16:05:27 <m3m0> #startmeeting openstack-freezer 05-10-2015
16:05:27 <freezerBot> Meeting started Thu Nov  5 16:05:27 2015 UTC and is due to finish in 60 minutes.  The chair is m3m0. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:05:27 <freezerBot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:05:27 <freezerBot> The meeting name has been set to 'openstack_freezer_05_10_2015'
16:05:29 <openstack> Meeting started Thu Nov  5 16:05:27 2015 UTC and is due to finish in 60 minutes.  The chair is m3m0. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:05:30 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:05:33 <openstack> The meeting name has been set to 'openstack_freezer_05_10_2015'
16:05:37 <m3m0> All: meetings notes available in real time at: https://etherpad.openstack.org/p/freezer_meetings
16:06:28 <m3m0> Hello guys, let's start the meeting
16:06:42 <m3m0> daemontool_ you're first
16:07:26 <daemontool_> ok
16:07:34 <daemontool_> https://review.openstack.org/#/c/242132/ this is important, master is broken
16:07:42 <daemontool_> afterwe removed the bin
16:07:47 <daemontool_> we didn't updated the setup.cfg
16:07:51 <m3m0> bin/freezerc?
16:07:59 <m3m0> or freezer/bin?
16:07:59 <daemontool_> the .exe
16:08:01 <daemontool_> yes
16:08:04 <daemontool_> freezer/bin
16:08:34 <m3m0> where are those .exe now?
16:08:35 <daemontool_> that neesd to go in asap
16:08:39 <daemontool_> there are no .exe
16:08:48 <daemontool_> they have been removed as per OS TC request in https://review.openstack.org/#/c/239668/
16:08:54 <m3m0> +2
16:09:07 <daemontool_> so m3m0  probably you need to review the documentation
16:09:22 <daemontool_> and improve the steps on how to use freezer on windows
16:09:31 <daemontool_> installing the deps/etc
16:09:40 <m3m0> yes, but those .exe came from cygwin
16:09:53 <daemontool_> ok so we have to add as a deps cygwin in the doc
16:10:02 <m3m0> should we put in the documentation that you need to install cygwin and extract from there?
16:10:03 <daemontool_> and explain how to make that bin available
16:10:07 <daemontool_> yes
16:10:08 <m3m0> ok, on it
16:10:09 <daemontool_> ty
16:10:24 <daemontool_> next: the freezr application to the big tent here https://review.openstack.org/#/c/239668/
16:10:36 <m3m0> wait
16:10:40 <daemontool_> ah ok
16:10:42 <daemontool_> sorry
16:10:45 <m3m0> can you please submit a patch
16:10:51 <daemontool_> where?
16:10:59 <m3m0> raising an error in windows
16:11:04 <m3m0> right now is broken
16:11:31 <daemontool_> ok we can do that together after the meeting
16:11:33 <m3m0> while we update the readme with the correct documentation?
16:11:35 <m3m0> yes
16:11:40 <daemontool_> ok
16:11:53 <m3m0> and we need to define all the windows steps as well
16:11:58 <daemontool_> yes please
16:12:10 <m3m0> because there is no "easy" way to deploy
16:12:19 <daemontool_> I understand that....
16:12:21 <m3m0> I'll work on that after the uo
16:12:27 <daemontool_> perfect
16:12:29 <m3m0> ui
16:12:50 <daemontool_> next: big tent application, I think we are doing well after a sloppy start
16:12:54 <m3m0> should we start using heat to deploy freezer? (windows, linux)?
16:13:09 <m3m0> cool, congratulations to everyone in this one :)
16:13:11 <daemontool_> I don't know we need to evaluate that
16:13:15 <m3m0> we need to celebrate this
16:13:26 <daemontool_> yes after we'll be accepted we will :)
16:14:15 <m3m0> do we need to do something related to the big tent
16:14:16 <m3m0> ?
16:14:18 <daemontool_> so we are on track with that, basically we are doing all the changes commented from the tc
16:14:29 <daemontool_> not for now, we just need a bit of patience
16:14:50 <daemontool_> next: we need to branch our repo and release tag
16:14:52 <m3m0> perfect
16:14:54 <daemontool_> as stable/kilo
16:15:12 <m3m0> for freezer, api and ui? or which repos will be involved?
16:15:27 <daemontool_> we can do that righ after windows is fixed, the web ui client_id are in and the api related changes
16:15:31 <daemontool_> yes the 3 repos
16:15:57 <daemontool_> we need to be consistent, thanks to Jokke_  today he gave ma a crush course on release mgmt in openstack
16:16:05 <daemontool_> that was really helpful
16:16:31 <daemontool_> so there's one change in the acl file
16:16:57 <daemontool> https://review.openstack.org/#/c/242025/
16:17:14 <daemontool_> after that change, web ui, windows and api we can branch stable/kilo
16:17:21 <daemontool_> and release a new feature version
16:17:48 <daemontool_> after that we should change the requirements in master to be in compliance with liberty
16:17:49 <m3m0> as soon as we have stable/kilo we should release a new version on pypi
16:17:53 <daemontool_> yes
16:18:02 <daemontool_> exactly that's part of the release process +1
16:18:06 <daemontool_> (thumbsup)
16:18:16 <m3m0> perfecto
16:18:26 <openstackgerrit> Merged openstack/freezer: Fix bug on setup.cfg after bin removal  https://review.openstack.org/242132
16:18:54 <daemontool_> I'm also creating a new project repo called openstack/python-freezerclient
16:19:10 <daemontool_> it will be empty at the beginning, we'll work on it after stable/kilo branching
16:19:15 <daemontool_> :)
16:19:21 <daemontool_> also we need to change the meeting time
16:19:32 <m3m0> to what time?
16:19:42 <m3m0> and what about openstack/freezer-bin
16:19:44 <m3m0> ?
16:20:14 <daemontool_> https://review.openstack.org/#/c/241039/
16:20:27 <daemontool_> m3m0,  freezer-bin it will not be create most probably
16:20:57 <m3m0> ok, then we need to find a way to ship those binaries in windows
16:21:06 <m3m0> the less steps required by the user, the better
16:21:10 <daemontool_> for now, let's provide the information in the doc on how to have freezer working on windows, then we'll find a better solution also in line with openstack tc
16:21:19 <daemontool_> but now we have too much things boiling
16:21:37 <daemontool_> I agree, but we don't have to host that bins in openstack, we can't
16:21:58 <m3m0> ok, let's have an offline discussion about this
16:22:04 <daemontool_> ok
16:22:08 <daemontool_> https://review.openstack.org/#/c/241039/
16:22:24 <daemontool_> meeting needs to be moved on a differnt time and use the openstack chans dedicated for that
16:22:36 <daemontool_> I was thinking about 5:00 pm GMT
16:22:40 <m3m0> are you familiar with those channels?
16:22:43 <daemontool_> yes
16:22:56 <daemontool_> like #openstack-meetings
16:22:59 <Jokke_> daemontool_: & all: https://wiki.openstack.org/wiki/StableBranch is really important to understand once we get into the stable branch model in Freezer
16:23:10 <m3m0> is 3pm  GMT better?
16:23:13 <daemontool_> yes Jokke_  thanks a lot for that
16:23:45 <daemontool_> yes 3pm is a bit early if people is in US west coast (i.e. california, seattle)
16:23:55 <daemontool_> Jokke_,  that is really important
16:24:08 <daemontool_> or we acn ask for a new room :)
16:24:40 <m3m0> then 5PM GTM should be ok
16:25:00 <daemontool_> ok
16:25:07 <daemontool_> it's all from me
16:25:12 <m3m0> ok, thanks a lot
16:25:31 <m3m0> reldan, you're next
16:26:03 <reldan> Thank you, m3m0
16:26:33 <reldan> Fixes for cinder are merged now. But I am not sure that it will work stable.
16:26:49 <m3m0> why not?
16:27:27 <reldan> Sometimes snapshot just go to error state. Sometimes during upload glance returns 500 error
16:28:00 <daemontool_> reldan, did you see the comments from Jokke_  about the error state?
16:28:41 <reldan> Yes, have fixed. The comments where about image state, but I offen have problem with snapshot state
16:29:46 <reldan> After our discussion, I have started work on config for parallel backup. I still cannot say how long will it take
16:30:06 <m3m0> but do you have any idea why is happening on cinder?
16:30:17 <m3m0> do you need help with that?
16:30:43 <openstackgerrit> Memo Garcia proposed openstack/freezer-web-ui: Refactoring api.py  https://review.openstack.org/236175
16:30:47 <Jokke_> reldan: do you have any data from those 500 in glance? I'd be interested to have a look on them
16:32:43 <reldan> Nope, It returned 500 on image-upload on the last portion of data it returns 500
16:32:56 <m3m0> so parallel wont be included in stable/kilo right?
16:33:10 <Jokke_> reldan: next time you see that, grab the logs for me
16:33:14 <Jokke_> from glance
16:33:27 <Jokke_> I'd like to see what causes is
16:33:29 <Jokke_> it even
16:34:37 <reldan> Ok, deal! But as far as I know config for parallel backup is number one priority now. Probably I should write test, but It will be very slow test. As far as I know I cannot create volume with less than 1GB size
16:34:53 <reldan> m3m0: I’m working on parallel bakup now.
16:36:05 <m3m0> ok, do you know who can help you with the tests?
16:36:37 <reldan> As far as I know they are not in our priority list. My priority is config for parallel backups.
16:36:47 <Jokke_> 1 gig volume backup does not sound like gateable test
16:37:15 <Jokke_> perhaps good to have out of tree to verify for now, but I wouldn't include it ...
16:37:31 <m3m0> Jokke_ any alternative?
16:38:08 <Jokke_> m3m0: tons of mocking and unit tests
16:38:10 <reldan> For integration tests I would prefer to have a couple of instances of freezer continiously backuping and restoring something
16:38:31 <reldan> 1) It would be great for statistics and performance
16:38:56 <reldan> 2) Some errors are from OS API and it can be discovered in on in 10 runs
16:39:17 <reldan> 3) It allows us to have information about windows backuping
16:39:35 <Jokke_> at least have a chat with infra before introducing anything like that ... they might just be nice and kill all freezer testing if you enable something like that out of bushes
16:39:35 <m3m0> where would you like to have this running?
16:39:50 <reldan> 4) It will allow us test freezer on real data that may cause memery overflow or something like that
16:40:27 <reldan> In our lab, I know that we have a lot of machines but they are still useless to us
16:40:56 <m3m0> I agree with your points, and lets make sure we share this data with the community
16:41:06 <m3m0> we can create a spec for performance
16:42:17 <reldan> We just need give this task any priority
16:42:31 <reldan> Because now I suppose we have no such task in our backlog
16:43:12 <m3m0> you're right, we don't have that at the moment, are you planning to work on this at the same time as the parallel backups?
16:43:25 <m3m0> or better, let's create a blueprint first
16:43:34 <m3m0> to define the architecture of the testing environment
16:43:42 <reldan> Nope, I currently work on parallel backup. I actually even don’t have experience work with our lab
16:43:47 <m3m0> and from them we can have more people involved in this
16:46:13 <reldan> Let’s create, you are right
16:46:40 <daemontool_> +1
16:46:49 <daemontool_> brilliant conversation
16:47:43 <m3m0> do you anything more to say reldan?
16:48:07 <reldan> So I will create the blueprint and going to show parallel backups config next week. It is all. Thank you m3m0
16:48:44 <m3m0> thanks a lot reldan and nice job, and please don't hesitate to reach anyone if you need help
16:48:50 <reldan> Thank you!
16:50:07 <openstackgerrit> Memo Garcia proposed openstack/freezer-web-ui: Refactoring api.py  https://review.openstack.org/236175
16:50:58 <vannif> so
16:51:39 <daemontool_> reldan. put it in an rst file also, in specs, soon well create a freezer-specs repo for that, and we can have use also wiki.openstack.org/wiki/Freezer
16:51:48 <daemontool_> if that's ok for you
16:51:57 <reldan> Ok!
16:52:01 <vannif> sorry, I've been disconnected
16:52:24 <daemontool_> +1 here Team, power of will : ) https://review.openstack.org/#/c/239668/
16:54:41 <vannif> 'm working on the support for editable actions by the web ui. That means that the api will extract actions from the jobs submitted by the user and create any new action accordingly, so that they will be available for use by the web ui
16:54:52 <vannif> Should be ready today
16:55:13 <vannif> I just like to have new features tested and with proper coverage
16:55:27 <vannif> Immediately after that, I switch back to integrations gates
16:55:34 <daemontool_> vannif,  can we create an integratino tests that check if python setup.py install fails?
16:55:43 <m3m0__> is the api going to change in order to support this?
16:55:46 <daemontool_> like what happened when we removed the bin dir?
16:55:48 <m3m0__> or is an internal change?
16:55:56 <daemontool_> sorry to interrupt
16:56:56 <vannif> m3m0: no. simply the creation of new jobs will result in actions being created in the relative db type
16:57:24 <m3m0__> so, all actions will be stored in the db as a result
16:57:28 <m3m0__> ?
16:57:41 <m3m0__> and any edit of an action will result in a new action as we agreed?
16:57:44 <vannif> yes. with valid action_id
16:58:01 <m3m0__> that sounds greate
16:58:13 <m3m0__> there is something that we need to discuss
16:58:19 <m3m0__> the backup_id
16:58:31 <m3m0__> I will set an offline meeting for that
16:58:39 <m3m0__> is really important in the ui
16:58:40 <vannif> ok
16:59:11 <m3m0__> regarding to the integration tests, are we using tempest?
17:02:09 <vannif> daemontool: not sure. the devstack plugin scripts install freezer components as devstack ... well .. plugins ^^
17:02:23 <vannif> that should fail if the setup fails
17:02:31 <vannif> s/should/will
17:04:04 <vannif> m3m0: the first tests will be simple, but yes I think we'll be using tempest
17:04:27 <m3m0__> ok, that sounds great
17:04:34 <m3m0__> do you have anything more to say?
17:04:58 <vannif> no. thank you
17:05:08 <reldan> https://blueprints.launchpad.net/freezer/+spec/advanced-testing
17:05:17 <reldan> First version
17:05:29 <m3m0__> thanks a lot vannif
17:05:38 <m3m0__> does anyone have anything more to say?
17:07:18 <m3m0__> thanks a lot guys
17:07:21 <m3m0__> #endmeeting