14:00:02 #startmeeting freezer 14:00:03 Meeting started Thu May 12 14:00:02 2016 UTC and is due to finish in 60 minutes. The chair is ddieterly. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:04 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:07 The meeting name has been set to 'freezer' 14:00:08 #topic attendance 14:00:13 o/ 14:00:33 :) 14:00:44 hi 14:00:53 Hi ddieterly 14:01:06 bonjour! 14:01:50 ok, hopefully others will trickle in 14:01:56 let's get to it 14:02:11 #topic specs repo 14:02:15 what is that? 14:03:31 whatever it is, it will be finished by eod 14:03:54 :0 14:04:20 We have a freezer-specs repo 14:04:35 That should be used from now on to store the specs 14:04:49 It needs to be initialized which is what I'm doing atm 14:05:16 what is involved in initialization? 14:05:41 The readme will contain information about the workflow to follow (specs, blueprint, approval, ...) 14:05:54 oh, so you are adding content to it 14:06:07 Initialization is writing the readme/doc + create the file hierarchy 14:07:29 o/ 14:07:38 o/ 14:08:49 is there the Freezer meeting going on here? 14:09:02 daemontool: yes 14:09:09 ok, so slashme is adding a napoleonic twist to the freezer-spec repo by regimenting it 14:09:20 daemontool supposedly 14:10:03 We forgot to say that the agenda for the meeting is available here as usual : https://etherpad.openstack.org/p/freezer_meetings 14:10:14 Feel free to add any topic you want to discuss 14:10:14 slashme thanks 14:10:52 what do we want to discuss about the freezer-specs repo? 14:12:22 we need to move the dr bp there after the specs repo setup 14:12:29 nothing? 14:12:36 can we move to the next topic then? 14:12:42 Nothing right now. Just wanted to point out that when it'll ve ready, this will be the way to go to submit specs 14:12:45 sounds good for me 14:13:03 how does that affect blueprints? 14:13:12 the bp cannot be approved 14:13:17 right now 14:13:28 that does not mean the implementation cannot be done 14:13:33 but the bp needs to be there 14:13:45 so, the spec is the same thing as the bp? 14:14:03 the specs repo is where the project bp are placed 14:14:04 just under review process? 14:14:14 ok 14:14:14 ddieterly, nope 14:14:22 ddieterly: I'll point you to the workflow as soon as it is finished 14:14:23 is that the way other projects do it? 14:14:24 so check the freezer-specs repo layout 14:14:27 That will be simpler 14:14:37 and check the nova-specs layout 14:14:40 we have to do that 14:14:47 that's what we are referring to 14:14:57 ddieterly: That's how most project procede, you can check the nova-specs repo 14:15:07 ok 14:15:35 FYI: https://github.com/openstack/nova-specs 14:15:54 yep, looking now... 14:16:06 beautiful 14:16:27 thatsdone, nice name :) 14:16:28 beau 14:16:30 bello 14:16:32 ok 14:16:35 next? 14:16:51 #topic freezer DR 14:17:06 where is the repo? 14:17:17 Repo is created and available here https://github.com/openstack/freezer-dr 14:17:37 that is a great achievement 14:17:39 :) 14:17:47 we need to get approved this in governance https://review.openstack.org/#/c/314288/ 14:17:57 we have only one issue right now the code was written based on kilo release so we need to check with liberty and mitaka or at least mitaka 14:18:14 then add the diagrams form the austin preso to the repo 14:18:24 szaher, good, we need to tag it then 14:18:38 not a bad thing, as it works from kilo onwards 14:18:43 but I don-t think we can branch now 14:18:45 the gate jobs need to be added 14:18:49 slashme, ? 14:18:52 ddieterly, yes 14:19:10 ddieterly, they should be added 14:19:22 if you are referring to the governance repo 14:19:23 we need to think about it 14:19:27 daemontool: yes ? 14:19:38 slashme, about tagging dr repo 14:19:43 for kilo, liberty, mitaka 14:19:50 I don't think we can branch now 14:19:56 but definetely we can tag 14:20:00 as we need to have at least two compute nodes in our devstack deployment 14:20:23 ddieterly, ah yes, the dsvm gate job is disabled there oyu are right 14:21:13 did we get a lot of support for this project from the community? 14:21:28 i heard 2 people voice concerns that this was out of scope 14:21:39 ddieterly, from HP? 14:21:56 dan dyer and there were 2 others in the dev meetup before slashme arrived 14:22:05 ok 14:22:10 the 2 others were from different companies 14:22:22 I've heard 2 people too 14:22:24 from different companies 14:22:27 saying that 14:22:43 than I discovered they wanted to upstream their own new project in openstack 14:22:50 to do that 14:22:54 lol 14:22:55 :) 14:23:15 well, on the face of it, it does seem to go far afield of freezer's core mission 14:23:40 wanders into monitoring and recovery/mitigation 14:23:46 well if you look at the mission in governance, we are there https://governance.openstack.org/reference/projects/freezer.html 14:24:21 so 14:24:22 i think we would do better to focus on making recovery simpler 14:24:26 I agree 14:24:29 that's what i heard at the conference 14:24:37 ymmv 14:24:42 in my opinion, if we want to split it at some point 14:24:45 I'm OK 14:24:49 and I agree on the recovery also 14:25:31 ok 14:25:34 I agree with all of that 14:25:34 I agree on giving higher priority to abckup and restore though 14:25:41 daemontool: +1 14:25:53 daemontool: +1 14:25:56 that's definetely our main focus at least for now 14:26:09 but let's see if others are interested on contributing 14:26:20 sure, sounds like a plan 14:26:27 +1 14:26:28 anything else on that topic? 14:26:30 glad we all agree :) 14:26:33 :) 14:26:45 oh, i wish i could give fausto a big hug right now 14:26:46 +1 14:27:09 ddieterly, everytime I try to hug you in person you do not want that 14:27:12 haha 14:27:32 it takes a while for americans to get used to physical contact 14:27:38 ok 14:27:45 we have a large personal space 14:28:13 i'll have to come to italy and get immersed in all the touching 14:28:19 haha 14:28:21 ok 14:28:39 ok, anything else on dr? 14:28:41 next? 14:28:42 what's the next? 14:28:54 #topic ceph isssues 14:29:18 what's issue? 14:29:30 good question, glad you asked 14:29:43 who put that topic on the agenda? 14:29:51 I thing there was someone 14:29:52 ping Silviu 14:29:56 askign the IRC channel 14:30:06 about a problem related to that 14:30:13 don-t know 14:30:24 let-s go next and we can come back to this later? 14:30:30 slashme, maybe knows about it? 14:31:06 SIlviu meet a problem 14:31:29 ok, next topic 14:31:47 Yes, Silviu had a problem 14:31:47 #topic PyPI 14:32:09 We should implement something similar to https://review.openstack.org/gitweb?p=openstack/horizon.git;a=commitdiff;h=8201d65cf983f9c995dd653953ad1719f6df8459 14:32:54 #topic PyPI 14:33:05 In order to manage content lenght when doing swift upload 14:33:16 slashme, ok yes 14:33:30 For the PyPI topic, 14:33:36 reldan, are you interested on doing that? or anyone else? 14:33:43 We need to update our package in pypi 14:33:48 slashme, where you thinking about someone? 14:33:55 I noticed that most of them are outdated 14:34:11 daemontool: This is a pretty low hanging fruit 14:34:17 shouldn't be that update automatically by the bot? 14:34:25 slashme, yes ok 14:34:33 Anyone that has acces to a ceph cluster can fix this 14:34:54 ok 14:36:10 daemontool: I’m doing it right now 14:36:11 yes 14:36:11 we can ask to Infra 14:36:24 about why the pypi packages are not updated 14:36:25 anymore 14:36:27 slashme: we have a ceph cluster ,but it also in the performance test 14:36:54 It shouldn't affect performance at all 14:37:08 And it shouldn't cause any bug on the ceph side 14:37:10 well probably yangyapeng needs to ask permission and all that 14:37:26 Do you have a RadosGw ? 14:37:27 as that ceph instance is already used by another team 14:38:32 process check, we need to move along 14:38:48 we have not install radosgw. 14:38:49 the ceph cluster also used by another team. we can test it in next scrum. . 14:39:31 can we move to next topic? 14:39:33 What's worry with ceph? 14:39:41 Okay. You need the radosgw in order to emulate the swift api if you want to use ceph with freezer. 14:40:04 We should be able to find a ceph cluster here to test the patch. 14:40:12 ddieterly: okay, next topic 14:40:18 next 14:40:33 I got it, I will think about this Scenario. 14:40:39 #topic nova, cinder backup to local/ssh issues 14:40:58 szaher you have the floor 14:41:18 about backup nova and swift to local. 14:41:36 In the process of actual backup, there is no swift, but more NFS, etc. 14:41:42 is szaher in the mtg today? 14:41:52 so, we have more need to backup it 14:41:54 There is a patchset from yangyapeng that can do the backup but we won't be able to restore it as we didn't store the metadata 14:42:01 ddieterly: Yes I am here :) 14:42:23 szaher yay! 14:42:28 https://review.openstack.org/#/c/315085/ 14:42:31 how to store metadata? 14:42:43 database or file? 14:43:02 There is a patchset on freezer-metadata done by reldan I reviewed it with him to store the metadata with the backup file so we can restore everything according to this metadata file 14:43:38 the metadata patch https://review.openstack.org/#/c/314209/ we need to review it by today as we need it urgently for another patchsets 14:44:14 yangyapeng: This patchset we won't be able to restore the backup files https://review.openstack.org/#/c/315547/ ! as we don't have any metadata info available 14:45:06 ddieterly: We need to merge the metadata patchset first then me and reldan are going to work on building cinder engine and nova engines 14:45:25 ok 14:45:27 so we will have tar engine for fs related backups, cinder engine for OS cinder and nova for OS nova backups 14:45:52 this patch I have a backup about instance._info, we restore backupfile use it ,it is ok ? 14:45:56 Also daemontools is working or rsync engine 14:47:06 so, is everybody who needs to going to do the reviews? 14:47:38 I have a json to store it in uuid-meta. Although this is so humble 。 14:47:57 everyone is welcome to review the patchset 14:48:10 thank you 14:48:12 suound like you want it approved 14:48:21 so, core members need to jump on it 14:48:28 yangyapeng: we have blueprints so if you want you can work on some of them, the metadata one is here https://blueprints.launchpad.net/freezer/+spec/freezer-metadata 14:48:54 next topic? 14:48:55 ddieterly: core and non core are more than welcomed :) we need more eyes 14:49:14 szaher of course 14:49:37 #topic tripleO bp 14:49:51 who owns that one? 14:50:24 who is asking for that? 14:50:45 why is it on the agenda? 14:50:45 This is anonymous ? 14:51:00 (i'll keep asking questions until somebody responsds ;-) ) 14:51:04 I added it 14:51:07 I guess this is daemontool 14:51:13 sorry one sec 14:51:14 ah ha! 14:51:18 someone come to by desk 14:51:22 哈哈 14:51:25 haha 14:51:26 I'll give update 14:51:33 to the freezer channel 14:51:49 basically we need to support 14:52:09 rh osp 14:52:16 or that kind 14:52:20 of os distributions 14:52:26 that uses tripleo 14:52:35 and ironic 14:52:38 Ah, that's why tripleo ? (OSP7 now uses tripleo) 14:52:50 is this something that errickson is asking for? 14:53:54 ddieterly, nope 14:53:58 ericsson uses Fuel 14:53:59 mirantis 14:54:02 that was requested 14:54:14 by users that use RH in their productino environment 14:55:00 ok 14:55:14 daemontool are you going to write the bp? 14:55:23 I should yes 14:55:31 also I have RH OSP available here 14:55:35 excellent 14:55:36 slashme, do you have the 14:55:42 contacts of that guy 14:55:49 Radoslaw if I remembe rwell 14:55:54 he had the ideas quite clear 14:55:59 he could write it 14:56:03 I can do it in case 14:56:19 Let me check 14:57:31 while he's off checking, how about the bp for dedup? 14:57:42 daemontool you planning on writing that too? 14:57:51 I think 14:57:56 Fabrizio could probably write that 14:57:57 first bp in freezer-spec 14:58:03 :) 14:58:11 zhangjn, :) 14:58:14 which fabrizio? 14:58:18 Fresco 14:58:40 2 mins remainaing 14:59:07 #topic pending reviews 14:59:24 please take a look at the reviews in the agenda and review if possible 14:59:42 ok 14:59:55 this one is particularly easy https://review.openstack.org/#/c/315250/ 15:00:23 ok, out of time 15:00:30 thanks everybody who participated 15:00:31 thanks ddieterly 15:00:36 thank you ddieterly 15:00:38 Time is always short, :) 15:00:39 can continue discussion in freezer room 15:00:45 ciao! 15:00:48 #endmeeting