| *** Rockyg has joined #openstack-relmgr-office | 00:12 | |
| *** Rockyg has quit IRC | 00:52 | |
| *** david-ly_ has joined #openstack-relmgr-office | 01:06 | |
| *** david-lyle has quit IRC | 01:09 | |
| *** dims has joined #openstack-relmgr-office | 02:14 | |
| *** dims_ has quit IRC | 02:15 | |
| *** dims has quit IRC | 02:20 | |
| *** dims has joined #openstack-relmgr-office | 02:22 | |
| *** dims has quit IRC | 03:01 | |
| *** AzherKhan has joined #openstack-relmgr-office | 03:01 | |
| *** AzherKhan has quit IRC | 03:10 | |
| *** mestery has joined #openstack-relmgr-office | 03:23 | |
| *** david-ly_ is now known as david-lyle | 03:51 | |
| *** mestery has quit IRC | 06:08 | |
| *** dims has joined #openstack-relmgr-office | 09:59 | |
| *** dims has quit IRC | 10:18 | |
| *** dims has joined #openstack-relmgr-office | 10:59 | |
| ttx | dhellmann: draft of the server versioning blogpost: http://paste.ubuntu.com/11778123/ -- let me know if you see anything missing | 11:38 |
|---|---|---|
| *** dims is now known as dimsum__ | 12:09 | |
| *** dims_ has joined #openstack-relmgr-office | 12:13 | |
| *** dimsum__ has quit IRC | 12:16 | |
| *** mestery has joined #openstack-relmgr-office | 12:52 | |
| *** mestery has quit IRC | 12:57 | |
| *** dimsum__ has joined #openstack-relmgr-office | 13:07 | |
| *** dims_ has quit IRC | 13:10 | |
| dhellmann | ttx: reading | 13:36 |
| dhellmann | ttx: two typos fixed in https://etherpad.openstack.org/p/BJXtVdm37N | 13:40 |
| dhellmann | ttx: do you want to mention the reason for not going with 12 for everyone? | 13:42 |
| ttx | the slow drift reason ? yeah, let me try to add that | 13:45 |
| dhellmann | yeah, I can see that being asked a bunch | 13:46 |
| dimsum__ | dhellmann: should mention big tent somewhere in there? Allow new projects to rapidly iterate? | 13:46 |
| ttx | dhellmann: what were the two typos ? copypasting from etherpad adds unwanted whitespace | 13:47 |
| dhellmann | ttx: conundrum on line 44 and got on line 37 of the etherpad | 13:47 |
| ttx | ack, thx | 13:47 |
| dhellmann | dimsum__: at this point we still want most projects to only release on the cycle boundaries, so I'm not sure that applies (yet) | 13:48 |
| *** dansmith is now known as superdan | 13:49 | |
| dimsum__ | dhellmann: right, this versioning scheme does give them a choice | 13:50 |
| dhellmann | yeah, but I don't think we're ready to advertise that too much, yet | 13:50 |
| dimsum__ | fair point :) | 13:50 |
| ttx | dhellmann: see etherpad | 13:57 |
| dhellmann | I'm not sure I would classify the range of versions we have as "radically-different" but I think I get what you're saying :-) | 13:58 |
| ttx | sufficnely-different ? | 13:58 |
| dhellmann | yeah, I like that better. Or purposefully distinct | 13:59 |
| dhellmann | ttx: ship it! | 13:59 |
| ttx | I'll go with purposefully distinct | 13:59 |
| * dhellmann heads offline for a bit | 14:00 | |
| ttx | http://ttx.re/new-versioning.html | 14:01 |
| dimsum__ | looks good ttx | 14:03 |
| ttx | dhellmann, dimsum__ thanks! | 14:04 |
| *** gordc has joined #openstack-relmgr-office | 15:04 | |
| *** mestery has joined #openstack-relmgr-office | 15:06 | |
| gordc | hi, had a release question regarding liberty-1 and reverting migrations | 15:07 |
| gordc | we merged this in liberty-1 to lower ids in our sql backend: https://review.openstack.org/#/c/183531/ | 15:08 |
| gordc | this apparently breaks heat (and maybe others) since heat generates ids beyond 128char | 15:08 |
| gordc | if i need to revert that migration, is it all good to just hit revert? or do i need to ensure some backward compatibility between liberty-1 and otherstuff | 15:09 |
| * gordc does not have a way to ensure this, that said, migration will not pass if you have data beyond 128char | 15:10 | |
| gordc | dhellmann: ^ in case you have feedback | 15:12 |
| *** dimsum__ has quit IRC | 15:13 | |
| *** mestery has quit IRC | 15:17 | |
| gordc | ttx: ^ | 15:25 |
| ttx | gordc: o/ | 15:54 |
| ttx | gordc: it's fine to break liberty-1 | 15:54 |
| ttx | I mean, it's not fine to break master generally, but it's not worse to break liberty-1 | 15:54 |
| ttx | technically it's not more "released" than any other commit in master | 15:55 |
| ttx | got to run | 15:56 |
| gordc | ttx: cool cool. thanks for confirmation :) | 15:57 |
| *** dimsum__ has joined #openstack-relmgr-office | 16:53 | |
| dhellmann | gordc: I think rather than just reverting that patch, you probably want a migration to change the size back, for anyone deploying from master | 17:18 |
| *** nikhil_k is now known as nikhil_k-eto | 17:33 | |
| *** nikhil_k-eto is now known as nikhil_k-away | 17:33 | |
| gordc | dhellmann: so the one issue is that if you have big ids, the migration will never pass. if you have 'small' ids it will... but i don't really have a consistent way to handle both cases. | 18:14 |
| dhellmann | gordc: in the follow up, introspect the column size and only change it back if it is small | 18:15 |
| gordc | dhellmann: but if i leave the original script in there, it will break. and you will never be able to get to followup script | 18:17 |
| dhellmann | gordc: update that one to skip? or replace its body with the logic to resize back up? | 18:17 |
| gordc | dhellmann: update the existing script to skip on error? | 18:18 |
| dhellmann | gordc: no, to not do anything -- are you using alembic or sqlalchemy-migrate? | 18:22 |
| dhellmann | gordc: either way, i think you'll confuse the migration tracking for any deployment that has successfully run that migration if you just delete it | 18:22 |
| gordc | sqlalchemy-migrate | 18:23 |
| dhellmann | gordc: with a database migration you have to assume that if it lands in master it has landed in production. You can't delete it, but you can make a new migration that undoes the change. | 18:23 |
| gordc | dhellmann: :( | 18:24 |
| dhellmann | gordc: yeah | 18:24 |
| gordc | dhellmann: i'll try to catch the error in existing script and add new script i guess. | 18:24 |
| dhellmann | gordc: that works. You could also delete its body to make it a noop. As long as the script still exists it will not break the tracking | 18:25 |
| gordc | dhellmann: yeah that works too. i'll give that a try. | 18:25 |
| gordc | dhellmann: thanks | 18:26 |
| *** mestery has joined #openstack-relmgr-office | 19:54 | |
| *** gordc has left #openstack-relmgr-office | 20:05 | |
| *** mestery has quit IRC | 20:57 | |
| *** openstack has joined #openstack-relmgr-office | 21:26 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!