| *** mridu has quit IRC | 01:35 | |
| *** jamesmcarthur has joined #storyboard | 01:41 | |
| *** jamesmcarthur has quit IRC | 02:20 | |
| *** jamesmcarthur has joined #storyboard | 02:55 | |
| *** persia has quit IRC | 04:20 | |
| *** persia has joined #storyboard | 04:21 | |
| *** udesale has joined #storyboard | 04:41 | |
| *** jtomasek has joined #storyboard | 07:18 | |
| *** tosky has joined #storyboard | 08:45 | |
| *** fatema_ has joined #storyboard | 09:40 | |
| fatema_ | Sorry for the delay There was some technical issues | 09:51 |
|---|---|---|
| fatema_ | the netstat command got me "tcp 0 0 0.0.0.0:8080 0.0.0.0:* LISTEN " | 09:52 |
| SotK | hey fatema_, I wrote some explanation yesterday before I noticed you'd dropped offline, you can see it in the logs here: http://eavesdrop.openstack.org/irclogs/%23storyboard/%23storyboard.2018-03-22.log.html#t2018-03-22T23:24:20 | 09:53 |
| fatema_ | SotK, oh yes it is clearer now | 09:58 |
| fatema_ | now it should be working fine and I shall go for the contribution part | 10:00 |
| *** udesale has quit IRC | 11:06 | |
| fungi | diablo_rojo: thanks for finding the osc migration issue, i'll see how far i get and maybe try to tackle the db reconfig today too | 12:14 |
| *** tosky has quit IRC | 12:15 | |
| *** tosky has joined #storyboard | 12:15 | |
| SotK | fungi: http://paste.openstack.org/show/709970/ is all the user ID columns I think | 12:27 |
| fungi | thanks SotK! i'll work that into the script i'm drafting and link it in here before i test on sb-dev.o.o | 12:28 |
| *** jamesmcarthur has quit IRC | 12:49 | |
| *** jamesmca_ has joined #storyboard | 12:49 | |
| *** jamesmca_ has quit IRC | 13:24 | |
| *** jamesmcarthur has joined #storyboard | 13:25 | |
| *** jamesmcarthur has quit IRC | 13:30 | |
| fungi | diablo_rojo: https://review.openstack.org/555787 Use utf8mb4 for MySQL database charset | 13:55 |
| fungi | that's mostly so we don't forget when deploying new sb servers, not that it happens especially often | 13:55 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/storyboard master: Use utf8mb4 for MySQL database charset https://review.openstack.org/555806 | 14:14 |
| fungi | SotK: mordred: with ^ i assume i also need a db migration, and that it's not kosher to simply edit the MYSQL_CHARSET value at http://git.openstack.org/cgit/openstack-infra/storyboard/tree/storyboard/db/migration/alembic_migrations/versions/001_initial_version.py#n37 ? | 14:14 |
| fungi | mordred: note that this is a blocker for the osc import from lp | 14:18 |
| fungi | (there is apparently some data in at least one osc bug which contains 4-byte characters) | 14:19 |
| mordred | fungi: oh wow, fun | 14:21 |
| fungi | it's also a blocker for heat and mistral, i'm told | 14:21 |
| fungi | so it's not like osc is the only one, just happens to be the only one so far which has requested an import | 14:21 |
| mordred | fungi: so - yeah - I *think* we'll need a migration script that does an alter on every column we care about | 14:23 |
| fungi | i think i also need to move sb.o.o and sb-dev.o.o to rebuilt trove instances using newer mysql? they're both running 5.6 at the moment | 14:23 |
| fungi | or is that new enough? | 14:24 |
| SotK | mordred: that matches my understanding | 14:24 |
| *** jamesmcarthur has joined #storyboard | 14:24 | |
| SotK | from when I was looking at fixing this a while ago | 14:24 |
| SotK | I didn't actually get any further than that | 14:25 |
| mordred | ALTER TABLE foo CONVERT TO CHARACTER SET utf8mb4 | 14:25 |
| mordred | is apparently the form we need | 14:25 |
| mordred | that should update the default as well as convert any columns as needed | 14:25 |
| fungi | so basically just loop through all the tables applying that and put it in a new migration file? | 14:26 |
| mordred | (which is nice and easy) | 14:26 |
| mordred | yup | 14:26 |
| fungi | is mysql 5.6 new enough for that? | 14:26 |
| mordred | yes | 14:27 |
| fungi | istr we had to upgrade the mysql version for at least some service we were doing that on previously | 14:27 |
| fungi | okay, cool | 14:27 |
| mordred | well - convert to is in 5. | 14:27 |
| mordred | 5.6 | 14:27 |
| mordred | not sure about utf8mb4 - looking | 14:27 |
| mordred | https://dev.mysql.com/doc/refman/5.6/en/charset-unicode-utf8mb4.html | 14:27 |
| mordred | 5.6 has utf8mb4 too | 14:27 |
| fungi | awesome | 14:27 |
| fungi | i'd still like to move these to mysql 5.7 at some point, but i guess there's no rush | 14:28 |
| fungi | and if the migration gets added to sb in the same patch as the storyboard/db/models.py adjustment then we shouldn't need to perform any manual operations server-side right? (i guess that's the entire point of having scripted db migrations?) | 14:29 |
| fungi | i'll still make sure to do a last-minute db dump before that change lands either way | 14:31 |
| -openstackstatus- NOTICE: zuul.o.o has been restarted to pick up latest code base and clear memory usage. Both check / gate queues were saved, be sure to check your patches and recheck when needed. | 14:53 | |
| *** jtomasek has quit IRC | 15:25 | |
| -openstackstatus- NOTICE: Gerrit will be temporarily unreachable as we restart it to complete the rename of some projects. | 15:50 | |
| *** fatema_ has quit IRC | 16:04 | |
| *** jamesmcarthur has quit IRC | 16:12 | |
| *** jamesmcarthur has joined #storyboard | 16:13 | |
| *** jamesmcarthur has quit IRC | 16:13 | |
| *** jamesmcarthur has joined #storyboard | 16:14 | |
| *** jamesmcarthur has quit IRC | 16:23 | |
| *** jamesmcarthur has joined #storyboard | 16:24 | |
| *** jamesmcarthur has quit IRC | 16:28 | |
| *** jamesmcarthur has joined #storyboard | 16:30 | |
| *** mridu has joined #storyboard | 16:51 | |
| mridu | hi, we can debug the issue. I wasn't able to run tox -e py35 successfully. Even after removing py3.6 from the system. | 16:54 |
| *** jamesmcarthur has quit IRC | 17:10 | |
| *** jamesmcarthur has joined #storyboard | 17:11 | |
| mridu | sorry this was by mistake | 17:11 |
| *** mridu has quit IRC | 17:16 | |
| *** diablo_rojo has quit IRC | 17:30 | |
| *** jamesmcarthur has quit IRC | 17:35 | |
| *** mridu_ has joined #storyboard | 17:38 | |
| *** mridu_ has quit IRC | 17:42 | |
| *** mridu has joined #storyboard | 17:42 | |
| *** mridu has quit IRC | 17:47 | |
| *** Mridu has joined #storyboard | 18:10 | |
| Mridu | On running tox command I am getting http://paste.openstack.org/show/710362/ as error | 18:11 |
| *** Mridu has quit IRC | 18:22 | |
| fungi | we just lost Mridu | 18:23 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/storyboard master: Use utf8mb4 for MySQL database charset https://review.openstack.org/555806 | 18:49 |
| fungi | SotK: mordred: ^ my first ever hackish attempt at an alembic migration | 18:49 |
| mordred | fungi: +2 on the patch-I think we can address the pg concern separately | 18:54 |
| fungi | i'm fine addressing it now too | 18:55 |
| fungi | mordred: is it just something like "if dialect.name=='mysql':" or is it more involved than that? | 19:11 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/storyboard master: Use utf8mb4 for MySQL database charset https://review.openstack.org/555806 | 19:12 |
| fungi | seems to match examples i've found | 19:12 |
| mordred | fungi: I'm assuming that matches mysqldb and pymysql drivers | 19:13 |
| TheJulia | out of curiosity, since I've not had time to really even read this channel this week... Is ironic still on track for the migration being run today? | 19:15 |
| *** mridu has joined #storyboard | 19:17 | |
| fungi | TheJulia: yeah, i plan on doing that later in my day | 19:20 |
| TheJulia | okay, thanks! | 19:20 |
| fungi | also working on 4-byte character support blocking some other imports (not a blocker for ironic afaik) and account deduplication (which also isn't a blocker but better done sooner than later) | 19:20 |
| clarkb | fungi: if storyboard is using trove make sure the instance has new enough mysql for 4 byte utf8 | 19:21 |
| *** diablo_rojo has joined #storyboard | 19:22 | |
| fungi | clarkb: yes, i checked and it's 5.6 | 19:23 |
| fungi | according to mordred 5.5 and later works for that | 19:23 |
| clarkb | cool | 19:23 |
| TheJulia | fungi: okay, thanks | 19:23 |
| fungi | though 5.7 apparently has workarounds for the "191-byte limit" on varchar(256) fields with utf8mb4 | 19:24 |
| fungi | based on my reading | 19:24 |
| fungi | er, varchar(255) i guess | 19:24 |
| fungi | skimming sb i think most of the varchar key fields are limited to 100 characters, so probably not an issue anyway | 19:27 |
| fungi | er, limited to 100 characters *or less* | 19:27 |
| fungi | 100 was the biggest i saw | 19:27 |
| *** openstackgerrit has quit IRC | 19:34 | |
| mridu | Hey, I am still unable to install tox | 19:35 |
| mridu | http://paste.openstack.org/show/710362/ | 19:35 |
| fungi | mridu: you came back! that looks like tox is properly installed, and even shows some passing tests | 19:36 |
| fungi | you just happen to have 272 failing tests too. hopefully someone here (not me i'm afraid) is aware of some local dependencies we may not express in packages alone | 19:37 |
| fungi | perhaps we have some database permission assumptions which the tests are relying on? | 19:37 |
| mridu | ohh... | 19:38 |
| mridu | what needs to be done then? | 19:38 |
| mridu | Hi all, can anyone up here help me with fixing the errors I am getting on installing tox -e py35 | 19:45 |
| diablo_rojo | fungi, I don't know of anything fancy we do. | 19:59 |
| fungi | i suppose i should try running tests locally | 20:00 |
| diablo_rojo | Thats what I was about to do :) | 20:00 |
| diablo_rojo | The only dependencies I see it installing are req.txt and test-reqs.txt | 20:01 |
| diablo_rojo | Huh. I actually have py35 failures too- 34 pass and 4 skipped with 272 failing | 20:01 |
| * diablo_rojo just pulled master and ran tests | 20:02 | |
| diablo_rojo | And the same with py27 | 20:02 |
| diablo_rojo | I'm seeing a lot of- 'can't connect to the mysql server 127.0.0.1' | 20:03 |
| mridu | same here | 20:04 |
| fungi | i expect our jobs also stand up mysql and postgresql instances listening on the loopback interface with fairly open permissions | 20:04 |
| fungi | and our unit tests probably don't opportunistically enable those, so just blindly assume they're present and then fail when they can't connect | 20:05 |
| diablo_rojo | sounds like something we need fixed. | 20:05 |
| *** diablo_rojo has quit IRC | 20:26 | |
| *** diablo_rojo has joined #storyboard | 20:26 | |
| fungi | i expect it's a pattern originally inherited from nova, but oslo.db eventually grew mechanisms for opportunistic db-specific testing | 20:33 |
| fungi | so for example if you run nova's unit tests locally and have mysql running it'll try mysql-specific tests but will skip them if you don't | 20:33 |
| diablo_rojo | fungi, so.. make sure mysql isn't running and the tests should run/skip to completion? | 20:35 |
| fungi | i think that's what the nova et al unit tests do, but this is not my field of expertise | 20:37 |
| fungi | also, basically all ci tests for my https://review.openstack.org/555806 change are dying with a retry_limit failure, so i'm probably a terrible person to attempt to answer these questions | 20:38 |
| fungi | hrm... why would even pep8 be hitting a retry_limit? | 20:45 |
| fungi | i guess the db stuff happens in the pre phase so gets retried when it breaks | 20:46 |
| fungi | that 'splains it | 20:46 |
| fungi | createdb: "utf8mb4" is not a valid encoding name | 20:46 |
| fungi | srsly? | 20:46 |
| diablo_rojo | Haha how is that not a valid name? | 20:48 |
| fungi | mordred: ^ any clue? should i really not be doing createdb -E utf8mb4 in tools/test-setup.sh after all? | 20:51 |
| fungi | i guess if that's getting run at the start, we want it on regular utf8 anyway and then the migration should change the tables to utf8mb4 once they get applied? | 20:51 |
| fungi | i'll try dropping that bit | 20:52 |
| fungi | oh! that was pgsql | 20:53 |
| fungi | no wonder :/ | 20:53 |
| * fungi sighs | 20:53 | |
| fungi | but still, i think i've talked myself into creating as utf8 so the migration gets correctly tested | 20:53 |
| *** openstackgerrit has joined #storyboard | 20:54 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/storyboard master: Use utf8mb4 for MySQL database charset https://review.openstack.org/555806 | 20:54 |
| mridu | hey, is there anything else I could take up meanwhile? | 20:59 |
| diablo_rojo | mridu, well figuring out how to run tests and actually create/mock out the db so they all can run locally would be super helpful. Otherwise there are the commands we need implemented in the cli that I've mentioned before. | 21:04 |
| openstackgerrit | Kendall Nelson proposed openstack-infra/storyboard master: Remove unnecessary _route method from users https://review.openstack.org/387080 | 21:36 |
| *** mridu has quit IRC | 21:51 | |
| fungi | well, at least my utf8mb4 change is only failing, not hitting retry_limit so i take that as a step forward | 22:01 |
| fungi | i'll see if i can figure out why and get the next iteration up before i get back to hacking on the deduplication script | 22:02 |
| fungi | aha, i see it | 22:04 |
| diablo_rojo | fungi, pep8 is failing cause you import things and don't use them, but I'm sure you saw that | 22:06 |
| fungi | that wasn't all | 22:07 |
| fungi | one of the things it complains about is also why the unit tests are failing | 22:08 |
| fungi | inspect() isn't a builtin function ;) | 22:08 |
| diablo_rojo | ah, that would do it | 22:08 |
| fungi | digging around in alembic docs now to find out what it's a method of... i was somewhat cribbing from a suggestion zzzeek made in #openstack-oslo earlier | 22:09 |
| fungi | the docs don't mention an inspect() method anywhere in alembic | 22:12 |
| fungi | oh! that's courtesy of sqlalchemy, not alembiv | 22:14 |
| fungi | alembic | 22:14 |
| diablo_rojo | Ha, solves the mystery of why you imported sqlalchemy in the first place | 22:14 |
| fungi | indeed | 22:15 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/storyboard master: Use utf8mb4 for MySQL database charset https://review.openstack.org/555806 | 22:15 |
| fungi | okay, back to account deduplication scripting while that bakes | 22:15 |
| *** EmilienM has quit IRC | 22:34 | |
| *** EmilienM has joined #storyboard | 22:36 | |
| EmilienM | hello | 22:38 |
| fungi | howdy | 22:39 |
| *** jamesmcarthur has joined #storyboard | 22:42 | |
| *** jamesmcarthur has quit IRC | 22:46 | |
| fungi | SotK: diablo_rojo: okay, after a bit of debugging, correcting table names and tweaking order of operations to avoid integrity violations, i have a dedup script which runs successfully on storyboard-dev.o.o | 22:48 |
| fungi | http://paste.openstack.org/show/710712/ is the script | 22:49 |
| fungi | i'm still double-checking the results in the db | 22:49 |
| fungi | tables/fields and recommended actions taken from SotK's list at http://paste.openstack.org/show/709970/ | 22:49 |
| fungi | only thing i didn't implement was deduplicating the subscriptions rows after rewriting user_id on them | 22:50 |
| fungi | since nobody can have actually logged into the duplicate accounts, and i don't see the import script setting subscriptions, i _think_ that's probably fine (i doubt there will be any hits at all for the duplicate accounts in that table, much less both versions of the account subscribing to the same things) | 22:51 |
| fungi | hrm, something's still not quite right. for some reason the sql update and delete commands executed on that cursor don't actually seem to take effect | 22:57 |
| fungi | aha, i needed to db_connection.commit() at the very end! | 22:59 |
| fungi | poking around in the database now, it seems to have done the right thing | 23:01 |
| fungi | a few extra eyes on the script and/or the server would be appreciated | 23:02 |
| fungi | sorry, that's actually http://paste.openstack.org/show/710718/ for the script | 23:05 |
| fungi | so anyway, if people are cool with the script and the state of the server i'll do a last-minute backup and then run that on the production storyboard.o.o | 23:08 |
| fungi | before i start doing new migrations | 23:08 |
| fungi | i've single-core approved https://review.openstack.org/554726 and will use it in the test and then production imports (after cleanups) | 23:13 |
| fungi | shifting gears, my utf8mb4 change is passing pep8 now, but the unit test jobs for it are timing out | 23:14 |
| fungi | yeah, something's still not right with my character set migration: http://logs.openstack.org/06/555806/5/check/tox-py27/7e01341/job-output.txt.gz#_2018-03-23_22_21_43_381913 | 23:16 |
| fungi | i'm not quite sure what to make of that | 23:18 |
| *** tosky has quit IRC | 23:19 | |
| openstackgerrit | Merged openstack-infra/storyboard master: Duplicating Users at Migration Time https://review.openstack.org/554726 | 23:24 |
| diablo_rojo_phon | I'm not sure what to make of it either. | 23:26 |
| diablo_rojo_phon | I'm fine with going ahead with things I think. | 23:27 |
| diablo_rojo_phon | Thanks for approving that fungi | 23:27 |
| fungi | diablo_rojo_phon: the ironic imports were just straight project-to-project mapping but tripleo-ui needs a bugtag filter, correct? | 23:28 |
| fungi | happen to know whether the project-config changes for all those merged yet? | 23:30 |
| fungi | i can double-check in a moment | 23:30 |
| diablo_rojo_phon | Correct fungi | 23:35 |
| fungi | yeah, looks like neither of the project-config changes have merged yet. i'll hunt them down and review next | 23:36 |
| diablo_rojo_phon | Sounds good. Happy to review and ping as well. | 23:36 |
| diablo_rojo_phon | fungi: I'll ping TheJulia | 23:37 |
| TheJulia | o/ | 23:38 |
| diablo_rojo_phon | We need the project config change up and merged for migration. | 23:38 |
| diablo_rojo_phon | Got the link? | 23:38 |
| fungi | https://review.openstack.org/554370 | 23:39 |
| TheJulia | fungi: yeah, project to project and as noted elsewhere I guess one big project group | 23:39 |
| fungi | that's ironic | 23:39 |
| TheJulia | fungi: thanks | 23:39 |
| fungi | reviewing it first | 23:39 |
| diablo_rojo_phon | Reviewed. | 23:43 |
| fungi | looks like a lot of them could stand to be added to the "ironic" project group | 23:43 |
| fungi | and we can drop some of the one-off project groups like "ironic-inspector" and "ironic-python-agent" | 23:44 |
| fungi | those were just being used to remap bug queries in lp | 23:45 |
| diablo_rojo_phon | Oh yeah they should all be added to the same group. | 23:45 |
| fungi | i'll take a quick shot at updating it | 23:45 |
| *** jamesmcarthur has joined #storyboard | 23:57 | |
| fungi | diablo_rojo_phon: TheJulia: i've updated https://review.openstack.org/554370 for ironic project grouping | 23:58 |
| fungi | while that runs the ci gauntlet i'll dig up the tripleo change | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!