14:00:40 <isviridov> #startmeeting magnetodb
14:00:41 <openstack> Meeting started Thu Feb  5 14:00:40 2015 UTC and is due to finish in 60 minutes.  The chair is isviridov. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:42 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:44 <openstack> The meeting name has been set to 'magnetodb'
14:00:46 <ikhudoshyn> o/
14:00:50 <aostapenko> o/
14:00:53 <isviridov> o/
14:00:53 <ajayaa> o/
14:01:02 <miqui_> o/
14:01:26 <isviridov> Today agenda https://wiki.openstack.org/wiki/MagnetoDB/WeeklyMeetingAgenda#Agenda
14:01:39 <isviridov> And we can start from first topic, I think
14:01:47 <isviridov> #topic  Go through action items isviridov
14:02:35 <isviridov> The first action charlesw clarify what are plans for row TTL in C* community
14:02:52 <isviridov> Not sure that charlesw around
14:02:57 <isviridov> Let us move on
14:03:04 <ikhudoshyn> +1
14:03:05 <isviridov> And discuss it later
14:03:09 <isviridov> ominakov file blueprint about migration
14:03:36 <isviridov> ominakov any news here?
14:04:28 <ikhudoshyn> move on 2: moveon returns
14:04:39 <isviridov> ikhudoshyn :)
14:05:01 <isviridov> Ok, I think it is important let me share some info with team about it
14:05:46 <ajayaa> that would be helpful. isviridov
14:06:23 <isviridov> With introducing custom LSI implementation we have an installed production instance running now, so in order to cover this operational gap we have started working on migration utility from old cassandra datastructure to new one
14:06:33 <isviridov> #link https://blueprints.launchpad.net/magnetodb/+spec/migration-script
14:07:02 <isviridov> It is going to be export/import via file or SWIFT
14:07:21 <isviridov> Sure it will introduce downtime during migration
14:08:20 <isviridov> In the future, I think, we have to do migration/upgrade on fly, but not this time
14:08:27 <isviridov> ajayaa does it help?
14:09:15 <ajayaa> isviridov I will go through the spec once it is written.
14:09:40 <ajayaa> Are we going to have one for this bp?
14:10:29 <isviridov> It is not yet, actually. I hope description in BP and code will document it good enought.
14:10:41 <charlesw> sorry I'm late, heavy snow right now in Boston
14:10:50 <isviridov> charlesw welcome
14:10:57 <miqui_> hello charlesw
14:11:07 <ajayaa> Hi charlesw
14:11:15 <isviridov> ajayaa actually it is in progress already and ominakov have to share the current state
14:11:16 <charlesw> Hi guys
14:11:26 <isviridov> ajayaa does it work for you?
14:11:36 <ajayaa> isviridov, yes.
14:11:40 <ajayaa> let's move on.
14:11:59 <isviridov> charlesw clarify what are plans for row TTL in C* community
14:12:05 <isviridov> Any news here?
14:12:51 <charlesw> The row TTL question has been asked many times in C* board
14:13:04 <charlesw> I dig into some of the hostory
14:13:30 <charlesw> They have no plan to open up the row marker as a TTL column
14:13:56 * isviridov just updated agenda https://wiki.openstack.org/wiki/MagnetoDB/WeeklyMeetingAgenda#Agenda. Sorry for late notice
14:14:20 <charlesw> So the working around is to have all columns except primary key cloumns to have TTL set
14:14:58 <ajayaa> What do you do incase of an update?
14:15:03 <charlesw> To update columns, you will have to preserve your TTL, then update the row with original TTL
14:15:22 <charlesw> basically a read before write
14:16:07 <ajayaa> How do you get rid of primary key values?
14:16:18 <ajayaa> They will still be there.
14:16:18 <isviridov> charlesw any way to influence C* community? Do we have answer like 'no we will not do it' or it is last known view on it?
14:16:33 <charlesw> it will be auto cleaned I believe
14:17:08 <charlesw> It's a last view on it. the same Q has been asked many times
14:17:14 <ajayaa> charlesw, you just said that all columns excepts primary keys will have ttl set.
14:17:43 <isviridov> charlesw thx
14:17:58 <charlesw> I don't believe you can set TTL on primary key
14:18:37 <miqui_> C* is cassandra right?
14:18:47 <ajayaa> charlesw, If you do an insert with a ttl then primary keys also have ttl associated with them.
14:19:11 <charlesw> but you can't get the ttl for primary key only
14:19:13 <isviridov> miqui_ yes C* == Cassandra
14:19:23 <isviridov> dukhlov you are our Casasndra wizard, what about implementation row TTL ourselves? Any thoughts?
14:21:02 <ajayaa> Primary keys are given some kind of special treatment in cassandra. If you did an insert with no ttl and updated some columns with ttl, then primary keys won't have ttl associated.
14:21:39 <ajayaa> And the row won't be removed. If you did an insert with ttl then primary keys would also get cleaned up.
14:22:10 <isviridov> dukhlov we can eather implement it and merge to upstream or keep it on our owns. Depends what will be easier. Let us think about it
14:22:56 <ajayaa> +1
14:22:59 <ajayaa> isviridov
14:23:22 <miqui_> once ttl is hit...the expectation is wht? the whole record is gone?
14:23:28 <dukhlov> It is interesting, but I don't like support separate casandra branch
14:24:06 <isviridov> miqui_ exactly #link https://blueprints.launchpad.net/magnetodb/+spec/row-expiration
14:24:11 <dukhlov> We need talk to cassandra developers first to get their opinion
14:24:25 <isviridov> dukhlov agree
14:24:38 <charlesw> We shouldn't allow update row with TTL. Will need to delete the row and re-insert with TTL
14:24:39 <dukhlov> if they say that this feature will never merged....
14:24:44 <miqui_> thanks isviridov
14:26:04 <isviridov> Let us consider, that it is one of the options. Not implementing and not merging are different things. If we can do it, it could cost us less that keeping this logic in mdb
14:26:49 <charlesw> Even though C* doesn't fully support row expiration, we can limit our scope to work with C* instead of branching out
14:27:06 <isviridov> #idea implement row TTL in C* by ourselves
14:27:43 <isviridov> charlesw agree, as we have this feature not in top of priorities
14:27:52 <isviridov> Moving on?
14:28:01 <dukhlov> yes, implement it ourselves looks much better to me instead branching
14:28:22 <dukhlov> isviridov: +1
14:28:24 <charlesw> +1
14:28:26 <isviridov> dukhlov I mean in C* code
14:29:11 <charlesw> -1 for changing C* code
14:29:40 <miqui_> this is a tough choice...
14:30:00 <miqui_> a separate C* branch is probably worse in the long term...
14:30:02 <miqui_> IMHO
14:30:29 <dukhlov> yes
14:30:29 <ajayaa> +1 miqui_
14:30:38 <isviridov> Yes, branch is always terrible
14:30:58 <dukhlov> anyway we have no resources for separate branch support now
14:30:58 <isviridov> Let ius move on to kilo-2 milestone
14:31:19 <isviridov> #topic k-2 status https://launchpad.net/magnetodb/+milestone/kilo-2
14:32:26 <isviridov> So this week we are releasing kilo-2 milestone, I've updated launchpad scope as you can see
14:33:20 <isviridov> charlesw it looks like we have postpond Push table-related metrics to statsd and Real Time Request Metrics
14:33:33 <charlesw> can you move the real time request metrics to kilo-3? The notification refactoring added quite some work and blocked the code review
14:33:59 <charlesw> yes,isviridov.
14:34:00 <isviridov> ikhudoshyn also Backup and restore for MagnetoDB and Backup Info data model and repository probably not going to happen this week
14:34:19 <isviridov> charlesw sure, thx
14:34:19 <ikhudoshyn> isviridov: yes, i'm afraid
14:34:36 <isviridov> ikhudoshyn ok
14:34:49 <ikhudoshyn> isviridov: hold on pls
14:34:53 <isviridov> Anything else?
14:34:57 <isviridov> aostapenko?
14:35:02 <ikhudoshyn> Backup Info data model and repository probably actually merged
14:35:20 <isviridov> Anything on the go and close to be finished and needs more attention?
14:35:23 <aostapenko> nothing interesting from my side
14:36:11 <isviridov> ikhudoshyn fixed
14:36:50 <isviridov> Ok, seems we have clarified it
14:37:03 <isviridov> Open discussions?
14:37:55 <isviridov> I would like to thank to aostapenko who made huge contribution this milestone
14:37:56 <miqui_> not much from me, just reading source, working on my mdb dev env...
14:38:00 <isviridov> !m aostapenko
14:38:00 <[o__o]> You're doing good work, aostapenko!
14:38:00 <openstack> isviridov: Error: "m" is not a valid command.
14:38:15 <isviridov> #topic Open discussion isviridov
14:38:16 <aostapenko> isviridov: thank you :)
14:38:50 <charlesw> isviridov, could you share your vision/roadmap to incubation?
14:38:51 <isviridov> miqui_ nice
14:39:05 <ominakov> hi, guys I have early poc for migration, but I think i should write spec first
14:39:42 <isviridov> ominakov we have just agreed that spec is not teh most important thing for one time using tool
14:39:44 <charlesw> ominakov, would love to see spec
14:39:56 <isviridov> charlesw :)
14:40:30 <isviridov> ominakov could you share you code first?
14:40:35 <isviridov> * you
14:40:41 <ominakov> yep, sure
14:41:09 <isviridov> but if you have any spec we really appreciate it as well :)
14:41:26 <isviridov> ominakov do you think you will finish with migration this week?
14:41:38 <isviridov> ominakov the question is kilo-2 or kilo-3
14:41:54 * isviridov remember teh question about incubation
14:43:49 <isviridov> So, charlesw we at the middle of reform now
14:44:43 <ominakov> i guess i'll push working poc this week and in future we can add some features and optimize it
14:45:05 <ajayaa> isviridov, I kind of fee bad when I don't see MagnetoDB on side of sahar, zaqar, murano etc...
14:45:11 <ajayaa> feel*
14:45:26 <miqui_> why would that be the case?
14:46:04 <isviridov> Just TC agree list of tags and the way how it will be given to projects all projects will be a part of openstack
14:47:08 <isviridov> ajayaa status of  zaqar and other incubated projects will be changed and all projects will receive their tags
14:47:50 <isviridov> integrated-release tag has been merged already #link https://review.openstack.org/#/c/146818/
14:47:51 <ajayaa> isviridov, waiting for the day when we get a tag. :D
14:48:36 <isviridov> ajayaa it will be list of tags, the most valuable I think is smth like 'used in production'
14:50:02 <isviridov> charlesw ajayaa have I answered? Can I describe smth else?
14:50:35 <isviridov> BTW ajayaa sahara is integrated, murano is not incubated. JFYI
14:50:54 <charlesw> What are the other valuable tags? How can we all help?
14:51:01 * isviridov probably has network lag
14:52:07 <ajayaa> isviridov, my intention was to just list out some projects with which MagnetoDB should stand shoulder to shoulder. :)
14:53:57 <isviridov> charlesw the list is still under definition, but it is build based on incubation, integration requirements. And we are doing our best to follow it, we have gate, functional, unit testing, great team and working process. What is most valuable now is making mdb better knows and improve it as a software
14:55:02 <isviridov> ajayaa reasonable note, in some aspects we are higher than our mates, but sure not in all of them :)
14:57:16 <isviridov> 3 mins left
14:57:36 <isviridov> Any other question or topic
14:58:33 <isviridov> Looks like we are done for today
14:58:40 <isviridov> Thank you all for comming
14:58:59 <isviridov> #endmeeting