| *** r-daneel has quit IRC | 00:23 | |
| *** Dinesh__Bhor has joined #openstack-glance | 00:28 | |
| *** Dinesh__Bhor has quit IRC | 00:31 | |
| *** zhurong has joined #openstack-glance | 00:42 | |
| *** wxy has joined #openstack-glance | 01:08 | |
| *** dalgaaf has quit IRC | 01:13 | |
| *** dalgaaf has joined #openstack-glance | 01:15 | |
| *** links has joined #openstack-glance | 01:28 | |
| *** links has quit IRC | 01:29 | |
| *** Dinesh__Bhor has joined #openstack-glance | 01:38 | |
| *** Dinesh__Bhor has quit IRC | 01:40 | |
| *** openstackgerrit has joined #openstack-glance | 01:55 | |
| openstackgerrit | Xianjin Shao proposed openstack/glance master: Description about the parmeters 'op', 'path' and 'value' when update image property or metadata https://review.openstack.org/541508 | 01:55 |
|---|---|---|
| *** masber has quit IRC | 01:59 | |
| *** gyee has quit IRC | 02:01 | |
| *** links has joined #openstack-glance | 02:12 | |
| *** itlinux has joined #openstack-glance | 02:14 | |
| openstackgerrit | wangdequn proposed openstack/glance master: correct grammar, duplicate a found https://review.openstack.org/541514 | 02:19 |
| *** Dinesh__Bhor has joined #openstack-glance | 02:23 | |
| *** Dinesh__Bhor has quit IRC | 02:23 | |
| *** Dinesh__Bhor has joined #openstack-glance | 02:28 | |
| *** Dinesh__Bhor has quit IRC | 02:29 | |
| openstackgerrit | wangdequn proposed openstack/glance master: correct grammar, duplicate the found https://review.openstack.org/541518 | 02:39 |
| *** Supun has joined #openstack-glance | 02:45 | |
| *** Supun has quit IRC | 02:48 | |
| *** Supun has joined #openstack-glance | 02:48 | |
| *** gaurangt has quit IRC | 02:57 | |
| *** gaurangt has joined #openstack-glance | 03:00 | |
| *** namnh has joined #openstack-glance | 03:18 | |
| *** itlinux has quit IRC | 03:25 | |
| *** pdeore has joined #openstack-glance | 03:38 | |
| *** Supun has quit IRC | 03:42 | |
| *** bkopilov has quit IRC | 03:51 | |
| *** udesale has joined #openstack-glance | 03:56 | |
| *** openstackgerrit has quit IRC | 04:04 | |
| *** abhishekk has joined #openstack-glance | 04:06 | |
| *** mine0901 has quit IRC | 04:06 | |
| *** pdeore has quit IRC | 04:07 | |
| *** itlinux has joined #openstack-glance | 04:08 | |
| *** rosmaita has quit IRC | 04:14 | |
| *** mvk has quit IRC | 04:21 | |
| *** harlowja has quit IRC | 04:32 | |
| *** nicolasbock has quit IRC | 04:32 | |
| *** pgadige__ has joined #openstack-glance | 04:38 | |
| *** namnh has quit IRC | 04:42 | |
| *** aniketh has quit IRC | 04:46 | |
| *** pooja-jadhav is now known as neha_alhat | 04:49 | |
| *** ratailor has joined #openstack-glance | 05:03 | |
| *** pdeore has joined #openstack-glance | 05:04 | |
| *** itlinux has quit IRC | 05:06 | |
| *** bkopilov has joined #openstack-glance | 05:33 | |
| Shilpa | abhishek: hi | 05:37 |
| abhishekk | Shilpa, hi | 05:37 |
| *** harlowja has joined #openstack-glance | 05:37 | |
| Shilpa | Kindly go through Alan's comment on https://review.openstack.org/#/c/433934/ | 05:38 |
| Shilpa | what i understood is during exit, expects either 0 or 1 as status | 05:38 |
| Shilpa | 0 for success and 1 for failure | 05:39 |
| Shilpa | if exit with message, it considering as failure case | 05:39 |
| *** zhurong has quit IRC | 05:39 | |
| abhishekk | Shilpa, I guess he is right | 05:42 |
| abhishekk | this exit message triggers as failure but in actual its not a failure case | 05:43 |
| abhishekk | to get rid of this you need to submit a new patch | 05:43 |
| abhishekk | and add this bug as related-bug tag https://bugs.launchpad.net/tripleo/+bug/1747690 | 05:44 |
| openstack | Launchpad bug 1747690 in tripleo "master promotion: Failed to call refresh: glance-manage db_sync" [Critical,Triaged] | 05:44 |
| abhishekk | instead using sys.exit you should print that message and return from their | 05:45 |
| Shilpa | so we need to print the user friendly message and need to return 0 for sucess and 1 for failure | 05:46 |
| abhishekk | Shilpa, no just print that messge as it is instead of sys.exit() and add return statement (without 0 or 1) | 05:48 |
| abhishekk | I guess that would be sufficient | 05:49 |
| abhishekk | and after pushing that patch, add Alen as a reviewer | 05:49 |
| abhishekk | Shilpa, push that patch as early as possible, even it is trivial for glance it's critical for tripleO | 05:52 |
| Shilpa | yes | 05:52 |
| *** zhurong has joined #openstack-glance | 05:53 | |
| Shilpa | one more thing | 05:53 |
| Shilpa | we have sys.exit for expand / migrate and contract also | 05:53 |
| Shilpa | there also we need to handle this | 05:53 |
| abhishekk | yes | 05:55 |
| abhishekk | but don't make change in _validate_engine | 05:55 |
| Shilpa | there are some failure case, so there we should exit with 1, right | 05:55 |
| abhishekk | Shilpa, that's not required | 05:56 |
| abhishekk | if you are exiting with error message then it considered as it is exited with 1 | 05:56 |
| Shilpa | ok, understood | 05:56 |
| Shilpa | so in Success case, we need to return and print the message | 05:57 |
| abhishekk | but the problem is then line pointed out by Alen it's not a failure so it should not exit with 1 or error message | 05:57 |
| abhishekk | correct | 05:57 |
| Shilpa | got it, will push the patch ASAP, thanks | 05:57 |
| abhishekk | so in expand, migrate and contract wherever you are exiting with message 'Database is up to date. No migrations needed.' you should print that message and return from their | 05:58 |
| abhishekk | Shilpa, just do one thing, print the message and instead of return just use system.exit() | 06:00 |
| Shilpa | with 0? | 06:00 |
| Shilpa | right | 06:00 |
| *** openstackgerrit has joined #openstack-glance | 06:03 | |
| openstackgerrit | Merged openstack/glance master: Adds 'web-download' import method https://review.openstack.org/532501 | 06:03 |
| abhishekk | Shilpa, sys.exit() is equivalent to sys.exit(0) | 06:04 |
| *** e0ne has joined #openstack-glance | 06:11 | |
| *** e0ne has quit IRC | 06:17 | |
| *** mosulica has joined #openstack-glance | 06:36 | |
| *** e0ne has joined #openstack-glance | 06:48 | |
| *** harlowja has quit IRC | 06:52 | |
| *** e0ne has quit IRC | 06:54 | |
| openstackgerrit | Xianjin Shao proposed openstack/glance master: show the default value of limit with the description of API when we use this API to show the information of image but the count of images is more than 25 and we don't know the default limit from API information. https://review.openstack.org/541577 | 06:57 |
| *** udesale has quit IRC | 07:02 | |
| *** udesale has joined #openstack-glance | 07:05 | |
| *** udesale has quit IRC | 07:06 | |
| *** udesale has joined #openstack-glance | 07:06 | |
| *** alexchadin has joined #openstack-glance | 07:10 | |
| *** AlexeyAbashkin has joined #openstack-glance | 07:22 | |
| *** AlexeyAbashkin has quit IRC | 07:42 | |
| *** udesale_ has joined #openstack-glance | 07:48 | |
| *** rcernin has quit IRC | 07:48 | |
| *** udesale has quit IRC | 07:51 | |
| *** pliu has quit IRC | 07:51 | |
| *** AlexeyAbashkin has joined #openstack-glance | 07:54 | |
| openstackgerrit | Shilpa Devharakar proposed openstack/glance master: Exiting with user friendly message and SystemExit() https://review.openstack.org/541601 | 07:59 |
| *** pliu has joined #openstack-glance | 08:03 | |
| *** alexchadin has quit IRC | 08:04 | |
| *** alexchadin has joined #openstack-glance | 08:16 | |
| *** threestrands has quit IRC | 08:22 | |
| *** tesseract has joined #openstack-glance | 08:24 | |
| *** zhurong has quit IRC | 08:25 | |
| *** zhurong has joined #openstack-glance | 08:26 | |
| Shilpa | abhishek: Hi | 08:35 |
| Shilpa | Kindly review https://review.openstack.org/#/c/541601/, pused for LP bug https://bugs.launchpad.net/glance/+bug/1747690 (master promotion: Failed to call refresh: glance-manage db_syn) | 08:36 |
| openstack | Launchpad bug 1747690 in Glance "master promotion: Failed to call refresh: glance-manage db_sync" [Undecided,In progress] - Assigned to Shilpa Devharakar (shilpasd) | 08:36 |
| *** tshefi_ has joined #openstack-glance | 08:37 | |
| *** pcaruana has joined #openstack-glance | 08:38 | |
| *** alexchadin has quit IRC | 08:39 | |
| rabel | hey there. could we merge https://review.openstack.org/#/c/534863/ ? it's just an indentation fix for the docs | 08:40 |
| Shilpa | Thank you Abhishek for reviewing https://review.openstack.org/#/c/541601/ | 08:40 |
| *** ratailor has quit IRC | 08:49 | |
| *** ratailor has joined #openstack-glance | 08:49 | |
| *** e0ne has joined #openstack-glance | 09:26 | |
| *** neha_alhat is now known as pooja_jadhav | 09:27 | |
| *** pcaruana has quit IRC | 09:57 | |
| *** namnh has joined #openstack-glance | 09:57 | |
| *** alexchadin has joined #openstack-glance | 10:00 | |
| *** pcaruana has joined #openstack-glance | 10:13 | |
| *** krypto has joined #openstack-glance | 10:19 | |
| krypto | hi all is there a way to see all the config values used by running glance process | 10:20 |
| *** dtantsur|afk is now known as dtantsur | 10:26 | |
| *** gcb has quit IRC | 10:29 | |
| *** zhurong has quit IRC | 10:35 | |
| *** mosulica has quit IRC | 10:40 | |
| *** abhishekk has quit IRC | 10:41 | |
| *** udesale_ is now known as udesale | 10:41 | |
| *** mosulica has joined #openstack-glance | 10:42 | |
| pbourke | krypto: vi /etc/kolla/glance/* | 10:51 |
| krypto | its not running as docker process , started as " python /usr/bin/glance-api --config-file=/etc/glance/glance-api.conf --log-file=/var/log/glance/glance-api.log" is there a way to find out the variables its using . | 10:55 |
| *** aniketh has joined #openstack-glance | 10:58 | |
| aniketh | yay! I was successful in setting up devstack. I wanted to know, like do we pull and update the code every time with master while developing with devstack? Also, can I be pointed towards some beginner level bugs that I could dig into? | 11:01 |
| aniketh | clarkb: ^ | 11:02 |
| rabel | krypto: wouldnt that be just to have a look into /etc/glance/glance-api.conf ? in addition you could have a look at the config reference in the docs to see the default values for parameters not explicitly mentioned | 11:02 |
| *** pcaruana has quit IRC | 11:09 | |
| *** namnh has quit IRC | 11:28 | |
| *** udesale has quit IRC | 11:30 | |
| *** pdeore has quit IRC | 11:35 | |
| *** mvk has joined #openstack-glance | 11:36 | |
| *** nicolasbock has joined #openstack-glance | 11:41 | |
| *** ratailor has quit IRC | 11:43 | |
| krypto | Thanks rabel i was looking for default values of glance,config reference documentation helped | 11:45 |
| *** bkopilov has quit IRC | 11:55 | |
| *** edmondsw has joined #openstack-glance | 12:02 | |
| *** pcaruana has joined #openstack-glance | 12:13 | |
| *** alexchadin has quit IRC | 12:16 | |
| *** alexchadin has joined #openstack-glance | 12:16 | |
| *** alexchadin has quit IRC | 12:21 | |
| *** alexchadin has joined #openstack-glance | 12:21 | |
| *** AlexeyAbashkin has quit IRC | 12:25 | |
| jokke_ | aniketh: cool ... so yes rebasing to the master and cherrypicking you own change on top of that + glance restart would be the correct workflow to test your changes manually | 12:37 |
| jokke_ | aniketh: I would advice not to work on the devstack's repo when you are actually doing your work, as you might need/want to blow that every now and then to get fresh devstack | 12:38 |
| *** AlexeyAbashkin has joined #openstack-glance | 12:40 | |
| jokke_ | aniketh: I for example have local repo on my laptop I do the development on and I have 2-3 devstacks running in containers on headless workstation I push the changes into. Obviously how and where you run your devstacks is matter of preference, just remember that you might want to blow that up every now and then :P | 12:40 |
| Shilpa | jokke: Hi | 12:41 |
| jokke_ | Shilpa: \o | 12:41 |
| Shilpa | Thank you for reviewing https://review.openstack.org/#/c/541601/ | 12:41 |
| jokke_ | np, it was pretty straight forward and made lots of sense within it's scope | 12:41 |
| jokke_ | I'd have been way more hesitant on approving that if those calls were introduced year ago :D | 12:42 |
| Shilpa | ok | 12:44 |
| *** pgadige__ has left #openstack-glance | 12:49 | |
| jokke_ | it's one of those things that even if it's wrong, but it's established behaviour that people's workflows might be depending on | 12:51 |
| jokke_ | but in this case it was pretty clear that no-one should be depending on those cases returning 1 so lets fix it while we still can :D | 12:51 |
| *** alexchadin has quit IRC | 12:53 | |
| *** alexchadin has joined #openstack-glance | 12:54 | |
| bhagyashris | jokke_, Hi, I just want guidance regarding the Add hooks or to make task configurable (reference: https://review.openstack.org/#/c/528754/) as in configuration file we are giving the plugin module name to load that task so is it possible we can give the class name or class path of tasks. | 12:54 |
| bhagyashris | jokke_, Actually I am trying to achive is like under plugins dir there are separate modules for each task so i want to define multiple task in one module and as per the user choice if he/she want to excute or load specific task with the help of configuration option then is it possible we can configure that by giving the class path etc like that | 12:56 |
| *** udesale has joined #openstack-glance | 12:58 | |
| *** alexchadin has quit IRC | 12:58 | |
| jokke_ | bhagyashris: so it should be the name defined for it in the setup.cfg | 12:58 |
| jokke_ | not exactly sure I do understand your question | 12:58 |
| jokke_ | the plugin config option is list of the plugins defined in the setup.cfg | 12:59 |
| jokke_ | unless I messed something seriously up | 12:59 |
| bhagyashris | jokke_, yeah that are the list of plugins defined in setup.cfg | 13:02 |
| bhagyashris | : get_flow method | 13:03 |
| jokke_ | and the get_import_plugins in glance/async/flows/plugins/__init__ takes that list and returns the flows of those ones defined in that config opt list | 13:04 |
| bhagyashris | yeah | 13:05 |
| jokke_ | did that answer your question or did I miss something? | 13:05 |
| *** rosmaita has joined #openstack-glance | 13:06 | |
| bhagyashris | jokke_, actually what exactly my question is for example there is no_op.py module in that there is only one _Noop task so to load that task user will need to give the no_op but if in no_op.py there are 3 different task for example _Noop1, _Noop2 and _Noop3 then user want to configure only _Noop1 and _Noop2 and not _Noop3 then is it possible we can load first to task only using the configuration option | 13:23 |
| bhagyashris | s/to/two | 13:24 |
| bhagyashris | i mean to say is it possible we can load task by giving the class name in the configuration | 13:29 |
| aniketh | jokke_: Oh, okay! So you have Glance setup in your host system itself to do your work? and Use devstack just for functional testing? | 13:33 |
| *** AlexeyAbashkin has quit IRC | 13:34 | |
| *** AlexeyAbashkin has joined #openstack-glance | 13:35 | |
| *** bkopilov has joined #openstack-glance | 13:35 | |
| rosmaita | jokke_ do we need to do 14.0.1 and 15.0.1 releases from the stable branches? I am thinking that the changes in the glance/cmd files make this not simply a testing issue | 13:37 |
| *** AlexeyAbashkin has quit IRC | 13:40 | |
| *** Supun has joined #openstack-glance | 13:41 | |
| openstackgerrit | Merged openstack/glance_store master: Imported Translations from Zanata https://review.openstack.org/538162 | 13:44 |
| bhagyashris | rosmaita, Hi Good morning! want to discuss regarding comments given on patch https://review.openstack.org/#/c/455837/ by you and some replies given by abhishekk so i would like to see your opinion on it. | 13:46 |
| rosmaita | bhagyashris sure, looking now | 13:48 |
| bhagyashris | rosmaita: ok thank you :) | 13:48 |
| *** alexchadin has joined #openstack-glance | 13:54 | |
| *** alexchadin has quit IRC | 13:55 | |
| *** AlexeyAbashkin has joined #openstack-glance | 13:55 | |
| *** alexchadin has joined #openstack-glance | 13:56 | |
| *** alexchadin has quit IRC | 13:56 | |
| *** alexchadin has joined #openstack-glance | 13:56 | |
| *** markvoelker has joined #openstack-glance | 13:57 | |
| *** alexchadin has joined #openstack-glance | 13:57 | |
| *** alexchadin has quit IRC | 13:57 | |
| *** alexchadin has joined #openstack-glance | 13:58 | |
| *** alexchadin has quit IRC | 13:58 | |
| *** alexchadin has joined #openstack-glance | 13:59 | |
| *** alexchadin has quit IRC | 13:59 | |
| jokke_ | bhagyashris: ahh, now I got you ... so you will put the ones that needs to be separate, to separate files and configure them individually | 14:01 |
| *** alexchadin has joined #openstack-glance | 14:02 | |
| jokke_ | bhagyashris: that way they stay properly decomposable and clear | 14:02 |
| jokke_ | rosmaita: i think we should | 14:03 |
| rosmaita | jokke_ ok, i will get patches up | 14:03 |
| rosmaita | jokke_ not sure if this is relevant to bhagyashris ' | 14:03 |
| rosmaita | s question or not, but https://review.openstack.org/#/c/527635/12/glance/opts.py@33 | 14:03 |
| jokke_ | rosmaita: btw, thanks for the web-download merge ...looks like gate is behaving as well finally :D | 14:03 |
| rosmaita | i'm still wondering about that | 14:03 |
| rosmaita | jokke_ the 'check' queue got up to around 500 yesterday | 14:04 |
| *** r-daneel has joined #openstack-glance | 14:04 | |
| rosmaita | infra must have really speeded things up for everything to get through | 14:04 |
| rosmaita | i didn't expect it to have merged yet! | 14:04 |
| jokke_ | yeah | 14:05 |
| rosmaita | mriedem did a really thorough review of the stable backports yesterday evening, btw | 14:05 |
| jokke_ | I guess they finally found the reason what was slowing everything down | 14:05 |
| jokke_ | I saw | 14:05 |
| jokke_ | rosmaita: so answer to that question, that's unfortunately what is needed for the plogin engine to work properly ... I guess we could build up some hack that will import and create those entries dynamically, but it would be really messy to read and try to follow | 14:07 |
| rosmaita | yeah, and i was also wondering about how operators modify setup.cfg to tell stevedore about their custom plugins? | 14:08 |
| jokke_ | I'm not exactly worried that a person who needs to write a taskflow specifically interacting with glance stuff in a specific format needs to add import and couple of definitions for it in 2 other files as well | 14:08 |
| rosmaita | well, it means that you must patch glance, can't just drop in your files | 14:09 |
| jokke_ | we could obviously move those plugin entries from opts.py to their own file to make that easier and cleaner | 14:09 |
| *** links has quit IRC | 14:09 | |
| rosmaita | yeah, so do you think it's possible to say 2.6 is current, but the plugin backend is experimental in Queens? | 14:09 |
| rosmaita | the plugin backend doesn't have a defined api | 14:10 |
| jokke_ | the mechanism is integrated to the import flow | 14:10 |
| jokke_ | and that | 14:10 |
| jokke_ | it's not part of the API per se | 14:10 |
| jokke_ | so those are kind of interlocked and same time decoupled | 14:11 |
| jokke_ | I'm currently working on documentation for all that plugin stuff. Also changing how those plugins are enabled in future will not change the API | 14:11 |
| *** r-daneel has quit IRC | 14:12 | |
| rosmaita | yeah, so i was thinking we can say that the end-user-facing API is locked down, but the operator-facing-plugin interface is open for a comment period | 14:12 |
| jokke_ | I'd say even that is not really gonna change per se as even if we introduce dynamic importing for those plugins the current model will still work | 14:13 |
| jokke_ | as said, I was looking into it and it just gets very messy very quickly when you start doing dynamic imports and is really breaking all the principles what PEP8 lays down for imports | 14:15 |
| rosmaita | ok, i'll put up a draft of releasenotes in a bit | 14:15 |
| openstackgerrit | Merged openstack/glance master: Exiting with user friendly message and SystemExit() https://review.openstack.org/541601 | 14:15 |
| jokke_ | that was quick | 14:15 |
| rosmaita | yeah, about the dynamic plugin imports, maybe we can grab dhellmann at the PTG and get some ideas | 14:16 |
| jokke_ | that could be good option | 14:16 |
| jokke_ | I'd be also more than happy to get a bug up and separate those definitions from opts.py so that there is dedicated file for the stuff possible 3rd party plugin deployer needs to touch having none of the glance internals coupled in it | 14:17 |
| *** e0ne has quit IRC | 14:17 | |
| jokke_ | if you are ok with that approach/it makes you sleeping better? | 14:17 |
| rosmaita | that's exactly what i'd like to shoot for | 14:18 |
| jokke_ | rosmaita: ok, let me look into that and get it done | 14:19 |
| jokke_ | _should_ be super trivial refactoring | 14:19 |
| rosmaita | before you do that, do you have an opinion on https://review.openstack.org/#/c/455837/11 ? | 14:19 |
| bhagyashris | rosmaita, jokke_, i got idea after reading the comment https://review.openstack.org/#/c/527635/12/glance/opts.py@33 so i will try experimenting on my best | 14:22 |
| bhagyashris | rosmaita, jokke_, yeah waiting for the opinions on https://review.openstack.org/#/c/455837/11 :) | 14:23 |
| rosmaita | bhagyashris if i am reading abhishek's comments correctly, we have broken postgresql support (not in your patch, in the other one)? | 14:25 |
| bhagyashris | rosmaita, yeah | 14:26 |
| bhagyashris | rosmaita, that is also one important point that my patch https://review.openstack.org/#/c/455837/11 is not breaking | 14:27 |
| jokke_ | rosmaita: I haven't revied the implementation yet ... will look through it today | 14:31 |
| *** r-daneel has joined #openstack-glance | 14:31 | |
| rosmaita | bhagyashris i'd like you to update the release note | 14:33 |
| rosmaita | since your feature isn't necessarily tied to rolling upgrades, you can remove "rolling" | 14:33 |
| rosmaita | jokke_ my question about https://review.openstack.org/#/c/455837/11 is whether the deviation from the spec-lite is ok | 14:35 |
| rosmaita | (i think it is) | 14:35 |
| bhagyashris | rosmaita, ok i will update soon and thank you for review :) | 14:36 |
| rosmaita | bhagyashris ty | 14:36 |
| jokke_ | rosmaita: bug 1747926 I targeted it to RC1 | 14:36 |
| openstack | bug 1747926 in Glance "Import plugins requires deployer to poke Glance code with internal details " [Critical,New] https://launchpad.net/bugs/1747926 - Assigned to Erno Kuvaja (jokke) | 14:36 |
| jokke_ | I'd assume having patch ready today | 14:36 |
| rosmaita | ok | 14:36 |
| jokke_ | feel free to triage :D | 14:36 |
| rosmaita | we release RC-1 tomorrow, is that right? | 14:37 |
| jokke_ | I think it should be tomorrow, yes | 14:37 |
| *** shilpa_ has joined #openstack-glance | 14:37 | |
| jokke_ | and that will be the point of cutting the branch so anything we can get in before means we don't need to backport | 14:37 |
| rosmaita | ok, let's be optimistic about targeting that bug for rc-1 | 14:37 |
| rosmaita | bhagyashris add your idea to https://bugs.launchpad.net/glance/+bug/1747926 if you have some time before erno starts working on it | 14:39 |
| openstack | Launchpad bug 1747926 in Glance "Import plugins requires deployer to poke Glance code with internal details " [Critical,Triaged] - Assigned to Erno Kuvaja (jokke) | 14:39 |
| *** alexchadin has quit IRC | 14:44 | |
| bhagyashris | rosmaita: sure, Just one question as the review comments or we can say it as a concern on patch https://review.openstack.org/#/c/455837/11 so is it bug that glance will not support postgresql | 14:46 |
| *** shilpa_ has quit IRC | 14:46 | |
| rosmaita | bhagyashris your patch does not care what the db engine is, am i right about that? | 14:47 |
| rosmaita | you just need to know what scripts exist and what the alembic_version is in the db | 14:48 |
| bhagyashris | rosmaita: yes | 14:48 |
| rosmaita | ok, i will put a comment clarifying that the comments about postgres don't apply to your patch (even though they are important) | 14:49 |
| bhagyashris | rosmaita: just for curiosity i mean is it bug? yeah definitely my patch is not the cause of this | 14:49 |
| bhagyashris | rosmaita: ok thank you :) | 14:49 |
| rosmaita | bhagyashris oh yes, definitely a bug, we will discuss at meeting tomorrow | 14:50 |
| aniketh | jokke_, rosmaita : Really sorry to ping you once in a while with similar doubts. Like, I'm confused right now. Let me break down things myself and see. | 14:50 |
| rosmaita | aniketh don't worry about pinging me, that's fine, just be patient as it's release candidate time and i may not get back to you right away | 14:51 |
| bhagyashris | rosmaita: ok sure will discussed tomorrow's meeting | 14:51 |
| aniketh | So, basically we use Devstack for functional testing alone right? For developing and changing the code - we have the direct clone of the Glance from the Git mirror right? | 14:51 |
| aniketh | rosmaita: Thanks! Yes, I will be :) | 14:52 |
| openstackgerrit | Bhagyashri Shewale proposed openstack/glance master: Implementation of db check command https://review.openstack.org/455837 | 14:53 |
| rosmaita | aniketh you can change the code in your devstack installation to see what happens "live" (after you restart whatever service you have modified) ... i think jokke_ 's point is that occasionally you may want to throw out a devstack and start over, so you may want to have your local code repository outside of your devstack machine | 14:55 |
| rosmaita | but everyone has different development workflows, you just need to experiment and see what works best for you | 14:55 |
| bhagyashris | jokke_, rosmaita : just push patch https://review.openstack.org/455837 after fixing review comments. thank you for review :) | 14:55 |
| *** e0ne has joined #openstack-glance | 14:56 | |
| aniketh | rosmaita: Yes, Even I'm a bit comfortable more on having codebase in my Host system rather than working on the VM where the VM tends to be sloppy at times. Could you help me with building Glance locally? | 14:58 |
| openstackgerrit | Erno Kuvaja proposed openstack/glance master: Decouple Image Import Plugin Opts https://review.openstack.org/541740 | 15:02 |
| jokke_ | rosmaita: bhagyashris ^^ | 15:02 |
| rosmaita | jokke_ that was quick | 15:03 |
| jokke_ | aniketh: literally my local where I initially do my changes is `git clone git://github.com/openstack/glance` and then the needed configs by your own preference | 15:04 |
| jokke_ | I have it clone under ~/src/openstack.org/ and I do have git config at the root there having my openstack details so when ever I do clone repo I haven't touched before I don't need to setup everything again to get `git review` working | 15:05 |
| jokke_ | rosmaita: I told you I had pretty good idea what it should need to decouple :P | 15:06 |
| *** krypto has quit IRC | 15:07 | |
| jokke_ | rosmaita: and it's even documented :P | 15:07 |
| *** krypto has joined #openstack-glance | 15:07 | |
| *** Supun has quit IRC | 15:14 | |
| aniketh | jokke_: Alright! So, did that - all those you told. Now what is the change to be made in the config files and stuff? | 15:14 |
| aniketh | jokke_: Also, Should I install Tox and it's dependencies? | 15:15 |
| rosmaita | jokke_ got a grammar nit on your patch | 15:16 |
| rosmaita | aniketh yes, you'll need tox so that you can run the tests, build the docs, etc | 15:16 |
| aniketh | rosmaita: Alright! | 15:17 |
| aniketh | rosmaita, jokke_: Just putting out, the basic thing what I understood, about the whole Glance/ OpenStack as such - We are making dev changes in the local repo, we do code tests there with tox. I'm guessing that there is no provision for functional testing from the local clone itself. So, for that, we are using Devstack. We cherrypick the commit from Gerrit and test it out within the devstack. Please do | 15:20 |
| aniketh | correct me if I'm wrong somewhere or I'm missing out on certain things? | 15:20 |
| rosmaita | aniketh actually glance does have functional tests independent of devstack | 15:21 |
| aniketh | Also, we are following python3 right? | 15:22 |
| rosmaita | we do both py2 and py3 | 15:22 |
| aniketh | rosmaita: Oh, how? I would like to follow that. Or maybe try to find which is better for me. | 15:22 |
| aniketh | rosmaita: Alright! Encountered an error while running tox command on my local repo - just to build the docs and check things out. | 15:23 |
| aniketh | https://www.irccloud.com/pastebin/7NlhnLxB/ | 15:23 |
| rosmaita | aniketh devstack has 2 primary uses: it gives you a contained openstack installation to experiment with, and it is also used for testing in the CI gate | 15:24 |
| *** AlexeyAbashkin has quit IRC | 15:24 | |
| *** AlexeyAbashkin has joined #openstack-glance | 15:24 | |
| rosmaita | aniketh what kind of system are you running that command on ? | 15:25 |
| *** jdillaman has quit IRC | 15:26 | |
| aniketh | rosmaita: Kubuntu 17.10 | 15:30 |
| aniketh | >> aniketh devstack has 2 primary uses: it gives you a contained openstack installation to experiment with, and it is also used for testing in the CI gate | 15:31 |
| aniketh | rosmaita Alright, So how we are using it and how it helps us? | 15:31 |
| openstackgerrit | Erno Kuvaja proposed openstack/glance master: Decouple Image Import Plugin Opts https://review.openstack.org/541740 | 15:32 |
| rosmaita | aniketh you are using it mostly as an experimental environment | 15:32 |
| jokke_ | rosmaita: fixed those and the misleading part referring to gonfiggen | 15:32 |
| aniketh | rosmaita: Oh okay! | 15:33 |
| jokke_ | aniketh: 1) you can use devstack to just learn using Glance & OpenStack in general 2) you can use it to experiment your changes, reproducing bugs and ensuring your changes behave as you want them to before submitting them 3) you can use it to verify the code you're reviewing that it actually behaves as it claims/intended | 15:35 |
| *** jdillaman has joined #openstack-glance | 15:36 | |
| jokke_ | aniketh: it's just lightweight environment to tinker around that is easy to throw away and restart from clean when needed | 15:36 |
| rosmaita | aniketh looks like you need to install python3-dev (I guess kubuntu uses same package names as ubuntu?) | 15:37 |
| jokke_ | aniketh: https://docs.openstack.org/infra/manual/developers.html should help you get started with the actual repos, gerrit etc. | 15:37 |
| jokke_ | aniketh: then I just moved from glance/.git/config the [user] section to ~src/openstack.org/.gitconfig file so I don't need to run those git config commands every time (I don't use the --global) flag | 15:39 |
| aniketh | rosmaita: Yes - It is just another Ubuntu distro :) | 15:41 |
| aniketh | rosmaita: It is already installed | 15:42 |
| aniketh | The version my system is using is 2.7 | 15:42 |
| aniketh | I might to use the alternative | 15:43 |
| rosmaita | i'm not sure why tox can't find it | 15:43 |
| aniketh | I believe it is basically the system python version conflict - I will google that up. It is not a problem :D | 15:44 |
| aniketh | jokke_: So, this is the reference{https://docs.openstack.org/infra/manual/developers.html } to be used to build a project in OpenStack locally right? | 15:45 |
| *** idlemind has quit IRC | 15:46 | |
| *** idlemind has joined #openstack-glance | 15:47 | |
| jokke_ | aniketh: no, that's just fairly simple guide how to setup your dev environment with all the needed accounts etc, to make your first commit | 15:56 |
| aniketh | jokke_: Okay! That is almost done. Is there something specific to Glance? | 15:57 |
| jokke_ | https://docs.openstack.org/glance/pike/contributor/index.html | 15:58 |
| jokke_ | if you find any inconsistencies there, please let us know | 16:00 |
| aniketh | jokke_: Yea! Sure | 16:03 |
| *** pcaruana has quit IRC | 16:03 | |
| *** Supun has joined #openstack-glance | 16:09 | |
| *** mosulica has quit IRC | 16:09 | |
| *** tshefi_ has quit IRC | 16:10 | |
| *** krypto has quit IRC | 16:12 | |
| jokke_ | rosmaita: so what do we _need_ to get in before tomorrow? | 16:13 |
| rosmaita | just looking at the buglist to make sure i'm not missing what kairat mentioned | 16:14 |
| rosmaita | pretty sure it was the ovf thing | 16:14 |
| rosmaita | so we need to get your 2.6 patch in | 16:15 |
| rosmaita | which i am about to approve | 16:15 |
| aniketh | jokke_: I will go through the links given and google some stuff up and see what I can come up with. If I am stuck - I will ping you by then | 16:16 |
| rosmaita | jokke_ this just got filed | 16:16 |
| rosmaita | https://bugs.launchpad.net/glance/+bug/1747869 | 16:17 |
| openstack | Launchpad bug 1747869 in Glance "./stack.sh doesn't work with postgres" [Critical,Triaged] - Assigned to Brian Rosmaita (brian-rosmaita) | 16:17 |
| rosmaita | i am going to put up a quick fix patch in a few minutes | 16:17 |
| rosmaita | we can talk in glance meeting tomorrow about longer term fix | 16:17 |
| kukacz | rosmaita: Hi, we had a chat about changing glance backend last week. I'm in the phase now attempting to add a new location url (the new backend = swift) to an existing image. failing here though - Invalid location error | 16:18 |
| kukacz | I've enabled debug and can see this error in glance-api.log: " Store glance_store._drivers.swift.store.SingleTen | 16:19 |
| kukacz | antStore doesn't support updating dynamic storage capabilities. Please overwrite 'update_capabilities'" | 16:19 |
| kukacz | I wonder if these features (location-add, location-update) are actually implemented in the swift store? | 16:20 |
| rosmaita | they should be | 16:20 |
| rosmaita | you said you are using ocata? | 16:20 |
| kukacz | yes | 16:20 |
| kukacz | the current store is "file" and I'm trying to add "swift+config" (with v3 auth) as an additional url record to existing image | 16:21 |
| kukacz | eg. grepping "update_capabilities" in my python glance_store path finds no occurences in the _drivers directory | 16:22 |
| kukacz | neither I could find it on github master branch though, so maybe my thinking in this way is wrong | 16:23 |
| rosmaita | kukacz it's defined in glance_store/capabilities.py and is inherited by the Store class in driver.py, and then the drivers in _drivers inherit from that Store class | 16:26 |
| kukacz | yes, I expected I should find then the "update_capabilities" implementation somewhere in the _drivers/swift/store.py which I did not | 16:29 |
| rosmaita | i dont' know if any of the drivers support that, actually | 16:32 |
| rosmaita | what that's for is if you want to change your swift from read-write to read-only, something like that | 16:32 |
| rosmaita | how are you adding the location? | 16:33 |
| kukacz | ok, maybe I don't need that. it's just an error in the log when trying to add the url as additional location | 16:33 |
| kukacz | using curl patch call | 16:33 |
| kukacz | can paste an example | 16:34 |
| rosmaita | that would be helpful | 16:34 |
| kukacz | https://gist.github.com/kukacz/3506af03d46c40e7e21dd722ad7ee451 | 16:35 |
| kukacz | also here are some lines from the log: https://gist.github.com/kukacz/87f6d0b6b18b3685bc13dd5810eff7f0 | 16:41 |
| rosmaita | ty | 16:41 |
| *** gyee has joined #openstack-glance | 16:45 | |
| *** Supun has quit IRC | 16:45 | |
| rosmaita | kukacz here's your problem: http://git.openstack.org/cgit/openstack/glance/tree/glance/common/store_utils.py?h=stable/ocata#n124 | 16:46 |
| *** markvoelker_ has joined #openstack-glance | 16:47 | |
| rosmaita | it's a security risk to allow users to set a swift+config:// location | 16:47 |
| *** itlinux has joined #openstack-glance | 16:49 | |
| rosmaita | and i don't think there are any policy checks around there to see if you're an admin or anything like that | 16:49 |
| *** e0ne has quit IRC | 16:49 | |
| *** markvoelker has quit IRC | 16:49 | |
| kukacz | hmm, ok. I've justed used this form (prefix) because I've seen it created so by glance itself - I've created one image using the "new" config to see how the location records look like | 16:49 |
| rosmaita | yes, and i think that is probably the correct thing to do | 16:50 |
| rosmaita | i'm not sure how to address this. can you come to the glance meeting tomorrow at 1400 UTC? | 16:51 |
| rosmaita | or if that's a bad time i can bring it up and see if anyone has any ideas | 16:52 |
| kukacz | the time is ok, I will be happy to join. just would be grateful for your guidance | 16:53 |
| rosmaita | i have an idea, but am hoping someone else has a better one. my idea is that since you're adding the locations on an isolated glance that only an admin can use, you can hack the code to remove swift+config from the restricted schemas at http://git.openstack.org/cgit/openstack/glance/tree/glance/common/store_utils.py?h=stable/ocata#n31 | 16:55 |
| rosmaita | but only on that restricted node | 16:56 |
| jokke_ | ouch, that does sound hacky ... likely also the only option to actually do that | 16:56 |
| rosmaita | make sure you read through that bug so you understand the implications | 16:56 |
| rosmaita | jokke_ yeah, was hoping you might have a better idea | 16:57 |
| openstackgerrit | Matthew Treinish proposed openstack/glance master: Specify a python executable for prlimits when using uwsgi https://review.openstack.org/531498 | 16:57 |
| kukacz | rosmaita: sounds great, love these hacks :-) | 16:57 |
| rosmaita | kukacz just be careful! | 16:58 |
| jokke_ | rosmaita: kukacz: one can always also circumvent all the fences and poke the db directly :P | 16:58 |
| kukacz | I will. perhaps this is still better than doing it by database record update | 16:58 |
| kukacz | :-) | 16:58 |
| kukacz | I'll collect votes which one is less "hacky" | 16:59 |
| jokke_ | kukacz: tbh, as long as you understand what you are doing, patching that one node to go around artificial restriction is lots safer | 17:00 |
| kukacz | jokke_: it's also my preferrence, to rely on the validations you great guys have already put into the code | 17:02 |
| kukacz | I'll try this method | 17:03 |
| rosmaita | looks like maybe we overdid the validations :P | 17:03 |
| *** pbourke has quit IRC | 17:04 | |
| *** pbourke has joined #openstack-glance | 17:05 | |
| jokke_ | rosmaita: that's always the double edged sword with security. There is always legit usecases for something you want to prevent to plug a hole | 17:06 |
| rosmaita | jokke_ about https://bugs.launchpad.net/glance/+bug/1747869 ... short term fix is to put a 'pass' right at the beginning of https://github.com/openstack/glance/blob/master/glance/cmd/manage.py#L146-L156 (though i don;'t know if that would pass pep8, but you know what i mean) | 17:08 |
| openstack | Launchpad bug 1747869 in Glance "./stack.sh doesn't work with postgres" [Critical,Triaged] - Assigned to Brian Rosmaita (brian-rosmaita) | 17:08 |
| rosmaita | or is that too stupid? | 17:08 |
| jokke_ | oh my | 17:09 |
| rosmaita | i think we need to get a patch up right away | 17:09 |
| jokke_ | so we just did glance manage to always use EMC and we still have that condition existing :o | 17:09 |
| jokke_ | now if we drop that check we will push EMC to the DBs that are actually not supported | 17:10 |
| *** harlowja has joined #openstack-glance | 17:11 | |
| rosmaita | yeah | 17:11 |
| rosmaita | i think its not a problem for the offline migrations | 17:11 |
| rosmaita | but i am not 100% sure | 17:12 |
| jokke_ | so I think either we admit that devstack won't work with postgre or ... | 17:12 |
| rosmaita | exactly | 17:13 |
| jokke_ | we rever the patch that caused it and back to the table to figure out how we do this the right way | 17:13 |
| *** udesale has quit IRC | 17:13 | |
| *** markvoelker_ has quit IRC | 17:14 | |
| jokke_ | so clearly there was valid reason why it had not been made such as default | 17:14 |
| *** markvoelker has joined #openstack-glance | 17:14 | |
| jokke_ | clearly we do not test on postgre so I'd be actually fine either way (hurst to say) | 17:14 |
| jokke_ | but I really do not like just bypassing validation and hoping the world does not explode bocause of it | 17:15 |
| rosmaita | agree | 17:15 |
| rosmaita | ok, i will put up a revert patch, and we'll figure out what to do for RC-2 | 17:16 |
| jokke_ | so why fixing postgre on devstack is such high importance when we do not test it (as in it's fundamentally broken) | 17:17 |
| jokke_ | do we officially state that we still support postgre? | 17:17 |
| *** harlowja has quit IRC | 17:17 | |
| rosmaita | looking to see | 17:18 |
| rosmaita | (asking in the tc channel) | 17:21 |
| kukacz | rosmaita, jokke_: the "hack" worked! I got 2 locations at my image now | 17:21 |
| kukacz | one thing i still don't understand: the "direct_url" parameter now points to the "file" backend which I'm about to remove. do I need to take care of updating this parameter too or is it just dynamically constructed? | 17:25 |
| *** AlexeyAbashkin has quit IRC | 17:26 | |
| jokke_ | I can't remember out of my head how that gets done | 17:26 |
| *** tesseract has quit IRC | 17:27 | |
| jokke_ | rosmaita: I do not understand wtf is going on here http://logs.openstack.org/40/541740/2/check/build-openstack-sphinx-docs/ec83d77/job-output.txt.gz#_2018-02-07_15_36_29_343487 as the sphinx run is the only one failing | 17:28 |
| rosmaita | kukacz i believe the direct_url is generated | 17:28 |
| rosmaita | it's the one indicated by the location strategy | 17:29 |
| jokke_ | rosmaita: and that import works just fine | 17:30 |
| rosmaita | is it glance.async.flows.plugins ? | 17:31 |
| rosmaita | http://git.openstack.org/cgit/openstack/glance/tree/glance/async/flows/plugins | 17:31 |
| openstackgerrit | Erno Kuvaja proposed openstack/glance master: Decouple Image Import Plugin Opts https://review.openstack.org/541740 | 17:32 |
| *** dtantsur is now known as dtantsur|afk | 17:33 | |
| jokke_ | it was | 17:35 |
| rosmaita | ok | 17:35 |
| rosmaita | based on dicsussion in tc channel, i think the thing to do is revert the change and reconsider how to handle this | 17:35 |
| rosmaita | i don't think it's going to be a clean revert | 17:38 |
| rosmaita | maybe we just keep it in rc-1 and acknowledge the bug | 17:38 |
| jokke_ | ouch, there was something merged that was based on it? | 17:39 |
| rosmaita | yeah, i think the system(exit) thing that got merged this morning | 17:40 |
| jokke_ | I'd say lets acknowledge it, keep it in rc1 and buy few more days to figure out why that's the case | 17:40 |
| jokke_ | oh yeah it was direct patch against that so they would need both to go | 17:41 |
| rosmaita | ok, will go with the keep-and-acknowledge for now | 17:41 |
| jokke_ | ++ | 17:41 |
| jokke_ | the docs build worked at least locally once I fixed that path | 17:42 |
| rosmaita | cool | 17:42 |
| jokke_ | so unless you hate the change that should be good to go :P | 17:42 |
| jokke_ | then hopefully abhishek is ok with it at the morn | 17:43 |
| jokke_ | rosmaita: another thing I'm bit worried about is that the requirements change does not merge | 17:45 |
| rosmaita | which requirements change? | 17:46 |
| jokke_ | 537045 | 17:46 |
| rosmaita | that's weird, so tempest-full passed on check but failed on gate | 17:47 |
| jokke_ | and that was second time in row | 17:49 |
| rosmaita | timeout during a volume test | 17:50 |
| rosmaita | i gave it another recheck, let's see what happens | 17:52 |
| jokke_ | would be good to have them up to date for branching | 17:54 |
| *** Supun has joined #openstack-glance | 17:59 | |
| *** cfriesen has left #openstack-glance | 18:02 | |
| openstackgerrit | Erno Kuvaja proposed openstack/glance master: Add validation to check if E-M-C is already in sync https://review.openstack.org/540736 | 18:04 |
| *** e0ne has joined #openstack-glance | 18:04 | |
| jokke_ | just rebase ^^ | 18:07 |
| *** harlowja has joined #openstack-glance | 18:11 | |
| *** Supun has quit IRC | 18:11 | |
| *** harlowja has quit IRC | 18:18 | |
| *** r-daneel_ has joined #openstack-glance | 18:28 | |
| *** Supun has joined #openstack-glance | 18:28 | |
| *** r-daneel has quit IRC | 18:29 | |
| *** r-daneel_ is now known as r-daneel | 18:29 | |
| openstackgerrit | Merged openstack/glance master: Enable Image Import per default and make current https://review.openstack.org/532502 | 18:36 |
| *** Supun has quit IRC | 18:38 | |
| *** Supun has joined #openstack-glance | 18:38 | |
| *** Supun has quit IRC | 18:57 | |
| *** e0ne has quit IRC | 18:59 | |
| *** Supun has joined #openstack-glance | 19:00 | |
| *** Supun has quit IRC | 19:09 | |
| *** e0ne has joined #openstack-glance | 19:10 | |
| *** r-daneel_ has joined #openstack-glance | 19:13 | |
| *** r-daneel has quit IRC | 19:14 | |
| *** r-daneel_ is now known as r-daneel | 19:14 | |
| *** Supun has joined #openstack-glance | 19:28 | |
| *** aniketh has quit IRC | 19:32 | |
| *** markvoelker has quit IRC | 19:33 | |
| *** markvoelker_ has joined #openstack-glance | 19:33 | |
| *** AlexeyAbashkin has joined #openstack-glance | 19:37 | |
| *** pcaruana has joined #openstack-glance | 19:52 | |
| *** Supun has quit IRC | 20:04 | |
| *** pcaruana has quit IRC | 20:07 | |
| *** AlexeyAbashkin has quit IRC | 20:33 | |
| *** mvk_ has joined #openstack-glance | 20:34 | |
| *** mvk has quit IRC | 20:36 | |
| openstackgerrit | Merged openstack/glance master: Updated from global requirements https://review.openstack.org/537045 | 20:57 |
| openstackgerrit | Brian Rosmaita proposed openstack/glance master: Regenerate sample configuration files https://review.openstack.org/541897 | 21:02 |
| openstackgerrit | Matthew Treinish proposed openstack/glance master: Specify a python executable for prlimits when using uwsgi https://review.openstack.org/531498 | 21:10 |
| *** e0ne has quit IRC | 21:23 | |
| *** e0ne has joined #openstack-glance | 21:27 | |
| *** threestrands has joined #openstack-glance | 21:39 | |
| *** Sukhdev has joined #openstack-glance | 21:59 | |
| *** pdr_ has quit IRC | 22:07 | |
| *** r-daneel has quit IRC | 22:08 | |
| openstackgerrit | Brian Rosmaita proposed openstack/glance master: Release note for Queens metadefs changes https://review.openstack.org/541923 | 22:10 |
| *** rcernin has joined #openstack-glance | 22:25 | |
| *** pdr_ has joined #openstack-glance | 22:35 | |
| *** e0ne has quit IRC | 22:47 | |
| *** itlinux has quit IRC | 22:50 | |
| *** Sukhdev has quit IRC | 23:23 | |
| *** masber has joined #openstack-glance | 23:27 | |
| *** Sukhdev has joined #openstack-glance | 23:36 | |
| *** edmondsw has quit IRC | 23:37 | |
| *** MattMan has quit IRC | 23:37 | |
| *** MattMan has joined #openstack-glance | 23:38 | |
| *** itlinux has joined #openstack-glance | 23:43 | |
| *** markvoelker_ has quit IRC | 23:45 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!