Wednesday, 2014-03-12

*** rossk has quit IRC00:23
*** rektide has quit IRC00:57
*** jraim has quit IRC00:57
*** jraim has joined #openstack-marconi00:58
*** openstack has joined #openstack-marconi01:05
*** russell_h has joined #openstack-marconi01:05
*** seiflotfy_ has joined #openstack-marconi01:19
*** oz_akan_ has joined #openstack-marconi01:23
*** reed has quit IRC01:30
*** nosnos has joined #openstack-marconi01:39
*** flwang has quit IRC02:42
*** haomaiwang has quit IRC02:51
*** haomaiwa_ has joined #openstack-marconi02:51
*** haomaiwa_ has quit IRC03:05
*** flwang has joined #openstack-marconi03:15
*** rwsu has quit IRC03:19
*** haomaiwang has joined #openstack-marconi03:20
*** haomaiwang has quit IRC03:36
*** haomaiwang has joined #openstack-marconi03:37
*** reed has joined #openstack-marconi03:37
*** haomaiw__ has joined #openstack-marconi03:40
*** haomaiwang has quit IRC03:43
*** rwsu has joined #openstack-marconi03:50
*** fifieldt has joined #openstack-marconi03:51
*** oz_akan_ has quit IRC04:15
*** sld_ is now known as sld04:16
*** oz_akan_ has joined #openstack-marconi04:18
*** oz_akan_ has quit IRC04:23
*** reed has quit IRC05:32
*** sld has left #openstack-marconi06:20
*** saju_m has joined #openstack-marconi06:52
*** flaper87|afk is now known as flaper8707:24
*** jamieh has joined #openstack-marconi08:16
*** jamieh is now known as Guest810608:16
*** saju_m has quit IRC08:41
*** ekarlso has quit IRC08:47
*** ekarlso has joined #openstack-marconi08:47
openstackgerritFengqian Gao proposed a change to openstack/marconi: Keep python 3.X compatibility for xrange  https://review.openstack.org/7987908:52
*** saju_m has joined #openstack-marconi09:09
*** flwang has quit IRC09:14
*** anamalagon_ has joined #openstack-marconi09:16
*** anamalagon has quit IRC09:19
*** tmu has quit IRC09:48
*** tmu has joined #openstack-marconi09:49
*** flwang has joined #openstack-marconi10:06
*** amarsharma has joined #openstack-marconi10:28
amarsharmaHey people I am new here can anybody help to setup Openstack or provide resources to help me?10:34
*** saju_m has quit IRC10:49
flaper87amarsharma: o/10:50
flaper87welcome to the most amazing channel in the whole IRC world10:50
flaper87so, the easiest way to install marconi is:10:50
flaper87$ git clone $MARCONI_REPO10:50
flaper87$ cd marconi/ && pip install -e .10:50
flaper87that will setup marconi (you should use a virtualenv)10:50
flaper87once the pip install command ends, you just run10:51
flaper87$ marconi-server -d -v10:51
flaper87amarsharma: brb in a bit, hope that helps10:51
amarsharmaHi Flavio thanks for help I am having some issues with my PC :/ I'll try and ask if it works :)11:05
*** ChanServ changes topic to "test/gate jobs are queuing now in preparation for gerrit maintenance at 12:00 utc (eta to resume is 12:30 utc)"11:27
*** amarsharma has left #openstack-marconi11:36
*** fifieldt has quit IRC11:42
*** ykaplan has joined #openstack-marconi11:44
*** amarsharma has joined #openstack-marconi11:55
*** mwagner_lap has quit IRC11:59
amarsharmaOk Flavio I'll try running in a virtual machine :)12:10
flaper87amarsharma: awesome, are you using linux or windows?12:11
amarsharmaI have a dual boot setup plus VMware installed in windows12:11
amarsharmaHey can I independently run marconi without whole openstack installed..12:13
amarsharmaI mean if I just install marconi with Mangodb I am good to go right ?12:15
*** oz_akan_ has joined #openstack-marconi12:16
flaper87amarsharma: sure, marconi is integrated with other openstack services but it's a stand-alone service, nontheless.12:16
amarsharmaOk thanks :)12:17
openstackgerritSascha Peilicke proposed a change to openstack/marconi: Fix config file name in README instructions  https://review.openstack.org/7992212:21
amarsharmaActually I am a GSOC aspirant and interested in Marconi Project I am trying to get around with Marconi code and currently seeing the bug tracker webpage of it :)12:22
*** ChanServ changes topic to "gerrit on review.openstack.org is down for maintenance (revised eta to resume is 13:00 utc)"12:24
*** ykaplan has quit IRC12:37
*** saju_m has joined #openstack-marconi12:37
*** oz_akan_ has quit IRC12:45
openstackgerritSascha Peilicke proposed a change to openstack/marconi: Fix functional config file instructions.  https://review.openstack.org/7992512:48
openstackgerritSascha Peilicke proposed a change to openstack/marconi: Rename logging.conf-sample to logging.conf.sample  https://review.openstack.org/7992612:48
*** amarsharma has quit IRC12:50
*** sriram has joined #openstack-marconi12:53
*** sriram has quit IRC12:53
*** oz_akan_ has joined #openstack-marconi12:53
*** sriram has joined #openstack-marconi12:53
*** openstackgerrit has quit IRC12:54
*** openstackgerrit has joined #openstack-marconi12:54
*** oz_akan_ has quit IRC12:55
*** oz_akan_ has joined #openstack-marconi12:56
*** alcabrera has joined #openstack-marconi12:59
alcabreraGood morning! :D13:00
*** cpallares has joined #openstack-marconi13:02
*** mpanetta has joined #openstack-marconi13:03
flaper87alcabrera: morning :)13:06
alcabreraflaper87: heeeeyyy!13:06
alcabreraflaper87: how are you?13:06
flaper87alcabrera: I'm doing fine, how are you doing?13:06
alcabreraawesome. I accidentally got some extra sleep last night and I feel pretty energized, as a result!13:07
*** openstackgerrit has quit IRC13:08
flaper87alcabrera: lol @ accidentally13:08
flaper87:D13:08
*** openstackgerrit has joined #openstack-marconi13:08
flaper87sounds like: "I was walking around my place and then I fell asleep without noticing"13:09
alcabreralol13:09
alcabreranot quite ;)13:09
alcabreraI woke up at 6am, then rolled over lazily, and fell asleep again 'til 8am13:10
flaper87alcabrera: ah ok, that sounds safer than my example :D13:10
flaper87alcabrera: hey, I wanted to ask you yday whether you've thought about possible tasks for the next OPW intership cycle13:11
alcabreraI hadn't. I've had that tab open for the past week. @_@13:11
*** mpanetta_ has joined #openstack-marconi13:15
*** mpanetta has quit IRC13:15
cpallaresalcabrera, flaper87: o/13:21
alcabreracpallares: good morning! :)13:22
alcabrerahow are you, cpallares?13:22
cpallaresalcabrera: I'm sleepy :P13:23
cpallaresalcabrera: how are you doing?13:23
alcabreraenergized. As I was telling flaper87 a little bit ago, I accidentally got two more hours of sleep than I intended. Rolled over at 6am and fell asleep 'til 8am. :P13:24
*** jcru has joined #openstack-marconi13:26
alcabreraflaper87: ping13:29
*** balajiiyer has joined #openstack-marconi13:29
flaper87alcabrera: pong13:29
alcabrerasaikrishna_ is interested in working on (https://bugs.launchpad.net/marconi/+bug/1239834) and was wondering if we could reassign it13:30
alcabrerasince it's been sitting around for some time, I figured a reassignment would be okay. What do you think, flaper87? :)13:31
flaper87alcabrera: that sounds good to me!13:32
alcabreracool, cool. I'll do that reassign magic then. :)13:33
alcabrerathanks, flaper87!13:33
*** amitgandhi has joined #openstack-marconi13:37
*** amitgandhi has quit IRC13:38
*** amitgandhi has joined #openstack-marconi13:39
*** mpanetta has joined #openstack-marconi13:41
*** mpanetta_ has quit IRC13:41
*** amarsharma has joined #openstack-marconi13:43
*** nosnos has quit IRC13:43
amarsharmaflaper87: Hey I am getting curl: (52) Empty reply from server after setting up marconi and sending curl -i -X PUT http://127.0.0.1:8888/v1/queues/samplequeue -H "Content-type: application/json" -d '{"metadata": "Sample Queue"}'13:44
*** oz_akan_ has quit IRC13:45
alcabreraamarsharma: creating a queue w/ PUT no longer takes metadata13:46
alcabrerathat was changed a few months ago. :)13:46
*** oz_akan_ has joined #openstack-marconi13:47
amarsharmaalcabrera: Then we should change this Guide on GitHub I followed the instruction there.13:47
amarsharma https://github.com/openstack/marconi13:47
amarsharmaHow can I test if my server is up and running ?13:48
alcabreraoh yes, that is out of date!13:48
alcabreraamarsharma: thanks for catching that. :)13:49
alcabreraso to test if the server is up13:49
alcabreratwo13:49
alcabreraerr13:49
alcabreratwo things13:49
alcabreraone is to issue a GET against the home url (host:port/v1)13:49
alcabrerathe other is to issue the PUT as above to create a queue, without the metadata13:49
alcabrera(/v1/queues/queuename)13:49
alcabrera(PUT /v1/queues/queuename)13:49
amarsharmaalcabrera: Your welcome :) and now I am getting HTTP/1.1 403 Forwarding Loop Detected13:54
amarsharmaServer: awarrenhttp/5.0.0.2.613:54
amarsharmaI guess I should learn more about marconi internals13:54
alcabreraoh wow - I've never seen HTTP 40313:54
alcabreraerr13:54
alcabreraThe Forwarding Loop issue13:55
alcabrerathat's new13:55
*** mpanetta_ has joined #openstack-marconi13:56
*** mpanetta has quit IRC13:56
amarsharmawait body says:13:56
amarsharma<html> <head> <title> Internal Error </title> </head> <body> <hr> <h1> ERROR </h1> <h2> Forwarding loop detected.13:57
amarsharma <BR> <BR> Reason: Source and Destination are same </h2> </body>13:57
alcabreranow I'm tempted to follow the guide from a fresh marconi install to see if it explodes in this fashion14:00
*** mwagner_lap has joined #openstack-marconi14:01
*** openstack has joined #openstack-marconi14:04
*** fungi has joined #openstack-marconi14:10
*** ChanServ sets mode: +o fungi14:10
*** fungi changes topic to "OpenStack Queuing and Notification Service || Smile :D || Meetings every Tuesday @ 15:00 UTC || Wiki: ttps://wiki.openstack.org/wiki/Marconi || Paste: http://paste.openstack.org/ || Send messages and make some noise :D"14:10
*** ChanServ sets mode: -o fungi14:10
*** fungi has left #openstack-marconi14:10
*** alcabrera is now known as alcabrera|afk14:11
*** malini has joined #openstack-marconi14:15
*** alcabrera|afk is now known as alcabrera14:16
*** Kuwagata has joined #openstack-marconi14:20
malinio/14:20
cpallaresmalini: \o14:21
*** chandan_kumar has quit IRC14:22
malinihello cpallares!!14:23
cpallareshello malini how are you?14:24
*** chandan_kumar has joined #openstack-marconi14:25
malinicpallares: good.ready for the devstack bug :)14:27
*** mpanetta_ is now known as mpanetta14:27
*** Guest8106 has quit IRC14:27
*** jamieh has joined #openstack-marconi14:28
*** jamieh is now known as Guest2411614:28
*** Anastas__ has joined #openstack-marconi14:29
*** Guest24116 is now known as jamiehannaford14:30
flaper87malini: soooo, any other good news ?14:30
maliniflaper87: no..I am going to add an experimental job in devstack14:31
malinitht'll help us move a lil faster14:31
maliniflaper87, alcabrera, flwang: I dont know if you have other priorities..But the next two days can we swarm on the stderr bug?14:31
malinisdague will definitely want to know abt tht during our graduation review on Tuesday14:32
* alcabrera thinks about priority things14:32
maliniI will get the experimental job submitted today14:32
alcabreramalini: I know where marconi-server is bratty and writes to stdout (or stderr?), but I don't know why devstack chokes on it14:33
alcabreratime allowing, I want to take 1 full day to study devstack14:34
malinialcabrera: we are interested in why marcobi opens the stderr fd14:34
alcabrerait should open it by default14:34
maliniwhy devstack does it a whole different topic14:34
alcabreraall processes do14:34
maliniit is not supposed to - 'other projects dont do it'14:34
alcabreraI mean14:35
alcabreraevery process on every operating system opens stderr. :(14:35
maliniredirecting the stderr to stdout will keep marconinrunning14:35
malinimarconi running*14:35
alcabreraI want to know why it writes to it, and why devstack explodes as a result14:35
alcabrerayes14:35
alcabrerathat14:35
alcabrera:D14:35
malinialcabrera: dont worry abt the devstack explosion for now ;)14:36
alcabrerakk14:36
malinii.e. until we know why marconi writes to it14:36
mpanettaWho is saying that things don't open stderr/stdout??14:36
alcabreramarconi-server writes to stderr because wsgiref does so by default14:36
malinimpanetta: devstack team14:36
mpanettaUh, do they unix? :P14:36
maliniflaper87, flwang: can you clarify how other openstack projects do this?14:37
mpanettaThe bloody kernel opens those file descriptors for you on exec, unless you purposfully close them, they will be open...14:37
mpanettaIt is just how things work... :(14:38
flwangmalini: i need a little bit time to pick up the background14:38
maliniwe need to figure out how to get marconi running on devstack w/o the redirection14:38
maliniWe should focus on how to do tht in the next couple of days14:38
mpanettamalini: what about the suggstion I had last week?14:38
alcabreraI'm cloning the nova repo to investigate whether it uses wsgiref14:38
maliniwe can argue abt if its the right approach or not next week ;)14:38
mpanettaOr maybe it was 2 weeks ago now...14:38
malinimpanetta: which suggestion?14:39
mpanettaTo do the daemonize thing14:39
maliniI have done a lot of stuff so far ;)14:39
flwangmalini: may I know how to reproduce it?14:39
maliniflwang: sure..you need to tun devstack with USE_SCREEN = false14:40
*** chandankumar_ has joined #openstack-marconi14:40
*** chandankumar_ has quit IRC14:40
maliniflwang: we current have a work around for the issue here https://github.com/openstack-dev/devstack/blob/master/lib/marconi#L15714:40
alcabreranova uses wsgiref in bare_metal_deploy, hmm...14:40
flwangmalini: so now Marconi has been associated with devstack, andI just need to install openstack with devstack on USE_SCREEN = false14:41
flwangmalini: to recreate this issue, is it?14:41
maliniflwang: we should be able to get rid of the workaround "2>&1" without causing marconi to die14:41
maliniflaper87: Yes14:41
maliniflwang: Yes14:41
maliniflaper87: sorry..tht was not for you14:41
flaper87malini: :D14:42
flwangflaper87 is busy now i think14:42
malininow I know how to get his attention ;)14:42
* flaper87 is fighting for what he thiks is right14:43
alcabrerago, flaper87, go!14:44
* alcabrera cheers flaper87 on14:44
flaper87<o/ <o> \o>14:44
alcabrerahmmm14:45
alcabreranone of [trove, heat, swift] use wsgiref14:45
malinialcabrera: intersting trove uses the same workaround14:45
mpanettaI saw that14:45
mpanettamalini: Didn't they say it wasn't technically needed anymore by trove tho? heh14:46
malinimpanetta: no.14:46
malinithey said they 'might not' need it now14:46
alcabreraheat uses eventlet as the default wsgi server14:46
mpanettaWell how come they get to do it when we can't?14:46
malinimpanetta: I spent a  good portion of my brain wondering abt tht..But tht argument wont help us .14:47
maliniIf we find out the root cause, we'll have a better argument14:47
mpanettaThis whole thing is confusing to me.  It almost feels to me like we are being punished for using falcon, and this is how they are punishing us.  Redirecting file descriptors is standard practice in the unix world. :(14:48
mpanettaSo we determined that wsgiref is using stderr right? alcabrera, malini?14:49
malinialcabrera: ?14:49
maliniwill adding a pdb help us figure out when stderr FD is getting opened?14:49
flaper87IIRC, none of the openstack projects use wsgiref14:49
flaper87am I wrong?14:49
alcabreraflaper87: that's what I'm seeing, except for maybe ceilometer and a small portion of nova14:50
alcabrerampanetta, malini: we've verified that wsgiref writes to stderr14:50
alcabreragonna reverify now14:50
alcabrera**I'm going to...14:50
*** AnastasiosAndron has joined #openstack-marconi14:51
*** amarsharma has quit IRC14:51
alcabreramost of the other projects seem to be using openstack/common/wsgi14:51
alcabrerawhich looks *really* awkward14:52
malinialcabrera: is it an easy change to try marconi with tht?14:52
cpallareshaha alcabrera14:52
flaper87that was the former wsgi lib in openstack, hence the migration to pecan14:52
*** Anastas__ has quit IRC14:52
flaper87alcabrera: ^14:52
flaper87that's already deprecated in the incubator14:52
alcabreraflaper87: good to know14:53
flaper87I'd expect *all* projects except from ceilo to be using that (or at least have used that)14:53
alcabreramalini: it wouldn't be an easy change14:53
malinialcabrera: ok..& since it is deprecated anyways..14:53
*** jmckind has joined #openstack-marconi14:54
mpanettaWait, so why can't we try and do http://legacy.python.org/dev/peps/pep-3143/ again?14:54
alcabrerare-verified: wsgiref logging writes to stderr14:55
mpanettaIf we did the things in that pep we would not be writing to any stdthings nymore14:55
mpanettaAnd we don't even have to do them all really.14:57
mpanettaHere is the list of things, and we only have to do 1,4,6 and 7.  And maybe not even 6... http://legacy.python.org/dev/peps/pep-3143/#correct-daemon-behaviour14:59
* alcabrera reads15:00
*** AnastasiosAndron has quit IRC15:00
alcabreraclose all open file descriptors (0, 1, 2), run in the background, disassociate from control terminal15:01
mpanettaYeah15:01
alcabreragonna run a quick test15:01
alcabrera:D15:01
alcabreraoh15:02
alcabrerawell that was fun15:02
alcabreraif you want to kill the python REPL15:02
alcabrera>>> import os15:02
alcabrera>>> os.close(0)15:02
alcabrerayup15:02
mpanettaSome programs reopen 0,1,2 and associate them with /dev/zero,/dev/null to do the dissassociation bit15:02
mpanettaWell, yeah you just closed your only source of input :P15:02
mpanettaIt does not work too well with interactive programs. ;)15:03
alcabrerahaha15:03
*** saju_m has quit IRC15:03
*** saju_m has joined #openstack-marconi15:04
malinialcabrera: other OS projects do not use wsgiref, rt?15:07
alcabreramalini: except for ceilometer and a part of swift, based on what I've seen in the past hour.15:08
alcabreraI git cloned a few of them. :)15:08
alcabrera*swift -> nova15:08
alcabreraI checked: marconi, ceilometer, heat, trove, swift, nova15:09
malinialcabrera: can you check if either of them have anything special to NOT use stderr?15:09
mpanettaalcabrera, malini would you like me to write a quick script that will wrap marconi-server and daemonize it, as a POC?15:09
malinimpanetta: sure..I can try tht in my devstack server15:09
maliniit will be awesome, in fact :)15:09
mpanettaIf it works we can move the important bits in to marconi-server15:10
flaper87alcabrera: what is that list?15:10
flaper87projects using wsgiref ?15:10
malinimpanetta: Yes15:10
alcabreraflaper87: the list of projects I cloned to check for wsgiref. I found that only marconi, ceilometer, and a small portion of nova use wsgiref.15:10
alcabreranova:bare_metal_deploy:api15:11
flaper87glance doesn't use wsgiref either15:11
flaper87nor does cinder15:11
alcabreraI wonder why...15:11
alcabrerait seems like all of these should be launchable in a wsgi server chosen by the user15:11
alcabreraand that there's no harm in having a command like wsgiref as part of the "play/dev" command for each project15:12
malinialcabrera:Can we swap out wsgiref with something else to make sure tht is really the problem?15:14
alcabrerahmmm15:14
maliniTht'll confirm we are chasing the right thing15:15
alcabrerahmmm15:15
alcabreraI feel like I've swapped it out already15:16
alcabreraI've deployed marconi with both gunicorn and marconi-server (wsgiref-style)15:16
alcabreraoutput is only written to stderr with wsgiref15:16
malinicool - tht confirms something15:18
malinican you give me ur gunicorn command , alcabrera?15:18
maliniI'll swapi it out in devstack script15:18
alcabreragunicorn -b BIND:PORT marconi.queues.transport.wsgi.app:app15:18
alcabrerawhere BIND is usually localhost for me15:18
maliniok..trying now15:19
alcabrera      PORT is usually 800115:19
malinialcabrera: it gets killed with gunicorn as well :(15:26
*** chandan_kumar has quit IRC15:26
alcabrera:/15:28
*** chandan_kumar has joined #openstack-marconi15:28
*** chandan_kumar has quit IRC15:28
mpanettatry daemonizing gunicorn15:29
mpanettaI think the switch is -D15:29
mpanettamalini: ^^15:29
malinimpanetta: trying now15:30
alcabrerathe switch is -D15:30
alcabreracomnfirmed. :)15:30
maliniI am running it with strace at the moment15:30
alcabreragunicorn -D -b BIND:PORT marconi.queues.transport.wsgi.app:app15:30
alcabreracool15:30
alcabrerakeep me updated15:30
mpanettaOh, strace will kill it probably15:30
maliniit did ;?15:30
maliniwith -D it works :)15:32
maliniso daemonizing might be our answer15:32
alcabrerasweet15:32
alcabreragunicorn it and call it done. :D15:33
mpanettaOf course it is :P15:33
maliniok..how do/can we daemonize marconi?15:33
mpanettaWill they allow us to pull in gunicorn?15:33
maliniI dont think so15:33
alcabreradang...15:33
alcabreraso15:33
alcabreradaemonizing marconi-server15:33
alcabrerathere's no command line option for that15:33
alcabrerabecause wsgiref doesn't understand daemonizing15:33
alcabreraso we'd have to roll our own15:34
malinihopefully mpanetta's wrapper will help15:34
malinilet me know when its ready , mpanetta15:34
mpanettaok15:34
maliniI'll give it a try15:34
mpanettacan you give me the full path to marconi-server please?15:34
alcabreramarconi/cmd/server.py15:35
alcabrerampanetta: ^15:35
mpanettaOh sorry I mean the installed executable15:36
mpanettaI am going to call it with os.exec15:36
mpanettaIf this works I will put the code in server.py15:36
malinimpanetta: $ which marconi-server15:36
malini/usr/local/bin/marconi-server15:36
mpanettaAnd that is on the devstack system correct?15:36
maliniyes15:36
mpanettacool15:37
*** saju_m has quit IRC15:40
malinimpanetta: shud I replace marconi/cmd/server.py with https://gist.github.com/anonymous/e32db6a5bc86bb2918a9 ?15:52
mpanettaHmm not replace15:53
mpanettaIt should go in to run() right before server.run I think15:54
malinimpanetta: ok..let me try15:54
flaper87quick question: What's the workaround we have now ?15:54
flaper87Was it applied to devstack ?15:55
maliniflaper87: yes15:55
flaper87damn15:55
flaper87I was about to ask if we could live with that15:55
flaper87anyway15:55
maliniThe 2&1 in https://github.com/openstack-dev/devstack/blob/master/lib/marconi#L15715:55
malinithat is temporary :(15:55
flaper87can we live with that ?15:55
flaper87:D15:55
maliniI can ..but not a few others ;)15:55
*** reed has joined #openstack-marconi16:00
malinimpanetta: I updated server.py16:02
maliniBut now I can no longer start with marconi-server16:02
*** balajiiyer has quit IRC16:06
maliniflaper87, alcabrera, flwang: mpanetta's fix solves the issue :)16:11
flaper87malini: yeah, lets call it fix...16:12
malinihttps://gist.github.com/anonymous/e32db6a5bc86bb2918a916:12
flaper87:D16:12
maliniis it acceptable to introduce this in marconi?16:12
*** AAzza1 has joined #openstack-marconi16:12
flaper87why not?16:12
maliniwooot woot16:12
flaper87the whole point behind wsgiref is to have a toy server to play with16:13
malinimpanetta: can you submit a patch with your changes?16:13
*** balajiiyer has joined #openstack-marconi16:13
flaper87not to use it in production16:13
maliniyes!!16:13
malini& the point of devstack is to aid development <sigh>16:13
mpanettawee16:13
mpanettaUm16:13
mpanettaI think I can ;)16:13
malinicool !16:13
maliniOur blocker now is the journaling patch..16:13
maliniI need some eyes on tht16:14
malinihttps://review.openstack.org/#/c/74088/16:14
maliniI am pinging some core folks for tht16:15
maliniI hope this will end all our problems :)16:16
AAzza1flaper87: Hello again! I'm shy girl who wants to try for OPW this year, Can you propose good easy bug for me to start with? :)16:16
cpallaresWelcome AAzza1 :)16:17
flaper87AAzza1: hey hey hey!16:17
flaper87Don't be shy, this is a fun channel, full of fun people16:18
flaper87AAzza1: people and robots16:18
alcabreraAAzza1: welcome!16:18
flaper87and some other kind of unknown beings16:18
flaper87AAzza1: let me help you with that bug hunting thing16:19
* cpallares wonders if flaper87 is an unknown being or a robot16:20
flaper87cpallares: and evolution of the mutation of both16:20
AAzza1okey, separate hi to all the uknown beings:) Yeah ,I just need something to start knowing the system16:20
flaper87s/and/an/16:20
AAzza1May I just want to ask the irrelevant question, that bothered me for years, how do you test you messaging service? I find people who should know the answer16:22
AAzza1just cann't imagine how to do this with all this transport etc16:23
flaper87AAzza1: It depends on the test. For example, if it's an unit test, you can mock the transports and test the "happy path"16:24
flaper87in the case of functional tests, you need to have a running service that the test can talk to16:24
flaper87In marconi's functional tests, it starts a marconi instance in the background and tests everyting against that instance16:25
*** rossk has joined #openstack-marconi16:26
AAzza1Hmm... I see, all bunch of tests should run for a while i think :)16:26
flaper87AAzza1: yeah16:26
alcabrerathe unit-ish suite manages to execute in about 200 seconds, last I tried the full run. :)16:26
cpallaresflaper87: someone this morning mentioned to alcabrera that the curl request in the readme was outdated, that would be a simple bug to fix.16:27
flaper87AAzza1: so, I heard the examples in the README are not up-to-date16:27
cpallareslol16:27
flaper87AAzza1: that's something you could contribute to, if you want16:27
openstackgerritMike Panetta proposed a change to openstack/marconi: Allow marconi-server to properly daemonize  https://review.openstack.org/7998616:27
flaper87cpallares: I told you, evolution of the mutation of (robot+unknown being)16:28
mpanettaOh yeah, I don't have to paste the link heh16:28
mpanettamalini: ^^16:28
cpallaresflaper87: does that give you psychic powers now?16:28
flaper87mpanetta: there are some trailing spaces in your patch16:28
mpanettaAh crap16:28
alcabrerampanetta: flaper87 beat me to it16:29
flaper87cpallares: oh you've no idea :D16:29
mpanettaIt is bad anyway, I forgot the import. :P  Let me fix it.16:29
alcabreraalso16:29
mpanettaWhere are the spaces?16:29
AAzza1Oh, thanks, I'll look :)16:30
alcabrerampanetta: gerrit shows them above. :)16:30
alcabreraalso, reviewed: python3 nit16:31
mpanettaYeh I should have got rid of that, sorry :(16:32
mpanettaShould I just completely remove that line?16:32
mpanettaNevermind, I got it.16:33
alcabrerakk16:33
alcabrerarun 'tox -e pep8' before sending in the patch. It'll catch all the silly whitespace errors. :)16:33
*** saikrishna_ has joined #openstack-marconi16:35
*** reed has quit IRC16:44
openstackgerritMike Panetta proposed a change to openstack/marconi: Allow marconi-server to properly daemonize  https://review.openstack.org/7998616:46
*** jamiehannaford has quit IRC16:47
alcabrerampanetta: +216:48
mpanettaYay!16:48
mpanettaI really hope this gets us past this hurdle.16:49
flaper87mpanetta: just a really small request16:50
flaper87sorry for not raising it before16:50
flaper87would you mind adding a comment saying that this is all wsgiref faults ?16:51
flaper87I mean, we're crazy but not stupid16:51
flaper87I don't want people to think the later16:51
flaper87:D16:51
mpanettahaha16:51
mpanettaWhere would you like me to make the comment?16:51
mpanettaJust ammend mt commit log?16:51
flaper87righ above this: # Open /dev/zero and /dev/null for redirection.16:51
mpanettaAh ok16:52
mpanettayep sec.16:52
flaper87mpanetta: just explain why we need that hack16:52
flaper87and how much we hate wsgiref now16:52
mpanettahaha ok16:52
flaper87and that malini will pay whoever fixes wsgiref16:52
mpanettaA million dollars right? :P16:53
mpanettaflaper87: How about:16:54
mpanettahttps://gist.github.com/anonymous/89ed46f9274664ec2df016:54
mpanettaflaper87: Does that look good, or is it too long?16:55
flaper87mpanetta: sounds good. Maybe add "This is specifically needed to allow marconi run under devstack but it may be also useful for other scenarios"16:55
alcabreraworks for me, mpanetta16:55
flaper87mpanetta: documentation is never long enough16:55
flaper87hauhauhauhua16:55
alcabreramap :: (a -> b) -> [a] -> [b]  -- most succinct docs16:56
alcabrera:P16:56
mpanettaI still fail at reading haskell :P16:56
alcabrerahehe16:58
openstackgerritMike Panetta proposed a change to openstack/marconi: Allow marconi-server to properly daemonize  https://review.openstack.org/7998616:59
alcabreraI need to host that haskell workshop at rax:atl16:59
alcabreraafter that, I'd aim for everyone who attended to be able to read most type signatures16:59
openstackgerritMike Panetta proposed a change to openstack/marconi: Allow marconi-server to properly daemonize  https://review.openstack.org/7998617:02
mpanettaSorry had one space issue.17:02
*** alcabrera is now known as alcabrera|afk17:02
*** sballe has joined #openstack-marconi17:03
*** balajiiyer has quit IRC17:03
*** balajiiyer has joined #openstack-marconi17:03
flaper87mpanetta: LGTM, thanks17:04
*** balajiiyer1 has joined #openstack-marconi17:04
*** balajiiyer has quit IRC17:08
*** sballe has quit IRC17:10
*** reed has joined #openstack-marconi17:10
*** saju_m has joined #openstack-marconi17:16
*** cpallares has quit IRC17:18
*** jamieh has joined #openstack-marconi17:21
*** jamieh is now known as Guest3561317:21
*** Guest35613 has quit IRC17:23
*** openstackstatus has quit IRC17:38
*** amarsharma has joined #openstack-marconi17:38
*** openstackstatus has joined #openstack-marconi17:38
*** reed has quit IRC17:57
*** reed has joined #openstack-marconi17:57
*** amarsharma has quit IRC17:58
*** amarsharma has joined #openstack-marconi17:58
*** amarsharma has quit IRC18:01
*** amarsharma has joined #openstack-marconi18:01
*** amarsharma has quit IRC18:03
*** amarsharma has joined #openstack-marconi18:03
*** amarsharma has quit IRC18:08
*** cpallares has joined #openstack-marconi18:08
amitgandhiflaper87: you will have to visit this place when you come to atlanta18:11
amitgandhihttps://www.google.com/maps/@34.089669,-84.25018,3a,15y,304.93h,95.58t/data=!3m4!1e1!3m2!1sDp7sguRtzaMdeROxOulbeg!2e018:11
mpanettaflaper87: No problem :)18:12
*** alcabrera|afk has quit IRC18:14
*** alcabrera has joined #openstack-marconi18:14
openstackgerritOz Akan proposed a change to openstack/marconi: Marconi Operations Document  https://review.openstack.org/7829418:24
*** saju_m has quit IRC18:30
alcabrerampanetta: I thought about the review a little more. Can you check out sys.exit instead of os._exit?18:33
alcabrerahttps://review.openstack.org/#/c/79986/4/marconi/cmd/server.py18:33
malinimy experimental job for devstack is now merged - now we can experiment on devstack without having to get it merged18:44
maliniflaper87: can you take a look at https://review.openstack.org/#/c/79986/ ?18:47
alcabreraflwang: ^18:50
alcabrera:D18:50
alcabrerait should be the minimum patch needed to have marconi-server be on good terms with devstack18:50
malini& we need to know soon, if tht will fix it18:59
maliniIt did locally on my devstack18:59
*** saikrishna_ has quit IRC19:03
*** malini is now known as malini_afk19:27
mpanettaalcabrera: You want me to change that?19:36
*** AAzza1 has quit IRC19:36
alcabrerampanetta: yes, please. Let's test with sys.exit. That'd be the correct thing to do, with the smallest fear of breakage across OSes and Python versions.19:36
mpanettaere is the note I was going on in the python docs: Note The standard way to exit is sys.exit(n). _exit() should normally only be used in the child process after a fork().19:36
alcabrerahmmmm19:37
mpanettaSince I was using it after the fork...19:37
alcabreraI see.19:37
alcabreraCould you link to the docs?19:37
mpanettahttp://docs.python.org/2/library/os.html#process-management19:37
flaper87mpanetta: there's got to be a better way to do it without calling a privatre function19:38
flaper87:/19:38
alcabreraFound it: http://docs.python.org/2/library/os.html#os._exit19:38
flaper87please, tell me there is19:38
mpanettaI don't mind calling sys.exit19:38
flaper87:D19:38
alcabrerahaha19:38
alcabrerathat's what the docs say, flaper87. :P19:38
flaper87damn :(19:39
mpanettaYeah it isn't private.  Well if it is, it is well documented :P19:39
alcabrerampanetta: all functions with '_' prefixing them are considered private, by python convention19:39
mpanettaI originally has os.exit(), but changed it based on that note, and that the daemon library uses that too.19:39
mpanettaalcabrera: Ah!19:39
mpanettaThank you :)19:39
alcabreranp. :)19:39
alcabreraso19:39
alcabreraI'm favorable towards os._exit19:40
alcabreraafter reading the docs19:40
alcabreraall the more so since there's a note stating that this is the way to go for forked children19:40
mpanettaI promise I didn't pull that usage out of my posterior.  I am not knowledgable enough in python (yet?) to do that :P19:41
alcabrerathe note persists into python 3-latest: http://docs.python.org/3/library/os.html#os._exit19:41
alcabreraso, it's a design decision that's here to stay, afaict19:41
mpanettaOk19:41
mpanettaWant me to keep it then?19:41
mpanettaYou guys know best19:42
alcabreraos._exit is my vote, mpanetta19:42
mpanettaok19:42
alcabreraI left the doc comments in the review19:43
alcabrerapointing to both py2-latest and py3-latest docs19:43
mpanettaOk19:43
*** sriram has quit IRC19:44
alcabreraflaper87: your blessing can save the world, though it can wait 'til tomorrow morning. ;)19:44
flaper87alcabrera: mpanetta done :)19:46
mpanettaYay!19:46
mpanettaThank you so much alcabrera and flaper8719:46
flaper87mpanetta: thank you for debugging / fixing that19:47
alcabrerathanks all - malini_afk will celebrate tomorrow. :)19:47
mpanettaflaper87: I had fun :)19:48
openstackgerritA change was merged to openstack/marconi: Allow marconi-server to properly daemonize  https://review.openstack.org/7998619:48
flaper87that was fast19:50
flaper87why doesn't that happen with my patches ?19:50
flaper87mpanetta: >.>19:50
alcabreralol19:56
alcabreraI'm out for the night, all. Take care!19:58
*** alcabrera has quit IRC19:58
mpanettaflaper87: I donno :P19:59
*** cpallares has left #openstack-marconi20:13
*** malini_afk is now known as malini20:57
maliniflaper87: the journaling thing doesnt seem to help https://review.openstack.org/#/c/74088/20:59
malini:(20:59
*** AAzza has joined #openstack-marconi20:59
maliniflaper87: But atleast we have an experimental job in devstack now..So we can iterate faster21:02
flaper87malini: mmh21:05
flaper87let me take a look21:05
AAzzaHi, can someone tell me, how to run tests. at least unittests?) just cannot figure out (shy)21:06
malinimeanwhile I'll add a  ' if [[ "$MARCONI_BACKEND" = 'mongodb' ]]; then' to look similar to how it was before the patch & run the experimental job.21:07
maliniJust to make sure mongo db is the issue21:07
maliniAAzza: tox -e py2721:07
maliniAAzza: we really need some docs on how to run the tests21:08
malinimaybe you could add them, once you figure out?21:08
AAzzamalini: maybe i can add info to readme about this?21:09
maliniAAzza: that would be awesome!!21:09
AAzzamalini: and also, should i install something other from requirements-test? tox command is not found(21:09
maliniyes..you need tox21:10
maliniI believe cpallares ran into some issues recently21:10
malini& had to use an older version of tox21:10
malinitry tox (1.6.1)21:11
AAzzamalini:  hmm...  i'll try, but there is no tox in test-requirements. is it ok?21:12
*** Kuwagata has quit IRC21:12
malinigood point..I dont know why, there's got to be a reason21:13
AAzzamalini: yeah, it runs with tox==1.6.1. thanks)21:14
flaper87malini: FWIW, something really bad happened to this job: http://logs.openstack.org/88/74088/4/experimental/check-tempest-dsvm-marconi/5400c48/console.html21:16
flaper87try running it again21:17
flaper87lol21:17
flaper87just read your comment21:17
flaper87nvm21:17
flaper87malini: are you following up with -infra about that failure ?21:19
flaper87that doesn't seem related to mongodb / marconi / whatsoever21:19
*** mpanetta has quit IRC21:27
maliniflaper87: why do you think its unrelated?21:28
malinihttp://logs.openstack.org/88/74088/4/experimental/check-tempest-dsvm-marconi/5400c48/logs/error.txt.gz21:29
malinithey recently changed devstack to quit in case of errors21:29
maliniSo I believe it just aborted coz of the marconi error21:29
flaper87malini: ah mmh, I had not seen that log21:30
flaper87:D21:30
malininp..Thts all I have been doing for a month now21:30
maliniSo I know where the error wud show up :D21:30
*** oz_akan_ has quit IRC21:33
*** balajiiyer1 has quit IRC21:49
*** fifieldt has joined #openstack-marconi21:54
*** jmckind has quit IRC21:58
*** AAzza has quit IRC22:02
*** mwagner_lap has quit IRC22:04
*** amitgandhi has quit IRC22:23
*** jergerber has joined #openstack-marconi22:36
*** malini is now known as malini_afk22:39
*** jergerber has quit IRC22:41
*** jcru has quit IRC22:49
*** balajiiyer has joined #openstack-marconi23:32
*** jqiu25 has joined #openstack-marconi23:53
*** oz_akan_ has joined #openstack-marconi23:54
*** mpanetta has joined #openstack-marconi23:57

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!