 | |
*** melaks has joined #openstack-sahara | 04:24 | |
*** Longgeek has quit IRC | 04:25 | |
*** melaks has left #openstack-sahara | 04:25 | |
*** Longgeek has joined #openstack-sahara | 04:25 | |
openstackgerrit | Ken Chen proposed openstack/sahara: Enable HDFS HA in Cloudera plugin https://review.openstack.org/199464 | 04:34 |
*** Nikolay_St has joined #openstack-sahara | 04:42 | |
*** saneax has quit IRC | 04:56 | |
*** saneax has joined #openstack-sahara | 04:59 | |
*** Poornima has quit IRC | 05:59 | |
*** nkrinner_afk is now known as nkrinner | 06:07 | |
*** degorenko has quit IRC | 06:10 | |
*** Nikolay_St has quit IRC | 06:12 | |
*** Poornima has joined #openstack-sahara | 06:12 | |
*** Longgeek has quit IRC | 06:51 | |
*** Longgeek has joined #openstack-sahara | 06:52 | |
*** pino|work has joined #openstack-sahara | 07:04 | |
*** Nikolay_St has joined #openstack-sahara | 07:21 | |
*** pino|work has quit IRC | 08:22 | |
*** witlessb has joined #openstack-sahara | 08:24 | |
*** pino|work has joined #openstack-sahara | 08:31 | |
openstackgerrit | Andrey Pavlov proposed openstack/sahara: [WIP] Devices formatting and mounting https://review.openstack.org/200483 | 08:34 |
*** tosky has joined #openstack-sahara | 08:51 | |
*** tsubic has joined #openstack-sahara | 09:07 | |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara: Migrate "flavor_id" to "flavor" in scenario tests https://review.openstack.org/202077 | 09:11 |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara: Add cluster verification https://review.openstack.org/196713 | 09:19 |
*** tsubic has left #openstack-sahara | 09:19 | |
openstackgerrit | Daniele Venzano proposed openstack/sahara: Add default templates for Spark 1.3.1 https://review.openstack.org/202456 | 09:22 |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara: Add cluster verification https://review.openstack.org/196713 | 09:29 |
openstackgerrit | Pino Toscano proposed openstack/sahara-image-elements: hadoop: restrict cloud-user home chmod for CentOS to v6 only https://review.openstack.org/202463 | 09:32 |
openstackgerrit | Andrey Pavlov proposed openstack/sahara: [WIP] Devices formatting and mounting https://review.openstack.org/200483 | 09:43 |
openstackgerrit | Pino Toscano proposed openstack/sahara-image-elements: hadoop: add vanilla/2.6 based on CentOS 7 https://review.openstack.org/202469 | 09:44 |
*** degorenko has joined #openstack-sahara | 09:46 | |
openstackgerrit | Daniele Venzano proposed openstack/sahara: Deprecate Spark 1.0.0 https://review.openstack.org/202471 | 09:56 |
openstackgerrit | Sergey Reshetnyak proposed openstack/sahara: [TEST COMMIT] Fix installing python-saharaclient https://review.openstack.org/202179 | 09:59 |
openstackgerrit | Pino Toscano proposed stackforge/sahara-ci-config: dib.sh: minor switch case refactor https://review.openstack.org/202474 | 10:01 |
openstackgerrit | Sergey Reshetnyak proposed openstack/sahara: [TEST COMMIT] Fix installing python-saharaclient https://review.openstack.org/202179 | 10:02 |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara-specs: Add cluster verification checks https://review.openstack.org/196576 | 10:03 |
openstackgerrit | Pino Toscano proposed stackforge/sahara-ci-config: Add a non-voting centos7 job for vanilla/2.6 https://review.openstack.org/202476 | 10:04 |
openstackgerrit | Daniele Venzano proposed openstack/sahara: Deprecate Spark 1.0.0 https://review.openstack.org/202471 | 10:15 |
openstackgerrit | Denis Egorenko proposed stackforge/sahara-ci-config: Set as voting jobs for Spark 1.3.1 plugin https://review.openstack.org/202487 | 10:29 |
*** coolsvap has quit IRC | 10:58 | |
*** coolsvap has joined #openstack-sahara | 10:58 | |
openstackgerrit | Sergey Reshetnyak proposed openstack/python-saharaclient: [TEST COMMIT] Test OS gate https://review.openstack.org/202499 | 10:59 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara: Implement recommendations for vanilla 2.6.0 https://review.openstack.org/177280 | 11:20 |
pino|work | hi degorenko, i guess i can ask you about the sahara ci? | 11:25 |
degorenko | pino|work, hi. Yep, sure | 11:25 |
pino|work | degorenko: is there a way to have https://review.openstack.org/#/c/202476/ tested? as in, do a s-i-e run with the modified configuration? | 11:26 |
pino|work | (thanks for reviewing it, btw!) | 11:27 |
openstackgerrit | Merged openstack/sahara: Remove the old scenario YAML files https://review.openstack.org/200949 | 11:27 |
degorenko | pino|work, for now it only manual. We have some map file for this, but it can be modified only manually. By me :) or skolekonov | 11:28 |
degorenko | pino|work, please see my comment here https://review.openstack.org/#/c/202474/1 | 11:30 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara: Implement recommendations for vanilla 2.6.0 https://review.openstack.org/177280 | 11:32 |
*** saneax has quit IRC | 11:33 | |
*** venza has quit IRC | 11:34 | |
pino|work | degorenko: oki, will change | 11:34 |
openstackgerrit | Sergey Reshetnyak proposed openstack/sahara: Fix problem with building sahara package https://review.openstack.org/202509 | 11:34 |
pino|work | degorenko: regarind the added job, is there a way to have it run, even manually? | 11:34 |
degorenko | pino|work, btw, yes, it can be tested. But you have added new job - and that is block - this patch should be merged | 11:34 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara-image-elements: Add elements for sync time on VM https://review.openstack.org/129611 | 11:34 |
degorenko | pino|work, but when you modified only scripts - that's easy to test. | 11:34 |
degorenko | pino|work, but you need ping me :) | 11:35 |
pino|work | that's what i'm doing? ;) | 11:35 |
*** venza has joined #openstack-sahara | 11:36 | |
degorenko | yes. i'm wait for your answer on first patch. If you agree - update please - and it'll be merged :) | 11:36 |
degorenko | both patches | 11:36 |
pino|work | ah, so i have to remove the Depends-On on the sahara-ci-config? | 11:37 |
*** skolekonov has joined #openstack-sahara | 11:37 | |
degorenko | also yes, depends-on to sahara-ci-config doesn't help you now. I'll add this feature later, when i will enough time for this | 11:38 |
pino|work | i see | 11:38 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara: Add recommendation support for Spark plugin https://review.openstack.org/193060 | 11:39 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara: Add recommendation support to Cloudera plugin https://review.openstack.org/193098 | 11:41 |
pino|work | degorenko: would it be possible to add some extra job in the zuul sahara-ci-config to trigger extra s-i-e checks? | 11:41 |
openstackgerrit | Andrey Pavlov proposed openstack/sahara: [WIP] Devices formatting and mounting https://review.openstack.org/200483 | 11:43 |
degorenko | no, because a) if you add new job - it should be merged at first; b) if you change some scripts - i'll add support depends-on for ci-config later | 11:43 |
degorenko | pino|work, ^^ | 11:43 |
pino|work | something like "check sie" as comment in a sahara-ci-config patch, would trigger the same checks as in s-i-e | 11:43 |
pino|work | ah ok | 11:44 |
*** skolekonov has left #openstack-sahara | 11:46 | |
*** skolekonov has joined #openstack-sahara | 11:46 | |
openstackgerrit | Pino Toscano proposed stackforge/sahara-ci-config: dib.sh: minor switch case refactor https://review.openstack.org/202474 | 11:47 |
openstackgerrit | Pino Toscano proposed stackforge/sahara-ci-config: Add a non-voting centos7 job for vanilla/2.6 https://review.openstack.org/202476 | 11:47 |
openstackgerrit | Sergey Reshetnyak proposed openstack/sahara: Fix problem with building sahara package https://review.openstack.org/202509 | 11:51 |
*** Longgeek has quit IRC | 11:54 | |
openstackgerrit | Sergey Reshetnyak proposed openstack/sahara: Fix installing python-saharaclient https://review.openstack.org/202179 | 11:59 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara: Add missing mako template for Spark 1.3.1 https://review.openstack.org/202519 | 12:00 |
*** _mattf is now known as mattf | 12:02 | |
*** mattf has quit IRC | 12:02 | |
*** mattf has joined #openstack-sahara | 12:02 | |
*** Longgeek has joined #openstack-sahara | 12:05 | |
*** venza has quit IRC | 12:10 | |
*** venza has joined #openstack-sahara | 12:11 | |
openstackgerrit | Merged stackforge/sahara-ci-config: dib.sh: minor switch case refactor https://review.openstack.org/202474 | 12:12 |
openstackgerrit | Merged stackforge/sahara-ci-config: Add a non-voting centos7 job for vanilla/2.6 https://review.openstack.org/202476 | 12:12 |
degorenko | tosky, ping | 12:12 |
pino|work | degorenko: thanks! | 12:16 |
*** tellesnobrega_ has joined #openstack-sahara | 12:16 | |
*** tellesnobrega_ has quit IRC | 12:16 | |
degorenko | pino|work, no problem | 12:16 |
openstackgerrit | Pino Toscano proposed openstack/sahara-image-elements: hadoop: add vanilla/2.6 based on CentOS 7 https://review.openstack.org/202469 | 12:17 |
openstackgerrit | Denis Egorenko proposed openstack/sahara: Derive Mako scenario templates for Spark 1.3.1 https://review.openstack.org/202526 | 12:18 |
*** nkrinner is now known as nkrinner_afk | 12:18 | |
openstackgerrit | Daniele Venzano proposed openstack/sahara: Deprecate Spark 1.0.0 https://review.openstack.org/202471 | 12:23 |
*** saneax has joined #openstack-sahara | 12:23 | |
openstackgerrit | Denis Egorenko proposed stackforge/sahara-ci-config: Set as voting jobs for Spark 1.3.1 plugin https://review.openstack.org/202487 | 12:24 |
tosky | degorenko: pong | 12:25 |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara: Added the ability to specify the name of the flavor_id https://review.openstack.org/201694 | 12:25 |
degorenko | tosky, hey, is it possible to add your patch with mako to stable/kilo? | 12:26 |
tosky | degorenko: uhm, aren't we using the tests from master for kilo? I remember a discussion before the summit and I think it was touched there ( SergeyLukjanov !!!) | 12:26 |
degorenko | tosky, we have only one branch :( | 12:27 |
tosky | degorenko: oh, I see, you checkout sahara and then reuse the same repository to run the tests; that's why the idea was to move the tests in another repository | 12:29 |
degorenko | tosky, i think, we need discussion with SergeyLukjanov | 12:30 |
SergeyLukjanov | what's up? | 12:30 |
degorenko | can we create kilo branch for ci-config? | 12:30 |
tosky | degorenko: yes; luckily the sahara meeting is less then two hours | 12:30 |
SergeyLukjanov | degorenko, do we really need it? | 12:31 |
tosky | SergeyLukjanov: so, support for mako templates in test scenario is in master only, which does not work for sahara from kilo, of course | 12:31 |
SergeyLukjanov | I would say - to keep kilo testing as is for now and wait for the separation to sahara-tests | 12:31 |
SergeyLukjanov | degorenko, we could make a branch if really needed | 12:31 |
SergeyLukjanov | I'm flying to SFO this weekend - will work on specs | 12:32 |
*** vgridnev has joined #openstack-sahara | 12:32 | |
degorenko | SergeyLukjanov, what do you think about this? | 12:32 |
SergeyLukjanov | degorenko, I prefer to avoid having branches in sahara-ci-config | 12:32 |
degorenko | we can always keep some hardcode :D | 12:32 |
SergeyLukjanov | degorenko, could we just have a few ifs? | 12:32 |
degorenko | SergeyLukjanov, then, we need keep old yaml | 12:32 |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara: Migrate "flavor_id" to "flavor" in scenario tests https://review.openstack.org/202077 | 12:32 |
degorenko | in sahara-ci-repo for example | 12:32 |
tosky | degorenko: uhm, the alternative could be to checkout master before running tests in the sahara repository - could it work? | 12:33 |
degorenko | tosky, how we can checkout to master on patch to kilo? | 12:33 |
SergeyLukjanov | degorenko, I was thinking that we're running yamls from stable/kilo | 12:33 |
degorenko | SergeyLukjanov, ah, yep | 12:33 |
degorenko | then we should keep some code in our sahara-ci | 12:34 |
degorenko | a few ifs :D | 12:34 |
SergeyLukjanov | yeah | 12:34 |
SergeyLukjanov | let's do it for now | 12:34 |
degorenko | okay, i'll do this | 12:34 |
tosky | degorenko: sorry, you will have to restore some of the code I killed :( | 12:35 |
degorenko | tosky, its okay, its my fault, i forged about kilo | 12:36 |
degorenko | forget* | 12:36 |
* tosky forgot too | 12:36 | |
degorenko | thanks :D | 12:36 |
*** saneax has quit IRC | 12:45 | |
*** saneax has joined #openstack-sahara | 12:53 | |
*** hdd has joined #openstack-sahara | 12:54 | |
*** Poornima has quit IRC | 12:59 | |
openstackgerrit | Sergey Reshetnyak proposed openstack/sahara: Fix installing python-saharaclient https://review.openstack.org/202179 | 13:10 |
*** _crobertsrh is now known as crobertsrh | 13:11 | |
*** coolsvap is now known as coolsvap|away | 13:13 | |
*** tmckay has quit IRC | 13:20 | |
*** tmckay has joined #openstack-sahara | 13:20 | |
*** tmckay has quit IRC | 13:21 | |
*** tmckay has joined #openstack-sahara | 13:22 | |
*** hdd has quit IRC | 13:24 | |
tmckay | SergeyLukjanov, ping | 13:24 |
vgridnev | Sahara code was moved to contrib in horizon! The change which aims that was merged | 13:24 |
tmckay | yay! that's good, right? | 13:25 |
*** hdd has joined #openstack-sahara | 13:26 | |
vgridnev | Maybe after that our changes will be merged faster | 13:27 |
crobertsrh | heh, I just noticed the contrib merge, vgridnev. Let the great rebase begin :) | 13:29 |
vgridnev | I already started that | 13:30 |
*** vgridnev has quit IRC | 13:33 | |
*** vgridnev has joined #openstack-sahara | 13:35 | |
*** hdd has quit IRC | 13:40 | |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara: Migrate "flavor_id" to "flavor" in scenario tests https://review.openstack.org/202077 | 13:46 |
*** esikachev has joined #openstack-sahara | 13:51 | |
*** hdd has joined #openstack-sahara | 13:55 | |
*** weiting has joined #openstack-sahara | 13:57 | |
tosky | esikachev: hi! Are you going to implement https://bugs.launchpad.net/sahara/+bug/1475301 ? It's an easy fix, I can do it (and sorry for that) | 14:01 |
openstack | Launchpad bug 1475301 in Sahara "[scenario][docs] Invalid command for run scenario tests" [Undecided,New] - Assigned to Evgeny Sikachev (esikachev) | 14:01 |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara: Fix README.rst in scenario dir https://review.openstack.org/202579 | 14:09 |
openstackgerrit | Chad Roberts proposed openstack/python-saharaclient: TEST commit, hoping to fix gate. https://review.openstack.org/202580 | 14:10 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara: Add missing mako template for Spark 1.3.1 https://review.openstack.org/202519 | 14:13 |
*** egafford has joined #openstack-sahara | 14:19 | |
*** skolekonov has quit IRC | 14:21 | |
*** ylobankov has joined #openstack-sahara | 14:29 | |
openstackgerrit | Ethan Gafford proposed openstack/python-saharaclient: Adding interface argument for job template and job https://review.openstack.org/190813 | 14:32 |
openstackgerrit | Sergey Reshetnyak proposed openstack/sahara: Fix problem with building sahara package https://review.openstack.org/202509 | 14:36 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara: Implement recommendations for vanilla 2.6.0 https://review.openstack.org/177280 | 14:38 |
*** skolekonov has joined #openstack-sahara | 14:39 | |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara: Migrate "flavor_id" to "flavor" in scenario tests https://review.openstack.org/202077 | 14:41 |
tosky | vgridnev: in order to not forget, do you think it make sense to file a bug about "add missing scenario tests for this feature" when your review is merged? | 14:42 |
*** weiting_ has joined #openstack-sahara | 14:42 | |
vgridnev | I will think about that, maybe it makes sense | 14:43 |
*** weiting has quit IRC | 14:43 | |
*** Longgeek has quit IRC | 14:58 | |
*** esikachev has left #openstack-sahara | 15:00 | |
*** weiting_ has quit IRC | 15:00 | |
*** Longgeek has joined #openstack-sahara | 15:00 | |
openstackgerrit | Denis Egorenko proposed stackforge/sahara-ci-config: Back using old format scenarion tests for Kilo support https://review.openstack.org/202621 | 15:02 |
openstackgerrit | Denis Egorenko proposed stackforge/sahara-ci-config: Do not use -s for spark plugin https://review.openstack.org/199150 | 15:05 |
*** coolsvap|away is now known as coolsvap | 15:06 | |
openstackgerrit | Denis Egorenko proposed stackforge/sahara-ci-config: Temporary use local repo for mapr eco https://review.openstack.org/191754 | 15:06 |
*** Longgeek has quit IRC | 15:07 | |
*** mattf is now known as _mattf | 15:17 | |
vgridnev | egafford, I think, that you can remove your -1 because related changes was merged | 15:28 |
vgridnev | https://review.openstack.org/#/c/193098/ | 15:28 |
egafford | vgridnev: True enough. | 15:31 |
egafford | Done; thanks. | 15:31 |
vgridnev | ok, thanks | 15:31 |
*** vgridnev has quit IRC | 15:40 | |
*** ig0r_ has quit IRC | 15:40 | |
*** coolsvap is now known as coolsvap|away | 16:05 | |
*** pino|work has quit IRC | 16:07 | |
*** vgridnev has joined #openstack-sahara | 16:24 | |
*** vgridnev has quit IRC | 16:30 | |
tosky | degorenko: missing resources on the Sahara CI? | 16:31 |
tosky | degorenko: I see some strange errors in few jobs | 16:31 |
degorenko | tosky, which one? | 16:32 |
tosky | degorenko: for example Quota exceeded for port: Requested 1, but available 0 here https://sahara.mirantis.com/logs/69/202469/2/check/dib-neutron-heat-vanilla_2.6.0-c6.6/064126c/console.html | 16:33 |
degorenko | thats probably problem with neutron ports | 16:33 |
degorenko | i'll clean them | 16:33 |
degorenko | we got this error second times :( | 16:33 |
tosky | I see also other errors about scaling step failed, maybe a glitch in the process, but I thought about resources | 16:34 |
degorenko | also, i think we need update devstack on our servers | 16:36 |
degorenko | but i don't want to do that now :D after weekend i think | 16:37 |
tosky | oh, sure :) | 16:37 |
degorenko | try to recheck your patch | 16:38 |
degorenko | it shouldn't have quotes issue now | 16:38 |
degorenko | quotas | 16:38 |
tosky | thanks, let's retry | 16:39 |
*** skolekonov has quit IRC | 16:49 | |
*** htruta_ has joined #openstack-sahara | 17:46 | |
*** henriquetruta has joined #openstack-sahara | 17:57 | |
*** henriquetruta has quit IRC | 17:58 | |
*** tosky has quit IRC | 18:07 | |
*** degorenko has quit IRC | 18:07 | |
*** htruta_ has quit IRC | 18:18 | |
*** Nikolay_St has quit IRC | 18:22 | |
openstackgerrit | Chad Roberts proposed openstack/sahara: Adding default templates for spark 1.3.1 https://review.openstack.org/202741 | 18:44 |
*** htruta_ has joined #openstack-sahara | 19:04 | |
*** hdd has quit IRC | 19:04 | |
openstackgerrit | Merged openstack/sahara: Fix README.rst in scenario dir https://review.openstack.org/202579 | 19:07 |
*** hdd has joined #openstack-sahara | 19:09 | |
*** htruta_ has quit IRC | 19:21 | |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara-image-elements: Add elements for sync time on VM https://review.openstack.org/129611 | 19:28 |
tellesnobrega | hey, is there a way to pass args when running MapReduce job? | 19:28 |
*** Nikolay_St has joined #openstack-sahara | 19:29 | |
*** hdd has quit IRC | 19:33 | |
elmiko | this keystone session thing is kicking my ass... | 19:46 |
elmiko | we do so many gymnastic leaps of faith with the context token information it's tough to lock them all down | 19:46 |
elmiko | egafford: i'm looking over the trust based creation stuff, got a minute to chat about it? | 19:52 |
egafford | elmiko: Sure; absolutely. | 19:52 |
elmiko | so, the reasoning for this is that some operations will take longer than the user is logged in for? | 19:53 |
egafford | There are two reasons for this. One is long-running operations, as you've cited. The other is cluster deletion outside the context of a tenant request (in periodic admin-initiated cleanup jobs.) | 19:54 |
elmiko | cluster deletion i get, no issue there | 19:54 |
elmiko | i'm wondering about cluster creation, because some of this session stuff involves auto re-authentication that is handled by keystoneclient | 19:55 |
egafford | I think you're going to suggest that we can refresh auth in keystone sessions for long-running ops, invalidating that need for trusts? | 19:55 |
egafford | Heh. | 19:55 |
elmiko | yea, i'm curious | 19:55 |
elmiko | does tosky have a test that will reproduceably show this creation error? | 19:55 |
egafford | Which creation error are we talking about? | 19:56 |
egafford | The error he found which drove me to address this change was r/t deletion. | 19:56 |
elmiko | ah, ok | 19:56 |
elmiko | so this is preventitive maintenance on the creation end? | 19:56 |
egafford | The repro steps for that are simple: | 19:57 |
egafford | 1) Create a cluster. Allow it to reach "configuring" state. | 19:57 |
egafford | 2) Stop Sahara. Twiddle the values in the periodic tasks module to cause the cluster cleanup job to occur quickly. Restart Sahara. | 19:57 |
egafford | 3) Observe explosion in logs. | 19:57 |
elmiko | but that's deletion right? | 19:58 |
egafford | On the creation end, yes, I believe this is preventive maintenance, but it's reasonable maint given that real clusters are likely larger than our tests. Yes, that is absolutely deletion. | 19:58 |
elmiko | and the whole stoping/starting sahara is definitely an issue | 19:58 |
egafford | But we don't know which clusters we'll need to delete, so we need to treat them all with suspicion. | 19:58 |
elmiko | right | 19:58 |
elmiko | argh | 19:58 |
egafford | Given that we need the trusts for deletion, do you have a suggestion that improves our lives on the creation side? | 19:59 |
elmiko | if i understand the issue, then i would think that sessions might help as they will reauthenticate when the token expires | 19:59 |
elmiko | which would mean that long creation times would be ok, as the user's token would re-auth as necessary | 20:00 |
egafford | I see where you're going, but I'm not sure what should be done about it. Particularly given your "argh" right there. We could certainly use sessions on the create side, and only use the trusts for deletion. That'd basically just be removing the trusts.use_admin_context_thingummy call. | 20:00 |
elmiko | yea | 20:00 |
elmiko | i'm just having issues bringing this all together because we muck with the values in the context seemingly at will | 20:01 |
egafford | elmiko: Yeah, that's a bit of a problem. | 20:01 |
elmiko | and since we rely on username/token authentication for everything it makes this difficult to share a connection and use the, imo, more powerful aspects of sessions | 20:01 |
egafford | I was kind of floored that our get_admin_context method just returned a new context that says it's an admin and has no auth token or service catalog. | 20:01 |
egafford | There's definitely some room for better encapsulation on that point. | 20:02 |
elmiko | that is_admin value gets passed in to the oslo.context object though, so i think it may be doing some magic there | 20:02 |
*** tosky has joined #openstack-sahara | 20:03 | |
elmiko | its tough to face having to abandon all this work... | 20:03 |
egafford | Hopefully yes. And it's probably very reasonable that an admin context can't automatically delete tenant nodes. | 20:03 |
egafford | To which work are you referring? | 20:04 |
elmiko | the sessions stuff | 20:04 |
egafford | Ah, I see. Well, we can use them for long ops, and keep the trusts for deletion, if there's an upside they give us. | 20:04 |
egafford | Reduced client creation is nice. | 20:04 |
egafford | Not sure if it's a driving need, I'll admit. | 20:04 |
elmiko | yea, reduced client creation, connection pooling, concentration of security updates | 20:05 |
egafford | Those are all nice things. | 20:05 |
elmiko | just not sure i can put all this stuff back in the box :/ | 20:05 |
egafford | I wouldn't mind a single thing in that list. Sure, absolutely. In the trust change, if we have sessions, we don't need to actually use the trust on the tenant request, so that frees you up there. | 20:06 |
elmiko | yea, just not sure i can get this done in time. the more i touch, the more this grows | 20:07 |
egafford | As for All The Rest of the context mucking we do, those may be more problematic on a case-by-case basis, but I don't think anything in the trusts change prohibits your change. Sure; sensible. | 20:07 |
elmiko | yea, agreed. | 20:08 |
elmiko | oh well, thanks for talking it out with me =) | 20:09 |
egafford | elmiko: I think just removing "trusts.use_os_admin_auth_token(cluster)" from _prepare_provisioning is your path to victory if you make yourself dependent on the trust change. | 20:09 |
egafford | That's the place where we actually futz with the context. | 20:10 |
elmiko | right, that would work | 20:10 |
elmiko | i need to modify how i'm authenticating the sessions | 20:10 |
elmiko | sadly, this gets into some deep keystone territory and i need to think through it a little more | 20:10 |
egafford | Sure. | 20:10 |
elmiko | it would also be nice to confer with AndreyPavlov, but he's sleeping currently. i'm gonna work on an email to the list i thinkg | 20:11 |
egafford | I do wonder in a large production context about the viability of having so many workers waiting on long ops. Feel like all this wants to be much more of a DAG where arbitrary workers wake up and take actions on the next cluster that needs action or testing. But that'd be a Whole Thing. | 20:12 |
egafford | And totally the opposite of what you're doing atm. :) | 20:12 |
elmiko | its a nice idea though, better streamlining of these ops | 20:13 |
elmiko | and i think it could compliment what i'm doing, as we would gain the ability for tokens to re-auth as necessary | 20:13 |
egafford | Anyway, sounds like you're on a sane and just course. Oh, okay, if we can do that, yeah, what you're doing is absolutely on the road there. | 20:14 |
egafford | (Great!) | 20:14 |
elmiko | heh, we'll see about sane... | 20:14 |
elmiko | ;) | 20:14 |
egafford | Okay, well, that leaves just regardless, and I don't care to associate with anyone who'd rather be sane than just, so all's well. | 20:15 |
elmiko | hahaha! | 20:15 |
elmiko | well played sir /me tips fedora | 20:16 |
elmiko | i think i need to follow more of the tmckay principle on this one; code like a mad man, then right the spec | 20:16 |
elmiko | write even... | 20:16 |
egafford | It is nice to have a bit of a POC before the spec. | 20:17 |
*** hdd has joined #openstack-sahara | 20:17 | |
egafford | Otherwise you end up being like me on the UJIM and dropping an entire table. | 20:17 |
elmiko | yea, i thought i did, but it wasn't deep enough | 20:17 |
elmiko | good point | 20:17 |
egafford | (Smooth move!) Yeah. It's hard to know where to draw the line. | 20:17 |
elmiko | c'est la vie | 20:17 |
egafford | C'est. | 20:17 |
*** hdd has quit IRC | 20:38 | |
*** hdd has joined #openstack-sahara | 20:58 | |
*** crobertsrh is now known as _crobertsrh | 21:06 | |
tmckay | elmiko, that is definitely the right plan | 21:19 |
elmiko | tmckay: which one? | 21:20 |
tmckay | oh, write the code first | 21:21 |
elmiko | yea, gonna try to practice that more frequently | 21:22 |
openstackgerrit | Merged openstack/sahara: Fix installing python-saharaclient https://review.openstack.org/202179 | 22:17 |
tosky | only if you write the test before | 22:34 |
tosky | :D | 22:34 |
elmiko | tosky: very nice sir | 22:43 |
*** hogepodge has quit IRC | 23:02 | |
*** witlessb has quit IRC | 23:17 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/sahara: Updated from global requirements https://review.openstack.org/202290 | 23:21 |
*** tosky has quit IRC | 23:21 | |
*** zigo has quit IRC | 23:21 | |
*** zigo has joined #openstack-sahara | 23:22 | |
*** openstackstatus has joined #openstack-sahara | 23:34 | |
*** ChanServ sets mode: +v openstackstatus | 23:34 | |
*** barra204 has joined #openstack-sahara | 23:54 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!