fluid-work IRC Logs-2014-01-12
[10:02:06 CST(-0600)] <colinclark> Bosmon: So reproducing the issue with multiple matchmakers
[10:02:10 CST(-0600)] <colinclark> I'm still a bit behind on today's emails
[10:02:15 CST(-0600)] <colinclark> so I should double check
[10:02:25 CST(-0600)] <colinclark> but as I saw it yesterday, we were throwing an exception somewhere in the stack
[10:02:29 CST(-0600)] <colinclark> unable to fire an event
[10:02:37 CST(-0600)] <Bosmon> yes
[10:02:43 CST(-0600)] <colinclark> I was able to reproduce it last night simply by switching to the .multipleMatchmakers configuration
[10:03:05 CST(-0600)] <colinclark> We just need to update the configuration-wherever it is-for the Flow Manager so that it points to the versions of the infrastructure that Avtar has deployed
[10:03:11 CST(-0600)] <colinclark> probably simplest to start with the STMM
[10:03:20 CST(-0600)] <colinclark> which is deployed at stmm.gpii.net on port 80
[10:03:54 CST(-0600)] <Bosmon> Ok, so in theory I could test this out without having to do it in a VM?
[10:06:09 CST(-0600)] <colinclark> Yes
[10:06:23 CST(-0600)] <Bosmon> Ok - can you let me know what I should do?
[10:06:54 CST(-0600)] <colinclark> I guess the first step is to find where the configurations for the Matchmaker URLs are
[10:07:06 CST(-0600)] <colinclark> and also find a preference set in the testData folder that requires the STMM
[10:10:05 CST(-0600)] <Bosmon> It appears the URLs are in proxy.json
[10:10:29 CST(-0600)] <Bosmon> But I don't see it has an entry for the stmm
[10:10:43 CST(-0600)] <Bosmon> At least the on in trunk
[10:10:53 CST(-0600)] <Bosmon> Is this in a branch?
[10:13:07 CST(-0600)] <colinclark> no
[10:13:11 CST(-0600)] <colinclark> I suspect Kasper configured it himself
[10:13:21 CST(-0600)] <colinclark> we'll presumably need to add entries for the other MMs
[10:13:24 CST(-0600)] <colinclark> + conditionPath: "catsAreDecent",
[10:13:27 CST(-0600)] <Bosmon> And I guess he is now off travelling .......
[10:13:40 CST(-0600)] <colinclark> Very close, anyway
[10:15:51 CST(-0600)] <colinclark> I'm just hopping back into my VM so I can follow along, Bosmon
[10:15:55 CST(-0600)] <colinclark> where is proxy.json?
[10:15:58 CST(-0600)] <Bosmon> So, at a rough guess, what we need to do is add in a new user mirroring the existing GPII-270-rbmm-demo.json
[10:16:04 CST(-0600)] <Bosmon> Only with an entry for a different matchmaker
[10:16:10 CST(-0600)] <Bosmon> and then add a corresponding entry into proxy.json
[10:16:31 CST(-0600)] <colinclark> yes
[10:16:35 CST(-0600)] <Bosmon> universal/gpii/node_modules/matchMaker/configs/proxy.json
[10:16:43 CST(-0600)] <Bosmon> Don't you have a facility to hit the F8 key in your editor? : P
[10:17:05 CST(-0600)] <colinclark> I'm completely lost in Sublime Text
[10:17:14 CST(-0600)] <colinclark> I don't even know how to search across the whole project
[10:17:16 CST(-0600)] <michelled> btw review2_demo2.json is a preference that asks for STMM
[10:17:31 CST(-0600)] <colinclark> thanks, fish
[10:17:32 CST(-0600)] <colinclark> that's great
[10:17:38 CST(-0600)] <Bosmon> michelled - I don't have that one
[10:17:41 CST(-0600)] <Bosmon> Which branch is it in?
[10:17:50 CST(-0600)] <michelled> it's in the review2 branch
[10:17:59 CST(-0600)] <Bosmon> ok
[10:18:08 CST(-0600)] <Bosmon> Just for reference, is there a pull request which put it there?
[10:18:15 CST(-0600)] <Bosmon> Or has it just been applied blind to the branch
[10:18:30 CST(-0600)] <colinclark> Kasper was issuing large pulls against review2
[10:18:35 CST(-0600)] <colinclark> and immediately pushing them
[10:18:42 CST(-0600)] <Bosmon> Oh yes
[10:18:46 CST(-0600)] <Bosmon> I saw a lot of that happening
[10:18:47 CST(-0600)] <michelled> this is the pull to universal: https://github.com/GPII/universal/pull/181
[10:18:51 CST(-0600)] <colinclark> Are you cool to give it a shot, Bosmon, while I continue with the slow demo preparations?
[10:18:53 CST(-0600)] <colinclark> nicely done, michelled
[10:18:55 CST(-0600)] <Bosmon> colinclark - yes
[10:18:57 CST(-0600)] <colinclark> thank you
[10:19:59 CST(-0600)] <michelled> Bosmon: if it helps, kasparnet has been tracking his pulls to review2 on this JIRA: http://issues.gpii.net/browse/GPII-505
[10:20:09 CST(-0600)] <Bosmon> Thanks, michelled
[10:20:12 CST(-0600)] <michelled> np
[10:20:27 CST(-0600)] <colinclark> The biggest thing I'm worried about, in addition to web sockets, is Avtar's email
[10:20:29 CST(-0600)] <colinclark> which I don't undersatnd
[10:20:47 CST(-0600)] <Bosmon> which one, colinclark
[10:20:50 CST(-0600)] <Bosmon> I don't see one recently
[10:21:12 CST(-0600)] <colinclark> it's off list
[10:21:16 CST(-0600)] <colinclark> i'll forward it to you
[10:21:49 CST(-0600)] <Bosmon> michelled - I'm noticing that the way that review2_demo2.json refers to its matchmaker type is completely different to the way that GPII-270-rbmm-demo.json refers to its one
[10:22:48 CST(-0600)] <michelled> Bosmon: kasparnet has not tested these preference sets yet, so there's a good change they aren't quite right
[10:22:54 CST(-0600)] <michelled> kasparnet: are you there?
[10:23:05 CST(-0600)] <Bosmon> michelled - well presumably he has tested them in some sense in order to paste his exception message
[10:23:13 CST(-0600)] <Bosmon> But presumably the versions he has tested are not these ones
[10:23:30 CST(-0600)] <michelled> I didn't double check that with him this morning, so I'm not sure
[10:23:30 CST(-0600)] <Bosmon> It would have helped a lot, of course, if he could have communicated what preferences he had tested in order to get his exception
[10:23:45 CST(-0600)] <kasparnet> michelled, Bosmon: i'm here - but gotta leave in 5 mins to go to the airport - but I'll be online there again (should take me 45 mins or so)
[10:23:57 CST(-0600)] <Bosmon> Similarly, if andreas had communicated what means he had used to verify that the stmm is working
[10:24:16 CST(-0600)] <michelled> quick question kasparnet - what preferences are you using when testing STMM and RBMM?
[10:24:21 CST(-0600)] <Bosmon> kasparnet - shall I just add a new MM entry in the proxy.json file like the current one?
[10:24:22 CST(-0600)] <kasparnet> yeah - I'm in the process of testing the last backup sets (bak_*)
[10:24:24 CST(-0600)] <Bosmon> Is that what you did?
[10:24:36 CST(-0600)] <michelled> also, do you have any uncommitted changes in your working directory?
[10:24:56 CST(-0600)] <kasparnet> michelled: no uncommitted changes no
[10:25:24 CST(-0600)] <Bosmon> kasparnet - how did you test the stmm?
[10:25:25 CST(-0600)] <kasparnet> Bosmon, michelled: I tried getting the damn proxy to work with the MMs (running everything but the MMs locally) but got an error message that I couldn't figure out
[10:25:32 CST(-0600)] <Bosmon> kasparnet - we know
[10:25:34 CST(-0600)] <Bosmon> What did you do!
[10:25:38 CST(-0600)] <kasparnet> Bosmon: only testing I did was with the flat MM
[10:25:43 CST(-0600)] <Bosmon> ah
[10:25:48 CST(-0600)] <Bosmon> You got the error even without the stmm configured?
[10:25:53 CST(-0600)] <kasparnet> which is our backup plan
[10:26:20 CST(-0600)] <Bosmon> Please say in your last 3 minutes exactly what you did to receive the error
[10:26:24 CST(-0600)] <kasparnet> well - in my attempt I added a "statistical" entry in the proxy.json file under the matchmaker/config dir
[10:26:27 CST(-0600)] <kasparnet> ok
[10:29:07 CST(-0600)] <kasparnet> so I tried to set it to use this config file: https://github.com/GPII/universal/blob/master/gpii/configs/fm.ps.sr.dr.mm.os.lms.multipleMatchMakers.json
[10:29:32 CST(-0600)] <kasparnet> (under windows that would be: SET NODE_ENV=fm.ps.sr.dr.mm.os.lms.multipleMatchMakers)
[10:29:45 CST(-0600)] <kasparnet> in the same console I run the node gpii.js command
[10:30:43 CST(-0600)] <kasparnet> and added the statistical MM (and modified RB MM) entries in this file https://github.com/GPII/universal/blob/master/gpii/node_modules/matchMaker/configs/proxy.json#L15
[10:30:49 CST(-0600)] <kasparnet> and that should've worked
[10:30:55 CST(-0600)] <kasparnet> ... except it didnt
[10:31:40 CST(-0600)] <kasparnet> and then I tried doing some debugging to figure out where it went wrong, but got lost
[10:32:28 CST(-0600)] <kasparnet> also tried running the a local version of the cloudbased FM and connect to a proxied MM pointing to the online MMs - but with just as little success
[10:32:57 CST(-0600)] <kasparnet> (following the instructions on this page: http://wiki.gpii.net/index.php/Flow_Manager_API)
[10:33:07 CST(-0600)] <kasparnet> Bosmon, michelled ^
[10:33:37 CST(-0600)] <Bosmon> kasparnet - thanks
[10:33:41 CST(-0600)] <Bosmon> So I have just tested this out now
[10:33:56 CST(-0600)] <Bosmon> And the universal logic seems to work
[10:34:02 CST(-0600)] <Bosmon> But in fact I get an ECONNREFUSED from http://stmm.gpii.net:80/match
[10:35:05 CST(-0600)] <Bosmon> colinclark - do you have any evidence that the stmm is in fact deployed, and what form does it take?
[10:35:10 CST(-0600)] <kasparnet> hmm... dont think I saw any ECONNREFUSED on my end
[10:35:20 CST(-0600)] <kasparnet> anyway, I really gotta run now.. will reconnect asap
[10:39:29 CST(-0600)] <Bosmon> yzen !!!
[10:39:35 CST(-0600)] <Bosmon> You just missed KASPARNET by 2 minutes
[10:39:42 CST(-0600)] <yzen> Bosmon: hi
[10:39:43 CST(-0600)] <Bosmon> We are trying to get the STMM working
[10:39:46 CST(-0600)] <colinclark> It's the great YURA
[10:40:06 CST(-0600)] <Bosmon> kasparnet claims that he gets an exception during event dispatching in Kettle middleware
[10:40:11 CST(-0600)] <Bosmon> Which so far I'm unable to reproduce
[10:40:25 CST(-0600)] <yzen> Bosmon: so does it mean it is actually working for you ?
[10:40:27 CST(-0600)] <Bosmon> I do, however, get an ECONNREFUSED from the real STMM at http://stmm.gpii.net/match
[10:40:42 CST(-0600)] <Bosmon> yzen - I am so far unable to demonstrate it working since I don't see that the STMM is really online
[10:41:31 CST(-0600)] <yzen> well it's hard to see what's going on without the access to logs or debugger, it looks like node is runnin
[10:41:36 CST(-0600)] <yzen> but not handling the match url
[10:41:44 CST(-0600)] <Bosmon> Apparently avtar is on his way shortly
[10:42:29 CST(-0600)] <yzen> Bosmon: do you want me to try to deploy it locally ?
[10:43:46 CST(-0600)] <Bosmon> yzen - locally for you, or for me
[10:43:54 CST(-0600)] <Bosmon> Do we have instructions on how to deploy it locally?
[10:44:05 CST(-0600)] <yzen> for me
[10:44:15 CST(-0600)] <yzen> well is it harder than the rule based ?
[10:44:42 CST(-0600)] <Bosmon> I guess it would help - colinclark, do you think that would be a good use of yzen's time, or should we try to deal with the socket.io issues until avtar arrives?
[10:44:54 CST(-0600)] <Bosmon> Do we even have enough information on Boyan about his socket.io problems?
[10:45:56 CST(-0600)] <yzen> Bosmon: well it looks like it does not work for some reason , the server doesn't even get hit when i try to open a socket
[10:46:03 CST(-0600)] <yzen> but still trying to investigate
[10:46:35 CST(-0600)] <Bosmon> Ok
[10:46:38 CST(-0600)] <Bosmon> that seems more urgent then
[10:46:54 CST(-0600)] <yzen> ok
[10:47:33 CST(-0600)] <Bosmon> So far the balance of evidence seems to be that our cloud version of the STMM isn't properly installed, so we'll wait for avtar
[10:47:48 CST(-0600)] <yzen> another thing i found is that for some reason cookie is not set when logging in even though the header is not there, would anybody know if firefox for example even let you save a cookie for localhost ?
[10:48:27 CST(-0600)] <Bosmon> yzen - we've been able to save cookie for localhost in general, since a lot of our UIO demos have relied on this
[10:48:38 CST(-0600)] <Bosmon> But there may be oddities with it
[10:48:40 CST(-0600)] <yzen> Bosmon: you mean from the client side
[10:48:45 CST(-0600)] <yzen> not through the header
[10:48:53 CST(-0600)] <Bosmon> For example I think that every localhost URL corresponds to a different "domain" from its point of view
[10:49:22 CST(-0600)] <Bosmon> yzen - can you explain what you mean?
[10:49:34 CST(-0600)] <yzen> yes but again , i m fairly certain you can't set cookies to localhost in chrome
[10:49:53 CST(-0600)] <yzen> http://code.google.com/p/chromium/issues/detail?id=56211
[10:50:16 CST(-0600)] <yzen> so I'm wondering if it's the same case for firefox
[10:51:11 CST(-0600)] <Bosmon> I believe firefox has different rules
[10:51:39 CST(-0600)] <Bosmon> But I think you can achieve consistent behaviour on both of them by using the workaround at the base of the Chrome issue report
[10:51:43 CST(-0600)] <Bosmon> I.e. a fake hosts entry
[10:51:53 CST(-0600)] <Bosmon> That would also be valuable for testing real CORS support : P
[10:52:17 CST(-0600)] <yzen> Bosmon: yep
[10:52:27 CST(-0600)] <Bosmon> Welcome, avtar!
[10:52:44 CST(-0600)] <Bosmon> We're having problems verifying that the cloud-based stmm is working correctly
[10:53:05 CST(-0600)] <Bosmon> In particular both http://stmm.gpii.net/match and http://stmm.gpii.net/ are giving ECONNREFUSED on port 80
[10:53:27 CST(-0600)] <avtar> hmm
[10:53:29 CST(-0600)] <avtar> one sec
[10:54:19 CST(-0600)] <colinclark> hi, sorry, i was just eating pancakes
[10:54:28 CST(-0600)] <colinclark> but it sounds like you guys are rocking it
[10:55:21 CST(-0600)] <Bosmon> I am eating "SRIRACHAMOLE (TM)"
[10:55:25 CST(-0600)] <Bosmon> It is a little intense ......
[10:55:52 CST(-0600)] <colinclark> what is it?
[10:56:00 CST(-0600)] <colinclark> something made with sriracha for breakfast?
[10:56:18 CST(-0600)] <Bosmon> yes
[10:56:28 CST(-0600)] <Bosmon> The stuff that is not sriracha appears to be guacamole : P
[10:56:38 CST(-0600)] <michelled> !
[10:57:10 CST(-0600)] <Bosmon> Urgh.... andreas is saying that he can reproduce kaspar's error
[10:57:43 CST(-0600)] <colinclark> I was able to see it yesterday, too
[10:57:58 CST(-0600)] <colinclark> But I'm sort of assuming it has to do with an ill-configured proxy.json
[10:58:24 CST(-0600)] <Bosmon> You mean, if it was well-configured, they would receive an ECONNREFUSED? : P
[11:00:39 CST(-0600)] <colinclark> yes, I suppose so
[11:04:58 CST(-0600)] <avtar> Bosmon: http://cloudshark.org/captures/ab498364e746
[11:05:24 CST(-0600)] <avtar> click anywhere and then select analysis tools > follow stream
[11:05:42 CST(-0600)] <avtar> that's a tcpdump capture on the vm where stmm is running
[11:06:21 CST(-0600)] <avtar> what type of request are you using?
[11:06:47 CST(-0600)] <Bosmon> Well, right now, I just type http://stmm.gpii.net/match into a browser and see what happens
[11:06:51 CST(-0600)] <Bosmon> What happens when you do the same?
[11:09:01 CST(-0600)] <Bosmon> Well, I appear to have a different IP resolved for that name than the one that appears in that capture log
[11:09:25 CST(-0600)] <Bosmon> I currently have Non-authoritative answer:
[11:09:25 CST(-0600)] <Bosmon> Name: stmm.gpii.net
[11:09:25 CST(-0600)] <Bosmon> Addresses: 205.211.169.180
[11:09:25 CST(-0600)] <Bosmon> 205.211.169.181
[11:10:10 CST(-0600)] <avtar> 180 and 181 are nginx instances
[11:10:14 CST(-0600)] <Bosmon> ok
[11:10:33 CST(-0600)] <avtar> i still see a 404
[11:10:43 CST(-0600)] <Bosmon> Yup
[11:11:04 CST(-0600)] <Bosmon> Me too
[11:11:18 CST(-0600)] <Bosmon> Sorry I was reporting ECONNREFUSED before, I was just reporting what node.js reported
[11:11:26 CST(-0600)] <Bosmon> But from the browser I see a 404
[11:11:38 CST(-0600)] <Bosmon> So - what kind of problem do we have?
[11:12:06 CST(-0600)] <Bosmon> So far I've seen no new activity in the tcpdump
[11:12:14 CST(-0600)] <Bosmon> So I assume that the nginx instances are broken
[11:13:36 CST(-0600)] <avtar> i should have mentioned that's not a live running capture
[11:13:58 CST(-0600)] <avtar> i ran tcpdump and uploaded the captured data so that it could be rendered using a browser
[11:14:15 CST(-0600)] <Bosmon> Ok
[11:14:18 CST(-0600)] <Bosmon> How does this help us?
[11:14:50 CST(-0600)] <avtar> i guess it's more so to rule out connection issues leading to the app
[11:15:19 CST(-0600)] <Bosmon> At the moment it's unclear to me even what paths the server is meant to be listening on
[11:15:24 CST(-0600)] <Bosmon> Can you shed any light on this?
[11:15:29 CST(-0600)] <Bosmon> Should /match be a valid URL?
[11:16:23 CST(-0600)] <avtar> i can't comment on stmm's functionality
[11:17:04 CST(-0600)] <avtar> it would help if someone could provide more info on what should be happening with the app so that i can troubleshoot the deployment
[11:19:15 CST(-0600)] <avtar> the log it's generated so far just says 'function'
[11:21:00 CST(-0600)] <avtar> brb, looking into michelled's pref server request
[11:21:44 CST(-0600)] <colinclark> kasparnet's back! Are you at the airport now?
[11:22:31 CST(-0600)] <kasparnet> not yet.. in the train
[11:22:35 CST(-0600)] <colinclark> cool
[11:22:53 CST(-0600)] <colinclark> it seems like avtar and Bosmon and yzen are starting to tackle the Matchmaker issues
[11:23:04 CST(-0600)] <kasparnet> ooh, glad to hear that
[11:23:08 CST(-0600)] <colinclark> so hopefully we'll have something working by the time you alnd
[11:23:27 CST(-0600)] <kasparnet> perfect - that'll give me some time to try out some of the NP sets
[11:23:30 CST(-0600)] <colinclark> kasparnet: What do you think are the odds of Gregg and Manuel being willing to do the dress rehearsal for the slow demo on Tuesday?
[11:24:07 CST(-0600)] <kasparnet> colinclark: I think they're pretty decent ... if they know we're pressed with the demos
[11:24:25 CST(-0600)] <colinclark> Can you mention it to Gregg if you see him tonight?
[11:24:32 CST(-0600)] <kasparnet> yup, I will
[11:24:35 CST(-0600)] <colinclark> I'll send him a note when I know more about how well I'm progressing
[11:24:46 CST(-0600)] <colinclark> tell him I'm deathly ill, making progress, but need some time
[11:24:58 CST(-0600)] <colinclark> fortunately I'm not actually deathly ill
[11:24:58 CST(-0600)] <kasparnet> yup will do
[11:25:04 CST(-0600)] <colinclark> thanks dude
[11:25:08 CST(-0600)] <colinclark> good luck
[11:25:14 CST(-0600)] <kasparnet> "gregg... Colin is dying"
[11:25:17 CST(-0600)] <colinclark> lol
[11:25:21 CST(-0600)] <kasparnet> then I'll leave that hanging for a few minutes
[11:25:23 CST(-0600)] <colinclark> yes
[11:25:24 CST(-0600)] <colinclark> do that
[11:25:30 CST(-0600)] <kasparnet> hehe
[11:25:55 CST(-0600)] <colinclark> safe travels, kasparnet
[11:26:05 CST(-0600)] <colinclark> and thank you for your dogged persistence
[11:26:10 CST(-0600)] <colinclark> catted persistence
[11:27:42 CST(-0600)] <Bosmon> avtar: I've created a pastie holding the HTTP request and response, as seen from the app end
[11:27:44 CST(-0600)] <Bosmon> http://pastie.org/8626851
[11:27:49 CST(-0600)] <Bosmon> Is this the info you were asking for?
[11:33:16 CST(-0600)] <avtar> Bosmon: no, i meant it would be helpful if someone could run the stmm locally and issue a request against it so that i could use that as a reference for its intended behaviour and troubleshoot the cloud deployment
[11:33:24 CST(-0600)] <Bosmon> pl
[11:33:26 CST(-0600)] <Bosmon> ok
[11:33:35 CST(-0600)] <Bosmon> perhaps someone can point me to instructions on how I can run the stmm locally
[11:33:42 CST(-0600)] <colinclark> Bosmon: It's just a node app
[11:33:48 CST(-0600)] <colinclark> clone it, npm install it, and then run it
[11:34:03 CST(-0600)] <avtar> i used colinclark's repo https://github.com/colinbdclark/GPII_StatisticalMatchmaker1
[11:34:24 CST(-0600)] <colinclark> updating your proxy.json config to point at localhost:8077
[11:34:40 CST(-0600)] <Bosmon> Thanks, colinclark and avtar
[11:34:40 CST(-0600)] <avtar> Bosmon: https://github.com/colinbdclark/GPII_StatisticalMatchmaker1/tree/GPII-275
[11:34:46 CST(-0600)] <avtar> that branch
[11:34:57 CST(-0600)] <colinclark> Either one is fine, in that tI accidentally committed to master first
[11:35:10 CST(-0600)] <colinclark> I believe Andy has merged my pull request into his repo, too
[11:35:31 CST(-0600)] <colinclark> Bosmon: Did you add extra logging statements to get this payload output?
[11:35:42 CST(-0600)] <Bosmon> colinclark - I did, yes
[11:38:51 CST(-0600)] <avtar> yzen: are you free to test flowmanager?
[11:39:08 CST(-0600)] <yzen> you mean sockets?
[11:39:19 CST(-0600)] <avtar> in its entirety
[11:39:37 CST(-0600)] <yzen> oh you mean the cloud based one ?
[11:39:40 CST(-0600)] <avtar> i'm assuming stmm relies on flowmanager as well
[11:39:43 CST(-0600)] <avtar> yup
[11:40:07 CST(-0600)] <yzen> yes i can help out with that
[11:40:17 CST(-0600)] <avtar> hugs
[11:58:28 CST(-0600)] <Bosmon> I've been running colinclark's GPII-275 branch of the STMM
[11:58:32 CST(-0600)] <Bosmon> And get the following errors:
[11:58:33 CST(-0600)] <Bosmon> http://pastie.org/8626927
[11:58:42 CST(-0600)] <Bosmon> It seems there is an undefined invoker at the STMM end
[11:58:48 CST(-0600)] <Bosmon> And the logs from that end are quite worrying
[11:58:54 CST(-0600)] <Bosmon> There's all kinds of irrelevant-seeming universal-like material
[11:59:13 CST(-0600)] <Bosmon> For example, an onOntologyServer call
[11:59:49 CST(-0600)] <Bosmon> And userLogin and userUpdate material
[11:59:56 CST(-0600)] <colinclark> yzen: ^
[12:00:13 CST(-0600)] <Bosmon> Did someone test this last night as working?
[12:00:28 CST(-0600)] <Bosmon> I'll look into what the undefined invoker is now
[12:00:32 CST(-0600)] <colinclark> No, no one has ever tested this
[12:00:38 CST(-0600)] <Bosmon> But I imagine there is still some kind of version mismatch
[12:00:47 CST(-0600)] <colinclark> Andy confirmed in email that he's never tested the app version of the STMM
[12:00:55 CST(-0600)] <colinclark> and the best we were able to do last night is get the thing running
[12:00:58 CST(-0600)] <Bosmon> yzen - have you ever tested a version of this repository?
[12:01:10 CST(-0600)] <colinclark> you've pushed further than anyone except yzen in this regard, Bosmon
[12:01:33 CST(-0600)] <yzen> Bosmon: no i have not touched stmm
[12:03:50 CST(-0600)] <colinclark> is there an example of a similarly-structured matchmaker that you have tested, yzen?
[12:03:58 CST(-0600)] <colinclark> so that we can compare?
[12:04:08 CST(-0600)] <yzen> colinclark: well rule based match maker afaik works ?
[12:04:19 CST(-0600)] <colinclark> They say it works, yes
[12:04:26 CST(-0600)] <colinclark> I don't believe avtar's quite gotten to a point where he can deploy it
[12:04:33 CST(-0600)] <colinclark> though we can try it on the server running in Greece
[12:06:46 CST(-0600)] <colinclark> Bosmon: Do you want to try against the RBMM running in Greece?
[12:07:37 CST(-0600)] <colinclark> It's apparently running here: http://160.40.50.130/
[12:07:40 CST(-0600)] <colinclark> on port 80
[12:08:00 CST(-0600)] <yzen> colinclark: well from what i can tell the statistical match maker code is not right
[12:08:00 CST(-0600)] <colinclark> So you'll just have to configure it in proxy.json and use one of Kasper's RBMM-based sets
[12:08:12 CST(-0600)] <colinclark> How so, yzen?
[12:08:26 CST(-0600)] <Bosmon> I guess that's possible, given it provides exactly the same output (i.e. empty) to GET requests as avtar's instance
[12:09:02 CST(-0600)] <Bosmon> I will work on the assumption that the payload sent to it by the review2 universal is correct and try prodding it with CURL
[12:09:15 CST(-0600)] <yzen> sure so the most basic thing i see is here: https://github.com/colinbdclark/GPII_StatisticalMatchmaker1/blob/GPII-275/lib/StatisticalMatchMaker.js#L10-L14 this invoker has args that are not what they supposed to be based on the function definition: https://github.com/colinbdclark/GPII_StatisticalMatchmaker1/blob/GPII-275/lib/StatisticalMatchMaker.js#L48
[12:10:15 CST(-0600)] <yzen> this is most likely a copy of the rbmm example but it is not what the stats matchmaker needs from what i can see
[12:10:31 CST(-0600)] <colinclark> Yes, Andy mentioned completely copy-pasting the RBMM implementation
[12:10:37 CST(-0600)] <colinclark> and said "I don't know what it means"
[12:12:18 CST(-0600)] <colinclark> yzen: Is it much work to fork his repo and fix it up?
[12:12:22 CST(-0600)] <yzen> no
[12:12:26 CST(-0600)] <colinclark> phew
[12:12:27 CST(-0600)] <Bosmon> colinclark - I've verified that the RBMM in Greece is working, if slow
[12:12:31 CST(-0600)] <colinclark> yes
[12:12:39 CST(-0600)] <colinclark> that's why I wanted to ensure that avtar deployed a copy
[12:13:02 CST(-0600)] <Bosmon> If we can get the STMM working locally, the ground is then cleared for avtar to get the cloud version working
[12:13:10 CST(-0600)] <colinclark> yep
[12:13:24 CST(-0600)] <avtar> it's almost there
[12:13:30 CST(-0600)] <avtar> waiting for npm to finish doing its thing
[12:13:47 CST(-0600)] <Bosmon> We're thinking, I think, of setting up an institutional mirror for npm
[12:13:51 CST(-0600)] <Bosmon> Or I could just try doing it privately
[12:13:58 CST(-0600)] <Bosmon> Since it seems the team collectively wastes more time on this than anything else
[12:14:37 CST(-0600)] <Bosmon> I'll hop in the shower whilst your npm finishes
[12:14:50 CST(-0600)] <colinclark> Ok, so the evidence is that the RBMM is indeed working
[12:14:58 CST(-0600)] <colinclark> and avtar, you're close to having it deployed on our servers?
[12:15:21 CST(-0600)] <colinclark> and yzen, you're going to fork the STMM and fix the erroneous binding to .match()
[12:15:26 CST(-0600)] <colinclark> is that correct?
[12:15:28 CST(-0600)] <avtar> colinclark: nick's node and java app have been deployed
[12:15:41 CST(-0600)] <avtar> i'm deploying review2 on the preferences server
[12:15:41 CST(-0600)] <colinclark> wonderful
[12:15:43 CST(-0600)] <colinclark> ah, great
[12:16:04 CST(-0600)] <yzen> colinclark: yes
[12:16:09 CST(-0600)] <avtar> any thoughts on the matchmaker port issue that i mentioned in my email?
[12:16:13 CST(-0600)] <colinclark> so Bosmon should be able to equally "prod" http://rbmm.gpii.net with curl and see the same result when he tried with the server in Greece?
[12:16:18 CST(-0600)] <colinclark> I didn't understand it, avtar
[12:16:20 CST(-0600)] <colinclark> can you elaborate?
[12:16:31 CST(-0600)] <avtar> i need to make matchmaker use a port other than 8080
[12:16:34 CST(-0600)] <avtar> that it defaults to
[12:16:38 CST(-0600)] <colinclark> Which Matchmaker?
[12:16:47 CST(-0600)] <colinclark> the Java side of the RBMM?
[12:17:43 CST(-0600)] <avtar> no, the o.g. matchmaker that flowmanager uses
[12:17:44 CST(-0600)] <avtar> https://github.com/GPII/universal/tree/master/gpii/node_modules/matchMaker
[12:19:07 CST(-0600)] <avtar> this is the config that it's been using since it was first deployed for kasper and yzen https://github.com/GPII/universal/blob/master/gpii/node_modules/matchMaker/configs/development.json
[12:19:36 CST(-0600)] <yzen> avtar: you are talking about the cloud based flow manager right ?
[12:19:38 CST(-0600)] <colinclark> avtar: You should be able to override the port it's running on here: https://github.com/GPII/universal/blob/master/gpii/node_modules/matchMaker/configs/development.json#L9
[12:19:49 CST(-0600)] <colinclark> is that correct, yzen?
[12:20:03 CST(-0600)] <Bosmon> avtar's version of the Greek RBMM I can verify as not working
[12:20:04 CST(-0600)] <avtar> yzen: yes, the cloud one
[12:20:06 CST(-0600)] <yzen> yes , avtar similar to what we did ..
[12:20:11 CST(-0600)] <Bosmon> It's behaviour is the same as his STMM one
[12:20:18 CST(-0600)] <Bosmon> That is, we get an ECONNREFUSED error in a payload
[12:20:35 CST(-0600)] <Bosmon> So I believe it must be that the nginx layer is misconfigured in the same way
[12:20:43 CST(-0600)] <yzen> avtar: now what i m wondering i guess is do we expect the cloud based flow manager to work with the proxy ?
[12:20:54 CST(-0600)] <yzen> colinclark ^
[12:21:13 CST(-0600)] <yzen> or proxy demo is separate ?
[12:22:14 CST(-0600)] <colinclark> I suppose kasparnet or michelled might well be able to answer that question from the functional perspecitve--i.e. based on what the preference sets actually look like
[12:22:30 CST(-0600)] <colinclark> but I was assuming that the Cloud-based Flow Manager would work the same as the local Flow Manager
[12:22:42 CST(-0600)] <colinclark> i.e. that it would be capable of making requests to any of the three Matchmakers
[12:23:07 CST(-0600)] <colinclark> Do you think that is going to be difficult?
[12:24:00 CST(-0600)] <yzen> colinclark: well it is not , we just need to make sure that the matchmaker for the cloud based one is deployed as a proxy rather than in dev mode. (avtar ^)
[12:27:11 CST(-0600)] <colinclark> Will that require further config changes?
[12:30:36 CST(-0600)] <colinclark> I'm assuming, yzen, that this might involve some kind of merging of the cloudBased.json config with fm.ps.sr.dr.mm.os.lms.multipleMatchmakers.json?
[12:32:39 CST(-0600)] <yzen> colinclark: well not really, and avtar can confirm, we deploy cloud based flow manager and the match maker as 2 separate server (with independent configs). The flow manager points to the match maker and that's all it knows about it. Now if we deploy the match maker with correct proxy.json and not development.json that would do, it will just proxy to the real ones.
[12:33:13 CST(-0600)] <colinclark> Once you've had the chance to fix the STMM bug, can you adjust that config?
[12:33:13 CST(-0600)] <avtar> confirmed ^
[12:33:17 CST(-0600)] <colinclark> cool
[12:33:27 CST(-0600)] <colinclark> We're actually getting somewhere here
[12:33:29 CST(-0600)] <colinclark> thanks yzen
[12:34:35 CST(-0600)] <yzen> colinclark: well only the ips for the match makers would need to be updated so if avtar needs help with that (he'd know the actual ips), i m here
[12:34:57 CST(-0600)] <colinclark> avtar: Does that make sense to you, and do you know what you have to do?
[12:35:14 CST(-0600)] <avtar> yzen: so the config is getting changed to proxy.json?
[12:35:38 CST(-0600)] <yzen> avtar: yes for the alongside match maker yes
[12:36:33 CST(-0600)] <colinclark> Just so I'm clear
[12:36:39 CST(-0600)] <colinclark> are we changing the contents of proxy.json?
[12:36:49 CST(-0600)] <colinclark> or change some configuration TO proxy.json?
[12:37:08 CST(-0600)] <yzen> colinclark: yes we are slightly changing the contents of proxy.json only
[12:37:30 CST(-0600)] <colinclark> avtar: make sense? ^
[12:37:34 CST(-0600)] <yzen> changes related to the correct ips
[12:37:58 CST(-0600)] <yzen> colinclark: is there a request i can make to the local stats matchmaker to see if it works ?
[12:38:31 CST(-0600)] <colinclark> yes, Bosmon had a pastie with the request that he logged from the FM
[12:38:34 CST(-0600)] <avtar> colinclark, yzen: the original matchmaker has not been in the list of components that's been redeployed using the review2 branch. should that be included now as well?
[12:39:07 CST(-0600)] <colinclark> http://pastie.org/8626927
[12:39:23 CST(-0600)] <avtar> yzen: we used your repo for flowmanager. do we now switch to the official gpii/universal repo with the review2 branch?
[12:40:06 CST(-0600)] <yzen> avtar: that stuff is not in master of GPII/universal
[12:40:38 CST(-0600)] <yzen> colinclark: right did you do it with curl?
[12:41:06 CST(-0600)] <yzen> i was just wondering if you have the line saved in your terminal's history or something
[12:41:08 CST(-0600)] <colinclark> yes, Bosmon used the information in that pastie to construct a payload and sent it via curl
[12:41:15 CST(-0600)] <colinclark> It wasn't me, unfortunately
[12:41:24 CST(-0600)] <colinclark> he'll be back soonish
[12:42:30 CST(-0600)] <yzen> ok
[12:42:35 CST(-0600)] <michelled> so yzen, is there something related to the flow manager that needs to get into the review2 branch?
[12:43:10 CST(-0600)] <yzen> michelled: I'm not sure
[12:43:21 CST(-0600)] <yzen> you mean the cloud based flow manager?
[12:44:17 CST(-0600)] <colinclark> "yzen: avtar: that stuff is not in master of GPII/universal"
[12:44:20 CST(-0600)] <colinclark> what did you mean by that?
[12:44:26 CST(-0600)] <colinclark> that's what michelled is asking, I think
[12:44:27 CST(-0600)] <michelled> yes, it unnerves me that we have deployed your repo instead of the review2 repo. I'm hoping we can get to the point where there is a well known and consistent set of branches in github
[12:47:21 CST(-0600)] <yzen> colinclark: s/now/not
[12:47:25 CST(-0600)] <yzen> not/now
[12:47:30 CST(-0600)] <colinclark> oh
[12:47:32 CST(-0600)] <colinclark> phew
[12:47:47 CST(-0600)] <yzen> colinclark: i think the match maker is all good now
[12:47:51 CST(-0600)] <yzen> i sent you a pull
[12:47:54 CST(-0600)] <michelled> ah great
[12:48:02 CST(-0600)] <colinclark> So it sounds like we should indeed be deploying off the review2 branch
[12:48:02 CST(-0600)] <colinclark> thanks, yzen
[12:48:23 CST(-0600)] <avtar> so to summarize
[12:49:09 CST(-0600)] <avtar> i should be switching flowmanager to gpii/universal's review2 branch
[12:49:31 CST(-0600)] <michelled> yes
[12:49:33 CST(-0600)] <avtar> the same goes for the standard matchmaker in the universal repo
[12:49:46 CST(-0600)] <avtar> but it should now use the proxy config
[12:51:00 CST(-0600)] <colinclark> sound right, yzen? ^
[12:51:19 CST(-0600)] <yzen> colinclark: yes
[12:51:36 CST(-0600)] <colinclark> yzen: Why did you ditch matchPost and its match invoker from the STMM?
[12:51:43 CST(-0600)] <colinclark> Is it just boilerplate he didn't need?
[12:52:08 CST(-0600)] <michelled> avtar: does using the proxy config require code changes?
[12:52:08 CST(-0600)] <yzen> colinclark: i yes from what i could tell his match signature is the same as the default one so he needs no other args
[12:52:28 CST(-0600)] <colinclark> You're sure he doesn't need transformation to occur at the end of the process?
[12:52:59 CST(-0600)] <avtar> michelled: i don't think so
[12:53:32 CST(-0600)] <avtar> michelled: this though http://pastie.org/pastes/8627054/text
[12:53:59 CST(-0600)] <avtar> those preference sets weren't uploaded to couch running on the pref server
[12:54:11 CST(-0600)] <avtar> wonky json
[12:54:31 CST(-0600)] <michelled> ok, I'll fix them up
[12:57:51 CST(-0600)] <avtar> yzen: https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json
[12:58:03 CST(-0600)] <avtar> does this mean matchmaker won't use port 8080 now?
[12:58:33 CST(-0600)] <avtar> or is 8080 still implied and now 8078 and 8079 will also be used?
[12:58:46 CST(-0600)] <yzen> 8080 is default
[12:58:58 CST(-0600)] <avtar> ok
[12:59:01 CST(-0600)] <avtar> that will need to be changed
[12:59:06 CST(-0600)] <yzen> yes 8079 only, flat and rule based depends on where you put them
[12:59:12 CST(-0600)] <yzen> avtar: that should be fine
[12:59:54 CST(-0600)] <colinclark> avtar: Some invalid JSON in kasparnet's preference sets?
[13:00:02 CST(-0600)] <avtar> colinclark: yes
[13:00:15 CST(-0600)] <avtar> yzen: could you please expand on that?
[13:00:37 CST(-0600)] <avtar> so it's going to be 8080 and just 8079?
[13:00:45 CST(-0600)] <avtar> or 8080, 8078, and 8079?
[13:00:59 CST(-0600)] <avtar> ah nm
[13:01:11 CST(-0600)] <avtar> https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json#L15 this is nick's app?
[13:01:31 CST(-0600)] <yzen> avtar: ok so the default for this server: https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json#L10 is 8080
[13:01:44 CST(-0600)] <yzen> if you want to change it you can do it under https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json#L8
[13:01:44 CST(-0600)] <kasparnet> hmmm.. weird - are those from the review2 or master branch?
[13:02:10 CST(-0600)] <yzen> avtar: the default for https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json#L22 is 8079 you can change it right there to anything you want
[13:02:27 CST(-0600)] <avtar> yzen: ok
[13:02:28 CST(-0600)] <yzen> mind you that that port should correspond to https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json#L14
[13:02:43 CST(-0600)] <avtar> colinclark: i feel the port option should be in the config by default ^
[13:02:48 CST(-0600)] <kasparnet> hmmm.. weird - are those from the review2 or master branch? <-- colinclark / avtar
[13:02:50 CST(-0600)] <avtar> could that be included at some point?
[13:03:03 CST(-0600)] <yzen> rule based https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json#L15 should point to wherever you have the rule based matchmaker, local or external. the same goes for the stats match maker
[13:03:06 CST(-0600)] <michelled> I just looked at review 2 and jsonlint finds them valid
[13:03:10 CST(-0600)] <michelled> kasparnet: ^
[13:03:11 CST(-0600)] <avtar> kasparnet: review2
[13:04:33 CST(-0600)] <avtar> yzen: do you mean line 16?
[13:04:58 CST(-0600)] <avtar> sorry, line 14 for the stats matchmaker
[13:05:09 CST(-0600)] <avtar> ?
[13:05:38 CST(-0600)] <yzen> avtar: not sure i understand
[13:05:50 CST(-0600)] <avtar> " the same goes for the stats match maker"
[13:05:55 CST(-0600)] <yzen> stats matcha maker intro is not there yet
[13:06:05 CST(-0600)] <yzen> but i assume we would need to add it along flat and ruleBased
[13:06:24 CST(-0600)] <colinclark> we might as well configure the system with URLs for all three matchmakers
[13:06:27 CST(-0600)] <yzen> colinclark: so you weren't sure about the stats match maker ?
[13:06:27 CST(-0600)] <colinclark> since we know what they are
[13:06:35 CST(-0600)] <Bosmon> In case anyone is interested, I have created a pastie explaining how to quickly test a web-based matchmaker using CURL:
[13:06:36 CST(-0600)] <Bosmon> http://pastebin.com/hckBZxZQ
[13:06:41 CST(-0600)] <colinclark> thank you Bosmon
[13:06:43 CST(-0600)] <colinclark> you are a lifesaver
[13:07:02 CST(-0600)] <colinclark> yzen: I just don't know what Andy actually intends, versus what he has just cut and pasted
[13:07:40 CST(-0600)] <colinclark> I guess we will have to evaluate what kind of a payload it returns with this modification
[13:08:09 CST(-0600)] <yzen> colinclark: well it returns something valid from what i can tell , not sure if stats part does it's job though
[13:12:02 CST(-0600)] <michelled> avtar: as far as I can tell, those json files are valid
[13:12:14 CST(-0600)] <Bosmon> Hi yzen - to me, your pull request still fails - same invoker error as before
[13:12:17 CST(-0600)] <Bosmon> I'll try to get more details
[13:12:44 CST(-0600)] <yzen> Bosmon: can you give me the exact curl line i can try on my machine ?
[13:12:56 CST(-0600)] <Bosmon> yzen: http://pastebin.com/hckBZxZQ
[13:15:45 CST(-0600)] <yzen> Bosmon: works for me
[13:16:15 CST(-0600)] <colinclark> what does the output look like, yzen?
[13:16:23 CST(-0600)] <yzen> Bosmon: this http://pastie.org/8627099
[13:16:26 CST(-0600)] <yzen> colinclark: ^
[13:18:02 CST(-0600)] <colinclark> Ok, do we have any payloads that actually contain settings?
[13:18:25 CST(-0600)] <colinclark> michelled: Do you know of some preference sets that require the STMM?
[13:18:43 CST(-0600)] <Bosmon> I've improved the logging in Kettle so it tells me the following:
[13:18:54 CST(-0600)] <Bosmon> ASSERTION FAILED: Error in invoker record: could not resolve members func, funcName or method to a function implementation
[13:18:54 CST(-0600)] <Bosmon> - got undefined from { funcName: 'gpii.matchMaker.statistical.match',
[13:18:54 CST(-0600)] <Bosmon> args:
[13:18:54 CST(-0600)]
<Bosmon> [ '
.0',
[13:18:54 CST(-0600)]
<Bosmon> '
.1',
[13:18:55 CST(-0600)]
<Bosmon> '
.2',
[13:18:55 CST(-0600)]
<Bosmon> '
.options.inverseRules' ] } { type: 'dispatchListener',
[13:19:16 CST(-0600)] <yzen> Bosmon: how do you deploy the match maker ?
[13:19:41 CST(-0600)] <Bosmon> I run node bin/statisticalMatchMaker.js
[13:20:13 CST(-0600)] <yzen> you are sure you trying with my code, smb ?
[13:20:18 CST(-0600)] <yzen> Bosmon:
[13:20:21 CST(-0600)] <Bosmon> Something is odd though since this message doesn't seem to correspond to the code I have ......
[13:20:23 CST(-0600)] <Bosmon> Yes
[13:20:27 CST(-0600)] <Bosmon> I have your code
[13:20:32 CST(-0600)] <Bosmon> Let me see what is up
[13:20:49 CST(-0600)] <avtar> colinclark: i forget, which instance of infusion did you say should be kept?
[13:20:50 CST(-0600)] <avtar> ./node_modules/kettle/node_modules/infusion
[13:20:50 CST(-0600)] <avtar> ./node_modules/jqUnit/node_modules/infusion
[13:20:50 CST(-0600)] <avtar> ./node_modules/infusion
[13:20:51 CST(-0600)] <yzen> colinclark: i wonder if andy's stuff is out of date where we did not use inverse rules
[13:21:00 CST(-0600)] <colinclark> the top one, avtar
[13:21:03 CST(-0600)] <avtar> ok
[13:21:35 CST(-0600)] <avtar> michelled: one sec
[13:22:34 CST(-0600)] <Bosmon> Ok
[13:22:44 CST(-0600)] <Bosmon> I see that yours and colin's branches refer to different revisions of universal
[13:22:53 CST(-0600)] <Bosmon> Although somehow this didn't show up in the pull request diff
[13:23:44 CST(-0600)] <Bosmon> Which revision of univeral actually is this, yzen? "universal": "git://github.com/GPII/universal.git#a03b5e4512691cad032186434bd3e8151ad23a6d"
[13:24:06 CST(-0600)] <Bosmon> Ok
[13:24:08 CST(-0600)] <colinclark> They look the same to me, Bosmon
[13:24:15 CST(-0600)] <colinclark> that should be the latest universal master
[13:24:19 CST(-0600)] <Bosmon> I'm completely confused now since both of these files have colin as their most recent committer
[13:24:20 CST(-0600)] <colinclark> as of last night
[13:24:39 CST(-0600)] <Bosmon> colinclark - the one in your own GPII-275 branch reads as follows: "universal": "git://github.com/GPII/universal.git#d5bf97ffff988cb7cbe3f0ce40a3bce5b44b639d",
[13:25:03 CST(-0600)] <colinclark> FFFFF
[13:25:18 CST(-0600)] <colinclark> this is such a mess
[13:25:20 CST(-0600)] <colinclark> what did I do?>
[13:25:22 CST(-0600)] <Bosmon> ok, let me revert to yzen's revision and retest
[13:25:22 CST(-0600)] <colinclark> sad;fijasdfai;edf xdvlj
[13:25:56 CST(-0600)] <colinclark> ok
[13:26:03 CST(-0600)] <colinclark> my GPII-275 branch is correct
[13:26:20 CST(-0600)] <Bosmon> urgh
[13:26:23 CST(-0600)] <colinclark> d5bf9 is second-to-latest
[13:26:25 CST(-0600)] <Bosmon> Ok, worse things are happening
[13:26:45 CST(-0600)] <Bosmon> In that I checked out yzen's branch, and yet I got your revision number still
[13:26:46 CST(-0600)] <colinclark> yzen's GPII-275 branch seems consisted with this
[13:26:55 CST(-0600)] <colinclark> are you using his master or his GPII-275 branch?
[13:27:14 CST(-0600)] <Bosmon> Sorry, you're right, sorry
[13:27:16 CST(-0600)] <colinclark> consistent
[13:27:20 CST(-0600)] <Bosmon> I fucked up, both of your rev numbers are the same
[13:27:21 CST(-0600)] <colinclark> phew
[13:27:26 CST(-0600)] <colinclark> sorry for the random characters
[13:27:29 CST(-0600)] <colinclark> it felt good, anyway
[13:27:30 CST(-0600)] <Bosmon> d5bf9 it is
[13:27:31 CST(-0600)] <yzen>
[13:27:48 CST(-0600)] <Bosmon> So this now heightens the mystery of why my test fails and yzen's succeeds
[13:27:58 CST(-0600)] <yzen> colinclark, avtar , Bosmon , michelled what would be the next thing i should take a look at ?
[13:28:09 CST(-0600)] <colinclark> make sure you delete node_modules and npm install again, Bosmon
[13:28:31 CST(-0600)] <colinclark> Ok, so we need to resolve this issue, which it sounds like Bosmon is now looking at
[13:28:35 CST(-0600)] <avtar> yzen: i deployed review2 on the pref server and can't start the process now
[13:28:38 CST(-0600)] <avtar> could you help with that?
[13:28:58 CST(-0600)] <yzen> avtar: sure
[13:29:07 CST(-0600)] <yzen> anything i can test locally ?
[13:29:39 CST(-0600)] <avtar> i'm running with the stock config https://github.com/GPII/universal/blob/review2/gpii/node_modules/preferencesServer/configs/production.json
[13:29:59 CST(-0600)] <avtar> i just enabled logging https://github.com/GPII/universal/blob/review2/gpii/node_modules/preferencesServer/configs/production.json#L9
[13:30:13 CST(-0600)] <avtar> but even without that i would invoke the process like this:
[13:30:26 CST(-0600)] <avtar> NODE_ENV=production node /srv/nodejs/gpii_pref_server/node_modules/kettle/lib/init.js /srv/nodejs/gpii_pref_server/gpii/node_modules/preferencesServer/configs/
[13:30:45 CST(-0600)] <avtar> it outputs 'function' and then dies
[13:30:57 CST(-0600)] <avtar> any idea what i'm doing wrong?
[13:31:17 CST(-0600)] <colinclark> kettle has a magical talent for this when its configuration files aren't quite right
[13:31:30 CST(-0600)] <colinclark> i.e. outputting "function" and then silently stopping
[13:32:20 CST(-0600)] <Bosmon> As described by http://issues.fluidproject.org/browse/KETTLE-17 : P
[13:32:32 CST(-0600)] <Bosmon> Although it appears that the behaviour has improved from outputting nothing at all : P
[13:32:54 CST(-0600)] <colinclark>
[13:35:54 CST(-0600)] <avtar> Bosmon: i took a look at the description but in this case i haven't modified the relative path to the configs directory
[13:36:44 CST(-0600)] <avtar> npm install finished its work successfully then i deleted the two extra infusion directories that i mentioned above
[13:37:02 CST(-0600)] <Bosmon> two extra!
[13:37:31 CST(-0600)] <Bosmon> Surely there should be just one extra ....
[13:37:35 CST(-0600)] <yzen> avtar: my deploys but with this warnings:
[13:37:39 CST(-0600)] <yzen> function
[13:37:40 CST(-0600)] <yzen> connect.multipart() will be removed in connect 3.0
[13:37:41 CST(-0600)] <yzen> visit https://github.com/senchalabs/connect/wiki/Connect-3.0 for alternatives
[13:37:42 CST(-0600)] <yzen> connect.limit() will be removed in connect 3.0
[13:37:48 CST(-0600)] <Bosmon> Ok, I see
[13:38:06 CST(-0600)] <Bosmon> There's an extra jqUnit one now
[13:39:33 CST(-0600)] <Bosmon> To add to our problems - I've done a completely fresh install of yzen's STMM repo and it still fails
[13:40:05 CST(-0600)] <Bosmon> Could you also do an rm -rf node_modules on yours too, yzen?
[13:40:18 CST(-0600)] <yzen> yes
[13:40:23 CST(-0600)] <Bosmon> Thanks
[13:41:54 CST(-0600)] <colinclark> any further luck, avtar? or are you still stuck and waiting for more help?
[13:42:27 CST(-0600)] <avtar> i'm trying to think about what i missed
[13:42:35 CST(-0600)] <avtar> going to do a fresh deploy again
[13:43:02 CST(-0600)] <yzen> yes i get the error
[13:43:09 CST(-0600)] <yzen> i guess i had masters package.json
[13:43:12 CST(-0600)] <Bosmon> yzen - the invoker error?
[13:43:13 CST(-0600)] <Bosmon> Cool
[13:43:22 CST(-0600)] <avtar> yzen: were you using 0.10.24 when you just tested?
[13:43:39 CST(-0600)] <yzen> avtar: 0.10.24 of what , sorry ?
[13:43:41 CST(-0600)] <Bosmon> Although I'm baffled at how you managed to do a npm install in only 3 minutes
[13:43:42 CST(-0600)] <avtar> nodejs
[13:45:03 CST(-0600)] <colinclark> and what version of nom nom nom are you both using?
[13:45:46 CST(-0600)] <avtar> 1.3.21
[13:46:22 CST(-0600)] <colinclark> ok, so that explains why you're getting infusion multiples, avtar
[13:46:52 CST(-0600)] <colinclark> and I think, based on the silence that you're experiencing, it's definitely an issue with the configuration
[13:46:52 CST(-0600)] <avtar> i love them all equally
[13:46:57 CST(-0600)] <colinclark>
[13:47:09 CST(-0600)] <colinclark> would it help to screenshare with yzen so he can see the server you're working on?
[13:47:11 CST(-0600)] <Bosmon> Sadly you will need to choose a favorite one : P
[13:47:27 CST(-0600)] <avtar> yzen: screenhero time?
[13:47:40 CST(-0600)] <yzen> give me 10min
[13:47:45 CST(-0600)] <yzen> brb
[13:47:47 CST(-0600)] <avtar> 10-4
[13:48:33 CST(-0600)] <avtar> Bosmon: forcing me to choose a favourite is just cruel
[13:48:59 CST(-0600)] <avtar> i want to nurture each one and see what they bring into the world
[13:49:11 CST(-0600)] <avtar> alright i'll stop now
[13:49:12 CST(-0600)] <avtar> brb
[13:52:25 CST(-0600)] <Bosmon> If there were any differences between them other than their FS location, there might be something to that
[13:52:30 CST(-0600)] <Bosmon> But sadly, they are all just clones : P
[13:52:57 CST(-0600)] <avtar> Bosmon: i should know in a min or two
[14:06:05 CST(-0600)] <Bosmon> Ok
[14:06:10 CST(-0600)] <Bosmon> I've resolved the error with the STMM
[14:06:19 CST(-0600)] <Bosmon> It seems like a simple mishandling of the global namespaces
[14:06:28 CST(-0600)] <Bosmon> Although as one often says, "I can't see how this ever worked" : P
[14:06:54 CST(-0600)] <Bosmon> That is, I can't see how this ever ran with any possible version of universal
[14:08:48 CST(-0600)] <colinclark> can you elaborate, Bosmon?
[14:09:36 CST(-0600)] <Bosmon> I've pushed a version which "seems to work for me" to https://github.com/amb26/GPII_StatisticalMatchmaker1/tree/GPII-275
[14:10:44 CST(-0600)] <Bosmon> colinclark so, in the following sequence: https://github.com/yzen/GPII_StatisticalMatchmaker1/blob/79449e33d88d1c441c1d24b9fa9ea9d209e18640/lib/StatisticalMatchMaker.js#L4-L16
[14:10:59 CST(-0600)] <Bosmon> This is in yzen's "fixed, condensed" version which he committed a few hours ago
[14:11:34 CST(-0600)] <colinclark> ah yes
[14:11:38 CST(-0600)] <Bosmon> He i) allocated the global namespace "gpii.matchMaker.statistical", ii) constructs a component with the same global name "gpii.matchMaker.statistical" but then iii) continues to use the SAME object reference he received in step i)
[14:11:45 CST(-0600)] <colinclark> ok
[14:11:57 CST(-0600)] <Bosmon> However, step ii) has necessarily modified the value at the namespace since it now consists of a component creator function
[14:12:04 CST(-0600)] <Bosmon> Rather than the original {} which was allocated
[14:12:13 CST(-0600)] <Bosmon> You'll be glad to hear that the fix involves a step away from single-var-ism : P
[14:12:34 CST(-0600)] <colinclark> michelled: Were you able to find any preference sets that a) actually need the STMM and b) has some preferences in them?
[14:13:11 CST(-0600)] <avtar> i'm seeing the same issue with a new deployment of the pref server
[14:13:12 CST(-0600)] <michelled> ah, sorry colinclark yes the prefs for demos 2 and 3 require STMM
[14:13:25 CST(-0600)] <colinclark> which ones are those, michelled?
[14:13:30 CST(-0600)] <Bosmon> colinclark - I am testing with the prefs set called "review2_demo2.json"
[14:13:35 CST(-0600)] <colinclark> ah, ok
[14:13:42 CST(-0600)] <avtar> should i deploy Bosmon's stmm?
[14:13:55 CST(-0600)] <Bosmon> And I do get some kind of intelligible JSON response out of it at the end
[14:14:00 CST(-0600)] <michelled> https://github.com/GPII/universal/blob/review2/testData/preferences/review2_demo2.json
[14:14:01 CST(-0600)] <colinclark> can you paste the reponse?
[14:14:17 CST(-0600)] <Bosmon> avtar - you could deploy it, yes - but so far the issues with every MM you've deployed seem the same, and different from any issues we have seen locally
[14:14:19 CST(-0600)] <colinclark> avtar: I'm thinking the answer is yes, but let me just get some further clarity on what we're seeing
[14:14:28 CST(-0600)] <avtar> ok
[14:14:29 CST(-0600)] <michelled> https://github.com/GPII/universal/blob/review2/testData/preferences/review2_demo3b_blind.json
[14:14:36 CST(-0600)] <colinclark> Bosmon: and you think it's the nginx reverse proxy that is at issue?
[14:14:44 CST(-0600)] <Bosmon> colinclark - that is my belief
[14:15:02 CST(-0600)] <Bosmon> The response actually consists of a JSON payload holding the string "ECONNREFUSED"
[14:15:13 CST(-0600)] <Bosmon> Which was it what confused me at first into believing that was the direct HTTP error
[14:15:40 CST(-0600)] <avtar> i could make stmm's port accessible directly to the internet so we could test that instead
[14:15:41 CST(-0600)] <Bosmon> But in fact there is no HTTP error from the proxy front-end, it in fact returns a valid JSON payload presumably indicating that it received an upstream error
[14:15:50 CST(-0600)] <colinclark> avtar, that would be a good place to start
[14:16:00 CST(-0600)] <colinclark> deploy Bosmon's branch and expose it directly to the world
[14:16:19 CST(-0600)] <colinclark> It seems to me that if you're getting a JSON error response, wouldn't it actually be coming from Kettle?
[14:16:30 CST(-0600)] <colinclark> I can't imagine that nginx displays its errors as JSON
[14:16:49 CST(-0600)] <Bosmon> colinclark - yes, that is certainly suspicious
[14:17:01 CST(-0600)] <colinclark> let's simplify the situation a bit and see what happens, avtar
[14:17:05 CST(-0600)] <colinclark> if it isn't too much trouble
[14:17:17 CST(-0600)] <avtar> let's try it with Bosmon's version first as-is
[14:17:19 CST(-0600)] <Bosmon> But in that case, what on earth could Kettle be trying to connect to that it is refused?
[14:17:23 CST(-0600)] <colinclark> ok
[14:17:29 CST(-0600)] <avtar> and then if that doesn't work i'll open the port
[14:17:51 CST(-0600)] <colinclark> can you paste me the response you're getting from your local STMM, Bosmon?
[14:18:02 CST(-0600)] <Bosmon> colinclark, michelled: Here is my payload response from my local STMM for the prefs set review2_demo2.json
[14:18:06 CST(-0600)] <avtar> something to keep in mind though is that the preferences server is currently offline
[14:18:07 CST(-0600)] <Bosmon> http://pastebin.com/cSC6LjF5
[14:18:15 CST(-0600)] <avtar> it wasn't when we were trying this earlier
[14:18:22 CST(-0600)] <colinclark> avtar: that should be fine for this particular component
[14:18:26 CST(-0600)] <avtar> ok
[14:18:47 CST(-0600)] <colinclark> we're still only getting easy one, Bosmon
[14:18:50 CST(-0600)] <colinclark> which has no settings
[14:19:03 CST(-0600)] <Bosmon> ok
[14:19:09 CST(-0600)] <colinclark> let's just see
[14:19:12 CST(-0600)] <Bosmon> Let me know which other set I can try
[14:19:37 CST(-0600)] <colinclark> yeah, review2_demo2 is just a collection of random preferences that we don't have easy access to the physical hardware for
[14:19:50 CST(-0600)] <colinclark> i.e. the pixel sense and android
[14:20:01 CST(-0600)] <colinclark> let's see if "blind" is more appropriate
[14:20:04 CST(-0600)] <colinclark> yes
[14:20:13 CST(-0600)] <colinclark> can you try this one, Bosmon? https://github.com/GPII/universal/blob/review2/testData/preferences/review2_demo3b_blind.json
[14:20:17 CST(-0600)] <Bosmon> ok
[14:20:40 CST(-0600)] <colinclark> My impression, though, is that none of these preferences sets actually exercise the Matchmaker's capacity to, well, match things
[14:20:46 CST(-0600)] <colinclark> which actually serves us very well
[14:21:03 CST(-0600)] <colinclark> michelled: Are those the only two preference sets that include "matchMakerType": "statistical"?
[14:21:14 CST(-0600)] <michelled> no, there are a couple more - one sec
[14:21:38 CST(-0600)] <colinclark> thank you
[14:21:41 CST(-0600)] <michelled> https://github.com/GPII/universal/blob/review2/testData/preferences/review2_demo3c_elderly.json
[14:22:02 CST(-0600)] <michelled> https://github.com/GPII/universal/blob/review2/testData/preferences/review2_demo3d_adult.json
[14:22:09 CST(-0600)] <avtar> Bosmon: your version is up
[14:22:14 CST(-0600)] <michelled> https://github.com/GPII/universal/blob/review2/testData/preferences/review2_demo3d_young.json
[14:23:00 CST(-0600)] <colinclark> the orca one should be very effective
[14:23:13 CST(-0600)] <colinclark> since it has application-specific preferences for a solution that runs only on Linux
[14:23:33 CST(-0600)] <colinclark> Bosmon, given that your payload specifies that you're on Windows, it should return back settings for NVDA if it is successful
[14:23:44 CST(-0600)] <colinclark> real, full-on application-specific preferences
[14:24:05 CST(-0600)] <Bosmon> So yes, it seems that these "cannot call method fire of undefined" messages are caused by having a corrupt specification for the matchmaker in these profiles
[14:24:24 CST(-0600)] <colinclark> the wrong name, you mean?
[14:24:26 CST(-0600)] <Bosmon> This profile, review2_demo3b_blind.json simply has the entry as "matchMakerType": "statistical",
[14:24:29 CST(-0600)] <Bosmon> Which is not correct
[14:24:34 CST(-0600)] <colinclark> what should it be?
[14:24:44 CST(-0600)]
<Bosmon> The entry should read "http://registry.gpii.org/common/matchMakerType": [
],
[14:25:05 CST(-0600)] <Bosmon> "don't ask me why" : P
[14:25:16 CST(-0600)] <Bosmon> Or how it is that matchMakerType has come to be considered a "common term".....
[14:25:55 CST(-0600)] <colinclark> yes, this was an issue I started to try to talk through before the holidays
[14:25:57 CST(-0600)] <colinclark> but it's too late
[14:26:21 CST(-0600)] <Bosmon> Urgh
[14:26:23 CST(-0600)] <colinclark> Bosmon: should michelled go ahead an prepare a pull request to fix this?
[14:26:28 CST(-0600)] <colinclark> in the sets that we've got?
[14:26:29 CST(-0600)] <Bosmon> Now I am getting the ECONNREFUSED message locally...............
[14:26:33 CST(-0600)] <colinclark> wooah
[14:26:55 CST(-0600)] <Bosmon> Oh
[14:26:59 CST(-0600)] <Bosmon> It's because it is not running
[14:27:05 CST(-0600)] <Bosmon> Well, that seems straightforward enough.......
[14:27:08 CST(-0600)] <avtar> heh
[14:27:21 CST(-0600)] <Bosmon> And that also explains the source of avtar's messages
[14:27:39 CST(-0600)] <avtar> nothing can explain the source of my messages :|
[14:27:44 CST(-0600)] <Bosmon> shush
[14:28:01 CST(-0600)] <colinclark> ooooh, avtar just got shushed!
[14:28:03 CST(-0600)] <Bosmon> Well ok in this case it is a REAL upstream ECONNREFUSED
[14:28:13 CST(-0600)] <colinclark> unshush, AVTAR!
[14:28:30 CST(-0600)] <colinclark> upstream to where, Bosmon?
[14:28:31 CST(-0600)] <Bosmon> But interestingly the returned JSON payload from the overall system is exactly the same as the one I get from avtar's deployed MMs
[14:28:39 CST(-0600)] <avtar> Bosmon: i haven't seen any traffic on port 8077 yet
[14:28:52 CST(-0600)] <Bosmon> The flowManager, talking to my local STMM, receives a REAL ECONNREFUSED
[14:29:12 CST(-0600)] <Bosmon> Which then, from the point of view of user login, reports exactly the payload which avtar's MMs seem to return
[14:29:12 CST(-0600)] <colinclark> but when you poke the STMM directly, this isn't the case?
[14:29:20 CST(-0600)] <Bosmon> No, it genuinely wasn't running
[14:29:28 CST(-0600)] <Bosmon> Sorry avtar, still testing locally
[14:29:31 CST(-0600)] <Bosmon> Give me a minute or so
[14:29:40 CST(-0600)] <avtar> np
[14:29:46 CST(-0600)] <avtar> i'm about to put all sorts of tacos in my mouth
[14:29:55 CST(-0600)] <Bosmon> ok
[14:29:56 CST(-0600)] <avtar> typing will affected
[14:30:00 CST(-0600)] <avtar> be
[14:30:06 CST(-0600)] <avtar> already it seems
[14:30:14 CST(-0600)] <Bosmon> colinclark - I can't test review2_demo3_blind.json since I don't have NVDA installed
[14:30:30 CST(-0600)] <Bosmon> During logon the flowmanager bombs out with Error: ENOENT, no such file or directory 'C:\Users\Bosmon\AppData\Roaming\nvda\nvda.ini'
[14:30:46 CST(-0600)] <Bosmon> This of course is also completely unacceptable behaviour since this results in a completely incomprehensible message from user logon
[14:30:49 CST(-0600)] <colinclark> can you at least probe the STMM directly with this preference set?
[14:31:05 CST(-0600)] <colinclark> I JUST SAID PROBE, GUYS
[14:31:18 CST(-0600)] <Bosmon> simmer down, chaps : P
[14:31:25 CST(-0600)] <avtar> lol
[14:32:09 CST(-0600)] <colinclark> so serious and british
[14:36:01 CST(-0600)] <avtar> 'british bosmon'
[14:36:04 CST(-0600)] <avtar> colinclark: you came up with that one
[14:36:15 CST(-0600)] <colinclark> ye-ah
[14:36:16 CST(-0600)] <Bosmon> Unfortunately my local STMM, given the payload from review2_demo3b_blind.json fails with a new kind of error
[14:36:37 CST(-0600)] <Bosmon> Which looks a bit like this:
[14:36:39 CST(-0600)] <Bosmon> 13:35:07.339: INFO urlExpander expanded: file://E:\Source\gits\GPII_StatisticalMatchmaker1\node_modules\universal\gpii\nod
[14:36:40 CST(-0600)] <Bosmon> e_modules\matchMaker/../../../testData/solutions/undefined.json
[14:36:40 CST(-0600)] <Bosmon> TypeError: Cannot read property 'from' of undefined
[14:36:40 CST(-0600)] <Bosmon> at gpii.ontologyServer.development.set (E:\Source\gits\GPII_StatisticalMatchmaker1\node_modules\universal\gpii\node_modu
[14:36:40 CST(-0600)] <Bosmon> les\ontologyServer\src\ontologyServer.js:89:19)
[14:36:43 CST(-0600)] <colinclark> it's not your best, though, avtar
[14:36:44 CST(-0600)] <colinclark> or my best
[14:37:09 CST(-0600)] <colinclark> Why is the ontology server involved at all?
[14:37:22 CST(-0600)] <Bosmon> I can't explain that
[14:37:47 CST(-0600)] <Bosmon> But it seems somehow "left behind" because yzen's strategy in these proxy integrations seems to be to reuse some parts of the MatchMaker server back in universal
[14:38:04 CST(-0600)] <Bosmon> This was no doubt the reason that in earlier days he was agitating for us to break these out as separate modules
[14:38:19 CST(-0600)] <Bosmon> But it seems somehow that a bit of the ontology server workflow is "left behind" in this matching process
[14:38:52 CST(-0600)] <colinclark> I don't know what to do
[14:39:33 CST(-0600)] <Bosmon> I think we will just have to axe this part of the demo
[14:39:35 CST(-0600)] <colinclark> ok
[14:39:40 CST(-0600)] <colinclark> Can we spend some time with the RBMM?
[14:39:48 CST(-0600)] <Bosmon> Especially if yzen is gone, it doesn't seem feasible to get the STMM working on this evidence
[14:39:51 CST(-0600)] <colinclark> They say they've tested it with the whole system
[14:40:48 CST(-0600)] <colinclark> is it possible to spend some time trying to get our server deployment of it working, and to verify that their version is working somehow?
[14:43:12 CST(-0600)] <Bosmon> Oh wait, I think my payload for the last test is wrong
[14:43:19 CST(-0600)] <Bosmon> So I have to back up and answer your previous question with "no"
[14:43:28 CST(-0600)] <Bosmon> That is, "no, I can't probe the STMM directly with this preference set"
[14:44:04 CST(-0600)] <Bosmon> No wait
[14:44:09 CST(-0600)] <Bosmon> let me back up completely
[14:44:16 CST(-0600)] <avtar> colinclark: i'm waiting on yzen at the moment for some assistance with the pref server
[14:44:17 CST(-0600)] <Bosmon> Sorry to be so confused
[14:44:23 CST(-0600)] <Bosmon> The answer to your question is YES
[14:44:33 CST(-0600)] <avtar> but i'll move onto redeploying flowmanager and matchmaker like we discussed earlier
[14:44:43 CST(-0600)] <avtar> from the official gpii repo
[14:45:42 CST(-0600)] <avtar> yay!
[14:46:01 CST(-0600)] <avtar> yzen: screenhero time?
[14:46:49 CST(-0600)] <colinclark> Bosmon: Can you elaborate, sorry?
[14:46:57 CST(-0600)] <Bosmon> colinclark - so here is the response from the STMM to review2_demo3b_blind preferences:
[14:46:58 CST(-0600)] <Bosmon> http://pastebin.com/L2AsetBS
[14:47:38 CST(-0600)] <colinclark> this looks correct
[14:47:47 CST(-0600)] <colinclark> michelled: Do you still have your windows laptop?
[14:47:52 CST(-0600)] <michelled> yes
[14:47:59 CST(-0600)] <Bosmon> Which, if I'm not mistaken, reflects far more the results of our own matchmaking processing than anything done by the STMM : P
[14:48:14 CST(-0600)] <colinclark> yes
[14:48:15 CST(-0600)] <colinclark> How hard would it be for you to get to the point where Bosmon is and test it with a full stack and with NVDA installed?
[14:48:16 CST(-0600)] <Bosmon> yzen - I corrected the error in your STMM pull and have pushed it to my own GPII-275 fork
[14:48:34 CST(-0600)] <colinclark> thank you thank you Bosmon
[14:49:00 CST(-0600)] <colinclark> my impression is that the STMM does indeed need the transformation stage here
[14:49:10 CST(-0600)] <Bosmon> Right
[14:49:15 CST(-0600)] <Bosmon> But it is being done, right?
[14:49:16 CST(-0600)] <colinclark> since the results, as they're returned, won't set correctly
[14:49:20 CST(-0600)] <colinclark> no, yzen removed it
[14:49:24 CST(-0600)] <Bosmon> ah ok
[14:49:43 CST(-0600)] <colinclark> Bosmon: Any chance you could look at what yzen did and undo most of it?
[14:50:05 CST(-0600)] <colinclark> My impression is that we need the matchPost function that he removed
[14:50:05 CST(-0600)] <michelled> Bosmon: what do I need to do to get to point where you are?
[14:51:09 CST(-0600)] <Bosmon> michelled - all you really need is my fork of the STMM
[14:51:20 CST(-0600)] <Bosmon> And then you need a set of slightly hacked preferences sets
[14:51:28 CST(-0600)] <Bosmon> And modifications to the proxy.json inside your universal
[14:52:10 CST(-0600)] <Bosmon> It is times like these that one has sympathy for KASPARNET's tendency to work with a "random bunch of uncommitted files" .....
[14:52:42 CST(-0600)] <Bosmon> I should probably commit my lightly hacked universal fork as well, but only about 3 lines have been edited
[14:54:14 CST(-0600)] <Bosmon> michelled - my STMM branch is here
[14:54:15 CST(-0600)] <Bosmon> https://github.com/amb26/GPII_StatisticalMatchmaker1/tree/GPII-275
[14:54:25 CST(-0600)] <michelled> thx
[14:56:28 CST(-0600)] <colinclark> Bosmon: So as I read this, Andy didn't quite have his matchPost component wired up properly
[14:57:28 CST(-0600)] <colinclark> and as a result, no transformation
[14:57:56 CST(-0600)] <colinclark> I don't quite even understand the function of the matchPost component
[14:58:25 CST(-0600)] <colinclark> as opposed to directly binding the statistical matchmaker's match invoker to gpii.matchMaker.statistical.matchPostMatch
[14:58:50 CST(-0600)] <michelled> Bosmon: once I've got your branch and run npm install, what do I need to do?
[14:59:27 CST(-0600)] <Bosmon> michelled - the next step is to adjust your file proxy.json within your universal to point to this matchmaker
[14:59:59 CST(-0600)] <Bosmon> On line 16, it should have:
[15:00:00 CST(-0600)] <Bosmon> "statistical": "http://localhost:8077"
[15:00:16 CST(-0600)] <Bosmon> After that all you need to do is to correct the entries in the preferences sets as I mentioned above
[15:00:35 CST(-0600)] <Bosmon> Incorrectly, on line 2, these read
[15:00:36 CST(-0600)] <Bosmon> "matchMakerType": "statistical",
[15:00:38 CST(-0600)] <Bosmon> Whereas they should have:
[15:00:46 CST(-0600)]
[15:01:05 CST(-0600)] <yzen> colinclark: i don't thing that you need any of the code i deleted, actually
[15:01:11 CST(-0600)] <yzen> only rule based match maker needed it
[15:01:23 CST(-0600)] <Bosmon> Oh yes, final step: you need to set your NODE_ENV environment variable
[15:01:30 CST(-0600)] <colinclark> yzen: ok, I'm just confused about the output that Bosmon was getting from the STMM
[15:01:37 CST(-0600)] <colinclark> maybe I'm just being terribly forgetful
[15:01:45 CST(-0600)] <colinclark> but doesn't it look like it needs to be transformed?
[15:01:53 CST(-0600)] <Bosmon> set NODE_ENV=fm.ps.sr.dr.mm.os.lms.multipleMatchMakers
[15:01:58 CST(-0600)] <Bosmon> Or Linux equivalent thereof
[15:01:58 CST(-0600)] <colinclark> this is it, yzen: http://pastebin.com/L2AsetBS
[15:02:25 CST(-0600)] <Bosmon> I believe that is the total set of requirements
[15:02:33 CST(-0600)] <colinclark> I mean, yzen, it's doing the right thing in terms of mapping orca to nvda
[15:02:55 CST(-0600)] <Bosmon> i) proxy.json ii) corrected preference sets, iii) NODE_ENV variable
[15:03:12 CST(-0600)] <colinclark> oh, perhaps I'm wrong, yzen
[15:03:18 CST(-0600)] <colinclark> line 351
[15:03:33 CST(-0600)] <colinclark> I guess we'll have to wait to see what michelled finds in practice
[15:03:39 CST(-0600)] <colinclark> but these look more or less right, don't they?
[15:04:13 CST(-0600)] <colinclark> in the paste bin file, I mean
[15:07:09 CST(-0600)] <Bosmon> avtar
[15:07:12 CST(-0600)] <Bosmon> Sorry for the delay
[15:07:22 CST(-0600)] <Bosmon> colinclark has reminded me that I am now due to test your STMM deployments
[15:07:26 CST(-0600)] <Bosmon> Could you tell me the details?
[15:07:42 CST(-0600)] <avtar> sure
[15:07:48 CST(-0600)] <avtar> i'm using your repo and branch
[15:08:08 CST(-0600)] <avtar> this one https://github.com/amb26/GPII_StatisticalMatchmaker1/tree/GPII-275
[15:08:11 CST(-0600)] <Bosmon> ok
[15:08:26 CST(-0600)] <avtar> currently it still has nginx sitting in front of it
[15:08:31 CST(-0600)] <Bosmon> ok
[15:08:32 CST(-0600)] <avtar> accepting requests on port 80
[15:08:36 CST(-0600)] <avtar> stmm is using 8077
[15:08:40 CST(-0600)] <Bosmon> Ok
[15:08:48 CST(-0600)] <Bosmon> Port 80 is still giving the ECONNREFUSED payload
[15:08:52 CST(-0600)] <Bosmon> Port 8077 doesn't seem to be open
[15:08:58 CST(-0600)] <avtar> i'll open 8077 now
[15:10:22 CST(-0600)] <avtar> http://205.211.169.179:8077/match
[15:11:21 CST(-0600)] <Bosmon> Ok
[15:11:28 CST(-0600)] <Bosmon> This shows the same payload as the front end
[15:11:37 CST(-0600)] <Bosmon> So I guess we can free the nginx layer from all blame
[15:11:47 CST(-0600)] <colinclark> vindication, avtar!
[15:12:00 CST(-0600)] <Bosmon> Can you get me some logs from the node server?
[15:12:01 CST(-0600)] <avtar> sweet sweetness
[15:12:37 CST(-0600)] <avtar> sure
[15:12:41 CST(-0600)] <avtar> cat /var/log/nodejs/statistical_matchmaker.log
[15:12:41 CST(-0600)] <avtar> function
[15:12:41 CST(-0600)] <avtar> function
[15:12:50 CST(-0600)] <Bosmon> glah
[15:12:54 CST(-0600)] <Bosmon> That's it?
[15:12:59 CST(-0600)] <avtar> yup :/
[15:13:03 CST(-0600)] <Bosmon> Well clearly it never started : P
[15:13:07 CST(-0600)] <Bosmon> Much like your preferences server
[15:13:12 CST(-0600)] <avtar> hey
[15:13:15 CST(-0600)] <avtar> you know what started now?
[15:13:20 CST(-0600)] <avtar> the pref server!
[15:13:22 CST(-0600)] <Bosmon> oh cool
[15:13:23 CST(-0600)] <colinclark> yay
[15:13:34 CST(-0600)] <colinclark> Bosmon: How do you know it didn't start?
[15:13:38 CST(-0600)] <avtar> i had to delete infusion from the kettle directory
[15:13:41 CST(-0600)] <Bosmon> Because there would be something in the logs!
[15:13:42 CST(-0600)] <colinclark> Because it didn't emit the connect deprecation warning?
[15:13:47 CST(-0600)] <Bosmon> And all the other junk, yes
[15:14:04 CST(-0600)] <avtar> deleted these ones
[15:14:05 CST(-0600)] <avtar> ./node_modules/kettle/node_modules/infusion
[15:14:05 CST(-0600)] <avtar> ./node_modules/jqUnit/node_modules/infusion
[15:14:06 CST(-0600)] <Bosmon> And these things always end with the annotation "Kettle server listening on port Xxxxx"
[15:14:12 CST(-0600)] <avtar> and left ./node_modules/infusion
[15:14:21 CST(-0600)] <colinclark> ah, because even production mode doesn't silence the logs for the STMM
[15:14:37 CST(-0600)] <Bosmon> I never actually tested production mode
[15:14:46 CST(-0600)] <Bosmon> I can't say I know what the difference would be in this case
[15:14:48 CST(-0600)] <Bosmon> Hopefully nothing
[15:14:50 CST(-0600)] <avtar> yzen: http://preferences.gpii.net/user/carla
[15:14:52 CST(-0600)] <colinclark> they look identical to me
[15:14:58 CST(-0600)] <avtar> looks like we won't have to chat after all
[15:15:16 CST(-0600)] <yzen> looks good
[15:15:33 CST(-0600)] <Bosmon> ok - avtar - all that description just now was for the prefs server? Rather than for the STMM?
[15:15:38 CST(-0600)] <yzen> avtar: wait so any other deployment stuff you need from me?
[15:16:01 CST(-0600)] <michelled> sigh. I'm back to a misbehaving system I'm reckoning
[15:16:10 CST(-0600)] <avtar> Bosmon: yes. only the function, function was related to stmm
[15:16:11 CST(-0600)] <michelled> re-cloning. and going to eat some lunch
[15:16:15 CST(-0600)] <colinclark> michelled just said reckoning
[15:16:18 CST(-0600)] <colinclark> which is a great word
[15:16:37 CST(-0600)] <Bosmon> ok - so how did you start the STMM, avtar?
[15:16:57 CST(-0600)] <colinclark> michelled: Do you want me to order you a pizza?
[15:17:04 CST(-0600)] <Bosmon> And I'm wondering whether you might need to do infusion deduplication on that one too
[15:17:15 CST(-0600)] <colinclark> or anyone else?
[15:17:22 CST(-0600)] <avtar> Bosmon: NODE_ENV=production node /opt/statistical_matchmaker/bin/statisticalMatchMaker.js
[15:17:29 CST(-0600)] <colinclark> bicycle tacos?
[15:17:30 CST(-0600)] <Bosmon> ok
[15:17:47 CST(-0600)] <Bosmon> avtar - could you just change that to NODE_ENF=development and see if we get any logs
[15:17:51 CST(-0600)] <avtar> sure
[15:17:55 CST(-0600)] <Bosmon> And also check if you have any duplicate infusions as with prefs
[15:18:11 CST(-0600)] <Bosmon> I fear you will have to go on a cruel mission once more
[15:18:35 CST(-0600)] <avtar> Bosmon: only ./node_modules/universal/node_modules/infusion
[15:18:57 CST(-0600)] <Bosmon> ok
[15:19:15 CST(-0600)] <Bosmon> Well let me know once you have started up development configuration and lets see if we get any logs out of it
[15:19:20 CST(-0600)] <avtar> Bosmon: looking better http://pastie.org/pastes/8627423/text
[15:19:23 CST(-0600)] <Bosmon> ok
[15:19:39 CST(-0600)] <Bosmon> Seems reasonable
[15:19:41 CST(-0600)] <Bosmon> I will try to connect now
[15:19:45 CST(-0600)] <Bosmon> And it should generate further logs
[15:19:56 CST(-0600)] <Bosmon> UGGHHHH!
[15:19:57 CST(-0600)] <Bosmon> It worked
[15:19:58 CST(-0600)] <Bosmon> Ok
[15:20:03 CST(-0600)] <Bosmon> So fault is somewhere in production config
[15:21:29 CST(-0600)] <avtar> Bosmon: http://pastie.org/pastes/8627431/text
[15:21:49 CST(-0600)] <Bosmon> avtar thanks, yes
[15:21:54 CST(-0600)] <Bosmon> It seems to have worked
[15:22:05 CST(-0600)] <Bosmon> And I can confirm that the "production" config is broken locally too
[15:22:08 CST(-0600)] <Bosmon> Although it gives a different error
[15:22:13 CST(-0600)] <yzen> Bosmon, avtar are you deploying stats match maker ?
[15:22:18 CST(-0600)] <Bosmon> yzen - yes
[15:22:30 CST(-0600)] <Bosmon> Only "development" config is working
[15:22:36 CST(-0600)] <avtar> colinclark, yzen: could we agree on a config for flowmanager and/or universal's matchmaker?
[15:22:42 CST(-0600)] <yzen> Bosmon: how come you want to use production ?
[15:22:58 CST(-0600)] <Bosmon> yzen - avtar thought it seemed a reasonable choice
[15:23:00 CST(-0600)] <avtar> there was some talk about missing matchmaker entries in the current flowmanager config
[15:23:03 CST(-0600)] <Bosmon> I had only tested with development
[15:23:20 CST(-0600)] <Bosmon> Since we are deploying a live service, it seems that "production" should be non-broken
[15:23:27 CST(-0600)] <avtar>
[15:23:30 CST(-0600)] <yzen> it should be development, but reference should be to the latest correct universal
[15:23:32 CST(-0600)] <colinclark> avtar: Yes, it sounds like we need to make a commit to universal and push it also to review2
[15:23:56 CST(-0600)] <yzen> production requires live ontology server
[15:24:02 CST(-0600)] <Bosmon> I see
[15:24:03 CST(-0600)] <Bosmon> Good to know
[15:24:10 CST(-0600)] <avtar> i feel this should be in readmes though
[15:24:15 CST(-0600)] <Bosmon> Ok well, that seems we have made some progress
[15:24:16 CST(-0600)] <colinclark> yes
[15:24:20 CST(-0600)] <colinclark> avtar: you hit the nail on the head
[15:24:36 CST(-0600)] <colinclark> so could we get away with deploying in development mode?
[15:24:42 CST(-0600)] <colinclark> and stifling the logs with a setLogging: false?
[15:25:21 CST(-0600)] <yzen> as long as universal is the one that everyone is using
[15:25:25 CST(-0600)] <Bosmon> Yes
[15:25:28 CST(-0600)] <colinclark> A Matchmaker server need not make any requests to anything other than the ontology server, is that right?
[15:25:32 CST(-0600)] <Bosmon> To be honest I don't see the logging as hazardous anyway
[15:25:38 CST(-0600)] <colinclark> so it doesn't matter if the ontology server is local
[15:25:39 CST(-0600)] <yzen> solutions registry
[15:25:40 CST(-0600)] <Bosmon> I'm sure avtar's disks will not fill up over the week
[15:25:58 CST(-0600)] <avtar> Bosmon: they're slow though
[15:26:21 CST(-0600)] <colinclark> I don't see any reason not to turn logging off once we've verified that this works
[15:26:35 CST(-0600)] <avtar> but i guess these urls aren't going to show up on hacker news
[15:26:58 CST(-0600)] <colinclark> no
[15:27:12 CST(-0600)] <colinclark> yzen: Okay, so the matchmaker also accesses the solutions registry
[15:27:24 CST(-0600)] <colinclark> which presumably is populated with identical data to the deployed version?
[15:27:59 CST(-0600)] <yzen> colinclark: it would as long as the universal has the same revision
[15:28:13 CST(-0600)] <colinclark> ok
[15:28:34 CST(-0600)] <colinclark> avtar: My impression is that we should just go ahead with using the development configuration
[15:28:37 CST(-0600)] <avtar> Bosmon: i guess it's okay for me to remove public access to port 8077 then?
[15:28:40 CST(-0600)] <colinclark> for both MMs
[15:28:42 CST(-0600)] <colinclark> avtar: yes
[15:28:44 CST(-0600)] <Bosmon> avtar - yes
[15:28:45 CST(-0600)] <Bosmon> Go ahead
[15:28:48 CST(-0600)] <colinclark> back to your lovely reverse proxy
[15:28:50 CST(-0600)] <Bosmon> Let me just verify the front-end URL
[15:29:11 CST(-0600)] <Bosmon> avtar - yup, verified
[15:29:22 CST(-0600)] <avtar> nooice
[15:29:24 CST(-0600)] <colinclark> avtar: So next, let's get to your question
[15:29:48 CST(-0600)] <colinclark> So, it sounds like we will redeploy both the STMM and the RBMM in the development configuration
[15:29:58 CST(-0600)] <colinclark> You've already done the STMM, I guess
[15:30:03 CST(-0600)] <avtar> yes
[15:30:09 CST(-0600)] <avtar> one sec
[15:30:10 CST(-0600)] <colinclark> does that, theoretically, mean we are cool as far as the Matchmakers themselves?
[15:30:15 CST(-0600)] <avtar> no
[15:30:21 CST(-0600)] <colinclark> ok
[15:30:23 CST(-0600)] <avtar> there's matchmaker.gpii.net
[15:30:35 CST(-0600)] <colinclark> what is matchmaker.gpii.net?
[15:31:08 CST(-0600)] <colinclark> I'm dumb, I know
[15:31:19 CST(-0600)] <colinclark> I'm clearly missing one component in how our services are laid out
[15:31:21 CST(-0600)] <avtar> colinclark: https://github.com/GPII/universal/tree/v0.2/gpii/node_modules/matchMaker
[15:31:22 CST(-0600)] <colinclark> is it the proxy itself?
[15:31:37 CST(-0600)] <avtar> the v0.2 branch
[15:32:03 CST(-0600)] <colinclark> yzen: It's the thing we call a "matchmaker," but which actually uses the Matchmaker Proxy to delegate to a remote Matchmaker (STMM, etc.) to do the matching, is that correct?
[15:35:19 CST(-0600)] <yzen> colinclark, avtar so we have several so called stand alone match makers and one match maker proxy that should be deployed alongside the cloud based match maker
[15:35:57 CST(-0600)] <colinclark> Ok, so this thing avtar is calling "the matchmaker" is actually the Matchmaker Proxy?
[15:35:57 CST(-0600)] <yzen> avtar: I'm not sure what v0.2 is about
[15:36:05 CST(-0600)] <colinclark> And needs to be deployed itself as a standalone service?
[15:36:11 CST(-0600)] <yzen> yes
[15:36:19 CST(-0600)] <colinclark> ok
[15:36:20 CST(-0600)] <yzen> but i m confused by 0.2 thing
[15:36:49 CST(-0600)] <avtar> yzen: kasper and you asked me to deploy 0.2 when we first set up flowmanager.gpii.net and matchmaker.gpii.net
[15:36:58 CST(-0600)] <avtar> i believe when you were in europe
[15:37:29 CST(-0600)] <avtar> the workshop in portugal, i think
[15:38:19 CST(-0600)] <yzen> avtar: right but now that branch is irrelevant. we need to deploy the match maker with proxy.json from the latest universal and i m guessing that is review2 branch, colinclark ?
[15:39:54 CST(-0600)] <avtar> which brings me to the config and possible changes https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json
[15:40:15 CST(-0600)] <colinclark> yes, definitely the review2 branch
[15:40:32 CST(-0600)] <colinclark> avtar: So in proxy.json, you're saying we should commit configuration for the two matchmakers, right?
[15:40:43 CST(-0600)] <avtar> yes
[15:41:08 CST(-0600)] <colinclark> ok
[15:41:08 CST(-0600)] <avtar> and preferably include a reference to the default port that the service expects to bind to
[15:41:20 CST(-0600)] <colinclark> ok
[15:41:50 CST(-0600)] <colinclark> yzen: to confirm, the latter issue-the default port-would just entail adding a "port" option to the matchMakerProxy's options?
[15:42:14 CST(-0600)] <colinclark> avtar: By default, would you like this file to be configured to hit http://rbmm.gpii.net and http://stmm.gpii.net
[15:42:14 CST(-0600)] <colinclark> ?
[15:42:27 CST(-0600)] <colinclark> And is the flat one configured correctly?
[15:42:51 CST(-0600)] <avtar> i'm not sure what the flat one is
[15:43:19 CST(-0600)] <yzen> colinclark: yes if you want to specify port you just add "port": XXXX options as a "kettle.server" component's option
[15:44:24 CST(-0600)] <colinclark> yura, so couple more questions
[15:44:47 CST(-0600)] <colinclark> When Kasper is using a flow manager with the "multiple matchmakers" configured, he will have a flat matchmaker running on the local host with it?
[15:44:50 CST(-0600)] <yzen> colinclark: https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json#L25 i m hoping it is correct
[15:45:02 CST(-0600)] <yzen> colinclark: yes correct
[15:45:23 CST(-0600)] <yzen> both are listed in proxy.json config
[15:45:42 CST(-0600)] <colinclark> ok
[15:46:01 CST(-0600)] <colinclark> so avtar, you're basically just asking us to update review2 and/or master to include the declarations for the ruleBased and statistical matchmakers
[15:46:11 CST(-0600)] <colinclark> such that we won't all have to do so manually each time we redeploy, is that correct?
[15:46:38 CST(-0600)] <colinclark> and to add a default port to the matchmaker proxy server, yes?
[15:46:50 CST(-0600)] <colinclark> such that it won't bind to 8080 by default
[15:47:00 CST(-0600)] <avtar> it's more so for documenting the config so that we don't wonder what's happening a few months from now
[15:47:24 CST(-0600)] <avtar> i'll still have to change the port numbers based on where and how the components are being hosted
[15:47:54 CST(-0600)] <Bosmon> avtar - right idea!
[15:48:00 CST(-0600)] <Bosmon> Always best to avoid surprising our future selves
[15:48:27 CST(-0600)] <Bosmon> I just came upon a fairly long JIRA I'd written only 9 months ago and I couldn't remember a thing about it
[15:48:48 CST(-0600)] <Bosmon> I read it with extreme interest
[15:48:49 CST(-0600)] <avtar> you gave birth to knowledge, Bosmon
[15:48:52 CST(-0600)] <Justin_o_> cindyli: do you think we should make a new pcp demo file like we did for pmt.. that is to make pcp.html file in the prefsEditor directory of the demos folder
[15:49:37 CST(-0600)] <avtar> colinclark, yzen: after the matchmakers i'll redeploy flowmanager from the review2 branch
[15:49:38 CST(-0600)] <yzen> avtar: colinclark if match maker related ports change cloud based one for the flow manage needs to change too https://github.com/GPII/universal/blob/master/gpii/node_modules/flowManager/configs/cloudBased.json#L22
[15:50:28 CST(-0600)] <cindyli> good idea, Justin_o_, to keep the original pcp demo code clean, even though we've touched internal scripts
[15:50:34 CST(-0600)] <avtar> yzen: only the matchmaker proxy edition gets mentioned there?
[15:50:40 CST(-0600)] <Justin_o_> cindyli: yes
[15:50:45 CST(-0600)] <yzen> avtar: yes
[15:50:49 CST(-0600)] <avtar> ok cool
[15:51:03 CST(-0600)] <Justin_o_> I figure it will be easier for people to know what's being demoed that way too
[15:51:12 CST(-0600)] <Justin_o_> cindyli: i'll try to quickly make that
[15:51:17 CST(-0600)] <cindyli> that's right. thanks, Justin_o_
[15:51:32 CST(-0600)] <avtar> colinclark, yzen: i'll be back in 5
[15:51:48 CST(-0600)] <avtar> gonna stretch out my old man body
[15:53:09 CST(-0600)] <Justin_o_> colinclark: can you remind me what PCP stands for
[15:53:23 CST(-0600)] <colinclark> Bosmon: So I'm just trying to tie up mental loose ends here
[15:53:42 CST(-0600)] <colinclark> Are you willing to make a pull request that modifies proxy.json based on avtar's requests?
[15:53:57 CST(-0600)] <colinclark> presumably it would including giving a real, non-8080 port to the matchmakerProxy
[15:54:11 CST(-0600)] <colinclark> and adding configurations for STMM and RBMM that point to his deployed services
[15:54:14 CST(-0600)] <colinclark> Do have that correct?
[15:54:58 CST(-0600)] <Bosmon> The latter two I understand
[15:55:08 CST(-0600)] <Bosmon> The former I don't
[15:55:14 CST(-0600)] <Bosmon> "presumably it would including giving a real, non-8080 port to the matchmakerProxy"
[15:55:30 CST(-0600)] <colinclark> sorry, it was ungrammatical
[15:55:38 CST(-0600)] <colinclark> I'm rushing and I shouldn't
[15:55:39 CST(-0600)] <colinclark> ok
[15:55:43 CST(-0600)] <Bosmon> The grammar is fine
[15:55:49 CST(-0600)] <Bosmon> It's just the underlying meaning I have problems with
[15:55:56 CST(-0600)] <Bosmon> Or more precisely, the underlying purpose
[15:55:58 CST(-0600)] <colinclark> So, I believe that by default, a Kettle server will bind to port 8080
[15:56:06 CST(-0600)] <colinclark> if a port isn't explicitly defined for the server
[15:56:14 CST(-0600)] <Bosmon> ok
[15:56:28 CST(-0600)] <colinclark> If you look here, you'll see that no port is defined: https://github.com/GPII/universal/blob/master/gpii/node_modules/matchMaker/configs/proxy.json#L9
[15:57:43 CST(-0600)] <Bosmon> Oh yes
[15:57:49 CST(-0600)] <Bosmon> I think I am beginning to understand
[15:57:59 CST(-0600)] <Bosmon> All this time, these matchmakers have already been dispatched internally via HTTP?
[15:58:25 CST(-0600)] <Bosmon> I guess I never read this configuration carefully
[15:58:34 CST(-0600)] <colinclark> The Matchmaker Proxy is itself a server that is dispatched to from the Flow Manager
[15:58:37 CST(-0600)] <Bosmon> Ok
[15:58:46 CST(-0600)] <Bosmon> But somehow THIS one defaults to 8080?
[15:58:48 CST(-0600)] <colinclark> The Proxy, in turn, resolves and dispatches to a particular MM instance
[15:58:54 CST(-0600)] <colinclark> yes, that's right, Bosmon
[15:58:59 CST(-0600)] <colinclark> I believe because it fails to define any port at all
[15:59:01 CST(-0600)] <Bosmon> And the Flow Manager doesn't
[15:59:10 CST(-0600)] <Bosmon> That is, doesn't default, but is fixed to 8081
[15:59:10 CST(-0600)] <Bosmon> Ok
[15:59:17 CST(-0600)] <colinclark> Am I making sense?
[15:59:22 CST(-0600)] <Bosmon> So I assume that there will need to be a yet further modification
[15:59:26 CST(-0600)] <colinclark> What's that?
[15:59:33 CST(-0600)] <Bosmon> Affecting the binding of the FlowManager itself onto this matchMakerProxy
[15:59:41 CST(-0600)] <colinclark> ah yes
[15:59:43 CST(-0600)] <colinclark> that must be true
[16:00:10 CST(-0600)] <Bosmon> At present I see this in the file named "cloudBased.json"
[16:00:30 CST(-0600)] <colinclark> in the top-level configs directory?
[16:00:40 CST(-0600)] <Bosmon> in flowManager's configs
[16:00:46 CST(-0600)] <colinclark> ah yes
[16:00:51 CST(-0600)] <colinclark> now I see it
[16:01:02 CST(-0600)] <Bosmon> NOW
[16:01:03 CST(-0600)] <Bosmon> YOU
[16:01:04 CST(-0600)] <colinclark> matcha matcha matchmaker
[16:01:07 CST(-0600)] <colinclark> ME
[16:01:07 CST(-0600)] <Bosmon> SEE
[16:01:10 CST(-0600)] <Bosmon> IT
[16:01:21 CST(-0600)] <colinclark> yes, those are the changes that need to be made
[16:01:23 CST(-0600)] <colinclark> to summarize
[16:01:31 CST(-0600)] <colinclark> bind to something other than 8080 here https://github.com/GPII/universal/blob/master/gpii/node_modules/flowManager/configs/cloudBased.json#L22
[16:02:02 CST(-0600)] <Justin_o_> cindyli: pushed up the new demo file
[16:02:03 CST(-0600)] <colinclark> 2. define said port here https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json#L9
[16:02:13 CST(-0600)] <cindyli> thanks, Justin_o_
[16:02:15 CST(-0600)] <colinclark> 3. add proper MM entries here https://github.com/GPII/universal/blob/review2/gpii/node_modules/matchMaker/configs/proxy.json#L15
[16:02:17 CST(-0600)] <cindyli> will merge in
[16:02:21 CST(-0600)] <Justin_o_> cindyli: thanks
[16:02:26 CST(-0600)] <Bosmon> Ok
[16:02:26 CST(-0600)] <cindyli> np
[16:02:27 CST(-0600)] <colinclark> and 4. issue a pull request to both master and review2
[16:02:34 CST(-0600)] <Bosmon> I will produce a pull request of such a kind
[16:02:43 CST(-0600)] <colinclark> But I'm worried about one thing
[16:02:50 CST(-0600)] <colinclark> what about non-cloud-based flow managers?
[16:03:04 CST(-0600)] <colinclark> don't they also use a Matchmaker Proxy
[16:03:05 CST(-0600)] <colinclark> ?
[16:03:12 CST(-0600)] <colinclark> presumably it's defined locally
[16:03:13 CST(-0600)] <Bosmon> I'm not sure I even understand the term "cloudBased"
[16:03:19 CST(-0600)] <Bosmon> So you will have to unpack this question
[16:03:29 CST(-0600)] <colinclark> it's an instance of the Flow Manager that runs entirely on a server
[16:03:53 CST(-0600)] <colinclark> Such that a web application or a Java "simple phone" or other non-native-local-Flow Manager device can be personalized
[16:04:14 CST(-0600)] <Bosmon> So by my reading, all the other configurations take their port numbers from "base.json"
[16:04:22 CST(-0600)] <colinclark> Very soon, we will cut off direct access to the Preferences Server by clients who aren't the Flow Manager or the preferences editors
[16:04:23 CST(-0600)] <Bosmon> Which causes them all to form a single server, listening on port 8081
[16:04:29 CST(-0600)] <colinclark> ok
[16:04:39 CST(-0600)] <colinclark> meaning they're all running locally within that flow manager instance
[16:05:02 CST(-0600)] <Bosmon> This is what I seem to understand
[16:05:03 CST(-0600)] <colinclark> I'm not entirely sure why the Cloud-Based instance doesn't use this same approach
[16:05:10 CST(-0600)] <colinclark> with the exception, of course, of the MMs themselves
[16:07:49 CST(-0600)] <colinclark> Bosmon: Are you okay to make this pull request?
[16:08:06 CST(-0600)] <Bosmon> colinclark - yes
[16:08:16 CST(-0600)] <colinclark> thanks
[16:08:16 CST(-0600)] <Bosmon> I will also correct those profiles which make corrupt references to the statistical matchmaker
[16:08:22 CST(-0600)] <colinclark> thanks
[16:08:32 CST(-0600)] <colinclark> Assuming that squares us away at least until michelled can test that the STMM is indeed working correctly on Windows
[16:08:38 CST(-0600)] <colinclark> which means I will head back to my slow demo
[16:08:51 CST(-0600)] <colinclark> and stop panicking
[16:10:48 CST(-0600)] <Bosmon> colinclark - what am I to bind the reference to the rulebased matchmaker onto?
[16:11:15 CST(-0600)] <colinclark> in the preference sets or in proxy.json?
[16:11:23 CST(-0600)] <Bosmon> !!!!
[16:11:35 CST(-0600)] <colinclark> four exclamation marks!
[16:11:44 CST(-0600)] <colinclark> here's five
[16:11:46 CST(-0600)] <colinclark> IIIII
[16:11:52 CST(-0600)] <Bosmon> 1111111!
[16:12:03 CST(-0600)] <colinclark> here are some pipes
[16:12:05 CST(-0600)] <colinclark> |||||||
[16:12:13 CST(-0600)] <Bosmon> ONEONEONEONEONEONEONEONEONE!
[16:12:27 CST(-0600)] <Bosmon> so, I'm not sure what you meant, "in the preference sets"
[16:12:37 CST(-0600)] <Bosmon> But I did mean proxy.json
[16:12:48 CST(-0600)] <colinclark> rbmm.gpii.net port 80
[16:12:54 CST(-0600)] <Bosmon> thanks
[16:12:58 CST(-0600)] <colinclark> presumably once Avtar redeploys it in development mode, it will actually work
[16:13:38 CST(-0600)] <colinclark> I guess we should also make a pull request to Andy's STMM repo once we've confirmed that it works
[16:13:55 CST(-0600)] <colinclark> But I am going to stop, now, being that guy who lists things that need to get done
[16:14:06 CST(-0600)] <Bosmon> It's a completely vital function
[16:14:16 CST(-0600)] <Bosmon> Woe betide us if we ever find ourselves devoid of a guy who lists things that need to get done
[16:16:19 CST(-0600)] <colinclark> michelled is back!
[16:16:23 CST(-0600)] <colinclark> how was your late late lunch?
[16:16:51 CST(-0600)] <michelled> it is good
[16:17:06 CST(-0600)] <michelled> so, I hate to ask this again, but what version of node and npm should I have?
[16:17:23 CST(-0600)] <michelled> I've got 0.8.23 and 1.2.18
[16:17:46 CST(-0600)] <michelled> and npm is complaining with the clean review branch
[16:18:52 CST(-0600)] <colinclark> that should be just fine on windows, michelled
[16:18:56 CST(-0600)] <colinclark> complaining how?
[16:20:57 CST(-0600)] <Bosmon> colinclark - I have been trying to prepare the pull request, but have unfortunately run into something mysterious
[16:21:03 CST(-0600)] <colinclark> oh, ok
[16:21:06 CST(-0600)] <colinclark> what is it?
[16:21:08 CST(-0600)] <Bosmon> In that it seems that a server corresponding to the "matchMakerProxy" is never started
[16:21:29 CST(-0600)] <Bosmon> I don't see any log messages for it, and netstat shows no server listening either on the port I chose (8083) or the default port 8080
[16:21:31 CST(-0600)] <colinclark> which config are you testing with?
[16:21:39 CST(-0600)] <Bosmon> multipleMatchMakers
[16:21:54 CST(-0600)] <colinclark> yzen: Does this mean anything to you? ^
[16:22:18 CST(-0600)] <colinclark> Are we doing some kind of local dispatch in this configuration instead? As I always dreamed we might?
[16:22:21 CST(-0600)] <michelled> colinclark: http://pastie.org/8627616
[16:22:46 CST(-0600)] <colinclark> michelled: Can you find out which package has the wrong version number>?
[16:23:01 CST(-0600)] <colinclark> perhaps it's listed in C:\Users\colin\code\gpii\windows\npm-debug.log
[16:24:03 CST(-0600)] <yzen> Bosmon: are you actually modifying multipleMatchMakers? because that is not what avtar will be deploying
[16:24:23 CST(-0600)] <Bosmon> yzen - I modifying proxy.json
[16:24:29 CST(-0600)] <avtar> Bosmon, colinclark, yzen: rbmm is running development mode now
[16:24:30 CST(-0600)] <Bosmon> In order to add a port number to the matchMakerProxy server
[16:24:34 CST(-0600)] <Bosmon> But actually I see no such server starting up
[16:25:07 CST(-0600)] <colinclark> thank you, avtar
[16:25:55 CST(-0600)] <yzen> Bosmon: let met try
[16:26:11 CST(-0600)] <Bosmon> yzen - thanks
[16:26:20 CST(-0600)] <Bosmon> I'll commit my pull request anyway, since it seems to be working
[16:26:24 CST(-0600)] <Bosmon> But not for reasons I can understand
[16:26:30 CST(-0600)] <colinclark> ok
[16:26:42 CST(-0600)] <Bosmon> So which config will avtar be deploying, if not this one?
[16:26:48 CST(-0600)] <colinclark> michelled: Bosmon's pull request may help with your testing, in that it sets up proxy.json for you
[16:27:00 CST(-0600)] <colinclark> cloudBased, Bosmon
[16:27:07 CST(-0600)] <yzen> Bosmon: he will deploy 2 proxy for match maker and cloudBased for flowmanager
[16:27:13 CST(-0600)] <yzen> ports in both need to be in sync
[16:28:58 CST(-0600)] <yzen> so Bosmon when i do NODE_ENV=proxy node node_modules/kettle/lib/init.js gpii/node_modules/matchMaker/configs/ i see in log
[16:28:59 CST(-0600)] <yzen> 17:28:06.761: Kettle Server is running on port: 8080
[16:29:00 CST(-0600)] <yzen> 17:28:06.761: Firing event onListen of component with typename kettle.server and id 8stdvg62-7 to list of 1 listeners
[16:29:01 CST(-0600)] <yzen> 17:28:06.762: Kettle Server is running on port: 8079
[16:29:02 CST(-0600)] <yzen> 17:28:06.762: Firing event onListen of component with typename kettle.server and id 8stdvg62-39 to list of 1 listeners
[16:29:38 CST(-0600)] <Bosmon> yzen - never had any idea about NODE_ENV=proxy since I don't see such an entry in gpii/configs
[16:29:38 CST(-0600)] <Bosmon> How does this work?
[16:30:21 CST(-0600)] <Bosmon> Ah I see, it is just a standalone matchmaker
[16:30:30 CST(-0600)] <yzen> Bosmon: yes
[16:30:36 CST(-0600)] <Bosmon> But my question is, why do I not see this server start up when I fire up multipleMatchMakers
[16:31:02 CST(-0600)] <Bosmon> Since as far as I see, that one includes matchMaker/configs/proxy.json
[16:31:10 CST(-0600)] <yzen> Bosmon: because in that config it is hosted alongside the flowManager
[16:31:14 CST(-0600)] <yzen> the same server
[16:31:16 CST(-0600)] <yzen> instance
[16:31:47 CST(-0600)] <Bosmon> yzen - how does it know to to do that, and ignore the server material in proxy.json?
[16:32:55 CST(-0600)] <yzen> Bosmon: because of how configs become grades for the top level config grade
[16:33:09 CST(-0600)] <Bosmon> I see
[16:33:14 CST(-0600)] <yzen> so the server subcomponent material is just merged without change
[16:34:24 CST(-0600)] <colinclark> So, Andy just responded
[16:34:33 CST(-0600)] <colinclark> with an answer that is worrisome
[16:35:01 CST(-0600)] <colinclark> Oh, thanks for that. I must admit that I just tried to redo what yura set up for the rbmm, without a documentation what those functions actually do =^^= Yep, the version of the smm (there are 2) that we send to the review requires the transformers. I thought that the matchpost function does that.
[16:35:02 CST(-0600)] <Bosmon> He just sent a followup
[16:35:10 CST(-0600)] <colinclark> PHEW
[16:35:12 CST(-0600)] <colinclark> thank you
[16:35:27 CST(-0600)] <colinclark> ok
[16:35:40 CST(-0600)] <colinclark> well, if michelled is able to overcome her environmental difficulties, I think we're very very close
[16:35:51 CST(-0600)] <colinclark> I guess we still need to PROBE the RBMM
[16:35:58 CST(-0600)] <colinclark> which I can do right now
[16:37:11 CST(-0600)] <michelled> sigh. the environmental difficulties are confusing me. so package.json has version '0.1' for gpii which seems to be what the error is complaining about.
[16:37:21 CST(-0600)] <colinclark> oh dear
[16:37:25 CST(-0600)] <michelled> but I just did a fresh checkout
[16:37:25 CST(-0600)] <colinclark> that seems very strange
[16:37:36 CST(-0600)] <colinclark> 0.1 isn't a valid semver format, is it?
[16:37:47 CST(-0600)] <colinclark> is your fresh checkout still complaining?
[16:37:57 CST(-0600)] <Bosmon> It should read "0.1.0"
[16:38:01 CST(-0600)] <Bosmon> I wonder what has happened to michelled
[16:38:56 CST(-0600)] <michelled> well, it says 0.1 in the windows package.json
[16:39:03 CST(-0600)] <colinclark> avtar: your version of the RBMM is running at http://rbmm.gpii.net, is that correct?
[16:39:08 CST(-0600)] <michelled> I too wonder what has happened
[16:39:09 CST(-0600)] <avtar> yes
[16:39:24 CST(-0600)] <colinclark> does this curl command look correct:
[16:39:26 CST(-0600)] <colinclark> curl -X POST -d "@mmpost.txt" http://rbmm.gpii.net/match --header "Content-Type:application/json"
[16:39:36 CST(-0600)] <colinclark> ooh, it's kasparnet!
[16:39:43 CST(-0600)] <colinclark> Did you manage to get to your destination, dude?
[16:39:49 CST(-0600)] <kasparnet> heya colinclark
[16:39:51 CST(-0600)] <kasparnet> I did
[16:39:52 CST(-0600)] <colinclark> It sounds like a lot of flights were cancelled/delayed
[16:39:52 CST(-0600)] <Bosmon> colinclark - that looks correct
[16:39:56 CST(-0600)] <colinclark> that's great
[16:40:10 CST(-0600)] <kasparnet> colinclark: yeah - they had some problems with the fog I think
[16:40:11 CST(-0600)] <Bosmon> colinclark - to be honest, I don't fully understand what I am doing with these port numbers
[16:40:19 CST(-0600)] <colinclark> I'm getting a 502 Bad Gateway
[16:40:22 CST(-0600)] <colinclark> avtar: ^
[16:40:23 CST(-0600)] <Bosmon> But I have basically just converted anything I found without a port number specified to 8083
[16:40:31 CST(-0600)] <colinclark> kasparnet: We're getting really close
[16:40:38 CST(-0600)] <colinclark> anything related to the proxy, you mean?
[16:40:48 CST(-0600)] <Bosmon> colinclark - any matchMaker
[16:40:50 CST(-0600)] <kasparnet> colinclark: awesome! Glad to hear that
[16:41:20 CST(-0600)] <colinclark> Bosmon: I'm not sure that's right
[16:41:25 CST(-0600)] <colinclark> but perhaps I'm also puzzled
[16:41:36 CST(-0600)] <colinclark> I THINK the only server at issue is the Proxy
[16:41:44 CST(-0600)] <Bosmon> colinclark - perhaps the ones in matchMaker/configs/production.json and development.json should be left alone
[16:41:51 CST(-0600)] <colinclark> let me check
[16:42:03 CST(-0600)] <Bosmon> What was the reason we wanted to avoid 8080, in any case?
[16:42:13 CST(-0600)] <colinclark> kasparnet: Your matchMakerType preferences are in the wrong format
[16:42:15 CST(-0600)] <colinclark> which caused a few issues
[16:42:33 CST(-0600)] <colinclark> and there were a lot of tricky details with the remote Matchmakers and the Proxy which we are just nearly resolving
[16:43:02 CST(-0600)] <Bosmon> colinclark - I also get Bad Gateway from the rbmm
[16:43:08 CST(-0600)] <colinclark> ok
[16:43:23 CST(-0600)] <kasparnet> colinclark: shit - sorry to hear that ... not seeing nick/clauda or andy here, I'll blame them for that
[16:43:47 CST(-0600)] <colinclark> they've all been very helpful
[16:44:33 CST(-0600)] <colinclark> Bosmon: No, I think you are correct
[16:44:53 CST(-0600)] <Justin_o_> cindyli: I have to go again. i'm not sure I'll be able to come back online tonight. I didn't really get anywhere with the save button in pcp
[16:44:58 CST(-0600)] <Bosmon> colinclark - change all of them?
[16:45:00 CST(-0600)] <colinclark> yzen: We're still confused, I'm sorry to distract you
[16:45:21 CST(-0600)] <cindyli> Justin_o_: i know, that's not a easy task
[16:45:24 CST(-0600)] <colinclark> but is it correct that in the case of matchmaker/configs/production.json and development.json, the matchmaker server will bind to port 8080?
[16:45:25 CST(-0600)] <yzen> colinclark: np
[16:45:41 CST(-0600)] <colinclark> Justin_o_: You've been an immense help
[16:45:46 CST(-0600)] <colinclark> Don't sweat it
[16:46:13 CST(-0600)] <colinclark> Justin_o_: Just before you leave, does kasparnet have everything he needs to run the PCP/PMT as it is currently implemented?
[16:46:18 CST(-0600)] <colinclark> is there anything he needs to know?
[16:47:48 CST(-0600)] <yzen> so colinclark neither production nor development is used when we deploy proxy.json
[16:47:55 CST(-0600)] <colinclark> right
[16:48:06 CST(-0600)] <colinclark> but if we're changing one configuration, does it make sense to change them all to not bind on 8080?
[16:48:22 CST(-0600)] <colinclark> assuming in production and development, there is indeed a server that binds on 8080?
[16:48:23 CST(-0600)] <colinclark> I guess that's the part I'm still fuzzy on
[16:50:11 CST(-0600)] <yzen> colinclark: it probably makes sense but that won't make avtar's life easier with deployment i guess
[16:50:15 CST(-0600)] <Justin_o_> colinclark: thanks.. and I replied to one or your e-mails yesterday with the info about the node based web server i used to run the PMT/PCP. Although he could use any he likes.. Also i'm not sure if he knows that we created a new demo in the demos/prefsEditors directory for the review
[16:51:24 CST(-0600)] <colinclark> kasparnet: ^^
[16:51:58 CST(-0600)] <avtar> colinclark: just saw your bad gateway message
[16:52:04 CST(-0600)] <colinclark> ok, thanks for clarifying, yzen
[16:52:29 CST(-0600)] <avtar> my bad
[16:52:41 CST(-0600)] <avtar> development.json uses port 80
[16:52:43 CST(-0600)] <avtar> one sec
[16:53:24 CST(-0600)] <colinclark> ok
[16:53:25 CST(-0600)] <avtar> colinclark: please try again
[16:53:31 CST(-0600)] <colinclark> ok
[16:53:33 CST(-0600)] <Justin_o_> good everyone and good luck
[16:53:40 CST(-0600)] <colinclark> thanks, Justin_o_
[16:53:42 CST(-0600)] <colinclark> You are amazing
[16:53:48 CST(-0600)] <colinclark> I really, really appreciate all your hard work
[16:54:00 CST(-0600)] <Justin_o_> thanks, sorry i couldn't do more.. hit a bunch of walls today
[16:54:08 CST(-0600)] <colinclark> Hey, you did great
[16:54:17 CST(-0600)] <avtar> stand tall, gentle justin
[16:54:32 CST(-0600)] <kasparnet> yeah - thanks Justin_o_! I'll try everything out tonight
[16:54:34 CST(-0600)] <colinclark> avtar: Can you see what the CPU is up to on that RBMM box?
[16:54:39 CST(-0600)] <colinclark> this request is SLOOOOW
[16:55:02 CST(-0600)] <colinclark> 504 Gateway Time-out
[16:55:17 CST(-0600)] <Justin_o_> colinclark: thanks
[16:55:29 CST(-0600)] <avtar> colinclark: interesting. i just saw this "fs: missing callback Error: EACCES, mkdir '/logs'"
[16:55:41 CST(-0600)] <colinclark> kasparnet: Just so you have this, here's how to manually PROBE a matchmaker with curl: http://pastebin.com/hckBZxZQ
[16:55:44 CST(-0600)] <avtar> which wasn't being reported earlier
[16:55:45 CST(-0600)] <avtar> one sec
[16:55:48 CST(-0600)] <yzen> Bosmon: sweet
[16:55:51 CST(-0600)] <Bosmon> colinclark - I've issued my pull request
[16:55:54 CST(-0600)] <Bosmon> to the review2 branch
[16:56:03 CST(-0600)] <colinclark> thanks Bosmon
[16:56:05 CST(-0600)] <Bosmon> I have omitted changes to the base ports for the matchmakers
[16:56:08 CST(-0600)] <colinclark> have a great chat with Clayton
[16:56:08 CST(-0600)] <Bosmon> I will be off to TTEA now
[16:56:12 CST(-0600)] <Bosmon> Back in perhaps 90 mins
[16:56:16 CST(-0600)] <colinclark> cool
[16:56:24 CST(-0600)] <colinclark> So kasparnet, we're just about on the verge of a working system
[16:56:42 CST(-0600)] <colinclark> We believe the STMM is now deployed correctly, but michelled was unable to get her Windows system working to test it
[16:56:42 CST(-0600)] <Justin_o_> kasparnet: great.. hope you get it working.. and cindy is doing some more work now to try to get the pcp working a bit better.
[16:56:58 CST(-0600)] <colinclark> Bosmon's pull request should configure the Flow Manager correctly for both Matchmakers
[16:56:59 CST(-0600)] <kasparnet> thanks Justin_o_
[16:57:10 CST(-0600)] <colinclark> it looks like we need to resolve a small issue with the RBMM right now
[16:57:22 CST(-0600)] <yzen> Bosmon: kasparnet colinclark ill be back in 45min
[16:57:37 CST(-0600)] <colinclark> thanks, yzen
[16:57:43 CST(-0600)] <colinclark> but I think it may fall to you to verify that, in practice, your preferences sets work with both MMs
[16:58:07 CST(-0600)] <colinclark> you'll note that Bosmon's pull request also corrects the matchmaker preference for some of these key preference sets
[16:58:08 CST(-0600)] <kasparnet> avtar, colinclark: yeah - I seem to remember nick mentioning something about the /logs dir (that one might have to create it manually) - perhaps it was on his github page
[16:58:22 CST(-0600)] <colinclark> but you'll need to double-check them
[16:58:49 CST(-0600)] <colinclark> avtar: https://github.com/NickKaklanis/RuleBased_MatchMaker
[16:59:05 CST(-0600)] <colinclark> "When executing the Rule-based MatchMaker, a new directory (/logs) is automatically created. If there are limited permissions, an exception will be thrown. In this case, try to manually create the "logs" directory at the root of the hard drive."
[16:59:05 CST(-0600)] <kasparnet> colinclark: yeah - I was going to do that anyway
[16:59:22 CST(-0600)] <avtar> oh wow
[16:59:24 CST(-0600)] <avtar> that's horrible
[16:59:24 CST(-0600)] <michelled> kasparnet, colinclark: I'm off to a play soon. I'll be back online when I get back. kasparnet, I guess the issues I was seeing on my windows box yesterday were environmental. I'll need to look more deeply into how it's setup.
[16:59:25 CST(-0600)] <kasparnet> (actually I was going to have andy and claudia do it, but I'm not sure they're gonna see their mail tonight)
[16:59:42 CST(-0600)] <colinclark> ok
[16:59:44 CST(-0600)] <colinclark> avtar: yes
[16:59:52 CST(-0600)] <colinclark> yes it is
[17:00:01 CST(-0600)] <kasparnet> i +1 that
[17:01:02 CST(-0600)] <colinclark> avtar: if this ends up being too awful or fatal, we can just go with their version of the RBMM
[17:01:19 CST(-0600)] <avtar> well
[17:01:53 CST(-0600)] <avtar> i created /logs and it's owned by the nodejs user account
[17:02:20 CST(-0600)] <colinclark> shall I try again?
[17:02:35 CST(-0600)] <avtar> sure
[17:02:56 CST(-0600)] <colinclark> SUCCESS
[17:03:08 CST(-0600)] <colinclark> and it's about 1000x faster than the other server
[17:03:19 CST(-0600)] <colinclark> kasparnet: I'm curious about network latency from Lux
[17:03:22 CST(-0600)] <colinclark> can you do me a quick favour?
[17:03:33 CST(-0600)] <colinclark> can you compare the speeds of curling to both servers?
[17:03:40 CST(-0600)] <colinclark> curl -X POST -d "@mmpost.txt" http://160.40.50.130/match --header "Content-Type:application/json"
[17:03:48 CST(-0600)] <avtar> you know what
[17:03:48 CST(-0600)] <colinclark> and curl -X POST -d "@mmpost.txt" http://rbmm.gpii.net/match--header "Content-Type:application/json"
[17:03:50 CST(-0600)] <colinclark> what?
[17:03:59 CST(-0600)] <avtar> they probably either ran this as root or /logs was world writeable
[17:04:05 CST(-0600)] <colinclark> yes
[17:04:18 CST(-0600)] <avtar> because i think the java app wants to write to that directory as well
[17:04:22 CST(-0600)] <colinclark> I would not be even the tiniest bit surprised
[17:10:09 CST(-0600)] <colinclark> avtar: So the preferences server is now deployed too, right?
[17:10:19 CST(-0600)] <avtar> yes
[17:10:20 CST(-0600)] <colinclark> we've got STMM and RBMM running nicely
[17:10:29 CST(-0600)] <avtar> nicely heh
[17:10:36 CST(-0600)] <colinclark>
[17:10:37 CST(-0600)] <colinclark> Cloud-Based Flow Manager, what about him?
[17:10:49 CST(-0600)] <avtar> the java process does spike processor usage up to 99%
[17:10:54 CST(-0600)] <colinclark> lol
[17:11:32 CST(-0600)] <avtar> we might not be able to fit all this on one droplet
[17:12:41 CST(-0600)] <avtar> regarding flowmanager, i think i'm waiting on a configuration
[17:13:15 CST(-0600)] <colinclark> the one from Bosmon?
[17:13:30 CST(-0600)] <avtar> yes
[17:14:38 CST(-0600)] <kasparnet> avtar: shit man – that's quite the difference in server speed
[17:15:16 CST(-0600)] <colinclark> avtar: I'll review and push that change to the review2 branch now
[17:15:18 CST(-0600)] <colinclark> and then we can test it out
[17:18:52 CST(-0600)] <avtar> kasparnet: nothing but the best for you
[17:23:06 CST(-0600)] <colinclark> avtar: I'll be a bit slow in getting that pull in
[17:23:08 CST(-0600)] <colinclark> gregg is on the phone
[17:27:09 CST(-0600)] <avtar> no worries
[17:27:45 CST(-0600)] <avtar> i'm going to tackle tony's tasks until you get back
[17:28:02 CST(-0600)] <colinclark> ok
[18:06:44 CST(-0600)] <colinclark> kasparnet: are you in the midst of pushing stuff to review2?
[18:07:02 CST(-0600)] <kasparnet> colinclark: sort of :S
[18:07:15 CST(-0600)] <colinclark> I was going to push https://github.com/GPII/universal/pull/200/files
[18:07:23 CST(-0600)] <colinclark> but I don't know if i should wait for you to take care of yours?
[18:07:32 CST(-0600)] <colinclark> is everything ok, kasparnet?
[18:07:36 CST(-0600)] <kasparnet> nope - do it ... that one is the most important one
[18:07:46 CST(-0600)] <kasparnet> yeah - it's just device reporter and solution registry stuff
[18:07:48 CST(-0600)] <colinclark> ok
[18:07:50 CST(-0600)] <colinclark> here I go
[18:08:38 CST(-0600)] <kasparnet> colinclark: I'm so fucking happy with the result of all the hard work you guys have done – things are actually looking somewhat promising right now
[18:09:05 CST(-0600)] <colinclark> kasparnet: I'm so fucking happy you are the voice of optimism right now
[18:09:12 CST(-0600)] <kasparnet> hehe
[18:09:17 CST(-0600)] <colinclark> as avtar can attest, things were looking pretty bleak a few hours ago
[18:09:23 CST(-0600)] <colinclark> so it's nice to hear some positive energy
[18:09:27 CST(-0600)] <avtar> ^ this
[18:09:50 CST(-0600)] <colinclark> ok kasparnet, avtar
[18:09:51 CST(-0600)] <colinclark> it's in
[18:10:11 CST(-0600)] <kasparnet> yay
[18:10:11 CST(-0600)] <colinclark> kasparnet: It's a flow manager, so can I just test it once avtar redeploys by doing the usual login thingy?
[18:10:24 CST(-0600)] <colinclark> i.e. flowmanager.gpii.net/user/sammy/login or whatever?
[18:10:32 CST(-0600)] <colinclark> is there a particular preference set that you recommend I test with?
[18:11:15 CST(-0600)] <kasparnet> colinclark: while not sure, I think due to the configuration being for a cloudbased flowmanager it might behave a bit special
[18:11:23 CST(-0600)] <kasparnet> I'd start with testing:
[18:11:28 CST(-0600)] * kasparnet finding payload example
[18:12:10 CST(-0600)]
<kasparnet> http://127.0.0.1:8081/carla/settings/
[18:12:20 CST(-0600)] <kasparnet> ooops
[18:12:30 CST(-0600)] <kasparnet> that last url is a reference to where I got the payload from
[18:12:47 CST(-0600)] <kasparnet> and obviously the 127.0.0.1:8081 should be replaced with flowmanager.gpii.net
[18:13:12 CST(-0600)] <colinclark> ok
[18:13:33 CST(-0600)] <colinclark> avtar: ping me up when you're ready for me to fire off this amazing command that kasper gave me
[18:13:39 CST(-0600)] <avtar> ok
[18:13:57 CST(-0600)] <colinclark> kasparnet: I'm also super relieved to have Javi on the scene
[18:15:46 CST(-0600)] <kasparnet> yes, indeed! I've just been out for a smoke with him
[18:16:40 CST(-0600)] <kasparnet> after having spent the weekend (and week for that matter) physically alone with this stuff, I needed some company to chat about the craziness of this sprint
[18:16:45 CST(-0600)] <colinclark> yes
[18:16:53 CST(-0600)] <colinclark> I'm so sorry it's been so crazy for you
[18:16:56 CST(-0600)] <colinclark> and for everyone
[18:17:16 CST(-0600)] <kasparnet> yeah, everyone!!!
[18:35:23 CST(-0600)] <colinclark> crazy quotes
[18:44:21 CST(-0600)] <avtar> og matchmaker is online in proxy mode using the review2 branch
[18:44:59 CST(-0600)] <kasparnet> colinclark, avtar: jhernandez just tried some of the NP sets out on linux, and there are some good news and some bad news
[18:45:10 CST(-0600)] <colinclark> og
[18:45:13 CST(-0600)] <colinclark> kasparnet: ok
[18:45:16 CST(-0600)] <colinclark> what have you learned?
[18:45:30 CST(-0600)] <colinclark> what's the URL, avtar?
[18:45:31 CST(-0600)] <kasparnet> good news is that for one of the NP sets of test 1 (RB MM) things are working!!!! YAY
[18:45:40 CST(-0600)] <colinclark> wow
[18:45:41 CST(-0600)] <colinclark> awesome
[18:45:45 CST(-0600)] <kasparnet> Bad news is that the ST MM seems to be giving some problems
[18:46:03 CST(-0600)] <colinclark> ok, tell me about it
[18:46:32 CST(-0600)] <kasparnet> Javier Hernández: getting "connection refused" when using review2_demo3b_blind
[18:46:32 CST(-0600)] <kasparnet> Javier Hernández: looks like the stmm is not responding
[18:46:54 CST(-0600)] <kasparnet> colinclark: I'm fixing up my windows machine right now and will test things as well
[18:47:04 CST(-0600)] <colinclark> hmm
[18:47:10 CST(-0600)] <colinclark> ok, let's take a look at this
[18:48:57 CST(-0600)] <avtar> colinclark: one sec, going to do the same for flowmanager
[18:49:09 CST(-0600)] <colinclark> ok
[18:51:53 CST(-0600)] <avtar> colinclark, kasparnet: does og matchmaker need to be reachable directly via the internet or does it only communicate with the flowmanager?
[18:52:03 CST(-0600)] <colinclark> just with the Flow Manager
[18:52:12 CST(-0600)] <avtar> ok
[18:52:34 CST(-0600)] <kasparnet> avtar: what is this "og" matchmaker you're talking about?
[18:52:47 CST(-0600)] <avtar> kasparnet: the one in the universal repo
[18:52:51 CST(-0600)] <avtar> and not stmm or rbmm
[18:52:52 CST(-0600)] <kasparnet> ok
[18:52:59 CST(-0600)] <kasparnet> and why "og"
[18:53:22 CST(-0600)] <avtar> kasparnet: http://www.urbandictionary.com/define.php?term=OG
[18:53:34 CST(-0600)] <colinclark> avtar: I'm getting the same error that jhernandez saw with the STMM
[18:53:36 CST(-0600)] <kasparnet> hahahaha, that's awesome avtar
[18:53:48 CST(-0600)] <colinclark> it seems to be ECONNREFUSEing us again
[18:54:05 CST(-0600)] <colinclark> NOBODY ECONNREFUSES KASPARNET!
[18:54:09 CST(-0600)] <avtar> does it matter that flowmanager is offline at the moment?
[18:54:24 CST(-0600)] <colinclark> It shouldn't, I don't think
[18:54:31 CST(-0600)] <kasparnet> nope - I agree - it shouldn't
[18:55:13 CST(-0600)] <colinclark> the rbmm is still responding well
[18:55:20 CST(-0600)] <colinclark> so I think the answer must be no
[18:55:37 CST(-0600)] <colinclark> and so speedily
[18:57:25 CST(-0600)] <avtar> colinclark: could you try it again?
[18:57:38 CST(-0600)] <colinclark> yup
[18:58:06 CST(-0600)] <colinclark> same result, avtar
[18:58:14 CST(-0600)] <colinclark> does it look like it's running from your end?
[18:58:25 CST(-0600)] <colinclark> and started in development mode and all that?
[18:58:37 CST(-0600)] <colinclark> Could it accidentally be in production mode?
[19:00:45 CST(-0600)] <jhernandez> c'mon colinclark!!
[19:00:50 CST(-0600)] <jhernandez> xDDD
[19:00:59 CST(-0600)] <jhernandez> android wants to use the stmm too
[19:01:01 CST(-0600)] <jhernandez>
[19:01:04 CST(-0600)] <colinclark> I know, I suck
[19:01:08 CST(-0600)] <colinclark>
[19:01:12 CST(-0600)] <avtar> colinclark: i will check in one sec
[19:01:15 CST(-0600)] <colinclark> thanks avtar
[19:01:18 CST(-0600)] <avtar> just about wrapping flowmanager stuff
[19:01:21 CST(-0600)] <colinclark> great
[19:01:26 CST(-0600)] <colinclark> I'll test it while you look at the STMM
[19:07:46 CST(-0600)] <avtar> colinclark: how about now?
[19:08:40 CST(-0600)] <colinclark> does this look sane to you, avtar?
[19:08:42 CST(-0600)] <colinclark> curl -X POST -d "@stmmpost.txt" http://stmm.gpii.net/match --header "Content-Type:application/json"
[19:08:56 CST(-0600)] <colinclark> I'm still being ECONNREFUSED
[19:09:19 CST(-0600)] <avtar> k i'll stop stmm and start it from the command line
[19:09:47 CST(-0600)] <Bosmon> Hey there chaps
[19:09:51 CST(-0600)] <Bosmon> How's it going
[19:10:07 CST(-0600)] <colinclark> we're a bit back to where you were when you left
[19:10:16 CST(-0600)] <colinclark> the stmm seems to be not working again
[19:10:20 CST(-0600)] <colinclark> but hopefully its just some detail
[19:10:50 CST(-0600)] <colinclark> Bosmon: javi and kasper are here from the other side
[19:10:53 CST(-0600)] <colinclark> testing away
[19:10:53 CST(-0600)] <avtar> colinclark: what's in stmmpost.txt?
[19:11:11 CST(-0600)] <colinclark> i fabricated it, so it might be slightly wrong
[19:11:16 CST(-0600)] <colinclark> let me pastebin it
[19:11:21 CST(-0600)] <avtar> sure
[19:11:28 CST(-0600)] <avtar> i restarted it btw
[19:11:35 CST(-0600)] <Bosmon> It's working for me, avtar and colinclark
[19:11:41 CST(-0600)] <Bosmon> STMM
[19:11:46 CST(-0600)] <colinclark> http://pastebin.com/WPg4zLBg
[19:12:00 CST(-0600)] <colinclark> Bosmon: Do you still have the curl payload you were using for the STMM?
[19:12:07 CST(-0600)] <Bosmon> ECONNREFUSED will be a symptom of that misconfiguration yzen mentioned
[19:12:14 CST(-0600)] <avtar> colinclark, Bosmon: i'm seeing activity now
[19:12:19 CST(-0600)] <colinclark> i.e. production mode, right Bosmon?
[19:12:22 CST(-0600)] <Bosmon> That is, the production config causing it to try to connect to the nonexistent ontology server
[19:12:25 CST(-0600)] <Bosmon> colinclark - right
[19:12:25 CST(-0600)] <colinclark> oh yes, it works now
[19:12:30 CST(-0600)] <avtar> colinclark: one sec
[19:12:34 CST(-0600)] <avtar> let me stop it again
[19:12:37 CST(-0600)] <colinclark> ok
[19:14:00 CST(-0600)] <avtar> colinclark: how about now?
[19:14:10 CST(-0600)] <colinclark> it's a thing of beauty
[19:14:16 CST(-0600)] <Bosmon> It's hardly that
[19:14:18 CST(-0600)] <Bosmon> But it is working
[19:14:21 CST(-0600)] <avtar> so wtf
[19:14:26 CST(-0600)] <colinclark> contextual vs. universal
[19:14:27 CST(-0600)] <avtar> all i did was restart the process
[19:14:46 CST(-0600)] <avtar> hmm
[19:14:46 CST(-0600)] <colinclark> in the context of 2:15 am in Luxembourg, it's a thing of rare beauty
[19:14:55 CST(-0600)] <jhernandez> sweet!
[19:15:01 CST(-0600)] <colinclark> jhernandez: can you try it?
[19:15:02 CST(-0600)] <jhernandez> android worked with the stmm
[19:15:05 CST(-0600)] <jhernandez>
[19:15:05 CST(-0600)] <Bosmon> What is the first hour at which someone will actually do a demo? : P
[19:15:06 CST(-0600)] <colinclark> FUCK YES
[19:15:10 CST(-0600)] <colinclark> YES YES YES
[19:15:19 CST(-0600)] <jhernandez> now trying on linux, but it should work too
[19:15:20 CST(-0600)] <jhernandez>
[19:15:28 CST(-0600)] <colinclark> oh thank goodness
[19:15:38 CST(-0600)] <colinclark> Bosmon: We have positive results of both MMs, then
[19:15:46 CST(-0600)] <avtar> gals and guys, could you try the other components as well?
[19:15:56 CST(-0600)] <Bosmon> colinclark - I guess that means that we will avoid a significant quantity of BLAME : P
[19:15:57 CST(-0600)] <avtar> flowmanager, og matchmaker, et all
[19:15:59 CST(-0600)] <colinclark> sorry avtar
[19:16:05 CST(-0600)] <colinclark> i'll stop celebrating
[19:16:13 CST(-0600)] <colinclark> ok OG Flow
[19:16:35 CST(-0600)] <colinclark> Bosmon: We're still going to get blamed
[19:16:44 CST(-0600)] <Bosmon> Certainly we will
[19:16:48 CST(-0600)] <jhernandez> worked fine on linux too
[19:16:48 CST(-0600)] <colinclark> for not "releasing" the proxy
[19:16:51 CST(-0600)] <colinclark> for not documenting anything
[19:16:52 CST(-0600)] <Bosmon> But as you say, contextual vs. universal : P
[19:16:56 CST(-0600)] <colinclark>
[19:17:45 CST(-0600)] <kasparnet> oooh - what a fantastic IRC to catch up on
[19:18:00 CST(-0600)] <Bosmon> What fun for you, KASPARNETTT : P
[19:18:18 CST(-0600)] <kasparnet> my windows is all set (except of course for the million things I've surely forgotten), so I'll start going through the NP sets on that
[19:19:07 CST(-0600)] <jhernandez> avtar: both stmm and rbmm are working fine here
[19:19:19 CST(-0600)] <Bosmon> yzen - are you still looking at socket.io?
[19:19:28 CST(-0600)] <colinclark> avtar: I'm trying to look at the Flow Manager now
[19:20:12 CST(-0600)] <cindyli> colinclark: i'm about to leave for tonight. is there anything you want from me before I leave?
[19:21:01 CST(-0600)] <cindyli> have sent an update to Boyan
[19:21:08 CST(-0600)] <cindyli> and the list
[19:21:11 CST(-0600)] <colinclark> cindyli: you are completely amazing
[19:21:15 CST(-0600)] <colinclark> I mean, seriously
[19:21:20 CST(-0600)] <cindyli>
[19:21:24 CST(-0600)] <colinclark> I'm sorry you had to give up a huge chunk of your weeknd
[19:21:26 CST(-0600)] <colinclark> weekend
[19:21:29 CST(-0600)] <colinclark> and I really appreciate your help
[19:21:35 CST(-0600)] <colinclark> Have a really good evening
[19:21:41 CST(-0600)] <colinclark> and say sorry to your family for me
[19:22:21 CST(-0600)] <cindyli> no worries, you've been working longer. have a good evening too, colinclark
[19:22:25 CST(-0600)] <colinclark> thanks again, cindyli
[19:22:29 CST(-0600)] <colinclark> This is the request I'm using to test the OG Flow Manager with: http://pastebin.com/uvK2ATGw
[19:22:43 CST(-0600)] <colinclark> the response, avtar, is pure contextual hotness
[19:22:45 CST(-0600)]
<colinclark>
[19:23:04 CST(-0600)] <colinclark> kasparnet, jhernandez: seem like a reasonable response from it?
[19:23:30 CST(-0600)] <colinclark> kasparnet: Are you satisfied with the state of the Preferences Server at this point?
[19:24:13 CST(-0600)] <avtar> ah crap
[19:24:15 CST(-0600)] <avtar> the pref server
[19:24:28 CST(-0600)] <avtar> i forgot about those four sets that b0rked with kanso
[19:24:31 CST(-0600)] <colinclark> oh yes
[19:24:42 CST(-0600)] <colinclark> isn't kasper going to look at those?
[19:24:50 CST(-0600)] <colinclark> michelled said they were valid JSON
[19:24:59 CST(-0600)] <avtar> i believe michelle said it's valid json
[19:24:59 CST(-0600)] <kasparnet> colinclark: yeah - that response looks correct
[19:25:00 CST(-0600)] <avtar> yes
[19:25:11 CST(-0600)] <avtar> so it must be a combination of bash and kanso
[19:25:20 CST(-0600)] <kasparnet> avtar: and it was only those 4 np sets?
[19:25:24 CST(-0600)] <avtar> yup
[19:25:29 CST(-0600)] <colinclark> avtar: I could totally give you a big germy kiss right now
[19:25:44 CST(-0600)] <avtar> don't hear that often...
[19:25:45 CST(-0600)] <avtar> enough
[19:25:50 CST(-0600)] <kasparnet> haha
[19:27:32 CST(-0600)] <avtar> colinclark: i'm still seeing this on the vm that's running rbmm
[19:27:34 CST(-0600)] <avtar> fs: missing callback Error: EEXIST, mkdir '/logs'
[19:27:34 CST(-0600)] <avtar> fs: missing callback Error: EEXIST, mkdir '/logs'
[19:27:34 CST(-0600)] <avtar> fs: missing callback Error: EEXIST, mkdir '/logs'
[19:27:49 CST(-0600)] <colinclark> urg
[19:28:05 CST(-0600)] <colinclark> do we care if it is returning correct results?
[19:28:09 CST(-0600)] <colinclark> I can't believe I'm even asking that
[19:28:24 CST(-0600)] <avtar> ls /logs | wc -l
[19:28:24 CST(-0600)] <avtar> 48
[19:28:50 CST(-0600)] <avtar> no, i guess we don't :|
[19:28:57 CST(-0600)] <jhernandez> kasparnet: https://github.com/GPII/universal/blob/review2/testData/preferences/review2_demo4.2_large_print.json#L7
[19:29:22 CST(-0600)] <jhernandez> I guess you mean magnifierEnabled
[19:29:30 CST(-0600)] <jhernandez>
[19:30:50 CST(-0600)] <jhernandez> do you want me to create a pr of it?
[19:30:55 CST(-0600)] <kasparnet> jhernandez: indeed
[19:31:07 CST(-0600)] <kasparnet> and yes please
[19:31:13 CST(-0600)] <kasparnet> (to the review2 branch)
[19:31:55 CST(-0600)] <jhernandez> ACK
[19:32:00 CST(-0600)] <kasparnet> cheers javi
[19:34:53 CST(-0600)] <colinclark> kasparnet: So in terms of the Preferences Server and the few preferences sets that failed...
[19:34:57 CST(-0600)] <colinclark> Are you able to look at those?
[19:35:09 CST(-0600)] <colinclark> I'm hoping for us to get to the point where avtar can take his leave from us
[19:35:18 CST(-0600)] <kasparnet> colinclark: yes, will do so now
[19:35:36 CST(-0600)] <colinclark> avtar: So both RBMM and STMM test out, as well as the OG Flow Manager
[19:35:47 CST(-0600)] <colinclark> is there anything else that you need us to test or worry about?
[19:35:50 CST(-0600)] <kasparnet> avtar: would you happen to still have the output telling which NP sets they are
[19:35:54 CST(-0600)] <colinclark> How did it go with Tony's stuff?
[19:36:06 CST(-0600)] <avtar> colinclark: it's still going
[19:36:31 CST(-0600)] <avtar> this took longer than i anticipated
[19:36:43 CST(-0600)] <avtar> but i'll wrap up his tasks tonight
[19:36:57 CST(-0600)] <avtar> kasparnet: ye
[19:36:57 CST(-0600)] <avtar> s
[19:38:42 CST(-0600)] <avtar> http://pastie.org/pastes/8627054/text
[19:38:45 CST(-0600)] <avtar> kasparnet ^
[19:39:01 CST(-0600)] <kasparnet> thanks - my log has been eaten at some point during all my disconnects tonight
[19:39:35 CST(-0600)] <colinclark> avtar: Is there anything I can do to help with Tony's stuff?
[19:40:18 CST(-0600)] <colinclark> Bosmon: If you have a free moment, can you submit a pull request to Andreas' Statistical Matchmaker repository?
[19:40:59 CST(-0600)] <avtar> colinclark: nah, all good
[19:41:07 CST(-0600)] <colinclark> ok
[19:41:23 CST(-0600)] <avtar> i'm going to go have dinner and then finish that
[19:43:22 CST(-0600)] <colinclark> cool, enjoy
[19:43:24 CST(-0600)] <kasparnet> avtar, colinclark: looks like it has to do with us using double backslash in those files
[19:43:28 CST(-0600)] <kasparnet> grep -l '
' *
[19:43:28 CST(-0600)] <kasparnet> bak_review2_demo1.json
[19:43:28 CST(-0600)] <kasparnet> bak_review2_demo3b.json
[19:43:30 CST(-0600)] <kasparnet> os_win7.json
[19:43:32 CST(-0600)] <kasparnet> screenreader_nvda.json
[19:44:01 CST(-0600)] <kasparnet> (ok - my cygwin output aint pretty)
[19:44:15 CST(-0600)] <kasparnet> anyway - it's only in those 4 exact files
[19:47:46 CST(-0600)] <avtar> brb
[19:48:02 CST(-0600)] <kasparnet> colinclark, avtar: Considering that they're "just" backup sets (and standard sets we're not using for the review), I'll say that unless you have some straight forward magic up your sleeve, it's not something that is vital to fix tonight
[19:48:22 CST(-0600)] <kasparnet> I imagine we could always upload them with a little curl script or something
[19:48:41 CST(-0600)] <colinclark> kasparnet: Yes, that seems fine
[19:48:48 CST(-0600)] <colinclark> if we don't need them for the review, we won't sweat them
[19:49:04 CST(-0600)] <colinclark> and as you say, if you find yourself in a bind while doing the rehearsals tomorrow, you can just load 'em up via curl
[19:49:39 CST(-0600)] <colinclark> kasparnet: Do you think it's safe at this point to send out a status update telling people that the MMs are working awesomely?
[19:49:56 CST(-0600)] <yzen> Bosmon: yes
[19:50:04 CST(-0600)] <yzen> i can get to communicate to the server
[19:50:19 CST(-0600)] <yzen> but i have the issue with cookies still so thats what i m looking at
[19:50:28 CST(-0600)] <Bosmon> Oh cool
[19:50:35 CST(-0600)] <Bosmon> Have you tried the hosts file dodge?
[19:51:05 CST(-0600)] <yzen> Bosmon: i did i could not get it working just yet
[19:51:07 CST(-0600)] <Bosmon> hi colinclark - a pull request from my branch?
[19:51:32 CST(-0600)] <colinclark> Bosmon: yup, that's the version we have deployed currently
[19:51:40 CST(-0600)] <colinclark> so it might as well go back to its "main" repository
[19:51:48 CST(-0600)] <colinclark> to save you from having to be permanently responsible for it
[19:52:05 CST(-0600)] <kasparnet> colinclark: well - we haven't tested them in windows yet - but yeah, it's a good idea
[19:52:12 CST(-0600)] <colinclark> ok
[19:52:22 CST(-0600)] <colinclark> I'm just trying to save you the effort of having to explain stuff to people
[19:52:28 CST(-0600)] <colinclark> since I imagine you're pretty panicked doing the testing
[19:52:37 CST(-0600)] <kasparnet> perhaps just a mention that we have to run the tests
[19:52:55 CST(-0600)] <kasparnet> no, I agree, it's a good idea
[19:53:03 CST(-0600)] <kasparnet> and thanks
[19:58:35 CST(-0600)] <colinclark> yzen: I don't want to distract you if you're focusing. If it would help, I'd be happy to listen to the issue you're facing with the cookie
[19:58:48 CST(-0600)] <yzen> colinclark: please
[19:58:57 CST(-0600)] <yzen> so i can now send socket messages
[19:59:00 CST(-0600)] <colinclark> great
[19:59:03 CST(-0600)] <yzen> but in order for them to fully work
[19:59:11 CST(-0600)] <yzen> i need a session established
[19:59:21 CST(-0600)] <yzen> it's done by logging in a user
[19:59:23 CST(-0600)] <colinclark> on the Flow Manager sever?
[19:59:26 CST(-0600)] <yzen> yes
[19:59:28 CST(-0600)] <colinclark> ok
[19:59:40 CST(-0600)] <yzen> and i do send the right headers afaik
[19:59:55 CST(-0600)] <yzen> Set-cookie
[20:00:05 CST(-0600)] <yzen> i see headers in the firebug
[20:00:09 CST(-0600)] <yzen> but they are not saved
[20:00:25 CST(-0600)] <colinclark> Can you easily send me an actual example of the headers you are sending?
[20:00:33 CST(-0600)] <yzen> well sure
[20:00:34 CST(-0600)] <colinclark> Does the problem occur equally in Chrome?
[20:00:42 CST(-0600)] <yzen> Set-Cookie kettle.sid=s%3AM0vevq7TE0bGi5qkovFSMO15.Zzu0c0Egvc5Crci8J2zShbwb%2BFb616iJz%2BT0slRJw%2Fw; Path=/; HttpOnly
[20:00:46 CST(-0600)] <yzen> for example
[20:02:21 CST(-0600)] <yzen> i was thinking the issue is that I'm on localhost
[20:02:22 CST(-0600)] <colinclark> How is the cookie then used by the client?
[20:02:26 CST(-0600)] <colinclark> that's possible
[20:02:31 CST(-0600)] <colinclark> are you using a local Node.js server?
[20:02:34 CST(-0600)] <yzen> well it's just saved as a session cookie
[20:02:35 CST(-0600)] <colinclark> as Cindy and Justin recommended?
[20:02:36 CST(-0600)] <yzen> yes
[20:02:48 CST(-0600)] <colinclark> hmm
[20:03:13 CST(-0600)] <colinclark> It's saved as a session cookie, but then when do you expect the browser to use it?
[20:03:27 CST(-0600)] <colinclark> i.e. do you need it for subsequent requests to the Flow Manager?
[20:03:35 CST(-0600)] <colinclark> or is it read by client-side code?
[20:03:48 CST(-0600)] <yzen> so when we ll be doing the update
[20:04:01 CST(-0600)] <yzen> we would need to make sure that it's done as part of the session
[20:04:31 CST(-0600)] <colinclark> meaning the session cookie needs to be sent back to the Flow Manager?
[20:04:48 CST(-0600)] <colinclark> and that presumably is also done via the socket.io connection
[20:05:26 CST(-0600)] <colinclark> I guess perhaps it is a same origin policy issue...
[20:05:43 CST(-0600)] <colinclark> since your webpage is running on localhost:80 and the Flow Manager is on 8081?
[20:06:26 CST(-0600)] <colinclark> So I guess two things come to mind, neither of which may be very sensible
[20:06:35 CST(-0600)] <yzen> colinclark: was thinking that too let me see
[20:06:43 CST(-0600)] <colinclark> 1. Is it possible to serve up the PCP/PMT pages from the Flow Manager itself?
[20:07:07 CST(-0600)] <colinclark> 2. Is there some other way to send along the session ID to the Flow Manager rather than relying on a cookie?
[20:07:34 CST(-0600)] <colinclark> It seems like #1 might well be much simpler
[20:07:36 CST(-0600)] <yzen> colinclark: so the cookie is not actually set at all
[20:07:42 CST(-0600)] <colinclark> ok
[20:08:57 CST(-0600)] <kasparnet> colinclark, yzen: a bit of an odd question, but when I'm running a local instance of the GPII in multi-matchmaker mode
[20:09:02 CST(-0600)] <kasparnet> what solution registry is being used?
[20:09:12 CST(-0600)] <kasparnet> it seems it's not the one on my local file system
[20:09:21 CST(-0600)] <yzen> kasparnet: the one which the match maker is deployed with
[20:09:28 CST(-0600)] <colinclark> that's not an odd question at all
[20:09:33 CST(-0600)] <colinclark> we had a lot of confusion about that earlier
[20:09:41 CST(-0600)] <kasparnet> ok, so each matchmaker has it's own version of the repo
[20:09:45 CST(-0600)] <colinclark> yes
[20:09:48 CST(-0600)] <kasparnet> of the solutionregistry*
[20:09:50 CST(-0600)] <kasparnet> ok
[20:09:51 CST(-0600)] <colinclark> will you need access to this to modify it?
[20:09:57 CST(-0600)] <colinclark> or will that be a terrible problem?
[20:10:07 CST(-0600)] <kasparnet> no, no terrible problems
[20:10:24 CST(-0600)] <kasparnet> but there is one entry in the statistical MM that I would like updated
[20:10:32 CST(-0600)] <kasparnet> but it doesn't have to be tongith
[20:10:52 CST(-0600)] <kasparnet> just before the review
[20:11:05 CST(-0600)] <colinclark> ok
[20:11:19 CST(-0600)] <colinclark> I think the idea was that avtar was going to give you a key to the kingdom
[20:11:22 CST(-0600)] <colinclark> or at least part of one
[20:11:26 CST(-0600)] <kasparnet> cool
[20:11:28 CST(-0600)] <colinclark> so you could make updates while he sleeps and has a normal life
[20:11:33 CST(-0600)] <colinclark> don't quote me on that
[20:11:36 CST(-0600)] <colinclark> but I think that was the idea
[20:11:40 CST(-0600)] <yzen>
[20:11:48 CST(-0600)] <kasparnet> I cant imagine I'd need to touch anything but the solution registry files
[20:12:27 CST(-0600)] <colinclark> I believe both MMs are deployed with master universals
[20:12:33 CST(-0600)] <colinclark> but we can, tomorrow, change that
[20:12:41 CST(-0600)] <kasparnet> ah
[20:12:54 CST(-0600)] <colinclark> and you can at least push up your changes to review2 and then it's probably a matter of just git pulling and stopping/starting the server
[20:13:09 CST(-0600)] <kasparnet> the change in question should already be there
[20:13:26 CST(-0600)] <kasparnet> .. but maavis to wont make it in until tomorrow, so there is no rush with the update
[20:13:30 CST(-0600)] <colinclark> ah, ok
[20:13:46 CST(-0600)] <yzen> hah interesting colinclark , i think it's related to what your were saying
[20:13:50 CST(-0600)] <colinclark> so we'll need each matchmaker to refer to the latest version of universal's review2 branch
[20:13:54 CST(-0600)] <colinclark> SOP, yzen?
[20:14:20 CST(-0600)] <yzen> if i host the client on 8888, and the fm is on 8080 the 8080 can't set a cookie right ?
[20:14:29 CST(-0600)] <colinclark> yes
[20:14:34 CST(-0600)] <colinclark> that's likely the case
[20:14:37 CST(-0600)] <yzen> what about with CORS?
[20:14:40 CST(-0600)] <colinclark> I don't know
[20:14:44 CST(-0600)] <yzen> ok ill try
[20:14:44 CST(-0600)] <colinclark> but I wouldn't count on it
[20:15:14 CST(-0600)] <colinclark> it looks like it's possibly supported with CORS, yes
[20:16:01 CST(-0600)] <yzen> colinclark: do you have a source of that ?
[20:16:05 CST(-0600)] <colinclark> It looks like the server can indeed say whether cookies should be respected
[20:16:06 CST(-0600)] <colinclark> https://developer.mozilla.org/en/docs/HTTP/Access_control_CORS
[20:16:17 CST(-0600)] <colinclark> " Servers can also notify clients whether "credentials" (including Cookies and HTTP Authentication data) should be sent with requests."
[20:16:39 CST(-0600)] <colinclark> though it's not clear if this is only for XHR
[20:16:50 CST(-0600)] <colinclark> XHR appears to have this "withCredentials" flag
[20:17:12 CST(-0600)] <yzen> yep
[20:17:16 CST(-0600)] <yzen> will try now
[20:20:49 CST(-0600)] <colinclark> If that doesn't work, could you mount a Kettle app that will serve up the PMT/PCP static pages from the same origin as the Flow Manager?
[20:20:54 CST(-0600)] <colinclark> a bit like how Evgeni's original demo did
[20:23:35 CST(-0600)] <yzen>
!BGIBADVAOsn lk WORKS!
[20:23:43 CST(-0600)] <colinclark> WOOOOOOOOT
[20:23:45 CST(-0600)] <yzen> sakgner;k gw;lr
[20:23:47 CST(-0600)] <yzen> yes
[20:23:51 CST(-0600)] <colinclark> øøøøøøøøøøøøøøøøøøøøøø
[20:23:57 CST(-0600)] <colinclark> ØØØØØØØØØØØØØØØØØØØØØ
[20:24:19 CST(-0600)] <yzen> tbh i did not do anything special except this xhrField everything else is set up as expected
[20:24:52 CST(-0600)] <colinclark> SOP bites us again
[20:25:04 CST(-0600)] <yzen> haha
[20:25:19 CST(-0600)] <colinclark> I guess the more interesting issue is, do we know how to make this work with the PCP so Kasper can demonstrate it?
[20:25:46 CST(-0600)] <colinclark> YURA YOU RULE
[20:28:10 CST(-0600)] <yzen> well i think i can play with it in the evening
[20:28:17 CST(-0600)] <yzen> around 11ish
[20:28:37 CST(-0600)] <yzen> btw this is the client i did http://pastie.org/8628061
[20:29:09 CST(-0600)] <colinclark> awesome
[20:29:19 CST(-0600)] <colinclark> Do what you can later tonight, yzen
[20:29:47 CST(-0600)] <yzen> ok see you later, ill go offline for a bit
[20:29:52 CST(-0600)] <kasparnet> YAY!!!!!!!!!!! yzen
[20:29:58 CST(-0600)] <kasparnet> You're awesome!!
[20:30:04 CST(-0600)] <yzen>
[20:31:00 CST(-0600)] <kasparnet> colinclark: I'm gonna give 30 mins to an attempt to get PCP/PMT up and running - and then I'm gonna crash
[20:31:07 CST(-0600)] <colinclark> yes, you should
[20:31:11 CST(-0600)] <colinclark> it's late
[20:31:30 CST(-0600)] <colinclark> Manuel and Gregg will understand if the PCP/PMT aren't quite working
[20:31:33 CST(-0600)] <kasparnet> colinclark: in any case, with yzen's new progress it might make sense to hold back on figuring out what to demo until tomorrow evening
[20:31:38 CST(-0600)] <colinclark> yes, exactly
[20:31:41 CST(-0600)] <colinclark> I gave them a heads up about the problems
[20:31:50 CST(-0600)] <colinclark> you can tell them tomorrow "we're really close"
[20:32:03 CST(-0600)] <kasparnet> cool
[20:32:23 CST(-0600)] <kasparnet> gonna give it a quick go now anyway - I'm actually quite curious to see it in action
[20:35:02 CST(-0600)] <colinclark> I haven't seen it yet either
[20:36:25 CST(-0600)] <Bosmon> colinclark - shall I make the STMM pull to Andreas repo rather than to yours?
[20:36:39 CST(-0600)] <colinclark> Yes
[20:36:44 CST(-0600)] <colinclark> that makes sense
[20:36:45 CST(-0600)] <Bosmon> ok
[20:48:06 CST(-0600)] <avtar> just caught up on the channel log
[20:48:20 CST(-0600)] <avtar> kasparnet: you'll have keys to the kingdom
[20:48:32 CST(-0600)] <kasparnet> avtar: yay
[20:48:34 CST(-0600)] <kasparnet> but no rush
[20:48:36 CST(-0600)] <colinclark> hi avtar
[20:48:37 CST(-0600)] <colinclark> what'd you have for dinner?
[20:48:43 CST(-0600)] <avtar> gluten free pasta
[20:48:47 CST(-0600)] <colinclark> nice
[20:48:58 CST(-0600)] <colinclark> me too
[20:49:20 CST(-0600)] <colinclark> with beans and broccoli
[20:49:32 CST(-0600)] <avtar> yumm
[20:49:46 CST(-0600)] <Bosmon> I have some BEEF CHILLIE
[20:49:47 CST(-0600)] <avtar> i just ate the pasta out of the bag
[20:50:07 CST(-0600)] <avtar> Bosmon: you win this round
[20:51:11 CST(-0600)] <avtar> colinclark: it looks like there's over a gig of free ram on the vm that's running stmm and rbmm
[20:51:21 CST(-0600)] <colinclark> that's good
[20:51:25 CST(-0600)] <avtar> i think we should be able to squeeze in more node processes
[20:51:33 CST(-0600)] <colinclark> so the RBMM isn't eating up RAM
[20:51:35 CST(-0600)] <avtar> so digital ocean might still be a viable option
[20:51:37 CST(-0600)] <colinclark> ok
[20:51:40 CST(-0600)] <avtar> well
[20:51:46 CST(-0600)] <avtar> it isn't eating any while it's sitting idle
[20:51:52 CST(-0600)] <colinclark>
[20:52:10 CST(-0600)] <avtar> i'm wondering if it's safe to set up a cron job and delete files older than an hour from /logs on that vm
[20:52:21 CST(-0600)] <colinclark> I don't see any reason why not
[20:52:23 CST(-0600)] <avtar> i'm assuming the services don't need them
[20:52:29 CST(-0600)] <colinclark> One would hope so
[20:52:35 CST(-0600)] <colinclark> or hope not
[20:52:36 CST(-0600)] <colinclark> or whatever
[20:52:42 CST(-0600)] <avtar> a bit of both
[20:52:55 CST(-0600)] <avtar> so are we good here for now?
[20:53:06 CST(-0600)] <colinclark> We are
[20:53:12 CST(-0600)] <avtar> ok
[20:53:18 CST(-0600)] <colinclark> kasparnet is proving how insanely heroic he is, staying up until almost 4 am
[20:53:28 CST(-0600)] <colinclark> but I think he's essentially tested everything that needed to be tested
[20:53:33 CST(-0600)] <colinclark> and your servers are going strong
[20:53:56 CST(-0600)] <colinclark> The only thing we'll need to do sometime tomorrow is to redeploy the MMs with the latest review2 branch of universal
[20:54:06 CST(-0600)] <colinclark> but that doesn't have to happen on any real schedule
[20:55:10 CST(-0600)] <kasparnet> yeah colinclark, avtar - pretty much everything I can test have worked
[20:55:30 CST(-0600)] <avtar> great success
[20:55:35 CST(-0600)] <kasparnet> there are a few things that fail - but I it seem to be a problem in the ST MM
[20:55:37 CST(-0600)] <kasparnet> indeed
[20:55:47 CST(-0600)] <colinclark> kasparnet, if I haven't already said so, you are fucking amazing
[20:55:57 CST(-0600)] <colinclark> we would be completely and utterly hopeless without you
[20:56:05 CST(-0600)] <colinclark> you know, you and avtar must be brothers
[20:56:05 CST(-0600)] <kasparnet> I'm horribly unsuccessful in getting the PCP/PMT up and running though - so think I'll just sleep instead
[20:56:07 CST(-0600)] <colinclark> since he's amazing
[20:56:13 CST(-0600)] <colinclark> and we'd be utterly hopeless without him, too
[20:56:17 CST(-0600)] <kasparnet> and have boyan hotline me tomorrow
[20:56:20 CST(-0600)] <kasparnet> indeed we would
[20:56:23 CST(-0600)] <colinclark> kasparnet: That makes sense
[20:56:29 CST(-0600)] <colinclark> Yura will do what he can later
[20:56:45 CST(-0600)] <colinclark> and I think Boyan and Alex are probably in good shape to get it working tomorrow during their time
[20:56:59 CST(-0600)] <colinclark> thank you again, kasparnet
[20:57:01 CST(-0600)] <colinclark> sleep well
[20:57:04 CST(-0600)] <kasparnet> colinclark, avtar: well, thanks for all your work over the weekend and tonight!!!! tty tomorrow at some point
[20:57:06 CST(-0600)] <kasparnet> thanks
[20:57:14 CST(-0600)] <colinclark> well, that's it
[20:57:17 CST(-0600)] <colinclark> I call this a success
[20:57:37 CST(-0600)] <colinclark> avtar: I really am insanely indebted to me
[20:57:43 CST(-0600)] <colinclark> so you should consider ways that I can repay you
[20:58:11 CST(-0600)] <colinclark> wow, that came out wrong
[20:58:14 CST(-0600)] <colinclark> let me say that again
[20:58:20 CST(-0600)] <colinclark> avtar: I really am insanely indebted to YOU
[20:58:56 CST(-0600)] <avtar> right back at ya, buddy
[20:59:05 CST(-0600)] <avtar> the conversation last week settled my nerves
[20:59:19 CST(-0600)] <avtar> so let's just say we're even
[20:59:53 CST(-0600)] <colinclark> lol
[21:00:08 CST(-0600)] <colinclark> I'm hoping everyone might want to let me take them out for lunch this week
[21:01:17 CST(-0600)] <Bosmon> Next month it is I who will be taking YOU out
[21:01:26 CST(-0600)] <Bosmon> And you can choose ANY WINE FROM THE GIANT DIRECCKTORY!
[21:01:33 CST(-0600)] <colinclark> lol
[21:01:53 CST(-0600)] <Bosmon> I don't think we took you to Greenbriar yet?
[21:01:56 CST(-0600)] <Bosmon> It's a pretty amazing place
[21:02:00 CST(-0600)] <colinclark> no, I don't think so
[21:02:02 CST(-0600)] <Bosmon> Halfway to the BOGGDANNOVVITCH HILLS
[22:57:01 CST(-0600)] <yzen> Bosmon: