15:00:28 <rhochmuth> #startmeeting monasca
15:00:30 <ddieterly> hello
15:00:33 <openstack> Meeting started Wed Sep 28 15:00:28 2016 UTC and is due to finish in 60 minutes.  The chair is rhochmuth. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:34 <pratid> hello
15:00:35 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:36 <ddieterly> o/
15:00:37 <openstack> The meeting name has been set to 'monasca'
15:00:37 <rhochmuth> o/
15:00:38 <witek> hi
15:00:41 <kamil__> o/
15:00:45 <koji> o/
15:00:47 <arturbasiak> o/
15:01:19 <rhochmuth> I didn't have any topics for today in the meeting agenda
15:01:19 <ddieterly> bonjour/guten tag/Kon'nichiwa
15:01:26 <rhochmuth> https://etherpad.openstack.org/p/monasca-team-meeting-agenda
15:01:34 <rhochmuth> ungaya
15:01:44 <witek> cześć
15:01:46 <ddieterly> ungirla
15:02:15 <rhochmuth> are there any open disucssion that folks would like to have
15:02:25 <pratid> I'd like to comment something
15:02:31 <pratid> 1st of all
15:02:36 <ddieterly> well, we are working on cassandra with lucene
15:02:36 <rhochmuth> pratid: ok
15:02:42 <pratid> I'm new in these meetings
15:02:58 <ddieterly> welcome, pratid
15:03:03 <hosanai> o/
15:03:08 <pratid> i've registered a new blueprint
15:03:10 <rhochmuth> welcome aboard
15:03:17 <pratid> https://blueprints.launchpad.net/monasca/+spec/ceilosca-value-meta-marshalling
15:03:28 <pratid> although it could be considered a bug
15:03:42 <pratid> would you like to comment it now?
15:03:43 <rhochmuth> i'm reading now
15:03:58 <pratid> or we could leave if for next meeting
15:04:42 <pratid> this is a problem we are suffering right now in FIWARE
15:04:52 <pratid> and we have solved it
15:04:55 <rhochmuth> pratid: i think i'll need to read off-line
15:05:02 <rhochmuth> definitely would like to see the issue resolved
15:05:03 <pratid> perfect
15:05:20 <rhochmuth> we also have a team within hpe that works on ceilosca
15:05:30 <rhochmuth> but they don't usually attend the monasca weekly meeting
15:05:32 <pratid> feel free to contact me
15:05:38 <rhochmuth> there are also folks from cisos
15:05:46 <pratid> would you like my email?
15:05:46 <rhochmuth> so, i would like to get everyone connected
15:05:53 <rhochmuth> yes, that would be great
15:05:57 <ddieterly> pratid the client can do its own stringify  without monasca doing it
15:06:18 <pratid> don't understand
15:06:41 <pratid> what do you mean with "its own stringify"?
15:06:41 <ddieterly> the client is able to put anything into the meta value that it would like to
15:06:48 <pratid> no, it isn't
15:06:52 <pratid> as I said
15:06:55 <ddieterly> if you would like to stringify something, then that is ok
15:07:10 <pratid> currently monasca API raises an error
15:07:36 <ddieterly> you can stringify the data before you call the monasca client
15:07:37 <pratid> our solution is simply stringify any non-string metadata item
15:07:55 <pratid> but we don't call it directly
15:08:04 <pratid> it's ceilometer pipeline
15:08:08 <pratid> which does
15:08:23 <pratid> using the publisher
15:08:50 <pratid> rhochmuth: if you like
15:08:59 <pratid> please tell people from HPE
15:09:15 <ddieterly> should this be in monasca api is the question, i guess
15:09:18 <pratid> to contact me at pablo dot rodriguezarchilla at telefonica dot com
15:09:47 <pratid> and we could discuss it in detail
15:09:48 <rhochmuth> i'll send email out after this meeting
15:09:50 <rhochmuth> thx
15:09:55 <pratid> you're welcome
15:10:33 <rhochmuth> ddieterly: the floor is yours
15:10:38 <rhochmuth> for cassandra update
15:11:01 <ddieterly> we are working on cassandra schema and perf testing
15:11:16 <ddieterly> we have decided to try to use lucene for secondary indexes
15:11:42 <ddieterly> lucene seems to give us the facilities that cassandra lacks in lookup capability
15:12:10 <ddieterly> we will be publishing a schema and some performance metrics soon
15:12:29 <witek> how are you testing the performance?
15:12:36 <ddieterly> that is the gist of it
15:12:54 <ddieterly> 3-node cluster, loading measurements as fast as possible
15:13:06 <rhochmuth> hp dl380p
15:13:08 <ddieterly> we'll first try python
15:13:13 <rhochmuth> 24 physical cores per node
15:13:26 <rhochmuth> RAID 50 on 24 1 TB disks
15:14:15 <kamil__> that's a lot
15:14:21 <ddieterly> should hopefully get 3k-5K inserts per sec per node
15:14:28 <witek> ddieterly: do you plan on java as well? I hope not.
15:14:38 <ddieterly> yes, i believe so
15:14:42 <rhochmuth> 3k-5k per core
15:15:00 <ddieterly> yes, per core
15:15:08 <ddieterly> i keep fat fingering that
15:15:20 <rhochmuth> witek: we are looking at performance of python
15:15:31 <rhochmuth> if it looks good, then we will stick with python
15:15:43 <ddieterly> because we LOVE python
15:15:43 <rhochmuth> if not, then we'll look at the options
15:15:49 <witek> :)
15:15:52 <witek> thanks
15:16:00 <ddieterly> ;-)
15:16:08 <rhochmuth> i was thinking scala would be our first choice
15:16:18 <ddieterly> we really do love scala
15:16:29 <rhochmuth> golang would be second
15:16:30 <ddieterly> seriously
15:16:37 <ddieterly> not so much love there
15:16:59 <rhochmuth> no, i wasn't being serious
15:17:09 <rhochmuth> we aren't looking at scala or golang
15:17:20 <ddieterly> yet
15:17:21 <rhochmuth> let's see how the python performance looks
15:18:04 <rhochmuth> i need a sarcasm emoii in irc
15:18:08 <rhochmuth> do they have one of those?
15:18:21 <ddieterly> for those interested in lucene/cassandra https://github.com/Stratio/cassandra-lucene-index/blob/branch-3.0.8/doc/documentation.rst
15:19:14 <rhochmuth> just so folks understand, lucene gets added as a plugin to cassandra, and then can be used to build secondary indexes and searched
15:19:21 <witek> do you want to compare it with the results without lucene?
15:19:57 <ddieterly> we have difficulty modelling our functionality with just cassandra alone
15:20:21 <ddieterly> so, we need lucene to maintain the API as it is
15:21:01 <ddieterly> if someone can come up with a cassandra only implementation that efficiently does what the monasca api needs to do, please share it with us
15:21:38 <rhochmuth> Here is the etherpad which has all the links to the proposed schemas so far
15:21:39 <rhochmuth> https://etherpad.openstack.org/p/monasca_cassandra
15:22:14 <rhochmuth> ddieterly: is the one we are looking at right now listed?
15:22:18 <ddieterly> we should have a contest to see if someone can do it
15:22:35 <rhochmuth> someone can do it for sure
15:22:38 <ddieterly> rhochmuth not yet, i'm getting the lucene indexes in place first
15:22:46 <rhochmuth> the issue is performanace
15:23:54 <ddieterly> the price will be a trip to colorado and dinner with rhochmuth
15:24:07 <ddieterly> s/price/prize
15:24:16 <ddieterly> or maybe it is a price
15:24:20 <ddieterly> lol
15:24:22 <rhochmuth> not much of an incentive
15:24:22 <pratid> ha, ha
15:24:48 <pratid> pri?e
15:24:59 <rhochmuth> so, are there any other topics to discuss
15:25:04 <rhochmuth> release plans update
15:25:07 <rhochmuth> ?
15:25:19 <rhochmuth> that caused me a bit of pain last week
15:25:49 <witek> I've seen your discussion with Doug Hellmann
15:25:58 <witek> thanks for clarifying this
15:26:23 <rhochmuth> well, i hope in the ocata release i can get everything done better
15:26:29 <rhochmuth> more timely
15:26:44 <rhochmuth> i think evewrything would have been super easy if i had updated tags
15:27:02 <rhochmuth> prior to when python-monascaclient and monasca-statsd were branched
15:27:14 <rhochmuth> unfortunately, i missed the train on that
15:27:28 <rhochmuth> it looks like we are in good shaped with all the other repos
15:27:53 <rhochmuth> and we can try and merge master to newton after oct 6th
15:28:16 <rhochmuth> although, theere might be some push-back on some of the commits
15:28:20 <rhochmuth> i guess we'll see
15:28:35 <rhochmuth> witek: so are you ok with everthing so far
15:28:48 <witek> yes
15:28:52 <rhochmuth> it isn't ideal, but i don't think i can do better
15:29:00 <rhochmuth> other than tag on-time next time around
15:29:06 <rhochmuth> witek: thx
15:29:50 <rhochmuth> are there any other topics then?
15:30:11 <witek> could you share the schedule for design summit sessions, please?
15:30:23 <rhochmuth> i don't have it, but i should
15:30:30 <rhochmuth> i'll have to track that down
15:30:34 <rhochmuth> i'll send email
15:30:39 <witek> thanks
15:30:50 <ddieterly> who will be at the summit?
15:30:54 <rhochmuth> we also need to be preparing for the sessions
15:31:05 <witek> yes
15:31:44 <witek> Kamil, Tomasz and me from Fujitsu EST and Poland
15:31:55 <ddieterly> rhochmuth and ddieterly will be thee
15:31:59 <rhochmuth> shinya will be there
15:32:10 <witek> Kouji?
15:32:20 <rhochmuth> also hopefully a couple of folks form hpe bristol for monasca-analytics
15:32:51 <rhochmuth> hisashi and daisuke
15:33:09 <koji> probably i'll be
15:33:11 <hosanai> rhochmuth: we will be there.
15:33:30 <rhochmuth> hosanai: thx
15:33:54 <rhochmuth> i still don't have a definitive statement from luis and the team in bristol
15:34:03 <rhochmuth> hopefully tomorrow i'll know more
15:34:22 <hosanai> rhochmuth: i think so too :-)
15:35:11 <rhochmuth> anymore topics for today then?
15:35:40 <ddieterly> none from me
15:36:27 <rhochmuth> ok, then i'l going to end the meeting
15:36:35 <ddieterly> ciao! ciao!
15:36:39 <rhochmuth> bye everyone
15:36:42 <pratid> bye
15:36:44 <witek> thanks, bye
15:36:49 <koji> thanks
15:36:49 <rhochmuth> #endmeeting