| -openstackstatus- NOTICE: Zuul was restarted due to an error; events (such as approvals or new patchsets) since 23:01 UTC have been lost and affected changes will need to be rechecked | 00:04 | |
| *** mrodden has quit IRC | 00:21 | |
| *** britthouser has quit IRC | 01:30 | |
| *** britthouser has joined #openstack-ansible | 01:39 | |
| *** alejandrito has quit IRC | 01:55 | |
| *** britthouser has quit IRC | 02:52 | |
| *** fawadkhaliq has joined #openstack-ansible | 03:22 | |
| *** dabernie_ has joined #openstack-ansible | 03:23 | |
| *** dabernie has quit IRC | 03:26 | |
| *** sdake_ has joined #openstack-ansible | 03:35 | |
| *** sdake has quit IRC | 03:39 | |
| *** sdake_ is now known as sdake | 03:46 | |
| odyssey4me | o/ | 05:17 |
|---|---|---|
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Uses tox for automating documentation builds https://review.openstack.org/207577 | 05:32 |
| *** fawadkhaliq has quit IRC | 05:38 | |
| *** pellaeon has quit IRC | 05:38 | |
| odyssey4me | mattt andymccr hughsaunders as soon as you're in, please review https://review.openstack.org/207576 and https://review.openstack.org/207577, both of which are crucial to adding a docs check which will allow our docs to merge | 05:50 |
| *** fawadkhaliq has joined #openstack-ansible | 05:59 | |
| *** javeriak has joined #openstack-ansible | 06:13 | |
| *** sdake_ has joined #openstack-ansible | 06:35 | |
| *** sdake has quit IRC | 06:39 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Updated master for new dev work - Liberty-2 https://review.openstack.org/199126 | 06:46 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Updated master for new dev work - Liberty-2 https://review.openstack.org/199126 | 06:52 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Update kilo for new dev work - 10 Aug 2015 https://review.openstack.org/211265 | 07:07 |
| *** fawadkhaliq has quit IRC | 07:12 | |
| *** fawadk has joined #openstack-ansible | 07:13 | |
| *** javeriak has quit IRC | 07:19 | |
| *** javeriak has joined #openstack-ansible | 07:20 | |
| *** javeriak has quit IRC | 07:32 | |
| *** javeriak_ has joined #openstack-ansible | 07:40 | |
| *** javeriak_ has quit IRC | 07:46 | |
| *** javeriak has joined #openstack-ansible | 07:46 | |
| *** javeriak_ has joined #openstack-ansible | 07:50 | |
| *** javeriak has quit IRC | 07:50 | |
| *** javeriak_ has quit IRC | 07:54 | |
| *** javeriak has joined #openstack-ansible | 07:54 | |
| mattt | odyssey4me: is teh former the correct review? | 07:56 |
| odyssey4me | mattt that one's already fine - it needs the other one to merge before it can go through | 07:57 |
| openstackgerrit | Christian Berendt proposed stackforge/os-ansible-deployment: Run bash through /usr/bin/env https://review.openstack.org/211885 | 07:57 |
| odyssey4me | sorry, I didn't notice that it was already two +2's | 07:57 |
| odyssey4me | https://review.openstack.org/207577 is the one you want | 07:58 |
| odyssey4me | and we'll have to backport that to both kilo and juno, although juno will need some adjustment | 07:58 |
| odyssey4me | I'll handle the backports, but I'll need support to get them merged quickly as there's a new -infra gate job which will require those to be in place in order to pass gate checks | 07:58 |
| mattt | odyssey4me: there is an HP reference in https://review.openstack.org/#/c/207577/4/setup.py, is that correct? | 08:00 |
| odyssey4me | mattt yep, that matches where it came from | 08:00 |
| mattt | ok done | 08:02 |
| *** sdake_ has quit IRC | 08:04 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Update kilo for new dev work - 10 Aug 2015 https://review.openstack.org/211265 | 08:07 |
| odyssey4me | thanks mattt | 08:07 |
| *** javeriak_ has joined #openstack-ansible | 08:33 | |
| *** javeriak has quit IRC | 08:35 | |
| *** javeriak has joined #openstack-ansible | 08:41 | |
| *** javeriak has quit IRC | 08:43 | |
| *** javeriak_ has quit IRC | 08:44 | |
| openstackgerrit | Christian Berendt proposed stackforge/os-ansible-deployment: Create /openstack/log/instance-info prior to using it https://review.openstack.org/211906 | 08:49 |
| odyssey4me | mattt if you could please review https://review.openstack.org/209047 that'd be awesome so we can update master to liberty-2 in the follow on patch | 08:53 |
| *** metral is now known as metral_zzz | 08:55 | |
| *** metral_zzz is now known as metral | 09:04 | |
| *** javeriak has joined #openstack-ansible | 09:13 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Add iptables rule to neutron agents containers https://review.openstack.org/211918 | 09:14 |
| mattt | odyssey4me: you still there ? | 09:18 |
| mattt | odyssey4me: in https://review.openstack.org/#/c/209047/, what specifically is 'overcoming import race conditions'? | 09:19 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Add iptables rule to neutron agents containers https://review.openstack.org/211918 | 09:27 |
| *** javeriak_ has joined #openstack-ansible | 09:27 | |
| *** javeriak has quit IRC | 09:28 | |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Add iptables rule to neutron agents containers https://review.openstack.org/211918 | 09:35 |
| *** javeriak_ has quit IRC | 09:52 | |
| *** javeriak has joined #openstack-ansible | 09:53 | |
| *** javeriak has quit IRC | 09:59 | |
| *** javeriak has joined #openstack-ansible | 10:02 | |
| *** javeriak has quit IRC | 10:06 | |
| *** fawadk has quit IRC | 10:41 | |
| *** javeriak has joined #openstack-ansible | 10:42 | |
| *** javeriak has quit IRC | 10:55 | |
| evrardjp | good afternoon guys | 11:40 |
| evrardjp | is the rackspace ELK stack for kilo already released somewhere? It's still mentionned in OSAD docs | 11:48 |
| evrardjp | rsyslog is really a pain for debugging logs | 11:49 |
| evrardjp | ;) | 11:49 |
| *** britthouser has joined #openstack-ansible | 11:53 | |
| *** britthou_ has joined #openstack-ansible | 11:54 | |
| *** britthouser has quit IRC | 11:57 | |
| odyssey4me | evrardjp there are only RC's at this stage - https://github.com/rcbops/rpc-openstack/tree/r11.0.0rc13 is the latest | 12:01 |
| odyssey4me | mattt without that patch, oslo.middleware pukes occasionally when apache loads the keystone wsgi - which causes 500 errors in keystone | 12:03 |
| odyssey4me | that patch isolates the threads properly, resulting in no race conditions between the threads when trying to load their deps | 12:03 |
| odyssey4me | admin & service are also appropriately segregated into their own application groups | 12:04 |
| odyssey4me | I didn't get much out of the keystone guys about the details of the race conditions, but that patch was based on advice from the keystone cores | 12:04 |
| mattt | odyssey4me: ok cool, i'm pretty much g2g then ... but what about hugh's comment about stuff breaking if setup isn't run? | 12:05 |
| mattt | hughsaunders: what did you mean there? | 12:05 |
| odyssey4me | mattt fyi: https://github.com/openstack/keystone/blob/master/httpd/wsgi-keystone.conf | 12:05 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Add iptables rule to neutron agents containers https://review.openstack.org/211918 | 12:05 |
| hughsaunders | mattt: where? | 12:05 |
| mattt | hughsaunders: https://review.openstack.org/#/c/209047/ | 12:06 |
| odyssey4me | mattt hughsaunders I think he means the entries in the defaults | 12:07 |
| odyssey4me | we use the ansible_processor_vcpus but default them in case they don't exist | 12:07 |
| mattt | odyssey4me: yeah the dynamic values, but what is the implication of doing that? | 12:07 |
| odyssey4me | well, those were the values we previously had - so I've maintained the same pattern, but just put them into variables instead of the template | 12:08 |
| hughsaunders | mattt: keystone_wsgi_threads: "{{ ansible_processor_vcpus | default(2) // 2 }}" | 12:09 |
| odyssey4me | I agree with hugh that it probably doesn't make sense to have them there as those facts will always exist if someone runs ansible in a normal way, but hughsaunders mentioned that it is slightly possible that someone may run it without the gathering facts stage, in which case those defaults will take effect | 12:09 |
| odyssey4me | it's an edge case | 12:09 |
| hughsaunders | what odyssey4me said | 12:09 |
| odyssey4me | I do think that we should probably remove it at some stage, but I didn't want to change the values in this patch - I just wanted to have variables to override and adjust the config so that keystone works | 12:10 |
| openstackgerrit | Christian Berendt proposed stackforge/os-ansible-deployment: Adds a pep8 target to tox.ini https://review.openstack.org/207662 | 12:17 |
| mattt | i see, so preferably let ansible fail if the facts don't exist | 12:17 |
| mattt | ok, since you're not changing behaviour that seems ok to me | 12:17 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: [WIP] Add iptables rule to neutron agents containers https://review.openstack.org/211918 | 12:18 |
| odyssey4me | thanks mattt | 12:19 |
| *** woodard has joined #openstack-ansible | 12:20 | |
| mattt | np | 12:21 |
| openstackgerrit | Christian Berendt proposed stackforge/os-ansible-deployment: Removes trailing whitespace for bashate https://review.openstack.org/207663 | 12:25 |
| openstackgerrit | Christian Berendt proposed stackforge/os-ansible-deployment: Removes tab indents for bashate https://review.openstack.org/207664 | 12:25 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Uses tox for automating documentation builds https://review.openstack.org/207577 | 12:30 |
| *** bapalm has joined #openstack-ansible | 12:31 | |
| *** shausy has joined #openstack-ansible | 12:32 | |
| openstackgerrit | Christian Berendt proposed stackforge/os-ansible-deployment: Make the tox docs environment workable https://review.openstack.org/212002 | 12:37 |
| *** KLevenstein has joined #openstack-ansible | 12:49 | |
| *** tlian has joined #openstack-ansible | 12:50 | |
| *** wmlynch has joined #openstack-ansible | 12:53 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Uses tox for automating documentation builds https://review.openstack.org/212011 | 12:58 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Uses tox for automating documentation builds https://review.openstack.org/212011 | 12:58 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Uses tox for automating documentation builds https://review.openstack.org/212015 | 13:02 |
| odyssey4me | hughsaunders andymccr mattt palendae sigmavirus24_awa please review these backports asap as they are required to fix the doc gating: https://review.openstack.org/212011 https://review.openstack.org/212015 | 13:04 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Move ansible logging to gate-check script only https://review.openstack.org/211730 | 13:05 |
| *** wmlynch has quit IRC | 13:10 | |
| *** britthou_ has quit IRC | 13:14 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Improve Keystone Apache configuration https://review.openstack.org/209047 | 13:14 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Improve Keystone Apache configuration https://review.openstack.org/212021 | 13:18 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Add ebtables to neutron agent configuration https://review.openstack.org/210593 | 13:21 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Add ebtables to neutron agent configuration https://review.openstack.org/210593 | 13:21 |
| *** vdo_ has joined #openstack-ansible | 13:21 | |
| mattt | odyssey4me: i think i've heard you say that 5x in the past week :P | 13:31 |
| odyssey4me | mattt this time it's for real-real | 13:31 |
| mattt | done | 13:31 |
| odyssey4me | last time it was a juno patch, but -infra wanted us to use the standard docs tests instead of our home grown frankenscripts | 13:31 |
| odyssey4me | so now we do it right | 13:32 |
| odyssey4me | thanks | 13:32 |
| openstackgerrit | Matt Thompson proposed stackforge/os-ansible-deployment: Add iptables rule to neutron agents containers https://review.openstack.org/211918 | 13:33 |
| *** jaypipes has joined #openstack-ansible | 13:42 | |
| *** sdake_ has joined #openstack-ansible | 14:00 | |
| *** spotz_zzz is now known as spotz | 14:00 | |
| *** javeriak has joined #openstack-ansible | 14:01 | |
| odyssey4me | andymccr mattt you missed this one :p https://review.openstack.org/212015 | 14:01 |
| odyssey4me | hughsaunders palendae sigmavirus24_awa ^ (backport) | 14:01 |
| dstanek | odyssey4me: i just commented on the tox kilo backport - the change by itself is broken | 14:03 |
| odyssey4me | dstanek how so? | 14:04 |
| odyssey4me | ah dstanek so that's ok - we'll get that other patch backported | 14:04 |
| dstanek | k, that change should be an easy merge and backport | 14:05 |
| odyssey4me | the issue here is that no docs patches are able to merge because we have an infra-job that skips the commit/lint jobs for pure doc patches | 14:05 |
| *** javeriak has quit IRC | 14:05 | |
| odyssey4me | -infra did some fixups a little while ago which closed a loophole we happened to be using | 14:05 |
| odyssey4me | they fixed a bug, and since then our skip tells zuul to ignore our doc patches | 14:05 |
| odyssey4me | so I've submitted a patch in -infra to add a docs gate check, but for that to even run we need your patch in every branch | 14:06 |
| odyssey4me | I know that it will fail without the ceilometer fix, but that will give incentive to merge the ceilometer fix and backport it. :) | 14:06 |
| dstanek | odyssey4me: isn't that a chicken/egg problem since my patch won't run? | 14:07 |
| odyssey4me | that's why I remove the dependency | 14:07 |
| odyssey4me | so all we need to do is merge the ceilometer patch right after it | 14:07 |
| odyssey4me | that'll unblock the gate and the world will be a better place | 14:07 |
| odyssey4me | make sense? | 14:08 |
| dstanek | yeah, i'll just be happy to start seeing those reviews get out of my dashboard view | 14:09 |
| odyssey4me | :) me too | 14:09 |
| odyssey4me | dstanek this is what's needed to clear that docs queue: https://review.openstack.org/210498 | 14:12 |
| odyssey4me | I expect that we'll get the +w today in the next few hours | 14:12 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:13 | |
| *** shausy has quit IRC | 14:17 | |
| *** shausy has joined #openstack-ansible | 14:17 | |
| *** britthouser has joined #openstack-ansible | 14:21 | |
| *** yaya has joined #openstack-ansible | 14:21 | |
| dstanek | odyssey4me: nice. that looks more like what I'm used to | 14:21 |
| *** Bjoern_ has joined #openstack-ansible | 14:21 | |
| *** javeriak has joined #openstack-ansible | 14:22 | |
| *** Mudpuppy has joined #openstack-ansible | 14:24 | |
| odyssey4me | dstanek :) I'd like to see if we can use more of the existing templated jobs, perhaps initially for the bashate and pep8 linting, then later for other things | 14:26 |
| *** jmckind has joined #openstack-ansible | 14:26 | |
| *** Mudpuppy_ has joined #openstack-ansible | 14:27 | |
| *** Mudpuppy has quit IRC | 14:30 | |
| dstanek | odyssey4me: yes, that's good goal to have. no more (or at least less) snowflake! | 14:31 |
| *** cristicalin has joined #openstack-ansible | 14:31 | |
| *** Mudpuppy_ is now known as Mudpuppy | 14:32 | |
| cristicalin | hi, anybody have issues with building the mysql-galera containers with latest master branch of osad ? | 14:32 |
| cristicalin | it fails for me at pip install python-MySQL | 14:32 |
| odyssey4me | cristicalin are you aware, first of all, that master of osad tracks openstack liberty? | 14:33 |
| cristicalin | yes, I am aware of that | 14:33 |
| cristicalin | but this seems to be too basic to fail because of liberty | 14:33 |
| odyssey4me | ok, cool - no, I haven't seen that - nor is our gate check... do you have a pastebin of the install log to inspect? | 14:34 |
| cristicalin | give me a sec, it's my first days with osad | 14:34 |
| cristicalin | where is the install log ? | 14:34 |
| odyssey4me | cristicalin also, are you using setup-infrastructure, or are you using run-playbooks.sh ? | 14:34 |
| odyssey4me | cristicalin the install log is output to the console | 14:35 |
| cristicalin | odyssey4me, I did it like this ./scripts/bootstrap-ansible.sh ./scripts/bootstrap-aio.sh ; cd playbooks ; openstack-ansible setup-everything.yaml | 14:35 |
| odyssey4me | cristicalin ah, then you're using the setup-infrastructure playbook which does an rsync from an upstream repo | 14:36 |
| cristicalin | hmm, ok I have to run it again with a tee to get the log | 14:36 |
| odyssey4me | try running the playbook 'repo-build.yml' | 14:36 |
| odyssey4me | then try 'cd /opt/os-ansible-deployment; ./scripts/run-playbooks.sh' | 14:37 |
| cristicalin | ok | 14:37 |
| odyssey4me | the upstream repo for master is very likely broken at the moment - thanks for the reminder as I think I can rebuild it now | 14:37 |
| odyssey4me | ah, not yet - need ot get https://review.openstack.org/199126 merged first | 14:38 |
| *** javeriak_ has joined #openstack-ansible | 14:38 | |
| *** javeriak has quit IRC | 14:38 | |
| cristicalin | is there any way to clean the environment to try with the kilo branch ? | 14:39 |
| cristicalin | or i need to deploy another vm | 14:39 |
| odyssey4me | cristicalin there's a script in the scripts directory called 'teardown' which will do an almost good job of tearing everything down | 14:40 |
| odyssey4me | it'll also destroy your /etc/openstack_deploy directory - so back that up if you care about it | 14:40 |
| cristicalin | currently I'm trying to get an AIO to work so don't yet care about the openstack_deploy dir | 14:40 |
| cristicalin | I was hoping to try liberty and osad at the same time but ... one at a time is probably best | 14:41 |
| odyssey4me | cristicalin yeah - best would be to checkout 11.1.0 as a starting point | 14:41 |
| cristicalin | that is kilo based ? | 14:42 |
| odyssey4me | use teardown, then bootstrap-aio; bootstrap-ansible; run-playbooks | 14:42 |
| odyssey4me | cristicalin yep, that's the latest tag from the kilo branch | 14:42 |
| odyssey4me | as upstream is constantly changing in the upstream master branch, our master branch can behave differently from day to day | 14:43 |
| odyssey4me | but we track the upstream branch so that we can fix things week by week and release our liberty branch only hours after upstream releases | 14:43 |
| cristicalin | ok, I'll take that route first to understand how osad works | 14:43 |
| odyssey4me | cristicalin you may find http://osad.readthedocs.org/en/latest/ useful later too :) | 14:44 |
| cristicalin | i'm looking for a tool that can be used for the lifecycle of the cloud not just for the initial deployment and osad at the moment seems as clode as it can get to that | 14:44 |
| *** jmckind has quit IRC | 14:44 | |
| odyssey4me | we still need to do quite a bit of work on the docs, but they're helpful to understand some of the basics of how the config works | 14:44 |
| cristicalin | odyssey4me, already bookmarked that :) thanks | 14:44 |
| odyssey4me | cristicalin yep, OSAD is the basis for Rackspace Private Cloud and is deployed in many production deployments - both in RAX DC's and in customer DC's | 14:45 |
| cristicalin | odyssey4me, I got the log, are you still interested in it ? | 14:46 |
| odyssey4me | cristicalin sure | 14:46 |
| cristicalin | odyssey4me, http://pastebin.com/keYymJ5z | 14:49 |
| odyssey4me | cristicalin line 1377 shows the upstream clone | 14:51 |
| odyssey4me | a local repo build is preferred and advised, but let me fix the upstream repo - give me 30 mins or so | 14:52 |
| *** galstrom_zzz is now known as galstrom | 14:52 | |
| *** alextricity has joined #openstack-ansible | 14:53 | |
| *** yaya has quit IRC | 14:54 | |
| cristicalin | hmm, the teardown.sh leaves behind the loopback devices for swift | 14:57 |
| cristicalin | and for some reason losetup -d does not want to delete them :( | 14:58 |
| cristicalin | need to reboot the test vm | 14:58 |
| odyssey4me | cristicalin if it's a test vm and you can destroy and rebuild, that'll be better | 15:00 |
| odyssey4me | otherwise yes, the teardown script could do with a little love | 15:00 |
| odyssey4me | as I recall they can't be deleted because they're still mounted - but we haven't had much chance to look into it | 15:00 |
| odyssey4me | upstream master wheel repo has been rebuilt, fyi | 15:01 |
| odyssey4me | palendae sigmavirus24 andymccr mattt hughsaunders can we please get this review through the door? (update sha's and version for head of kilo) https://review.openstack.org/211265 | 15:02 |
| *** phalmos has joined #openstack-ansible | 15:05 | |
| *** meteorfox has quit IRC | 15:05 | |
| *** matt_____ has joined #openstack-ansible | 15:06 | |
| *** serverascode has quit IRC | 15:06 | |
| *** mattoliverau has quit IRC | 15:06 | |
| *** gus has quit IRC | 15:06 | |
| javeriak_ | hey guys, is it absolutely necessary to use actual target hostnames in openstack_user_config.yml, someone just mistakenly left them as infra1/infra2/infra3 etc, and the playbooks have completed fine, no actual testing has been done yet so I'm not sure if the deployment is in a healthy state | 15:06 |
| *** meteorfox has joined #openstack-ansible | 15:07 | |
| odyssey4me | javeriak_ no, as long as the IP's are fine then it's fine - the playbooks put hosts files everywhere with those values | 15:08 |
| *** gus has joined #openstack-ansible | 15:08 | |
| odyssey4me | it would make my ocd tendancies turn me inside out, but hey - it's your environment :p | 15:08 |
| sigmavirus24 | heh | 15:09 |
| sigmavirus24 | poor odyssey4me | 15:09 |
| *** serverascode has joined #openstack-ansible | 15:11 | |
| javeriak_ | odyssey4me, okay cool, this actually seems like a better way, instead of using the random hostnames people tend to give, it now has uniform containers prepended with infra1_ etc | 15:11 |
| *** jwagner_away is now known as jwagner | 15:17 | |
| sigmavirus24 | odyssey4me: https://mail.python.org/pipermail/cryptography-dev/2015-August/000481.html Cryptography 1.0 was released today | 15:21 |
| sigmavirus24 | We should remove the SHA pin we have for that for Keystone fernet tokens | 15:21 |
| alextricity | Hello :). How is image sync'ing done in Kilo now? rsync? | 15:22 |
| alextricity | for glance images | 15:22 |
| odyssey4me | sigmavirus24 every time I check, that crypto library still hasn't released a v1 - so I just let the sha update | 15:23 |
| odyssey4me | alextricity no change, a shared glance back-end is recommended | 15:23 |
| odyssey4me | so either NFS, or swift or cloudfiles | 15:23 |
| alextricity | interesting. I set up a 11.0.4 build a few weeks back and it looks like some kind of syncing is going on. | 15:24 |
| odyssey4me | glance store hacks are bad, mmmmk | 15:24 |
| alextricity | But maybe I did something and can't think of it right now | 15:24 |
| odyssey4me | oh sigmavirus24 I didn't see your message before that :) yes, in that case I think a follow-up patch to remove the pin to a sha is very necessary :) | 15:25 |
| odyssey4me | alextricity in an AIO build there's only one glance container - so no problem :p | 15:25 |
| javeriak_ | odyssey4me: while we're on the topic of syncing glance images, if I didnt want to setup a backend storage nodes, whats the recommended way to simply sync images to save on all three infra's | 15:25 |
| alextricity | odyssey4me: No this is a clustered build | 15:25 |
| openstackgerrit | Erik Wilson proposed stackforge/os-ansible-deployment: Archive Keystone to Keystone Federation rst content https://review.openstack.org/211747 | 15:26 |
| odyssey4me | javeriak_ there is no recommended way - the recommended way is to use a shared back-end of some sort | 15:26 |
| odyssey4me | you could do something like rsync and cron it, but you'd have no guarantee that a new image would be on all controllers in time | 15:26 |
| *** shausy has quit IRC | 15:27 | |
| odyssey4me | so you'd have to also set your load balancer for glance to use one glance container as a primary, and the others only in backup mode | 15:27 |
| odyssey4me | and if you lose your primary glance container, you'd have to make your new primary do the syncing | 15:27 |
| odyssey4me | it's a mess, which is why there's no hack in osad for it | 15:27 |
| javeriak_ | odyssey4me: why the later, if the image is present on all three, would it matter where the request landed? | 15:27 |
| odyssey4me | the right way to do it is to use nfs, swift, or ceph as a glance back-end | 15:27 |
| odyssey4me | javeriak_ rsync would still take time to sync from one node to the others - before it's finished then other node may get the request | 15:28 |
| javeriak_ | also i havnet had much time to look into any of these options, but I assume they each require setting up a separate storage node? | 15:29 |
| odyssey4me | javeriak_ we don't facilitate the setup of anything except swift - all we do at this stage is allow you to configure which nfs/ceph back-end you wish glance to use | 15:29 |
| odyssey4me | you would have to setup your nfs server or ceph cluster | 15:30 |
| alextricity | javeriak_: For the NFS option, you could use the first glance container to host the NFS export | 15:30 |
| odyssey4me | yep, that's an option - just make sure you're doing your backups :) | 15:30 |
| javeriak_ | okay yes I understand, customer deployments would of course use proper backends, for in-house testing, simply syncing and giving the images time to get to all containers seems the easiest atm :) | 15:31 |
| odyssey4me | javeriak_ sure, do whatever you like in your own lab :) | 15:35 |
| *** yaya has joined #openstack-ansible | 15:37 | |
| javeriak_ | haven't seen cloudnull around for a while? | 15:43 |
| odyssey4me | javeriak_ he's on leave in europe, enjoying a disconnected holiday | 15:44 |
| palendae | "disconnected" | 15:46 |
| palendae | He's dropped in and done a few reviews | 15:47 |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: WIP: Generic ini conf template https://review.openstack.org/212079 | 15:47 |
| javeriak_ | odyssey4me: lucky guy :) | 15:49 |
| *** bapalm has quit IRC | 15:49 | |
| *** bapalm has joined #openstack-ansible | 15:49 | |
| *** yaya has quit IRC | 15:49 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Uses tox for automating documentation builds https://review.openstack.org/212011 | 15:56 |
| *** alop has joined #openstack-ansible | 15:57 | |
| *** matt_____ is now known as mattoliverau | 16:03 | |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: WIP: Generic ini conf template https://review.openstack.org/212079 | 16:03 |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: WIP: Separate nova.conf vars from os_nova role https://review.openstack.org/212087 | 16:03 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Implement /usr/bin/env as the shebang in all bash scripts https://review.openstack.org/211885 | 16:07 |
| *** sdake_ is now known as sdake | 16:13 | |
| odyssey4me | Sam-I-Am good question on https://review.openstack.org/211747 - the docs were meant to cover both how to set things up, but also how some of the things we do automagically are done manually | 16:17 |
| odyssey4me | but I'd like to discuss how we can arrange the information in a way that's easier to consume | 16:17 |
| *** sigmavirus24 is now known as sigmavirus24_awa | 16:18 | |
| odyssey4me | I've learned that it's possible to use meetbot in this channel, so perhaps we should schedule a chat between myself, you and miguelgrinberg to discuss the content and cover any Q&A? Some time next week perhaps? | 16:18 |
| Sam-I-Am | i think the descriptions of things done automagically for end-users belong in devdocs | 16:19 |
| Sam-I-Am | the things end-users need to do go in pre- and post- steps in the install guide | 16:19 |
| palendae | Jumping in here, but maybe we should clearly delineate between 'dev docs' and 'user/deployer docs' | 16:19 |
| palendae | In the hierarchy | 16:19 |
| palendae | Right now it seems implicitly that everything not install guide is dev docs | 16:20 |
| odyssey4me | palendae yeah, good point | 16:20 |
| odyssey4me | Sam-I-Am the truth is that almost all that info is install guide material - the command line stuff is more operations guide material | 16:20 |
| odyssey4me | that could possibly go into upstream docs, actually | 16:21 |
| odyssey4me | but we could also include it in dev docs to illustrate what we're automating | 16:21 |
| odyssey4me | palendae I'm kinda of the opinion that that the in-repo docs should pretty much be dev docs... and we should have another repo for install docs which should target being added into the upstream install guide list | 16:22 |
| palendae | That's a reasonable stance | 16:23 |
| Sam-I-Am | odyssey4me: whats wrong with just the doc dir? | 16:23 |
| Sam-I-Am | i dont see a reason for a separate project | 16:23 |
| odyssey4me | it's pretty much the pattern for every other openstack project | 16:23 |
| Sam-I-Am | although, big-tentish... we could go in openstack-manuals | 16:23 |
| Sam-I-Am | if you look at fuel, their docs are in-tree | 16:23 |
| palendae | Then again, yeah | 16:23 |
| odyssey4me | different dir for now, but not a different project - just a different repo | 16:23 |
| *** erikmwilson has joined #openstack-ansible | 16:23 | |
| Sam-I-Am | there's the openstack devdocs too | 16:24 |
| Sam-I-Am | or devref, whatever you want to call it | 16:24 |
| odyssey4me | sure, so I'm happy with this case for now and a different directory is a good place | 16:24 |
| odyssey4me | but there is going to come a time when a split makes more sense | 16:25 |
| odyssey4me | it'll give the operators and docs team more autonomy over their own patches without getting lost between code patches | 16:25 |
| Sam-I-Am | the automated build process that dstanek is working on should help | 16:25 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 16:25 | |
| odyssey4me | Sam-I-Am yep, I'm waiting for the final +2+w from -infra for that to become a gate check | 16:26 |
| Sam-I-Am | there's probably three classes of things - install, operations, and devref | 16:26 |
| odyssey4me | next step will be to learn how to use the publishing process | 16:26 |
| Sam-I-Am | none of which should duplicate upstream content | 16:26 |
| odyssey4me | Sam-I-Am agreed | 16:26 |
| KLevenstein | so should we just go ahead and polish up the files in the federation docs patch, and worry about rearranging everything into different repos later? | 16:26 |
| Sam-I-Am | so... basically ops includes things specific to osad | 16:26 |
| odyssey4me | KLevenstein of course :) | 16:26 |
| Sam-I-Am | KLevenstein: polish as in make render nicely or polish as in make some sense of the blob | 16:27 |
| Sam-I-Am | right now there's no obvious rhyme or reason to the conent | 16:27 |
| KLevenstein | Sam-I-Am: both, ideally | 16:27 |
| Sam-I-Am | content | 16:27 |
| KLevenstein | organize the information, then make it render nicely | 16:27 |
| odyssey4me | Sam-I-Am I think the OSAD user-base is probably more interested in devref and install docs at this point | 16:27 |
| *** yaya has joined #openstack-ansible | 16:27 | |
| odyssey4me | new operator entrants want the install guide, new developer entrants want the devref | 16:28 |
| erikmwilson | Why are putting out information on things that don't work yet? | 16:28 |
| odyssey4me | erikmwilson the stuff actually does work ,it's just not very pretty | 16:28 |
| odyssey4me | it's a little klunky to use for end-users | 16:28 |
| odyssey4me | so we're putting it out there in order for more operators and developers to be able to try it out and be informed | 16:29 |
| erikmwilson | Ok. What we need for the federation stuff is an overarching procedure that gives the order of all the sub procedures. | 16:29 |
| *** phalmos has quit IRC | 16:29 | |
| Sam-I-Am | ideally the install guide is going to contain a section for configuring any features (including federation) | 16:29 |
| palendae | I think also putting it down so we're not lost | 16:29 |
| odyssey4me | it's part of driving the conversation about federation and what needs to happen for real people to use it | 16:29 |
| odyssey4me | yes, there is what palendae said too - we did a ton of work, and a lot of people are interested in how to make it go... and in a month we'll probably need our own docs to remember :p | 16:30 |
| KLevenstein | okay, cool. so then our next steps are to get an overview of the process that fits all the pieces in the patch together, fix up the content accordingly, then make it render nicely as part of OSAD’s in-tree install guide | 16:32 |
| erikmwilson | Sounds good to me. | 16:33 |
| KLevenstein | erikmwilson: for RST formatting conventions etc., model it on what’s already there in the OSAD doc repo | 16:34 |
| erikmwilson | Will do | 16:34 |
| palendae | the install guide portion, ideally | 16:34 |
| palendae | Most of the stuff outside the install guide doesn't match cause I just kinda wrote stuff | 16:34 |
| KLevenstein | rightyright | 16:34 |
| palendae | We'll need to make it all match. New stuff I write does | 16:34 |
| KLevenstein | palendae: that’s what they pay us the tech writer bucks for :) | 16:35 |
| palendae | yep! | 16:35 |
| Sam-I-Am | lol | 16:35 |
| odyssey4me | and away she goes: https://review.openstack.org/210498 expect to see a docs verify and merge check in zuul's list soon :) | 16:36 |
| Sam-I-Am | so... theres really no devref except for random readme fles? | 16:36 |
| palendae | \o/ | 16:36 |
| palendae | Sam-I-Am: Everything i wrote was intended to be devref | 16:36 |
| palendae | In some areas it's kind blurred | 16:36 |
| odyssey4me | palendae expect to see many gate failures until we fix the docs :) | 16:37 |
| palendae | odyssey4me: Fine with me | 16:37 |
| palendae | Though last I knew they were building just fine | 16:37 |
| odyssey4me | palendae there's a series of patches from dstanek we need to hit and backport pronto | 16:37 |
| odyssey4me | next step I'd like to use -infra's bashate and pep8 lint checks instead of our own custom one | 16:38 |
| *** vdo_ has quit IRC | 16:38 | |
| palendae | Depends on how their pep8 one works | 16:38 |
| palendae | Since we have to locate our python files | 16:38 |
| Sam-I-Am | palendae: what im really asking is, do we want a devref directory so we can structure this stuff | 16:39 |
| Sam-I-Am | rather than random files everywhere | 16:39 |
| odyssey4me | palendae no matter, we'll work it out and perhaps patch the upstream job to be a little more flexible | 16:39 |
| palendae | Sam-I-Am: Yes, that was what I was thinking. 2 branches, devdocs and user | 16:39 |
| odyssey4me | Sam-I-Am by 'random files' you mean the example files right? | 16:39 |
| palendae | odyssey4me: I think he means the top level files in doc/ | 16:39 |
| palendae | Again, they were intended to be dev docs so they were not random at all when it was made | 16:40 |
| Sam-I-Am | odyssey4me: anything .rst | 16:40 |
| Sam-I-Am | or .md i guess | 16:40 |
| palendae | Are there .mds? | 16:40 |
| Sam-I-Am | the example config files can be in the source | 16:40 |
| Sam-I-Am | or whatever | 16:40 |
| *** persia has quit IRC | 16:42 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adds ceilometer link in configure.rst https://review.openstack.org/207576 | 16:43 |
| Sam-I-Am | so, keystone seems to do this - https://github.com/openstack/keystone/tree/master/doc/source | 16:43 |
| Sam-I-Am | anything in there is devref | 16:43 |
| Sam-I-Am | obviously the install guide, config ref, and other bits (mostly for end users) live in other repos | 16:43 |
| Sam-I-Am | so, for now we can either make a devref dir... or move the install guide elsewhere | 16:44 |
| KLevenstein | any reason we can’t have devref live here: https://github.com/stackforge/os-ansible-deployment/tree/master/doc/source | 16:46 |
| KLevenstein | and keep the install guide as a subdir? | 16:46 |
| odyssey4me | KLevenstein Sam-I-Am at this point why don't we consider both the install guide and devref docs all to be devrefs, until such time that it changes? | 16:48 |
| KLevenstein | that seems reasonable to me | 16:48 |
| odyssey4me | the devrefs are just appendices? | 16:48 |
| *** cristicalin has quit IRC | 16:48 | |
| KLevenstein | doc re-architecting isn’t something to be done off the cuff anyway | 16:48 |
| odyssey4me | ie deeper info on particular topics beyond the common use-cases covered by the install guide | 16:48 |
| KLevenstein | sounds good | 16:48 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Uses tox for automating documentation builds https://review.openstack.org/212015 | 16:49 |
| odyssey4me | \o/ | 16:49 |
| KLevenstein | sweet | 16:49 |
| Sam-I-Am | odyssey4me: i dunno, i think the install guide is pretty end-user friendly to be a devref | 16:50 |
| odyssey4me | still waiting for the -infra patch to take effect... testing with a doc patch in review every now and again | 16:50 |
| odyssey4me | Sam-I-Am yes it is, but this group will be adding more docs in the future that will not necessarily be friendly | 16:50 |
| odyssey4me | the professional docs people can then patch the docs to make them friendly | 16:51 |
| KLevenstein | this is one of those cases where I’m okay with “work with what we have, redesign when we need to (but be aware that need for redesign is coming” | 16:51 |
| odyssey4me | we can't hold the operators who want to add bits back from adding bits, even if their bits are in broken english | 16:51 |
| odyssey4me | KLevenstein agreed, we iterate with what we have - when it becomes unmanageable then we do a sprint to overhaul it | 16:52 |
| *** javeriak has joined #openstack-ansible | 16:52 | |
| *** bapalm has quit IRC | 16:52 | |
| odyssey4me | hahaha, the first merge fail due to a doc fail | 16:52 |
| odyssey4me | we have docs checks in the gate people :) | 16:53 |
| Sam-I-Am | odyssey4me: which? | 16:53 |
| odyssey4me | Sam-I-Am it's still processing - you'll have to check the zuul status board | 16:54 |
| *** javeriak_ has quit IRC | 16:54 | |
| odyssey4me | http://status.openstack.org/zuul/ | 16:54 |
| odyssey4me | filter for stackforge/os-ansible-deployment and expand by default | 16:54 |
| odyssey4me | https://review.openstack.org/207576 is the key review to unblock the master gate | 16:55 |
| odyssey4me | and it's just waiting for resources | 16:55 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Document required repository hosts config info https://review.openstack.org/202258 | 16:57 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Documentation of the static-route feature https://review.openstack.org/207849 | 16:57 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Cinder_volumes_containers are automatically tagged with is_metal https://review.openstack.org/209861 | 16:58 |
| dstanek | odyssey4me: i forgot that when i looked at your change job this morning that the templated docs job uses a tox environment that i didn't include | 17:15 |
| dstanek | odyssey4me: so i just have to add that for the job to work | 17:15 |
| *** alop has quit IRC | 17:17 | |
| odyssey4me | dstanek oh dear, but for any change to merge it needs to include https://review.openstack.org/207576 | 17:20 |
| odyssey4me | are you able to do the update quickly? | 17:20 |
| openstackgerrit | David Stanek proposed stackforge/os-ansible-deployment: Adds venv environment to tox for -infra docs job https://review.openstack.org/212112 | 17:20 |
| odyssey4me | dstanek you'll have to merge https://review.openstack.org/207576 and https://review.openstack.org/212112 otherwise the docs check will not pass | 17:21 |
| dstanek | odyssey4me: yes.... ^ that won't work.... updating the ceilometer review | 17:21 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Adds venv environment to tox for -infra docs job https://review.openstack.org/212113 | 17:22 |
| odyssey4me | dstanek fyi, as we have no docs in-repo for juno that patch as it is will work just fine for that branch, so I've backported it | 17:23 |
| openstackgerrit | David Stanek proposed stackforge/os-ansible-deployment: Adds ceilometer link in configure.rst https://review.openstack.org/207576 | 17:23 |
| dstanek | odyssey4me: ^ updated | 17:24 |
| *** alop has joined #openstack-ansible | 17:24 | |
| odyssey4me | dstanek ta, oddly I don't see it coming into the queue - I guess it hit some of the -infra nodes which aren't yet updated | 17:24 |
| *** persia has joined #openstack-ansible | 17:24 | |
| *** persia has quit IRC | 17:24 | |
| *** persia has joined #openstack-ansible | 17:24 | |
| *** britthou_ has joined #openstack-ansible | 17:27 | |
| *** britthouser has quit IRC | 17:30 | |
| odyssey4me | it appears that zuul is having a little hernia | 17:32 |
| svg | let's cross our streams on him | 17:35 |
| odyssey4me | svg eh? | 17:36 |
| svg | to kill him? | 17:36 |
| odyssey4me | svg haha, no - one must be nice to zuul | 17:36 |
| svg | but... busting makes me feel good! | 17:37 |
| odyssey4me | zuul schedules all the testing, so it's better to buy him large quantities of the water of life | 17:37 |
| svg | I never understood how picking the name of a bad guy for such a crucial process seemed like a good idea :) | 17:38 |
| odyssey4me | svg nerd humour? | 17:38 |
| svg | euhm | 17:39 |
| svg | Aren't we talking about the Ghostbuster reference? | 17:39 |
| odyssey4me | svg in my case http://status.openstack.org/zuul/ but in your case, yes | 17:39 |
| svg | Well, the point of the pun is that those are the same :) | 17:40 |
| * odyssey4me pokes zuul with a pointed stick | 17:40 | |
| svg | odyssey4me: are you always that serious? | 17:41 |
| odyssey4me | svg very serious when I'm being held back from my scotch by a patch that I got merged :p | 17:42 |
| * svg is zipping from his scotch with a view on a pool and a part of southern France | 17:42 | |
| * svg ducks | 17:43 | |
| * odyssey4me throws a pointed stick in svg's general direction | 17:43 | |
| *** yaya has quit IRC | 17:43 | |
| odyssey4me | crapola, over 4000 events in the queue :/ | 17:44 |
| *** Bjoern_ has left #openstack-ansible | 17:54 | |
| *** woodard has quit IRC | 17:55 | |
| *** phalmos has joined #openstack-ansible | 17:56 | |
| *** yaya has joined #openstack-ansible | 17:56 | |
| *** bapalm_ has joined #openstack-ansible | 18:15 | |
| *** bapalm has joined #openstack-ansible | 18:23 | |
| *** bapalm has quit IRC | 18:24 | |
| *** woodard has joined #openstack-ansible | 18:39 | |
| openstackgerrit | Marc Gariépy proposed stackforge/os-ansible-deployment: lxc.mounts udev conflict with lxc.autodev=1 https://review.openstack.org/212142 | 18:45 |
| openstackgerrit | Christopher H. Laco proposed stackforge/os-ansible-deployment: Templateize neutron_api_workers https://review.openstack.org/212146 | 18:52 |
| *** Mudpuppy has quit IRC | 18:57 | |
| *** Mudpuppy has joined #openstack-ansible | 18:57 | |
| *** Mudpuppy has quit IRC | 18:57 | |
| *** Mudpuppy has joined #openstack-ansible | 18:58 | |
| *** javeriak_ has joined #openstack-ansible | 19:02 | |
| *** javeriak has quit IRC | 19:02 | |
| *** cristicalin has joined #openstack-ansible | 19:04 | |
| *** yaya has quit IRC | 19:06 | |
| *** cristicalin has quit IRC | 19:07 | |
| dstanek | odyssey4me: yay, i looks like that one is rolling along now. i still don't know how the tox review got out of order ;-) probably not worth looking into | 19:15 |
| odyssey4me | dstanek meh, it'll unblock soon and I'll sort the other branches out tomorrow | 19:16 |
| *** jmckind_ has joined #openstack-ansible | 19:20 | |
| *** javeriak_ has quit IRC | 19:27 | |
| *** javeriak has joined #openstack-ansible | 19:40 | |
| *** woodard_ has joined #openstack-ansible | 19:48 | |
| *** woodard has quit IRC | 19:51 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Adds ceilometer link in configure.rst https://review.openstack.org/207576 | 20:10 |
| *** jmckind_ has quit IRC | 20:35 | |
| *** woodard_ has quit IRC | 20:39 | |
| *** woodard has joined #openstack-ansible | 20:40 | |
| *** jwagner is now known as jwagner_away | 20:40 | |
| *** galstrom is now known as galstrom_zzz | 20:43 | |
| *** jmckind has joined #openstack-ansible | 20:55 | |
| *** jmckind_ has joined #openstack-ansible | 21:11 | |
| *** jmckind has quit IRC | 21:12 | |
| *** tlian has quit IRC | 21:31 | |
| *** fawadkhaliq has joined #openstack-ansible | 21:43 | |
| *** jmckind_ has quit IRC | 21:44 | |
| *** fawadkhaliq has quit IRC | 21:50 | |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Implement /usr/bin/env as the shebang in all bash scripts https://review.openstack.org/211885 | 21:56 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Small readme fix https://review.openstack.org/207700 | 21:57 |
| openstackgerrit | Jesse Pretorius proposed stackforge/os-ansible-deployment: Updated master for new dev work - Liberty-2 https://review.openstack.org/199126 | 21:57 |
| *** woodard has quit IRC | 22:01 | |
| *** woodard has joined #openstack-ansible | 22:01 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Small readme fix https://review.openstack.org/207700 | 22:20 |
| *** woodard has quit IRC | 22:24 | |
| *** KLevenstein has quit IRC | 22:35 | |
| *** marekd is now known as marekd_404 | 22:36 | |
| *** spotz is now known as spotz_zzz | 22:39 | |
| *** devlaps has joined #openstack-ansible | 22:44 | |
| *** jmckind has joined #openstack-ansible | 22:45 | |
| *** Mudpuppy has quit IRC | 22:48 | |
| *** galstrom_zzz is now known as galstrom | 22:51 | |
| *** markvoelker has quit IRC | 23:14 | |
| *** phalmos has quit IRC | 23:37 | |
| *** galstrom is now known as galstrom_zzz | 23:45 | |
| *** alop has quit IRC | 23:53 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Updated master for new dev work - Liberty-2 https://review.openstack.org/199126 | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!