14:01:12 <m3m0> #startmeeting freezer_2016_02_18 14:01:13 <openstack> Meeting started Thu Feb 18 14:01:12 2016 UTC and is due to finish in 60 minutes. The chair is m3m0. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:16 <openstack> The meeting name has been set to 'freezer_2016_02_18' 14:01:38 <m3m0> hello and welcome to the freezer meeting 14:02:00 <m3m0> as always our agenda is available here: https://etherpad.openstack.org/p/freezer_meetings 14:02:21 <m3m0> please raise your hand if you are here for the freezer meeting 14:02:22 <EinstCrazy> hello 14:02:23 <m3m0> o/ 14:02:41 <yangyapeng> :) 14:02:51 <zhangjn> :) 14:03:30 <slashme> Plop 14:04:10 <m3m0> first of all welcome to zhangjn and yangyapeng 14:04:18 <slashme> Welcome :) 14:04:42 <m3m0> first order of bussines: Freezer-scheduler scheduling method (relates to https://bugs.launchpad.net/freezer/+bug/1546269) 14:04:43 <openstack> Launchpad bug 1546269 in Freezer "When restarted, the freezer-scheduler stops executing scheduled jobs." [Undecided,New] 14:04:45 <yangyapeng> Very happy 14:05:29 <slashme> This one is a bad bug. 14:05:38 <m3m0> this is a critical issue that we need to address soon 14:05:50 <slashme> First, make sure you take note of it and the current workaround 14:06:19 <m3m0> are those workarounds the same as you propossed in the bug? 14:06:39 <slashme> Yes, but they are not great workarounds 14:08:02 <slashme> I think we need to rework a bit the scheduling behaviour of the freezer-scheduler. I wrote my ideas about that on this etherpad: https://etherpad.openstack.org/p/freezer_scheduler_scheduling 14:08:35 <m3m0> what's the effort do you think is need it for this? 14:08:42 <slashme> But that seems too big of a rework to implement and backport it to liberty 14:09:29 <slashme> Maybe we can manage a quickfix by just enabling the persistent stores of APScheduler and load back the store when the scheduler is started 14:09:58 <m3m0> does that mean adding a new requirenment? sqlite? 14:10:26 <slashme> Then question is do we try to implement the bigger picture (I really think we should), and if yes, in which timeframe. Mitaka ? Newton ? 14:10:34 <slashme> Yes, sqlqlchemy and sqlite 14:10:45 <m3m0> is mitaka still viable? 14:11:39 <slashme> I think so, it is not really a feature 14:11:58 <slashme> More a multi bugfix 14:12:33 <m3m0> for the api, besides updating the mappings is there any other change that you might have in consideration? 14:13:04 <slashme> Maybe small ones, version checking maybe. 14:14:00 <slashme> I think that if you try submiting a job with version that is egal or inferior the the one stored in the api, it should reply with error + job curently stored under that id 14:21:30 <m3m0> in order to avoid this kind s of issues in which a merge of jobs is need it 14:22:23 <m3m0> we need to disable the option to update jobs while they are in a running status, by doing that we avoid having versioning issues if a status and an edit happens at the same time 14:24:16 <m3m0> guys, anyone has input on this? or do you require more context? 14:25:06 <m3m0> slashme, anything else on this topic? 14:25:25 <slashme> Yup. 14:25:52 <EinstCrazy> Maybe, I need some context. I think I need to reproduce this bug first. Would you give me the whole step? 14:25:53 <slashme> So just to be clear: 14:25:54 <slashme> - We implement a quickfix first. this one gets backported to liberty 14:26:33 <slashme> - We implement the bigger solution on mitaka 14:26:50 <m3m0> EinstCrazy: https://bugs.launchpad.net/freezer/+bug/1546269 14:26:50 <openstack> Launchpad bug 1546269 in Freezer "When restarted, the freezer-scheduler stops executing scheduled jobs." [Undecided,New] 14:26:50 <slashme> EinstCrazy: You have the steps writen on the bug 14:27:37 <m3m0> do you want to take ownership of this? 14:28:37 <m3m0> or get the right people involved in this please because it has high priority 14:28:43 <slashme> Yes 14:28:52 <m3m0> thanks a lot 14:29:02 <m3m0> next item in the agenda 14:29:22 <m3m0> #topic Log Levels to refactor freezer-agent's code to oslo-logging 14:29:40 <m3m0> szaher: you here? 14:29:57 <szaher> Yes 14:30:11 <m3m0> could you expand a little bit more about this? 14:31:47 <szaher> after moving to oslo.config and oslo.log we need to change almost of freezer files to use oslo.log instead of python native logging so I think it's the correct time to define log levels 14:32:28 <szaher> we have some incorrect levels in freezer code like things that should be info and we log it as warn and so on 14:32:55 <szaher> we need to put a clear criteria how to define the level for the piece of code you're writing 14:32:58 <m3m0> have you identified all the cases in freezer? 14:33:32 <szaher> No, I think we need to do it together in mid-cycle meetup 14:34:27 <szaher> I can work on this and list all place we do logging in freezer and we discuss them in the meetup 14:34:32 <m3m0> so there 2 actions for this, define the logging level and update the files to work witht the new model right? 14:34:42 <m3m0> that could be great 14:34:44 <szaher> Yes 14:34:54 <m3m0> as well if you can work on the logging level could be great 14:35:00 <m3m0> something like -vvvv 14:36:37 <slashme> Is there any OpenStack guidelines on what each log level should contain ? For example critical is only something that prevents the app from working, and so on. 14:37:09 <szaher> slashme https://wiki.openstack.org/wiki/LoggingStandards 14:38:08 <szaher> at the end of this page you will find a link ( http://stackoverflow.com/questions/2031163/when-to-use-log-level-warn-vs-error ) defining log levels that we should use 14:38:13 <slashme> @szaher Perfect ! 14:38:16 <m3m0> grat 14:38:37 <m3m0> let's review them and try to follow them :) 14:38:57 <m3m0> so, you wont start working on this until the mid-cycle meeting right? 14:39:32 <szaher> Ok ... 14:40:09 <m3m0> do you have anything more to add? 14:40:35 <szaher> Yea one thing 14:41:15 <szaher> We need to test this patch https://review.openstack.org/#/c/280827/2 on windows and linux as well and we need to merge it asap 14:41:59 <m3m0> ok, I'll try to set up a windows machine 14:42:16 <szaher> We need this in Mitaka I think 14:45:44 <m3m0> ok, please @all review this 14:45:55 <m3m0> well moving to the next topic 14:46:12 <m3m0> #topic Midcycle meetup 14:46:57 <slashme> As daemontool is not present, we can't make big progress on this one 14:47:03 <m3m0> ddieterly zhangjn yangyapeng EinstCrazy szaher slashme 14:47:19 <ddieterl_> yes 14:47:24 <EinstCrazy> Yes 14:47:27 <yangyapeng> yes 14:47:30 <slashme> HPE is okay to host the meetup here in Galway 14:47:30 <szaher> Yes 14:47:54 <ddieterly> i'm able to travel to galway 14:47:54 <slashme> The most important thing now is to setup a date 14:48:02 <m3m0> we are planning to have a mid-cycle meetup in Galway, Ireland to discuss some changes for freezer 14:48:58 <m3m0> do you have have any issue if we do this in March? 14:49:18 <ddieterly> i'm unavailable on march 1 14:49:22 <ddieterly> other than that, no issues 14:50:13 <ddieterly> how many days are we thnking? 14:50:31 <m3m0> I was thinking around 17-18 to introduce people to st patricks day 14:51:02 <ddieterly> please have the guiness ready for consumption ;-) 14:51:24 <slashme> We will :) 14:51:42 <m3m0> it's on the requirenments.txt for the meeting 14:51:50 <zhangjn> :) 14:51:54 <ddieterly> excellent 14:52:00 <m3m0> so everyone is welcome to join 14:52:16 <slashme> Any other contributors (beside daemontool and ddieterly) think they will make the trip ? 14:52:17 <m3m0> but we'll try to update the irc 14:53:08 <ddieterly> is that a firm date? 17-18? 14:53:30 <slashme> No. We need to check daemontools availability 14:53:55 <m3m0> nop, but we need to setup this asap 14:53:55 <slashme> I guess we need our PTL for this kind of meetup ;) 14:53:56 <ddieterly> ok 14:54:08 <ddieterly> yea, that's only 4 weeks away 14:55:18 <m3m0> ok we'll sync again when daemontool is available 14:55:28 <zhangjn> we have no budget to this trip :( 14:55:48 <m3m0> where are you based zhangjn? 14:56:05 <zhangjn> so you'd better update meeting info in IRC. 14:56:22 <zhangjn> China 14:56:31 <m3m0> we will :) 14:56:37 <zhangjn> ths 14:57:15 <m3m0> ok guys we have 2 topics left and no time 14:57:31 <m3m0> should we discuss this in #openstack-freezer channel? 14:57:38 <ddieterly> sure 14:57:42 <yangyapeng> ok 14:57:50 <EinstCrazy> yes 14:57:55 <m3m0> and we need to send an email to the openstack community to let them know about the meeting 14:58:07 <m3m0> does anyone else has something to say? 14:59:12 <m3m0> that's a no :) 14:59:20 <m3m0> thanks all for participate 14:59:23 <m3m0> #endmeeting