15:00:08 <rhochmuth> #startmeeting monasca
15:00:09 <openstack> Meeting started Wed Oct 14 15:00:08 2015 UTC and is due to finish in 60 minutes.  The chair is rhochmuth. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:13 <openstack> The meeting name has been set to 'monasca'
15:00:14 <rhochmuth> role call
15:00:16 <ddieterly> o/
15:00:21 <rhochmuth> 0/
15:00:22 <rhochmuth> o/
15:00:28 <tsv> 0/
15:00:32 <rbak> o/
15:00:34 <mroderus> o/
15:00:35 <witek> o/
15:01:01 <rhochmuth> give it a minute or two
15:01:24 <rhochmuth> we only have one agenda topic listed
15:01:31 <rhochmuth> 1. Fujitsu:
15:01:32 <rhochmuth> a. paging issue for monasca-api
15:01:44 <rhochmuth> If anyone else has anthing to add please doo
15:01:52 <witek> Tomasz wanted to bring it up, but i don't see him :(
15:01:53 <rhochmuth> last week we were completely booked
15:02:01 <rhochmuth> Here is the link, https://etherpad.openstack.org/p/monasca-team-meeting-agenda
15:02:20 <rhochmuth> i think i saw some reviews for the paging issue
15:02:39 <rhochmuth> is thing something that we just need to get some reviers looking at, or is it something else
15:02:55 <mroderus> I think it's a bigger issue
15:03:01 <rhochmuth> ohhh
15:03:14 <mroderus> not really an "issue", rather an extension
15:03:31 <rhochmuth> so, basically new features
15:03:38 <mroderus> yes
15:03:56 <rhochmuth> do you want to try and discuss now or wait until tomasz is available
15:04:02 <mroderus> one use case in which it came up was when one wants to sort the alarms on the Horizon page, e.g. by severity
15:04:14 <mroderus> I think that makes sense
15:04:21 <rhochmuth> absolutely
15:04:27 <rhochmuth> we've had similar discussion here
15:04:44 <mroderus> I just remember that Tomasz mentioned that an API change would be necessary to get this sorting feature on the UI
15:04:47 <mroderus> ok
15:05:03 <rhochmuth> right now we can only filter on certain fields, like severity, but we can't order by
15:05:28 <rhochmuth> so, if you wanted to page through your alarm ordered by state and severity, you can't really do that
15:05:35 <rhochmuth> so, i think that would be extremely useful
15:05:45 <rhochmuth> currently, we are doing client side processing
15:05:45 <mroderus> have you ever discussed that before?
15:05:55 <rhochmuth> but that implies brining in the entire list
15:05:58 <mroderus> yes, that's the problem I guess
15:06:07 <rhochmuth> we have just had some discussion on the hp team locally
15:06:11 <rhochmuth> nothing to detailed
15:06:26 <rhochmuth> our ui/ux team has had some requests in this area
15:06:35 <rhochmuth> but we aren't working on them yet
15:07:03 <rhochmuth> so, i think we are all in agreement that this would be extremely useful
15:07:11 <rhochmuth> the only problem is getting resources to address it
15:07:23 <mroderus> ok.. on the Fujitsu side, this topic is not urgent. But I think that at some point we will start working on it
15:07:31 <mroderus> yes, I know this problem :)
15:07:42 <rhochmuth> ok, whoever gets to it first can do it
15:07:56 <mroderus> yes
15:07:58 <rhochmuth> i don't think it is a difficult area
15:08:06 <rhochmuth> and is probably low hanging fruit
15:08:13 <rhochmuth> easy to add, with a big benefit
15:08:26 <mroderus> that sounds good
15:08:47 <mroderus> I guess that's all I can say to it. Tomasz knows the details
15:08:56 <rhochmuth> ok
15:09:03 <rhochmuth> how about we move on then
15:09:20 <rhochmuth> no one has added more agenda topics
15:09:29 <rhochmuth> i think discussing status would be good
15:09:30 <mroderus> If you like, you can give us a status update on your work at HP on the logging stuff
15:09:36 <rhochmuth> sounds good
15:09:51 <mroderus> I remember that TSV wanted to start looking into it in September or October or so
15:10:11 <rhochmuth> yes, i think we are starting to free up
15:10:21 <rhochmuth> our original dates got pushed out
15:10:31 <mroderus> same thing on our side
15:10:40 <rhochmuth> so in another week or two i think we'll be available to start working on the logging stuff again
15:10:56 <rhochmuth> so, let's say after the tokyo summit
15:11:00 <mroderus> sounds great
15:11:00 <rhochmuth> to be safe
15:11:31 <rhochmuth> we are really incented to get this work completed
15:11:32 <mroderus> we can also discuss more about that in Tokyo.. split up work etc.
15:11:42 <rhochmuth> so i think you'll see a very strong commitment from hp in this area
15:11:52 <mroderus> ok, great
15:12:13 <rhochmuth> sure, yest we can discuss in tokyo
15:12:23 <mroderus> will TSV be there?
15:12:29 <rhochmuth> tsv won't be going as it turns out, but we'll have another representative from teh logging team
15:12:53 <rhochmuth> so we'll have enough representation
15:13:17 <rhochmuth> has anyone looked at devstack
15:13:29 <rhochmuth> or the Tempest tests
15:13:40 <rhochmuth> the tempest tests have been up for review for a while
15:13:48 <rhochmuth> i would like to get them merged
15:14:00 <rhochmuth> if folks can at least take a quick look and +1 taht would help
15:14:13 <rhochmuth> if not, we'll probably get them merged real soon anyway
15:14:32 <tsv> mroderus, will start looking at this from next week. I will not be there for the summit but would like to be involved in any mail/chat discussions if possible. thanks
15:14:51 <rhochmuth> thansk tsv
15:15:01 <rhochmuth> you'll be involved in everything of-course
15:15:40 <rhochmuth> Here is the review for the Tempest Test, https://review.openstack.org/#/c/228122/
15:15:40 <mroderus> tsv: sounds good. we'll surely involve you in the discussions
15:16:06 <rhochmuth> there are of-course lot's of enhancements to continue to add
15:16:15 <rhochmuth> but i think it is in reasonable state to merge
15:16:27 <rhochmuth> then it will be much easier to collaborate with others on with
15:17:04 <rhochmuth> hopefully, the overall organization and framework seems reasonable to folks, although much of this was dictated by the conventions that had already been put in place
15:17:31 <rhochmuth> There is also a README at, https://review.openstack.org/#/c/228122/23/monasca_tempest_tests/README.md, that descriebs how to run them
15:17:42 <rhochmuth> using the monasca vagrant environment
15:17:51 <rhochmuth> it should be really easy to run them
15:17:58 <rhochmuth> but if anyone has any issues please let me know
15:18:45 <rhochmuth> Devstack is also getting in good shape
15:19:15 <rhochmuth> However, I've been trying to use the Monasca Devstack with a Vagrant virtualbox vm, and hitting an issue
15:19:25 <rhochmuth> Everything is working for Deklan pergectly though
15:19:37 <rhochmuth> but he was smart enough to build his own VM from scratch
15:19:49 <rhochmuth> and i'm too stubborn to do that
15:20:10 <rhochmuth> anyway, it is very easy to get started with DevStack
15:20:28 <rhochmuth> you'll need to use a ubuntu trusty os
15:21:12 <rhochmuth> There is a nice README at, https://github.com/stackforge/monasca-api/tree/master/devstack, that describes the variables to set
15:21:44 <rhochmuth> We've been potentially discussing if the monasca devstack works whether it will maek sense to continue with the monasca-vagrant environment
15:21:54 <rhochmuth> so, that is a potential topic for the future
15:22:27 <mroderus> how many environments do we have at the moment? DevStack, Vagrant and Docker?
15:22:45 <rhochmuth> yes, i think that is all
15:22:50 <mroderus> ok
15:22:57 <bklei> \o
15:23:20 <rhochmuth> so, unless questions on tempest and devstack let's move on
15:23:46 <rhochmuth> Are there any reviews that folks are pushing to get reviewed?
15:24:01 <bklei> i'd like to see https://review.openstack.org/#/c/234266/ get merged..
15:24:17 <bklei> monasca-api change (Check entire set of value_meta key/value pairs for length)
15:24:23 <rhochmuth> yes
15:24:30 <rhochmuth> deklan did you take a look at that one
15:24:31 <ddieterly> i take a look at it today
15:24:36 <bklei> gracias
15:24:53 <rhochmuth> i looked at it and it looks fine to me
15:25:00 <rhochmuth> i guess i didn't +1
15:25:22 <rhochmuth> but i will soon
15:25:23 <bklei> thx -- we expanded our db today for value_meta to 8192
15:25:31 <rhochmuth> in production?
15:25:35 <bklei> yup
15:25:42 <rhochmuth> how did that go?
15:25:51 <bklei> no issues!
15:25:56 <rhochmuth> awesome
15:26:12 <rhochmuth> thanks for doing that
15:26:14 <bklei> now we can bring in new monasca-agent with all the goodness libvirt fixes next week
15:26:16 <bklei> np
15:26:23 <witek> :)
15:26:50 <bmotz> I'd quite like to progress https://review.openstack.org/228975 (or equivalent) as I'm otherwise I'm maintaining an internal fork
15:27:27 <rhochmuth> tomasz left some comments
15:27:37 <bmotz> but he was overall happy with it
15:28:25 <rhochmuth> ok, i'll take another look and get some eye on it here
15:28:31 <bmotz> great, thanks
15:28:33 <rhochmuth> i don't see any issues with it
15:28:38 <ddieterly> if someone could +2 this that would be great https://review.openstack.org/#/c/234464/
15:29:47 <rhochmuth> i just +1'd
15:29:54 <rhochmuth> straight-forward fix
15:30:04 <rhochmuth> if others agree, please +1 or +2
15:30:18 <rhochmuth> looks like a simple fix that will address some lingering issues in the python api
15:30:47 <ddieterly> yea, measurements list does not work in python api without this fix
15:31:31 <rhochmuth> so, did this work at one point, and then influxdb broke us
15:31:57 <ddieterly> this problem was not discovered because the python api was developed using the java persister
15:32:09 <rhochmuth> ohhh, yeah i recall now
15:32:15 <ddieterly> now that devstack is up and running, we can have both python persister and python api running at same time
15:32:45 <ddieterly> some incompatibilities were exposed
15:33:23 <bklei> ddieterly: let me know if my response to your comment in https://review.openstack.org/#/c/234252/ doesn't make sense
15:33:26 <rhochmuth> so, that is also worth pointing out with the devstack is that there is an easy way to select either the jave or python components
15:33:40 <ddieterly> bklei: ok
15:34:16 <rhochmuth> yes, the goal is to test for an upper bound
15:34:58 <ddieterly> i think we need to determine the actual encoding of the chars in vertica
15:35:04 <rhochmuth> i don't think it is worth converting the value meta to UTF32 and then testing for an exact fit
15:35:40 <bklei> the vertica support folks indicated UTF-16, but roland found some links that seemed to imply it could end up being UTF32
15:35:47 <bklei> just being safe
15:36:00 <bklei> (other reviews roland has up do X4)
15:36:04 <rhochmuth> sorry, i think you are correct on the utf16
15:36:14 <rhochmuth> but that still means it can be 4 bytes
15:36:25 <rhochmuth> similar to how utf16 can either be 1, 2, 3 or 4 bytes
15:36:46 <rhochmuth> well, maybe not 1
15:36:50 <bklei> gotcha, then as a rule i agree with chars X 4 for safety
15:36:58 <ddieterly> also, does the varchar(n) specify chars or bytes?
15:37:13 <rhochmuth> varchar is bytes
15:37:16 <bklei> right
15:37:19 <rhochmuth> that is what the document says
15:37:25 <rhochmuth> should have been varbytes
15:37:31 <rhochmuth> :-)
15:37:34 <bklei> lol
15:37:44 <ddieterly> that does not seem right to me
15:37:46 <ddieterly> sorry
15:37:57 <rhochmuth> what doesn't seem right?
15:38:02 <rhochmuth> that varchar is bytes?
15:38:12 <ddieterly> if i say varchar(1), and i stick a 2 byte char in there, then it would blow up?
15:38:24 <rhochmuth> that is my understanding
15:39:03 <ddieterly> doesn't that seem counterintuitive?
15:39:14 <rhochmuth> https://my.vertica.com/docs/5.0/HTML/Master/1231.htm
15:39:16 <rhochmuth> sorry old link
15:39:24 <rhochmuth> The maximum length parameter for VARCHAR and CHAR data type refers to the number of octets that can be stored in that field and not number of characters.
15:39:33 <rhochmuth> that is a quote
15:39:42 <ddieterly> well, if that is what is says...
15:39:48 <rhochmuth> When using multibyte UTF-8 characters, the fields must be sized to accommodate from 1 to 4 octets per character, depending on the data.
15:39:59 <rhochmuth> that is another quote
15:40:18 <ddieterly> those whacky vertica folks
15:40:33 <bklei> hey man -- your company
15:40:55 <rhochmuth> There si also the review i submitted at, https://review.openstack.org/#/c/231741/
15:41:14 <rhochmuth> i need to address tomasz comment
15:41:30 <rhochmuth> but it is a similar change
15:41:53 <rhochmuth> but in the persister
15:42:38 <bklei> I'll +1 that when you add comment as to why x4
15:42:47 <bklei> but agree with the change
15:43:07 <rhochmuth> i think i'll just introduce a static var with a good name
15:43:13 <bklei> cool
15:43:27 <rhochmuth> so unless more reviews to talk about
15:43:33 <rhochmuth> we can move on
15:43:43 <rhochmuth> i was wondering about grafana 2.0 support
15:43:54 <rhochmuth> has that progressed further?
15:43:55 <bklei> rbak: you here?
15:44:07 <rbak> yep, sorry
15:44:39 <rbak> I'm finally getting some time to actually work on grafana, but I don't have a whole lot to give an update on
15:45:04 <rhochmuth> sounds goo
15:45:24 <rhochmuth> do you think you'll be able to make some progress over the next few weeks then?
15:46:15 <rbak> Yeah, definitely
15:46:34 <mroderus> does it make sense to start talking to the Grafana project at this state or should we wait till we've implemented our changes and then try to get our changes merged?
15:46:50 <rbak> I would wait
15:46:54 <mroderus> ok
15:47:05 <rbak> It shouldn't take too long as long as I can focus some time on it
15:47:57 <rhochmuth> ok, i've run out of topics
15:48:05 <ddieterly> i have a question
15:48:12 <rhochmuth> #questions
15:48:18 <rhochmuth> #topic questions
15:48:21 <ddieterly> is anyone trying to run the python code?
15:48:43 <bklei> not yet at twc
15:48:49 <ddieterly> i've found 3 bugs that would prevent the api from working
15:49:02 <ddieterly> bmotz: aren't you doing something with the python?
15:49:38 <ddieterly> crickets out there
15:50:02 <ddieterly> well, that's all i got ;-)
15:50:04 <bklei> one question for rhochmuth -- any progress on caching in api?
15:50:12 <rhochmuth> no progress
15:50:15 <rhochmuth> sorry
15:50:34 <bklei> ok -- i hope to make https://review.openstack.org/#/c/221492/ more comprehensive -- this week
15:50:43 <bklei> (avoiding multiple inner joins)
15:50:54 <rhochmuth> sounds good
15:51:01 <ddieterly> bklei: sweet!
15:51:17 <bklei> starting weekly meetings with vertica folks -- they're pushing back on that, caching, etc.
15:51:54 <ddieterly> bklei: we eagerly await your submission for review ;-)
15:51:56 <bklei> as for VER-40005 -- that fix is a lon way out
15:51:59 <bklei> :)
15:52:32 <bklei> s/lon/long
15:52:33 <rhochmuth> InfluxDB is claiming 300K metrics per seconds
15:52:37 <rhochmuth> See, https://influxdb.com/blog/2015/10/07/the_new_influxdb_storage_engine_a_time_structured_merge_tree.html
15:52:56 <rhochmuth> Time Structured Merge Tree or TSM Tree for short
15:53:08 <bklei> would like to see if that's really true and the cluster stays up/stable
15:53:29 <bklei> but promising
15:53:32 <rhochmuth> Yes, I agree
15:53:45 <rhochmuth> Sounds like they are making progress though
15:54:24 <rhochmuth> Anyway, it is an interesting read
15:54:49 <bklei> i wonder if that's another complete re-write
15:55:27 <rhochmuth> ok, time is winding down
15:55:52 <rhochmuth> should we adjourn?
15:55:58 <bklei> +1
15:56:09 <bmotz> whoops, sorry - missed that
15:56:17 <rhochmuth> hold-on
15:56:20 <bmotz> I've been evaluating the python API, but not using it in anger
15:56:28 <bmotz> (in answer to ddieterly's question)
15:56:36 <rhochmuth> thanks bmotz
15:56:49 <ddieterly> bmotz: what's 'anger'?
15:56:54 <rhochmuth> ddieterly has made some recent changes
15:57:02 <bmotz> a real deployment
15:57:34 <bmotz> I found quite a few bugs using it casually, so didn't want to go further until tempest tests, etc, were in place
15:57:43 <bmotz> and there was some confidence around it
15:58:01 <ddieterly> i see
15:58:20 <ddieterly> we'll be getting it into shape more now
15:58:30 <bmotz> sounds good
15:58:36 <ddieterly> i'm fixing more bugs today
15:58:40 <bmotz> great :)
15:58:45 <bklei> ddieterly: vertica support in python?
15:58:57 <ddieterly> lol, no
15:59:06 <bklei> is helion going to switch to python?
15:59:19 <rhochmuth> that is the goal
15:59:25 <ddieterly> that is the long-term goal
15:59:45 <bklei> cool
15:59:51 <ddieterly> does vertica have python drivers?
15:59:51 <rhochmuth> so, our focus will start to transition
15:59:56 <rhochmuth> ok, winding down
16:00:09 <bklei> can research ddieterly
16:00:17 <rhochmuth> by everyone
16:00:21 <bklei> thx roland
16:00:24 <ddieterly> ciao
16:00:26 <mroderus> thanks, bye!
16:00:35 <Kamil___> Bye
16:00:40 <witek> bye
16:00:41 <rhochmuth> #endmeeting