Tuesday, 2017-10-24

ara-slack<liyuankui> Currently we are building a system to invoke playbooks, and we want to use ara to show the execution result.  But we record various other stuff for the playbook execution, such as params, users, etc...  So we want to link our record with ara record?  I suppose some post-run hook would be helpful.06:07
*** resmo has joined #ara06:16
*** egonzalez has joined #ara08:13
egonzalezdmsimard, when you around wanted to ask you a few questions about unique hosts identifiers https://storyboard.openstack.org/#!/story/200115908:14
*** egonzalez has quit IRC10:06
*** egonzalez has joined #ara11:41
dmsimardegonzalez: hi11:54
egonzalezdmsimard, regarding https://storyboard.openstack.org/#!/story/200115913:03
egonzalezis there any work on it?13:03
dmsimardegonzalez: not right now, no.13:04
dmsimardegonzalez: I don't think it would make the cut for 1.0 either, there's still a considerable backlog to work through for it13:05
dmsimardegonzalez: it's not an easy problem, what's your use case ?13:05
egonzalezdmsimard, in our internal work want to use ara data to create usage statistics about success and stuff like that filtered by hosts, but being diff ID per execution cannot match the same data13:07
dmsimardegonzalez: so here's the problem... in the ara early days (alpha/beta), hosts were considered unique across playbooks. This allowed you to do things like look at a playbook history for a particular host https://youtu.be/k3qtgSFzAHI?list=PLyLLwe4-L1ETFVoAogQqpn6s5prGKL5Ty&t=2313:09
dmsimardegonzalez: however, ansible doesn't have a similar "handshake" analogous to what puppet agent would do with a puppetmaster for example (certificate signature), so from playbook to playbook, a server called 'webserver' is not necessarily the same machine13:10
egonzalezdmsimard, makes sense now13:11
egonzalezdmsimard, thanks a lot13:11
dmsimardegonzalez: I've discussed this with several folks at the PTG and there are perhaps some paths worth exploring, I've also discussed this with some core Ansible and Tower folks13:11
dmsimardegonzalez: I have several ideas to work this out but it's still a lot of work, there's database schema work involved, a lot of frontend, etc.13:12
dmsimard1.1 maybe :)13:12
* dmsimard needs to get 1.0 out the door13:13
dmsimardegonzalez: if you're crafty, you could probably generate your own persistent IDs and use that13:15
dmsimardegonzalez: but that's probably not much different than some awk/grep or ~raw SQL13:15
egonzalezdmsimard, many thanks, will think a way to achieve that. Thanks13:28
*** ara-slack has quit IRC13:49
*** ara-slack has joined #ara13:49
*** ara-slack has quit IRC13:49
*** ara-slack has joined #ara13:49
*** ChanServ sets mode: +v ara-slack13:49
ara-slack<dmsimard> @liyuankui hey there, ara_record can be used to attach arbitrary data to your report, yes. Have you looked at the documentation ? http://ara.readthedocs.io/en/latest/usage.html#using-the-ara-record-module13:50
*** egonzalez has quit IRC15:03
*** resmo has quit IRC15:42
ara-slack<jcapps> hmmm is there a way to make diff output show up in ara?  (I swear we used to be able to see such diffs in 0.13.3, but now cannot reproduce)19:41
ara-slack<jcapps> also... Environment=ARA_LOG_LEVEL=DEBUG does not seem to pump out much info19:50
ara-slack<dmsimard> @jcapps not to my knowledge, adding diff support to ara is on the to-do list :slightly_smiling_face:19:56
ara-slack<jcapps> oh it is?  I will look for that/vote19:58
ara-slack<dmsimard> https://storyboard.openstack.org/#!/story/200092919:59
*** jparrill has joined #ara22:44
*** jparrill has quit IRC22:45
*** weshay|ruck is now known as weshay|PTO23:38

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!