| *** btully has quit IRC | 00:21 | |
| *** masber has joined #openstack-glance | 00:21 | |
| *** AlexeyAbashkin has joined #openstack-glance | 00:24 | |
| *** AlexeyAbashkin has quit IRC | 00:28 | |
| *** itlinux has joined #openstack-glance | 00:29 | |
| *** openstackgerrit has joined #openstack-glance | 01:51 | |
| openstackgerrit | Merged openstack/glance master: Add validation to check if E-M-C is already in sync https://review.openstack.org/540736 | 01:51 |
|---|---|---|
| *** btully has joined #openstack-glance | 02:06 | |
| *** btully has quit IRC | 02:11 | |
| *** harlowja has quit IRC | 02:52 | |
| openstackgerrit | Brian Rosmaita proposed openstack/glance master: Revise help text for uri filtering options https://review.openstack.org/546020 | 03:00 |
| *** bkopilov has quit IRC | 03:08 | |
| *** ykarel|away has joined #openstack-glance | 03:10 | |
| openstackgerrit | Brian Rosmaita proposed openstack/glance master: Correct length limit for custom property value https://review.openstack.org/546021 | 03:13 |
| *** AlexeyAbashkin has joined #openstack-glance | 03:24 | |
| *** AlexeyAbashkin has quit IRC | 03:29 | |
| openstackgerrit | Brian Rosmaita proposed openstack/glance master: Revise help text for uri filtering options https://review.openstack.org/546020 | 03:29 |
| *** konan has joined #openstack-glance | 03:45 | |
| *** konan is now known as abhishekk | 03:45 | |
| *** rosmaita has quit IRC | 03:49 | |
| *** btully has joined #openstack-glance | 03:55 | |
| *** btully has quit IRC | 03:59 | |
| *** ykarel|away has quit IRC | 03:59 | |
| *** udesale has joined #openstack-glance | 04:13 | |
| *** links has joined #openstack-glance | 04:21 | |
| *** ykarel has joined #openstack-glance | 04:21 | |
| *** pdeore has joined #openstack-glance | 04:21 | |
| *** Shilpa has joined #openstack-glance | 04:27 | |
| Shilpa | abhishekk, hi | 04:28 |
| abhishekk | Shilpa: hi | 04:28 |
| Shilpa | my patch related to validation got merged yesterday, so need to check any Test cases are failing on your patch of postgres | 04:30 |
| Shilpa | 540736 this got merged | 04:30 |
| abhishekk | Which my patch you are talking about? | 04:30 |
| Shilpa | 545211 | 04:31 |
| abhishekk | You can check that in your local environment | 04:32 |
| Shilpa | yes doing that | 04:32 |
| Shilpa | just pinged you to informed you the same | 04:32 |
| abhishekk | Ok, thank you | 04:33 |
| *** abhishekk is now known as Konan | 04:38 | |
| *** ying_zuo has left #openstack-glance | 04:39 | |
| *** bkopilov has joined #openstack-glance | 04:40 | |
| Shilpa | Konan, hi, yes unit test cases are failing | 04:41 |
| *** abhishekk has joined #openstack-glance | 04:41 | |
| *** dalgaaf has quit IRC | 04:47 | |
| *** dalgaaf has joined #openstack-glance | 04:49 | |
| *** harlowja has joined #openstack-glance | 04:54 | |
| *** ratailor has joined #openstack-glance | 05:06 | |
| *** pdeore_ has joined #openstack-glance | 05:45 | |
| *** pdeore has quit IRC | 05:47 | |
| abhishekk | Shilpa, hi | 05:55 |
| abhishekk | around 16 tests are failing, I am fixing the same | 05:55 |
| abhishekk | jokke_, you around? | 06:05 |
| Shilpa | abhishekk, hi | 06:12 |
| abhishekk | hi | 06:13 |
| Shilpa | thanks for this | 06:14 |
| Shilpa | one thing that check is also used during offline | 06:14 |
| Shilpa | so for check also we should not have check of validate engine | 06:14 |
| abhishekk | check should be use only for online | 06:16 |
| Shilpa | please explain why we can not have for offline? | 06:18 |
| abhishekk | because in case of offline could you run E-M-C separately? no | 06:20 |
| abhishekk | it will run as a part of db_sync only | 06:20 |
| abhishekk | offline you should run db_sync only and not expand, migrate, contract explicitly | 06:21 |
| openstackgerrit | Abhishek Kekane proposed openstack/glance master: Migration support for postgresql https://review.openstack.org/545211 | 06:21 |
| abhishekk | that is why we have decided to have it as a check command and not an option as --check to glance-manage db_sync | 06:24 |
| openstackgerrit | Abhishek Kekane proposed openstack/glance master: Triggers shouldn't be execute in offline migration https://review.openstack.org/544792 | 06:26 |
| abhishekk | I hope you got my point | 06:26 |
| *** pdeore_ has quit IRC | 06:42 | |
| *** Konan has quit IRC | 06:44 | |
| *** pdeore has joined #openstack-glance | 06:50 | |
| *** cschwede has joined #openstack-glance | 06:51 | |
| Shilpa | for offline, one can run EMC to db sync and it will work | 06:52 |
| abhishekk | what is the use of runnign EMC offline? | 06:56 |
| abhishekk | means what is the use of running 3 different commands offline where you can do it by only one | 06:57 |
| abhishekk | it does not make any sense | 06:57 |
| abhishekk | and if any one wants to use E-M-C offline then yes it will work but what is the use of having check command offline as the whole purpose of having check command is to help operator to understand the current version during rolling upgrades | 06:59 |
| *** tshefi has joined #openstack-glance | 06:59 | |
| *** threestrands has quit IRC | 07:10 | |
| *** harlowja has quit IRC | 07:10 | |
| *** rcernin has quit IRC | 07:12 | |
| openstackgerrit | Abhishek Kekane proposed openstack/glance master: Migration support for postgresql https://review.openstack.org/545211 | 07:15 |
| openstackgerrit | Abhishek Kekane proposed openstack/glance master: Triggers shouldn't be execute in offline migration https://review.openstack.org/544792 | 07:16 |
| *** btully has joined #openstack-glance | 07:32 | |
| abhishekk | jokke_, you around? | 07:35 |
| *** pcaruana has joined #openstack-glance | 07:35 | |
| *** btully has quit IRC | 07:37 | |
| *** alexchadin has joined #openstack-glance | 07:42 | |
| *** AlexeyAbashkin has joined #openstack-glance | 07:53 | |
| *** btully has joined #openstack-glance | 07:55 | |
| *** mosulica has joined #openstack-glance | 07:56 | |
| *** MattMan has quit IRC | 08:03 | |
| *** MattMan has joined #openstack-glance | 08:03 | |
| *** pdeore has quit IRC | 08:12 | |
| *** pdeore has joined #openstack-glance | 08:13 | |
| *** tesseract has joined #openstack-glance | 08:20 | |
| *** hoonetorg has quit IRC | 08:24 | |
| *** alexchadin has quit IRC | 08:24 | |
| *** alexchadin has joined #openstack-glance | 08:24 | |
| *** btully has quit IRC | 08:37 | |
| *** hoonetorg has joined #openstack-glance | 08:42 | |
| *** dtantsur|afk is now known as dtantsur | 08:47 | |
| *** e0ne has joined #openstack-glance | 09:18 | |
| openstackgerrit | Bhagyashri Shewale proposed openstack/glance-specs master: Introduce db sync check feature https://review.openstack.org/546096 | 09:57 |
| *** alexchadin has quit IRC | 10:27 | |
| *** d0ugal_ has joined #openstack-glance | 10:31 | |
| abhishekk | jokke_, are you around? | 10:32 |
| *** d0ugal has quit IRC | 10:34 | |
| *** abhishekk has quit IRC | 10:35 | |
| *** alexchadin has joined #openstack-glance | 10:49 | |
| *** d0ugal_ has quit IRC | 10:50 | |
| *** d0ugal has joined #openstack-glance | 10:51 | |
| *** d0ugal has quit IRC | 10:51 | |
| *** d0ugal has joined #openstack-glance | 10:51 | |
| *** mvk_ has quit IRC | 10:53 | |
| *** dtantsur is now known as dtantsur|brb | 10:54 | |
| *** alexchadin has quit IRC | 10:59 | |
| *** bkopilov has quit IRC | 11:18 | |
| *** pdeore has quit IRC | 11:20 | |
| *** ratailor has quit IRC | 11:29 | |
| *** pbourke has quit IRC | 11:39 | |
| *** vladiskuz has quit IRC | 11:40 | |
| *** pbourke has joined #openstack-glance | 11:41 | |
| *** vladiskuz has joined #openstack-glance | 11:41 | |
| *** udesale has quit IRC | 11:43 | |
| *** amoralej has joined #openstack-glance | 11:45 | |
| amoralej | could some core take a look into https://review.openstack.org/#/c/544792/ ? it's blocking promotions in RDO | 11:46 |
| *** Konan has joined #openstack-glance | 11:53 | |
| Konan | jokke_, hi you around? | 11:53 |
| jokke_ | yeah | 11:54 |
| jokke_ | hey | 11:54 |
| Konan | So i just need to remove that doc change? | 11:54 |
| jokke_ | sorry, has not ben my intetion to ignore you ... just not been good morning | 11:55 |
| jokke_ | so I think we should remove that doc change there, unless you can point out to me why it's related | 11:55 |
| jokke_ | also lets file a bug that we need to revisit the documentation of db_manage and specially sync based on these recent changes and target it to queens | 11:56 |
| Konan | actually if someone wants to use EMC then the need to set super privileges to glance user or set that parameter | 11:56 |
| Konan | Sure | 11:56 |
| Konan | I will remove that from this patch and file a new bug | 11:57 |
| Konan | That's why I have added a topic in PTG etherpad as well | 11:57 |
| jokke_ | Konan: so That is very true, I do not see the link between that and why db_sync should not run the triggers 'though ;) | 11:57 |
| jokke_ | gr8 thnx | 11:57 |
| jokke_ | I will consider Brian's +2 and just ninja it in as soon as zuul agrees | 11:58 |
| Konan | jokke_: give me 5 minutes I will push a patch | 11:58 |
| jokke_ | np | 11:58 |
| Konan | B4 that you need to approve it's parent patch as well | 11:58 |
| jokke_ | ohh my | 11:59 |
| jokke_ | oh, so that doc change should have been actually part of the parent already :) | 12:00 |
| jokke_ | now it makes so much more sense | 12:00 |
| openstackgerrit | Abhishek Kekane proposed openstack/glance master: Triggers shouldn't be execute in offline migration https://review.openstack.org/544792 | 12:00 |
| Konan | jokke_: IMO we need to revisit the entire document for rolling upgrades | 12:01 |
| jokke_ | Konan: ++ | 12:02 |
| jokke_ | and the approach | 12:03 |
| Konan | Yes agree | 12:03 |
| jokke_ | Looking these changes, I'm not convinced at all that this is the right way to do it anymore | 12:03 |
| jokke_ | god damn the manage is starting to become a mess | 12:04 |
| *** Supun has joined #openstack-glance | 12:04 | |
| Konan | https://bugs.launchpad.net/glance/+bug/1750555 | 12:04 |
| openstack | Launchpad bug 1750555 in Glance "Revisit database rolling upgrade documentation" [Undecided,New] | 12:04 |
| Konan | jokke_: yes, we need to revisit it in Rocky | 12:05 |
| Konan | jokke_: need to go out for 2 hours, please ping me if something is urgent | 12:08 |
| *** alexchadin has joined #openstack-glance | 12:10 | |
| *** btully has joined #openstack-glance | 12:11 | |
| jokke_ | Konan: one question in the parent | 12:11 |
| jokke_ | ah, k .... have a good one! | 12:11 |
| Konan | jokke_: answered | 12:15 |
| *** alexchadin has quit IRC | 12:16 | |
| *** btully has quit IRC | 12:16 | |
| jokke_ | ty | 12:16 |
| *** alexchadin has joined #openstack-glance | 12:16 | |
| *** Supun has quit IRC | 12:38 | |
| *** bkopilov has joined #openstack-glance | 12:52 | |
| *** alexchadin has quit IRC | 12:53 | |
| *** rosmaita has joined #openstack-glance | 12:54 | |
| rosmaita | jokke_ ping | 13:00 |
| *** Supun has joined #openstack-glance | 13:00 | |
| jokke_ | rosmaita: pong | 13:07 |
| rosmaita | just checking our rc-3 status | 13:08 |
| *** Supun has quit IRC | 13:11 | |
| jokke_ | there's those couple of db changes from Abhishek pending | 13:11 |
| rosmaita | and the gate seems slow | 13:12 |
| jokke_ | surprised? :P | 13:12 |
| *** alexchadin has joined #openstack-glance | 13:15 | |
| rosmaita | jokke_ did you get a chance to look at https://review.openstack.org/#/c/545649/ yet? | 13:16 |
| *** tshefi has quit IRC | 13:17 | |
| rosmaita | jokke_ and if you are ok with that one, need you to look at the follow-up, https://review.openstack.org/#/c/546020/ | 13:20 |
| jokke_ | so any idea why we need to do that in the _internal_plugins but the plugins seems to be just fine? | 13:21 |
| jokke_ | I just really dislike the idea to decouple the config options from the plugin | 13:22 |
| rosmaita | well, the plugin itself calls some other stuff that was causing a circular import | 13:23 |
| jokke_ | I thought we fixed that already moving the filtering script elsewhere | 13:24 |
| rosmaita | that only fixed one problem | 13:24 |
| rosmaita | next problem was that the util file in glance/common could not find config options, so they needed to be imported | 13:25 |
| rosmaita | when i tried to import them out of the plugin, i got the circular import problem agian | 13:25 |
| rosmaita | so i moved them so the only thing being imported was the options | 13:25 |
| rosmaita | it may be an ordering problem | 13:26 |
| rosmaita | i would have sworn i saw the uri filtering working in devstack | 13:26 |
| rosmaita | but when i cranked up a new one to verify how it worked to do the docs, i got the 500 on import when using web-download | 13:27 |
| jokke_ | hmm-m wtf ... 'cause so did I | 13:27 |
| rosmaita | i think the plugin code itself is fine, because it never calls common/util -- that's done by the images controller | 13:28 |
| jokke_ | did we merge wrong PS of that change or something? | 13:28 |
| rosmaita | i don't think so | 13:28 |
| rosmaita | in any case, the current version plus these patches seems to work | 13:29 |
| rosmaita | we obviously need more functional tests around this using a live devstack i think | 13:29 |
| rosmaita | anyway, the plugin code on its own is fine, and the util code on its own is fine ... it's when you combine them that you get the circular import | 13:30 |
| rosmaita | that's why you split them | 13:30 |
| jokke_ | ffs | 13:30 |
| rosmaita | so this is just a continuation of that same split | 13:30 |
| rosmaita | allow the util to import the config options | 13:31 |
| rosmaita | but they can't live in the plugin | 13:31 |
| rosmaita | without major refactoring | 13:31 |
| rosmaita | but yeah, i tried not to move them, but no luck | 13:31 |
| *** Supun has joined #openstack-glance | 13:31 | |
| jokke_ | ok, would you feel horribly against put the config options to plugin_opts.py or something ... I just feel super bad for some reason cluttering the __init__.py with them ... and I really don't know why, just feels wrong ;) | 13:32 |
| rosmaita | in any case, there's not a lot in that __init__ file, so we can just declare a convention that the config opts go there | 13:32 |
| *** dtantsur|brb is now known as dtantsur | 13:32 | |
| rosmaita | i think this experience shows that maybe that's the place they should be! (in __init__) | 13:32 |
| jokke_ | LOL ... perhaps you are right | 13:33 |
| rosmaita | let's stick with them in __init__ for now, anyway | 13:33 |
| rosmaita | i toyed actually with having them in a separate web_download_opts.py | 13:34 |
| rosmaita | and actually had a version doing that, but the __init__ seemed better somehow | 13:34 |
| jokke_ | then I might offend some people, quite a lot, but I absolutely hated the follow up patch ... it makes that glance-image-import.conf.sample look more and more like any other glance config file ... 7 options and 1000 lines of cluttered jargon | 13:34 |
| rosmaita | actually, i think it's shorter than the original | 13:35 |
| rosmaita | :) | 13:35 |
| jokke_ | can we like condense those help texts to about 30% and utilize the documentation | 13:35 |
| jokke_ | it adds like 100lines | 13:36 |
| rosmaita | well, the problem is that the options are coupled but the help text needs to be readable in isolation | 13:36 |
| rosmaita | the only thing we could do is say ... "This is a whitelist. RTFM" | 13:37 |
| rosmaita | that woudl def be shorter! | 13:37 |
| jokke_ | Why? And I'm serious about that question, they are grouped for a reason, they mention related options for a reason | 13:37 |
| jokke_ | If all help texts needs to be individually readable, lets get rid of all that clutter then ;) | 13:38 |
| rosmaita | well, i did the rewrite, so of course i think it's brilliant | 13:38 |
| rosmaita | :) | 13:38 |
| rosmaita | but that was like 8 hours ago, so i will take a look now | 13:39 |
| jokke_ | Like I honestly do like the content, it just makes me vomit a bit every time I see it in the sample config file | 13:39 |
| jokke_ | I would really really like to dense that ... specially as they are now well explained in the docs | 13:39 |
| rosmaita | well i am definitely wrong about them being shorter | 13:42 |
| *** amoralej is now known as amoralej|lunch | 13:43 | |
| rosmaita | and believe it or not, i actually agree with you | 13:43 |
| *** Supun has quit IRC | 13:43 | |
| *** Supun has joined #openstack-glance | 13:43 | |
| jokke_ | :) | 13:44 |
| rosmaita | i think the key points are to mention (a) what the list filters, (b) a whitelist trumps a blacklist, and (c) see the docs | 13:44 |
| jokke_ | Me goes and draws cross on the wall | 13:44 |
| rosmaita | ok, will put up a new patch | 13:44 |
| rosmaita | i think the example just clutters things up becasue it's hard to explain | 13:44 |
| rosmaita | anyway, we can see what it looks like in a few minutes | 13:44 |
| jokke_ | yeah | 13:44 |
| jokke_ | and I think I just found a thing I probably should propose to oslo_config for Rocky | 13:45 |
| rosmaita | ok, give me a few min to get some coffee and then i'll get a patch up | 13:45 |
| jokke_ | sure, coffeinating is important | 13:45 |
| rosmaita | :) | 13:45 |
| jokke_ | I think what would be great is actually having help text for the opt group | 13:45 |
| rosmaita | \_/? | 13:46 |
| jokke_ | how much easier it would be to explain that whitelists does take preference over blacklist on group level rather than repeating it in every single option | 13:46 |
| ykarel | jokke_, +1 from zuul, can you look now: https://review.openstack.org/#/c/544792/ Triggers shouldn't be execute in offline migration | 13:48 |
| ykarel | Actually it's required in queens also | 13:48 |
| rosmaita | jokke_ that's a really good point, to have the option to have help text associated with the group itself | 13:49 |
| *** rmcall has joined #openstack-glance | 14:00 | |
| openstackgerrit | Brian Rosmaita proposed openstack/glance master: Revise help text for uri filtering options https://review.openstack.org/546020 | 14:01 |
| *** rmcall has quit IRC | 14:08 | |
| *** abhishekk has joined #openstack-glance | 14:09 | |
| abhishekk | rosmaita, jokke_ anything to do for me? | 14:10 |
| *** links has quit IRC | 14:15 | |
| *** Konan has quit IRC | 14:16 | |
| rosmaita | hi abhishekk , maybe see what you think of the revision to https://review.openstack.org/546020 | 14:16 |
| abhishekk | rosmaita, looking | 14:16 |
| rosmaita | i think you answered all of jokke_ 's questions on your db patches | 14:16 |
| rosmaita | did you put a patch up adding the text to the docs about the glance user having "super" privileges if you're going to do zero-dowtime db upgrade? | 14:17 |
| rosmaita | i saw you removed it from that one patch at jokke_ 's request | 14:17 |
| abhishekk | rosmaita, yes, I am preparing a new patch now | 14:17 |
| rosmaita | but you are right, we need to make sure that gets into the docs | 14:17 |
| rosmaita | cool! | 14:17 |
| abhishekk | will be up in 10-15 minutes | 14:17 |
| rosmaita | excellent | 14:18 |
| rosmaita | that's a doc patch, so i can ninja-approve | 14:18 |
| jokke_ | so the 545211 | 14:18 |
| abhishekk | sounds good | 14:18 |
| *** ykarel is now known as ykarel|away | 14:19 | |
| jokke_ | my concern is that as we remove those checks, we also might break someone silently | 14:19 |
| abhishekk | jokke_, are you talking about validate_engine check? | 14:20 |
| rosmaita | i don't think so, the point of the check was to make sure you only install triggers for mysql, and we have a different way to handle that now | 14:20 |
| abhishekk | rosmaita, correct | 14:21 |
| jokke_ | lets assume IBM still uses db2 like they at least used to. And those migrations have been working so far and life is good. Does our alembic provide them the same still or is there a risk now when we moved to use emc and removed all contraints that the db2 blows up half way of the migration and the ops are working overtime restoring from backups? | 14:21 |
| rosmaita | well, if they were able to upgrade ocata -> pike, they should be ok pike -> queens | 14:22 |
| rosmaita | that's where we started using alembic | 14:22 |
| abhishekk | jokke_, what just rosmaita said | 14:22 |
| rosmaita | we don't test db2, so hopefully if there was a problem, they would have noticed and told us | 14:23 |
| rosmaita | let me add this comment here, that i may have left on the wrong patch: | 14:23 |
| rosmaita | The E-M-C strategy is to split a database migration into three stages, whereas we previously had monolithic migrations. Splitting them this way allows the possibility of "live" DB upgrades (which we support only for MySQL because of the triggers). The "offline" migration uses the E-M-C strategy so that we don't have to maintain separate online vs. offline scripts. But you are right that we need more documentation around | 14:23 |
| rosmaita | this. | 14:23 |
| rosmaita | except for the scripts, we just use alembic | 14:24 |
| rosmaita | the EMC just splits the stages into separate scripts | 14:24 |
| rosmaita | for easier maintenance | 14:24 |
| rosmaita | but we def need better docs, there's a lot of confusion about rolling upgrades (which is entire glance) and zero-downtime db upgrades (which we only support for mysql) and EMC strategy (which handles both online and offline db migrations) | 14:25 |
| rosmaita | i put a friday hackday proposal on the ptg planning etherpad suggesting hacking the db docs | 14:26 |
| jokke_ | ok, so essentially we come to the point that we should only ever try to put those triggers in if we are dealing with MySQL and that's the only scenario that can be done online | 14:27 |
| jokke_ | and regardless if the admin wants to utilize the step scripts individually as long as they are offline | 14:28 |
| rosmaita | right | 14:28 |
| *** tshefi has joined #openstack-glance | 14:28 | |
| jokke_ | and we do not have mechanism to know if the glance is offline or not, we just assume that the admin would not run db_sync that the do know utilizes emc online | 14:29 |
| jokke_ | I think we're going wrong direction with this patch removing the db detection and warning the op | 14:30 |
| rosmaita | we have warnings in some of the other patches | 14:31 |
| rosmaita | i tested this last night, if you try anything other than db_sync with postgres, you get an error message | 14:31 |
| rosmaita | (pretty sure that's what happened) | 14:31 |
| rosmaita | in any case, i got results that made sense | 14:32 |
| rosmaita | but , you are right, we need to be careful | 14:33 |
| abhishekk | jokke_, even keystone uses triggers for rolling upgrades and they don't have checks for engine (only thing is they have different triggers for mysql, postgresql and sqlite) | 14:33 |
| jokke_ | abhishekk: so they actually do have checks for different engines | 14:34 |
| abhishekk | yes but for executing the respective triggers based on engine | 14:34 |
| rosmaita | so i think we really need to do rc-3 today, do we want to hold off on the db stuff for rc-4 on thursday? | 14:35 |
| rosmaita | although, of course, it would be good if we can get the db stuff figured out today | 14:36 |
| jokke_ | and I think that's what we should do as well and if we do not have triggers for that specific engine we fault out and give the operator --offline flag whenre they take the responsibility that they are actually running the command glance offline | 14:36 |
| rosmaita | that is basically what we have | 14:37 |
| rosmaita | if you run db-sync, you are doing offline | 14:37 |
| *** amoralej|lunch is now known as amoralej | 14:37 | |
| rosmaita | if you do separate e m c steps, you are only your own (that's the experimental stuff) | 14:38 |
| *** ykarel|away is now known as ykarel | 14:38 | |
| rosmaita | *you are on your own | 14:38 |
| jokke_ | rosmaita: well my problem is _that_ claim is/has not been stated anywhere | 14:38 |
| abhishekk | jokke_, so should I mention that in the documentation patch? | 14:39 |
| rosmaita | https://docs.openstack.org/glance/latest/admin/db.html | 14:39 |
| rosmaita | https://docs.openstack.org/glance/latest/admin/zero-downtime-db-upgrade.html | 14:39 |
| rosmaita | that first page only mentions db sync (i think) | 14:40 |
| rosmaita | the second one needs to be updated to say "experimental in queens" | 14:40 |
| rosmaita | i may add something to the release notes like for uwsgi saying the reason it's still experimental is lack of resources (need in-gate testing set up) | 14:42 |
| *** alexchadin has quit IRC | 14:45 | |
| *** Supun has quit IRC | 14:47 | |
| *** Supun has joined #openstack-glance | 14:48 | |
| abhishekk | jokke_, so what exactly we should do now? | 14:52 |
| jokke_ | abhishekk: I'm trying to understand the situation ... like told many times before databases are not my strong field | 14:54 |
| *** hoonetorg has quit IRC | 14:54 | |
| abhishekk | jokke_, one thing is sure that with these patches it will be lot more stable than now (that is my opinion though ;) ) | 14:55 |
| openstackgerrit | Abhishek Kekane proposed openstack/glance master: Revisit database rolling upgrade documentation https://review.openstack.org/546172 | 14:56 |
| *** udesale has joined #openstack-glance | 14:56 | |
| rosmaita | i am getting confused with all these patches. if i have a postgres db, can i run expand, migrate, contract separately? | 14:58 |
| *** ykarel is now known as ykarel|away | 14:59 | |
| *** d0ugal has quit IRC | 14:59 | |
| abhishekk | rosmaita, now yes | 14:59 |
| abhishekk | rosmaita, with my patch yes, but on current master no | 15:00 |
| *** d0ugal has joined #openstack-glance | 15:00 | |
| rosmaita | so if you run db_sync, your patch https://review.openstack.org/#/c/544792/5 does not install triggers | 15:01 |
| rosmaita | is there a similar check in migrate (if postgres -> no triggers)? | 15:01 |
| abhishekk | for migrate call there are no triggers | 15:02 |
| abhishekk | triggers are only for expand and contract | 15:02 |
| rosmaita | oh right, they are installed in expand | 15:02 |
| rosmaita | sorry | 15:02 |
| abhishekk | yes and removed in contract | 15:02 |
| rosmaita | same question, except about expand | 15:02 |
| abhishekk | so if we don't want to execute EMC with postgres then need to change a code litile bit | 15:05 |
| abhishekk | s/litile/little | 15:05 |
| rosmaita | there are 2 separate questions here | 15:06 |
| rosmaita | (1) doing the e - m - c operations on a database | 15:06 |
| rosmaita | (2) giving an operator the ability to do them separately | 15:06 |
| rosmaita | we have to do all 3 operations | 15:06 |
| abhishekk | yes | 15:06 |
| rosmaita | it's just a question of whether we give an operator access to them individually | 15:06 |
| jokke_ | hopefully my comment in the review makes sense | 15:07 |
| rosmaita | we need to do (2) if we are going to have the online zero-downtime migrations | 15:07 |
| rosmaita | but only for mysql | 15:07 |
| rosmaita | since that's all we support | 15:07 |
| *** r-daneel has joined #openstack-glance | 15:07 | |
| rosmaita | i think (1) and (2) got a bit tangled up on Shilpa's patch | 15:07 |
| abhishekk | yes | 15:08 |
| rosmaita | jokke_ what are your thoughts about rc-3 ? | 15:09 |
| rosmaita | namely | 15:09 |
| rosmaita | should we do rc-3 today with web-upload and doc changes | 15:10 |
| rosmaita | and aim for rc-4 with the database glance-manage stuff and docs? | 15:10 |
| rosmaita | rc-4 on thursday | 15:10 |
| rosmaita | or, hold off on rc-3 | 15:10 |
| rosmaita | and get everything in at once? | 15:11 |
| *** dtantsur is now known as dtantsur|brb | 15:11 | |
| rosmaita | abhishekk ^^ your thoughts too | 15:11 |
| jokke_ | since rc2 we literally have nothing but translations and doc changes merged there | 15:11 |
| abhishekk | I guess we should have all in rc-3 and we can do it tomorrow | 15:12 |
| rosmaita | ok, works for me | 15:14 |
| abhishekk | to be sure we are on the same page 1) for db_sync no need of engine check and 2) for EMC should work only for MYSQL and sqlite (sqlite due to test cases) | 15:15 |
| rosmaita | abhishekk well, 1) only because you are turning off triggers completely | 15:15 |
| abhishekk | rosmaita, yes | 15:16 |
| rosmaita | and for 2) we are talking about the glance-manage expand, glance-manage migrate, glance-manage contract commands | 15:16 |
| abhishekk | yes | 15:16 |
| rosmaita | not the EMC operations | 15:16 |
| jokke_ | abhishekk: yes that works for me when the triggers are limited out db_sync | 15:16 |
| *** pdr_ has quit IRC | 15:16 | |
| abhishekk | So give me a hour I will come up with a patch | 15:16 |
| rosmaita | abhishekk you are fantastic! | 15:16 |
| *** itlinux has quit IRC | 15:16 | |
| jokke_ | abhishekk: sorry :( | 15:17 |
| abhishekk | also as https://bugs.launchpad.net/glance/+bug/1749640 bug is important for tripleo I will remove its dependency on parent patch (migration with postgre) and submit it as a separate patch so that we can merge it quickly | 15:18 |
| openstack | Launchpad bug 1749640 in Glance "db sync fails for mysql while adding triggers" [High,In progress] - Assigned to Abhishek Kekane (abhishek-kekane) | 15:18 |
| rosmaita | abhishekk you handle the code, i will patch the docs | 15:18 |
| abhishekk | rosmaita, great | 15:18 |
| abhishekk | jokke_, no issues | 15:18 |
| abhishekk | rosmaita, jokke_ If my above point make sense should I remove the dependency? | 15:18 |
| *** Supun has quit IRC | 15:18 | |
| rosmaita | ok. in the mean time, jokke_ let's get https://review.openstack.org/#/c/545649/ and https://review.openstack.org/#/c/546020/ worked out | 15:20 |
| abhishekk | great | 15:20 |
| rosmaita | abhishekk i think so, it applies to db-sync and the anti-postgres stuff is still in place for expand, contract, and migrate commands atm | 15:21 |
| abhishekk | rosmaita, yes | 15:21 |
| rosmaita | makes sense to remove the dependency, then | 15:22 |
| abhishekk | ok | 15:22 |
| *** hoonetorg has joined #openstack-glance | 15:25 | |
| *** ykarel|away has quit IRC | 15:35 | |
| *** d0ugal has quit IRC | 15:36 | |
| openstackgerrit | Abhishek Kekane proposed openstack/glance master: Triggers shouldn't be execute in offline migration https://review.openstack.org/544792 | 15:36 |
| *** d0ugal has joined #openstack-glance | 15:40 | |
| *** d0ugal has quit IRC | 15:52 | |
| *** takedakn has joined #openstack-glance | 16:02 | |
| *** takedakn has quit IRC | 16:05 | |
| *** dtantsur|brb is now known as dtantsur | 16:06 | |
| *** itlinux has joined #openstack-glance | 16:06 | |
| *** Supun has joined #openstack-glance | 16:06 | |
| *** mosulica has quit IRC | 16:06 | |
| *** d0ugal has joined #openstack-glance | 16:06 | |
| *** links has joined #openstack-glance | 16:15 | |
| *** d0ugal has quit IRC | 16:17 | |
| *** pcaruana has quit IRC | 16:21 | |
| *** Supun has quit IRC | 16:26 | |
| *** AlexeyAbashkin has quit IRC | 16:26 | |
| abhishekk | ready with the patch with testing, just running tox internally | 16:34 |
| *** d0ugal has joined #openstack-glance | 16:34 | |
| rosmaita | abhishekk great! | 16:36 |
| *** udesale has quit IRC | 16:36 | |
| openstackgerrit | Abhishek Kekane proposed openstack/glance master: Migration support for postgresql https://review.openstack.org/545211 | 16:40 |
| abhishekk | rosmaita, jokke_ ^^^ | 16:40 |
| rosmaita | ty | 16:41 |
| abhishekk | I have tested this patch with scrips http://paste.openstack.org/show/674262/ | 16:41 |
| rosmaita | abhishekk i am taking over your https://review.openstack.org/#/c/546172/ patch, will make you co-author | 16:41 |
| abhishekk | for postgresql, db sync will work as expected and exapnd, migrate and contract command will exit with error | 16:42 |
| abhishekk | rosmaita, no issues | 16:42 |
| rosmaita | abhishekk excellent | 16:42 |
| abhishekk | have a careful look and let me know if you have any suggestions | 16:43 |
| rosmaita | will do ... it's like 10pm for you now, probably time for you to go offline ... i will look and leave notes on the review | 16:45 |
| abhishekk | (I will be around for next couple of hours) | 16:46 |
| *** gyee has joined #openstack-glance | 16:46 | |
| rosmaita | ok, i will look now then | 16:46 |
| abhishekk | thank you | 16:47 |
| *** e0ne has quit IRC | 16:47 | |
| *** r-daneel has quit IRC | 16:55 | |
| *** d0ugal has quit IRC | 16:59 | |
| *** itlinux has quit IRC | 17:02 | |
| *** d0ugal has joined #openstack-glance | 17:03 | |
| *** tshefi has quit IRC | 17:04 | |
| *** itlinux has joined #openstack-glance | 17:05 | |
| *** r-daneel has joined #openstack-glance | 17:13 | |
| *** harlowja has joined #openstack-glance | 17:23 | |
| *** imacdonn has joined #openstack-glance | 17:30 | |
| rosmaita | abhishekk question for you on the review | 17:32 |
| abhishekk | rosmaita, looking | 17:33 |
| *** pcaruana has joined #openstack-glance | 17:36 | |
| *** links has quit IRC | 17:37 | |
| *** btully has joined #openstack-glance | 17:38 | |
| *** d0ugal has quit IRC | 17:38 | |
| abhishekk | rosmaita, replied | 17:41 |
| rosmaita | ty | 17:41 |
| abhishekk | let me know, so that I can push a new PS if you are satisfied | 17:41 |
| *** btully has quit IRC | 17:42 | |
| rosmaita | abhishekk that's what i thought, just wanted to verify | 17:43 |
| abhishekk | rosmaita, great, will upload a new PS in next 15 minutes | 17:43 |
| rosmaita | thanks, appreciate your patience! | 17:43 |
| *** Supun has joined #openstack-glance | 17:44 | |
| abhishekk | rosmaita, no issues! | 17:49 |
| abhishekk | rosmaita, I have tested this patch with postgresql and combinations (db_sync, expand, migrate and contract) | 17:51 |
| abhishekk | I am 100% sure it will work without issue for mysql but should I just test again with mysql for a safer side? | 17:52 |
| rosmaita | it's getting pretty late for you, i'll do it | 17:53 |
| rosmaita | how are you testing? | 17:53 |
| abhishekk | it will take another 20 minutes | 17:53 |
| abhishekk | I will push a patch now and then will perform a test again mysql | 17:54 |
| abhishekk | if something fails will update you | 17:54 |
| rosmaita | ok, cool | 17:54 |
| openstackgerrit | Abhishek Kekane proposed openstack/glance master: Migration support for postgresql https://review.openstack.org/545211 | 17:55 |
| *** AlexeyAbashkin has joined #openstack-glance | 17:57 | |
| *** d0ugal has joined #openstack-glance | 18:00 | |
| *** AlexeyAbashkin has quit IRC | 18:02 | |
| *** openstackgerrit has quit IRC | 18:03 | |
| abhishekk | rosmaita, could you please also review https://review.openstack.org/#/c/544792/ ? | 18:04 |
| *** d0ugal has quit IRC | 18:05 | |
| *** Supun has quit IRC | 18:05 | |
| rosmaita | abhishekk: looking | 18:06 |
| abhishekk | thank you | 18:06 |
| jokke_ | abhishekk: Thanks for the new rev of 545211 ... that looks great | 18:09 |
| jokke_ | lets hope zuul agrees | 18:09 |
| abhishekk | jokke_, it will :) | 18:09 |
| abhishekk | jokke_, just give me another 15 minutes, I am testing with mysql now (pretty sure it will work though) | 18:10 |
| *** d0ugal has joined #openstack-glance | 18:11 | |
| *** Supun has joined #openstack-glance | 18:13 | |
| *** AlexeyAbashkin has joined #openstack-glance | 18:13 | |
| *** pcaruana has quit IRC | 18:16 | |
| rosmaita | abhishekk your https://review.openstack.org/#/c/544792/ doesn't seem to address https://launchpad.net/bugs/1749640 any more ... i think you have fixed a different bug (that hasn't been filed yet), something like "db sync should not install/remove triggers" | 18:16 |
| openstack | Launchpad bug 1749640 in Glance "db sync fails for mysql while adding triggers" [High,In progress] - Assigned to Abhishek Kekane (abhishek-kekane) | 18:16 |
| rosmaita | if you want, i will file the bug, you basically explain the problem in your commit message on https://review.openstack.org/#/c/544792/6 | 18:16 |
| abhishekk | rosmaita, yes it will fix | 18:17 |
| *** AlexeyAbashkin has quit IRC | 18:17 | |
| rosmaita | ok, i take it back, you have a more general fix | 18:18 |
| abhishekk | yes | 18:18 |
| *** harlowja has quit IRC | 18:23 | |
| *** Supun has quit IRC | 18:29 | |
| abhishekk | rosmaita, jokke_ tested with mysql, no issues \o/ | 18:32 |
| rosmaita | \o/ | 18:33 |
| abhishekk | i will wait for jenkins results and then go offline | 18:34 |
| abhishekk | jokke_, could you look at 544792 patch (its blocking tripleo) | 18:36 |
| *** d0ugal has quit IRC | 18:38 | |
| *** d0ugal has joined #openstack-glance | 18:39 | |
| *** tesseract has quit IRC | 18:45 | |
| *** d0ugal has quit IRC | 18:51 | |
| *** harlowja has joined #openstack-glance | 18:55 | |
| *** d0ugal_ has joined #openstack-glance | 18:57 | |
| *** harlowja_ has joined #openstack-glance | 18:59 | |
| *** harlowja has quit IRC | 18:59 | |
| *** btully has joined #openstack-glance | 19:27 | |
| *** btully has quit IRC | 19:31 | |
| EmilienM | abhishekk: thanks. Indeed, jokke_ if you can look asap, it's blocking our production chain. Thanks a lot! | 19:34 |
| abhishekk | rosmaita, hey tempest-full-py3 job has failed on 545211 patch, need to recheck it | 19:35 |
| abhishekk | i will opt out now | 19:35 |
| rosmaita | ok, will do ... have a good evening, and thanks! | 19:36 |
| jokke_ | EmilienM: abhishekk yeah was waiting for zuul, wasn't finished yet when I was here last time | 19:36 |
| EmilienM | jokke_: no prob :) | 19:37 |
| abhishekk | rosmaita, jokke_ thank you, have a nice rest of the day | 19:37 |
| *** abhishekk has quit IRC | 19:37 | |
| *** d0ugal_ has quit IRC | 19:45 | |
| *** dtantsur is now known as dtantsur|afk | 19:46 | |
| *** amoralej is now known as amoralej|off | 19:47 | |
| *** tshefi has joined #openstack-glance | 19:52 | |
| *** tshefi has quit IRC | 19:55 | |
| *** tshefi has joined #openstack-glance | 19:55 | |
| *** d0ugal_ has joined #openstack-glance | 19:56 | |
| *** d0ugal_ has quit IRC | 20:03 | |
| *** d0ugal_ has joined #openstack-glance | 20:04 | |
| *** d0ugal_ has quit IRC | 20:10 | |
| *** e0ne has joined #openstack-glance | 20:11 | |
| *** d0ugal_ has joined #openstack-glance | 20:13 | |
| *** AlexeyAbashkin has joined #openstack-glance | 20:23 | |
| *** AlexeyAbashkin has quit IRC | 20:27 | |
| *** d0ugal_ has quit IRC | 20:30 | |
| *** pcaruana has joined #openstack-glance | 20:40 | |
| *** openstackgerrit has joined #openstack-glance | 20:44 | |
| openstackgerrit | Brian Rosmaita proposed openstack/glance master: Revise database rolling upgrade documentation https://review.openstack.org/546172 | 20:44 |
| *** e0ne has quit IRC | 20:47 | |
| openstackgerrit | Brian Rosmaita proposed openstack/glance master: Revise database rolling upgrade documentation https://review.openstack.org/546172 | 20:48 |
| *** e0ne has joined #openstack-glance | 20:51 | |
| *** d0ugal_ has joined #openstack-glance | 20:56 | |
| *** r-daneel_ has joined #openstack-glance | 20:59 | |
| *** r-daneel has quit IRC | 21:00 | |
| *** r-daneel_ is now known as r-daneel | 21:00 | |
| openstackgerrit | Brian Rosmaita proposed openstack/glance master: Revise help text for uri filtering options https://review.openstack.org/546020 | 21:06 |
| rosmaita | jokke_ : see what you think ^^ | 21:07 |
| jokke_ | that works :D | 21:08 |
| rosmaita | \o/ | 21:15 |
| *** AlexeyAbashkin has joined #openstack-glance | 21:23 | |
| *** AlexeyAbashkin has quit IRC | 21:27 | |
| *** tshefi has quit IRC | 21:28 | |
| *** pcaruana has quit IRC | 21:37 | |
| *** d0ugal_ has quit IRC | 22:03 | |
| *** jdurgin has joined #openstack-glance | 22:13 | |
| *** r-daneel has quit IRC | 22:22 | |
| *** r-daneel has joined #openstack-glance | 22:22 | |
| *** threestrands has joined #openstack-glance | 22:23 | |
| *** rosmaita has quit IRC | 22:27 | |
| *** rcernin has joined #openstack-glance | 22:33 | |
| *** itlinux has quit IRC | 22:49 | |
| *** rosmaita has joined #openstack-glance | 22:53 | |
| *** e0ne has quit IRC | 22:55 | |
| *** openstackgerrit has quit IRC | 23:04 | |
| *** NostawRm has quit IRC | 23:15 | |
| *** NostawRm has joined #openstack-glance | 23:19 | |
| *** AlexeyAbashkin has joined #openstack-glance | 23:23 | |
| *** dalgaaf has quit IRC | 23:26 | |
| *** AlexeyAbashkin has quit IRC | 23:27 | |
| *** dalgaaf has joined #openstack-glance | 23:28 | |
| *** rosmaita has quit IRC | 23:30 | |
| *** r-daneel has quit IRC | 23:34 | |
| *** r-daneel has joined #openstack-glance | 23:35 | |
| *** Tahvok has quit IRC | 23:46 | |
| *** Tahvok has joined #openstack-glance | 23:47 | |
| *** r-daneel has quit IRC | 23:50 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!