| *** andrewbogott is now known as andrewbogott_afk | 00:01 | |
| jlk | So.. if keystone goes behind apache, as keystone devs suggest, how does one "stop" keystone in order to do a db migration, without stopping other things running behind apache? | 00:11 |
|---|---|---|
| *** markvoelker has quit IRC | 00:26 | |
| *** Piet has joined #openstack-operators | 00:26 | |
| mgagne | jlk: let me know when you get an answer to that one | 00:27 |
| jlk | looks like we do a site disable for the keystone specific site and an apache reload | 00:27 |
| *** blair has joined #openstack-operators | 00:28 | |
| jlk | although I don't see if that kills existing connections. | 00:28 |
| *** sbfox has quit IRC | 00:34 | |
| mgagne | jlk: how do you do that with CMS that do not fully support orchestration? | 00:45 |
| jlk | The snarky answer is to not use a CMS that doesn't support orchestration | 00:45 |
| jlk | CMS == change management system ? | 00:46 |
| mgagne | Configuration Management System | 00:46 |
| mgagne | or software | 00:46 |
| jlk | the other answer would be that you figure out a way to stop connections before they hit apache | 00:46 |
| *** VW_ has quit IRC | 00:46 | |
| *** pasquier-s has quit IRC | 00:52 | |
| *** pasquier-s has joined #openstack-operators | 00:53 | |
| *** radez is now known as radez_g0n3 | 00:58 | |
| dvorak | jlk: apparently the app will check for a file to have changed every x time, and reload the app if it has been touched | 01:06 |
| *** Marga_ has joined #openstack-operators | 01:07 | |
| dvorak | this covers it - https://code.google.com/p/modwsgi/wiki/ReloadingSourceCode | 01:07 |
| dvorak | that said, I think I'd rather run each app under a separate nginx instance. I hate configuring apache and I'd rather be able to restart the entire web server if I want | 01:08 |
| jlk | yeah, we're trying to trim down memory use | 01:08 |
| dvorak | our control node and compute node hardware configs are the same, so that's not a problem we have :) | 01:08 |
| jlk | the point isn't reloading on changed file, the point is stopping the service when we do database migrations | 01:09 |
| *** markvoelker has joined #openstack-operators | 01:09 | |
| jlk | dvorak: our control nodes /are/ our compute nodes | 01:09 |
| dvorak | and yeah, then I'd say you just take it out of the LB | 01:10 |
| jlk | apache /is/ the LB :) | 01:10 |
| jlk | but yeah | 01:10 |
| dvorak | I think your config may not be the common case situation :) | 01:10 |
| jlk | that's what makes it fun! | 01:11 |
| *** markvoelker has quit IRC | 01:14 | |
| *** Marga_ has quit IRC | 01:33 | |
| *** pasquier-s has quit IRC | 01:34 | |
| mgagne | jlk: I feel it's a bit like running all openstack services in the same eventlet service. you wouldn't do that for obvious reasons. Ironically, we end up with the same state/challenge with apache/wsgi containers | 01:38 |
| jlk | yeah well... | 01:38 |
| jlk | moving to apache isn't exactly by choice | 01:38 |
| mgagne | I 100% agree | 01:39 |
| mgagne | my point is that we don't know (yet) how to properly handle the tasks you mentioned | 01:39 |
| mgagne | especially for people running all-in-one setup | 01:39 |
| *** alop has joined #openstack-operators | 01:41 | |
| *** pasquier-s has joined #openstack-operators | 01:42 | |
| *** j05h has quit IRC | 01:42 | |
| mgagne | people using fpm (or any other tools wrapping around fpm), how do you generate your .changes file? | 01:43 |
| *** alop has quit IRC | 01:44 | |
| *** Marga_ has joined #openstack-operators | 01:45 | |
| *** SimonChung1 has quit IRC | 01:46 | |
| *** VW_ has joined #openstack-operators | 01:53 | |
| *** reed has quit IRC | 02:00 | |
| *** VW_ has quit IRC | 02:00 | |
| *** VW_ has joined #openstack-operators | 02:00 | |
| *** yapeng_ has joined #openstack-operators | 02:01 | |
| *** dmsimard_away is now known as dmsimard | 02:08 | |
| *** markvoelker has joined #openstack-operators | 02:10 | |
| *** markvoelker has quit IRC | 02:15 | |
| *** VW__ has joined #openstack-operators | 02:16 | |
| *** VW__ has quit IRC | 02:16 | |
| *** VW_ has quit IRC | 02:16 | |
| *** VW_ has joined #openstack-operators | 02:16 | |
| *** Marga_ has quit IRC | 02:21 | |
| *** markvoelker has joined #openstack-operators | 02:27 | |
| *** blair has quit IRC | 02:32 | |
| *** markvoelker has quit IRC | 02:33 | |
| *** VW_ has quit IRC | 02:52 | |
| *** VW_ has joined #openstack-operators | 02:53 | |
| *** VW_ has quit IRC | 02:56 | |
| *** VW_ has joined #openstack-operators | 02:57 | |
| *** VW_ has quit IRC | 02:58 | |
| *** VW_ has joined #openstack-operators | 02:59 | |
| *** VW_ has quit IRC | 02:59 | |
| *** VW_ has joined #openstack-operators | 03:00 | |
| *** VW_ has quit IRC | 03:00 | |
| *** VW_ has joined #openstack-operators | 03:00 | |
| *** VW__ has joined #openstack-operators | 03:04 | |
| *** Piet has quit IRC | 03:05 | |
| *** VW__ has quit IRC | 03:06 | |
| *** VW__ has joined #openstack-operators | 03:06 | |
| *** VW_ has quit IRC | 03:07 | |
| *** harlowja_ is now known as harlowja_away | 03:21 | |
| *** blair has joined #openstack-operators | 03:21 | |
| *** markvoelker has joined #openstack-operators | 03:28 | |
| *** markvoelker has quit IRC | 03:33 | |
| *** sanjayu has joined #openstack-operators | 04:10 | |
| *** sbfox has joined #openstack-operators | 04:22 | |
| *** markvoelker has joined #openstack-operators | 04:29 | |
| *** dmsimard is now known as dmsimard_away | 04:29 | |
| *** SimonChung has joined #openstack-operators | 04:30 | |
| *** alaski has quit IRC | 04:30 | |
| *** mgagne has quit IRC | 04:31 | |
| *** markvoelker has quit IRC | 04:34 | |
| *** sbfox has quit IRC | 04:43 | |
| *** mgagne has joined #openstack-operators | 04:45 | |
| *** mgagne is now known as Guest75711 | 04:45 | |
| *** Marga_ has joined #openstack-operators | 04:57 | |
| *** yapeng_ has quit IRC | 05:04 | |
| *** blair has quit IRC | 05:20 | |
| *** markvoelker has joined #openstack-operators | 05:31 | |
| *** openstack has joined #openstack-operators | 05:35 | |
| *** markvoelker has quit IRC | 05:36 | |
| *** esker has joined #openstack-operators | 05:40 | |
| *** esker has quit IRC | 05:41 | |
| *** SimonChung1 has joined #openstack-operators | 05:43 | |
| *** SimonChung has quit IRC | 05:45 | |
| *** blair has joined #openstack-operators | 06:04 | |
| *** blair has quit IRC | 06:19 | |
| *** markvoelker has joined #openstack-operators | 06:32 | |
| *** markvoelker has quit IRC | 06:38 | |
| *** david-lyle_afk has joined #openstack-operators | 07:33 | |
| *** markvoelker has joined #openstack-operators | 07:34 | |
| *** markvoelker has quit IRC | 07:39 | |
| *** Marga_ has quit IRC | 07:40 | |
| *** Marga_ has joined #openstack-operators | 07:50 | |
| *** Marga_ has quit IRC | 08:11 | |
| *** VW__ has quit IRC | 08:17 | |
| *** blair has joined #openstack-operators | 08:23 | |
| *** matrohon has joined #openstack-operators | 08:28 | |
| *** markvoelker has joined #openstack-operators | 08:35 | |
| *** Miouge has joined #openstack-operators | 08:38 | |
| *** markvoelker has quit IRC | 08:39 | |
| *** yapeng_ has joined #openstack-operators | 09:05 | |
| *** racedo_ has joined #openstack-operators | 09:05 | |
| *** turnerg has joined #openstack-operators | 09:06 | |
| *** openstack has joined #openstack-operators | 15:27 | |
| mgagne | Anyone using anvil to package openstack? Are built rpm packages "compatible" with distro packages? (same names, init scripts, etc.) | 15:41 |
| *** michchap has quit IRC | 15:47 | |
| *** michchap has joined #openstack-operators | 15:49 | |
| *** mdorman has joined #openstack-operators | 15:54 | |
| *** esker has quit IRC | 15:55 | |
| *** andrewbogott_afk is now known as andrewbogott | 15:56 | |
| *** matrohon has quit IRC | 15:59 | |
| *** hj-hp has joined #openstack-operators | 16:00 | |
| *** andrewbogott is now known as andrewbogott_afk | 16:08 | |
| *** j05h has joined #openstack-operators | 16:24 | |
| *** blairo has quit IRC | 16:27 | |
| *** j05h has quit IRC | 16:27 | |
| *** blair has joined #openstack-operators | 16:27 | |
| *** SimonChung1 has quit IRC | 16:34 | |
| *** VW_ has quit IRC | 16:34 | |
| *** andrewbogott_afk is now known as andrewbogott | 16:36 | |
| *** VW_ has joined #openstack-operators | 16:37 | |
| *** j05h has joined #openstack-operators | 16:41 | |
| *** sbfox has joined #openstack-operators | 16:54 | |
| *** alop has joined #openstack-operators | 17:02 | |
| klindgren_ | mgagne, we are | 17:03 |
| klindgren_ | for the most part the spec files are based upon rdo spec files | 17:03 |
| mgagne | klindgren_: my question is related to the one I sent to the ML | 17:03 |
| klindgren_ | so we try to be compliant with distro packages | 17:03 |
| dvorak | mgagne: the big downside from my standpoint of the OS compatable packages are they they don't allow multiple versions to be easily installed, and everything has to agree on a shared set of dependencies. | 17:03 |
| klindgren_ | if nothing else so that that puppet modules work | 17:04 |
| klindgren_ | which we use anvil + upstream puppet modules without any issues | 17:04 |
| mgagne | klindgren_: oh, right | 17:04 |
| mgagne | klindgren_: cool, thanks! | 17:04 |
| mgagne | dvorak: I understand | 17:05 |
| mgagne | dvorak: we haven't think about multiple versions yet | 17:05 |
| dvorak | mgagne: so on my control nodes, if I upgrade cinder, it upgrades everything, which means I have to take a longer single outage instead of several pretty short outages | 17:05 |
| klindgren_ | though - you will have some issues openstack and redhat dependencies conflicting | 17:05 |
| klindgren_ | that need to be worked around | 17:06 |
| dvorak | That's why we're going down the path of virtualenvs | 17:06 |
| dvorak | we can switch versions by moving a symlink and preinstall the new versions | 17:06 |
| mgagne | dvorak: everything? you mean cinder-api, cinder-scheduler, etc. ? | 17:06 |
| dvorak | mgagne: no, I mean that if I upgrade cinder from icehouse to juno, it upgrades every service, because of the dependencies. | 17:06 |
| mgagne | dvorak: unless we move away from Puppet, that's gonna be challenging | 17:07 |
| dvorak | so neutron, nova, etc | 17:07 |
| dvorak | mgagne: I've got a patch up for designate that adds virtualenv support there | 17:07 |
| dvorak | most of the code could be refactored to go into openstacklib | 17:07 |
| dvorak | sorry, puppet-designate - https://review.openstack.org/#/c/157032/ | 17:08 |
| dvorak | we're still working on adding test coverage for the new code, but the puppet code is pretty close to final | 17:09 |
| mgagne | dvorak: IMO, the change looks too intrusive or opinionated. Furthermore, people might just opt for other installation methods like fpm anyway | 17:10 |
| mgagne | dvorak: I feel Puppet might not be the right tool to install/configure a service with venv | 17:11 |
| dvorak | Any change to those modules that supports anything but upstream packages is going to be fairly intrusive | 17:11 |
| dvorak | I don't think this is a puppet vs anything else issue, I think it's more that the existing modules are already heavily opinionated towards upstream packages. | 17:12 |
| mgagne | well, I think it deserves more design and engineering to make it flexible enough to allow other installation methods | 17:12 |
| mgagne | dvorak: yes they are and adding a switch case doesn't feel like the right way | 17:12 |
| mgagne | dvorak: but I could be wrong | 17:12 |
| dvorak | I'd love to have feedback on the patch. It's mostly intended as a starting point for talking about this issue. | 17:13 |
| dvorak | designate is a good place to experiment with this sort of thing, since it's lacking packaging in general | 17:13 |
| *** hj-hp has quit IRC | 17:15 | |
| *** SimonChung has joined #openstack-operators | 17:21 | |
| *** hj-hp has joined #openstack-operators | 17:22 | |
| *** matrohon has joined #openstack-operators | 17:24 | |
| klindgren_ | does nayone here have a script that they use to fixup quotas between cells? | 17:44 |
| *** Marga_ has joined #openstack-operators | 17:46 | |
| mgagne | klindgren_: aren't quota supposed to only enforced in the API cell? | 17:47 |
| klindgren_ | yea - but the issue we have is vm's that are deleted are still counting against quota | 17:47 |
| mgagne | klindgren_: or you mean actual usage vs quota usage in API cell? | 17:48 |
| mdorman | yeah quotas are disabled in the compute cells. i think what he means is reconcile quota usages with reality | 17:48 |
| mgagne | klindgren_: oh =( | 17:48 |
| klindgren_ | I know I heard more than one person talking about quotas getting out of sync and having to script it to put quota inline. I know Joe for sure talked about it. | 17:48 |
| *** andrewbogott has left #openstack-operators | 17:51 | |
| *** hj-hp has quit IRC | 18:02 | |
| *** hj-hp has joined #openstack-operators | 18:04 | |
| *** logan2 has quit IRC | 18:18 | |
| *** harlowja_away is now known as harlowja_ | 18:19 | |
| *** dboik has joined #openstack-operators | 18:25 | |
| *** SimonChung1 has joined #openstack-operators | 18:29 | |
| *** SimonChung has quit IRC | 18:29 | |
| *** SimonChung has joined #openstack-operators | 18:30 | |
| *** SimonChung1 has quit IRC | 18:30 | |
| *** SimonChung has quit IRC | 18:31 | |
| *** SimonChung has joined #openstack-operators | 18:32 | |
| *** VW_ has quit IRC | 18:42 | |
| *** VW_ has joined #openstack-operators | 18:59 | |
| *** turnerg has quit IRC | 19:00 | |
| *** sbfox has quit IRC | 19:08 | |
| *** sbfox has joined #openstack-operators | 19:10 | |
| *** hj-hp has quit IRC | 19:10 | |
| *** SimonChung1 has joined #openstack-operators | 19:12 | |
| *** SimonChung has quit IRC | 19:12 | |
| *** SimonChung has joined #openstack-operators | 19:12 | |
| *** SimonChung1 has quit IRC | 19:12 | |
| *** SimonChung has quit IRC | 19:29 | |
| *** SimonChung1 has joined #openstack-operators | 19:29 | |
| mgagne | I think I'm bad at math: __init__() takes at least 4 arguments (8 given) | 19:30 |
| *** matrohon has quit IRC | 19:46 | |
| *** turnerg has joined #openstack-operators | 19:49 | |
| *** Rockyg has joined #openstack-operators | 19:51 | |
| *** reed has joined #openstack-operators | 19:56 | |
| *** harlowja_ has quit IRC | 19:57 | |
| *** devlaps has joined #openstack-operators | 19:58 | |
| *** harlowja has joined #openstack-operators | 20:01 | |
| *** SimonChung1 has quit IRC | 20:03 | |
| *** SimonChung has joined #openstack-operators | 20:03 | |
| *** SimonChung has quit IRC | 20:03 | |
| *** SimonChung1 has joined #openstack-operators | 20:03 | |
| *** SimonChung has joined #openstack-operators | 20:04 | |
| *** SimonChung1 has quit IRC | 20:04 | |
| *** SimonChung1 has joined #openstack-operators | 20:05 | |
| *** SimonChung has quit IRC | 20:05 | |
| *** logan2 has joined #openstack-operators | 20:29 | |
| *** Piet has joined #openstack-operators | 20:39 | |
| *** esker has joined #openstack-operators | 20:44 | |
| *** dmsimard is now known as dmsimard_away | 20:53 | |
| *** Marga_ has quit IRC | 20:57 | |
| *** Marga_ has joined #openstack-operators | 20:57 | |
| *** turnerg has quit IRC | 20:58 | |
| *** turnerg has joined #openstack-operators | 21:01 | |
| *** VW_ has quit IRC | 21:03 | |
| *** VW__ has joined #openstack-operators | 21:03 | |
| *** VW__ has quit IRC | 21:03 | |
| *** VW_ has joined #openstack-operators | 21:03 | |
| *** sandywalsh has quit IRC | 21:23 | |
| *** Marga_ has quit IRC | 21:23 | |
| *** Marga_ has joined #openstack-operators | 21:23 | |
| *** sandywalsh has joined #openstack-operators | 21:24 | |
| *** Rockyg has quit IRC | 21:26 | |
| *** Marga_ has quit IRC | 21:34 | |
| *** Marga_ has joined #openstack-operators | 21:34 | |
| *** Marga_ has quit IRC | 21:38 | |
| *** Marga_ has joined #openstack-operators | 21:39 | |
| *** SimonChung1 has quit IRC | 21:51 | |
| *** turnerg has quit IRC | 21:52 | |
| *** harlowja has quit IRC | 21:58 | |
| *** belmoreira has joined #openstack-operators | 21:58 | |
| *** turnerg has joined #openstack-operators | 21:59 | |
| *** SimonChung has joined #openstack-operators | 22:06 | |
| *** SimonChung has quit IRC | 22:07 | |
| *** esker has quit IRC | 22:09 | |
| *** SimonChung has joined #openstack-operators | 22:11 | |
| *** radez is now known as radez_g0n3 | 22:15 | |
| *** blair has quit IRC | 22:20 | |
| *** pboros has quit IRC | 22:25 | |
| *** turnerg has quit IRC | 22:25 | |
| *** harlowja has joined #openstack-operators | 22:28 | |
| *** Marga_ has quit IRC | 22:31 | |
| *** Marga_ has joined #openstack-operators | 22:31 | |
| *** SimonChung1 has joined #openstack-operators | 22:39 | |
| *** SimonChung has quit IRC | 22:39 | |
| *** dboik has quit IRC | 22:46 | |
| *** SimonChung1 has quit IRC | 22:56 | |
| *** SimonChung has joined #openstack-operators | 22:57 | |
| *** Rockyg has joined #openstack-operators | 23:00 | |
| *** SimonChung has quit IRC | 23:01 | |
| *** SimonChung1 has joined #openstack-operators | 23:01 | |
| *** SimonChung1 has quit IRC | 23:02 | |
| *** SimonChung has joined #openstack-operators | 23:02 | |
| *** j05h has quit IRC | 23:02 | |
| *** j05h has joined #openstack-operators | 23:03 | |
| *** SimonChung has quit IRC | 23:03 | |
| *** SimonChung1 has joined #openstack-operators | 23:03 | |
| *** j05h has quit IRC | 23:03 | |
| *** VW_ has quit IRC | 23:34 | |
| *** signed8bit has joined #openstack-operators | 23:35 | |
| *** signed8bit has quit IRC | 23:43 | |
| *** belmoreira has quit IRC | 23:44 | |
| *** Piet has quit IRC | 23:51 | |
| *** dmsimard_away is now known as dmsimard | 23:58 | |
| *** dmsimard is now known as dmsimard_away | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!