Monday, 2014-09-08

« Sunday, 2014-09-07 Index Tuesday, 2014-09-09 »
*** alcabrera|afk is now known as alcabrera01:07
*** alcabrera is now known as alcabrera|afk01:24
*** flwang1 has joined #openstack-zaqar01:53
*** echevemaster has joined #openstack-zaqar02:03
*** vkmc has quit IRC02:47
*** echevemaster has quit IRC03:19
*** sgotliv has joined #openstack-zaqar05:32
*** flwang1 has quit IRC05:38
*** flaper87|afk is now known as flaper8705:39
*** sgotliv has quit IRC06:44
*** prashanthr_ has joined #openstack-zaqar07:15
*** flwang1 has joined #openstack-zaqar07:58
openstackgerritA change was merged to openstack/zaqar: Fix pep8 error in bench tool  https://review.openstack.org/11941608:04
*** prashanthr_1 has joined #openstack-zaqar08:04
*** prashanthr_ has quit IRC08:05
*** sgotliv has joined #openstack-zaqar08:07
*** flwang1 has quit IRC08:59
*** flwang1 has joined #openstack-zaqar09:00
*** dguerri`afk is now known as dguerri09:13
*** sgotliv has quit IRC09:57
*** AAzza_afk has quit IRC10:32
*** AAzza_afk has joined #openstack-zaqar10:34
*** AAzza_afk is now known as AAzza10:34
*** sgotliv has joined #openstack-zaqar10:38
*** prashanthr_ has joined #openstack-zaqar10:59
*** prashanthr_1 has quit IRC10:59
*** flwang1 has quit IRC11:02
*** prashanthr_ has quit IRC11:20
*** vkmc has joined #openstack-zaqar11:32
*** flwang1 has joined #openstack-zaqar11:32
vkmcmorniiing11:35
openstackgerritFlavio Percoco proposed a change to openstack/zaqar: Add a dox.yml config file  https://review.openstack.org/11974011:48
flaper87vkmc: gooooooooooooooooooooooooooooood morning11:50
flaper87how was your weekend ?11:50
vkmcflaper87, hellooooooooooo11:50
vkmcit was great :) productive and tons of resting11:52
vkmcyours?11:52
flaper87same here, spent time with family, friends, played guitar and coded too11:53
flaper87Weekends are way more productive than working-week days11:53
vkmcweeee :)11:54
openstackgerritFlavio Percoco proposed a change to openstack/zaqar: Add a dox.yml config file  https://review.openstack.org/11974011:55
vkmchow is the guitar treating you? already learn some Zeppelin pickings?11:55
* flaper87 loves his guitar11:56
flaper87Nah, I'm in an acoustic/classic period of my life :D11:57
flaper87vkmc: btw, I left some comments on the API docs review11:58
flaper87lemme know what you think11:58
vkmcflaper87, yup, fixing them11:58
vkmcI agree with that11:58
vkmcmore considering that we are migrating to Rally in the short term11:58
flaper87right11:59
flaper87cool11:59
flaper87is people getting tired of meetings? NO ITEMS HERE: https://wiki.openstack.org/wiki/Meetings/Zaqar12:00
*** alcabrera|afk is now known as alcabrera12:00
vkmcmaybe is that we are in such a good shape that we don't need meetings to get along12:02
vkmc:p12:02
*** flwang1 has quit IRC12:03
*** jchai has joined #openstack-zaqar12:08
openstackgerritVictoria Martínez de la Cruz proposed a change to openstack/zaqar: Adds the API reference to the devs guide  https://review.openstack.org/11639012:36
openstackgerritVictoria Martínez de la Cruz proposed a change to openstack/zaqar: Adds the API reference to the devs guide  https://review.openstack.org/11639012:37
vkmcflaper87, if you have a moment, could you try to compile the API ref ^12:39
vkmcit looks there is recursion with imports again... or something alike12:39
vkmcmaybe you can spot it... I'm trying to fix it12:40
flaper87sure12:42
vkmcthaaaaaaanks12:43
flaper87yeah, the import thing is back12:59
vkmcdarn12:59
flaper87I read/heard something the other day about deja-vu being some kind of brain-race between permanent memory and temporary. This is the proof that's just bullshit. I'm having a deja-vu with this bug and it's not my brain's fault13:00
flaper87I don't get why we have an utils and a helpers13:06
flaper87sounds like utils could be merged into helpers13:07
flaper87except for 1 or 2 functions13:07
vkmcyeah...13:07
vkmcthing is13:07
vkmcwe have helpers in common13:07
*** earnThis|away has joined #openstack-zaqar13:08
vkmc(we merged utils with helpers in common, the last time)13:08
vkmcand now we have helpers in common and utils in wsgi13:08
vkmc:p13:08
*** earnThis|away is now known as earnThis13:08
vkmcnot sure why is there a 'common' place for wsgi only13:09
flaper87vkmc: because that'll be helpful when we start working on the notification API13:09
vkmcit makes sense if there are common functions for 1++ transport drivers, but in this case it's wsgi only13:09
vkmcgood point13:10
flaper87there are some common functions for wsgi transports13:10
flaper87in queues.wsgi.utils we should have *just* queue specific helpers13:10
flaper87the rest should go to common/helpers.py13:10
*** sriram has joined #openstack-zaqar13:12
vkmcok :)13:12
flaper87vkmc: fixing it, btw13:15
flaper872 seconds13:15
flaper87almost there13:15
vkmcoh I was refactoring it myself :)13:16
vkmccool, looking forward for it13:17
*** mpanetta has joined #openstack-zaqar13:21
flaper87vkmc: sphinx builds again13:22
flaper87testing zaqar now13:22
flaper87I'll push as soon as tests are done13:23
vkmcflaper87, thanks13:23
*** mpanetta_ has joined #openstack-zaqar13:26
*** mpanetta has quit IRC13:26
*** mpanetta_ has quit IRC13:27
*** mpanetta has joined #openstack-zaqar13:28
openstackgerritFlavio Percoco proposed a change to openstack/zaqar: Remove recursive imports  https://review.openstack.org/11975413:33
flaper87vkmc: ^13:33
vkmc:D13:33
vkmcneat13:34
vkmcthere are not enough poptarts for you13:34
flaper87:D13:36
*** alcabrera is now known as alcabrera|afk13:45
*** alcabrera|afk is now known as alcabrera13:52
vkmcalcabrera, hey :)13:57
alcabrerahi hi, vkmc. :)13:57
vkmcalcabrera, how are you doing? how was your weekend?13:58
alcabreravkmc: I'm basically the happiest ever, my weekend was awesome, and I feel great (and tired)! :D14:00
alcabrerahow are you? :)14:00
* flaper87 ninja-approved his own patch... so fun14:00
flaper87:D14:00
flaper87alcabrera: glad to hear that :D14:00
alcabreraflaper87: heeeyyyy14:00
vkmcflaper87, lol14:01
vkmcalcabrera, happy to hear that too!14:01
vkmcI'm doing ok, great weekend as well14:01
alcabrerayay!14:02
*** itisit has joined #openstack-zaqar14:02
vkmcanxiously waiting the travel grant support news14:03
vkmcthere is nothing yet14:03
* flaper87 pictures vkmc like this on her email client: http://www.likecool.com/Gear/Pic/GifF5/GifF5.gif14:04
vkmcROFL14:05
vkmcexactly like that hahaha14:05
alcabrerahahaha14:06
alcabreraperfect, flaper87. :P14:06
flaper87:D :D :D :D14:08
vkmcI cannot wait to meet y' all14:08
flaper87vkmc: if you don't bring gummy bears with you (and obviously for me), you are sooooooooooo going to regret meeting me.14:09
alcabreraI barely survived meeting flaper87 when I lacked the gummy bears. ;p14:09
alcabreravkmc: ^14:09
flaper87alcabrera: exactly14:10
vkmcflaper87, I was planning to bring you argentinian candy, I can switch to gummybears though14:10
* flaper87 can get really mad and starts giving bear-hugs to people until he crashes their bones14:10
vkmcoh my14:10
vkmcI'm definitively switching to gummybears14:10
flaper87vkmc: mmh, ok ok, I'll take the argentinian candies :D14:10
* kgriffs sits back with some popcorn to watch the show14:10
flaper87as long as it's sweet and it's for me and not kgriffs I'm fine14:10
flaper87ooooooooooooopsssssssssssss14:10
* kgriffs sad panda14:11
vkmclol14:11
* kgriffs panda wants candy14:11
flaper87kgriffs: I'll share some, don't worry14:11
* kgriffs panda smiles14:11
kgriffshttp://i.imgur.com/sEnOU.jpg14:11
flaper87ROOOOOOOOOOOFL14:12
vkmchahaha14:12
alcabreraawww. <314:12
vkmcare there poptarts in Paris? or is it a US thing?14:14
vkmcif it's the latter... kgriffs you better bring some14:15
flaper87there are but with different names14:15
flaper87but I deffinitely don't know how they're called so14:16
flaper87:D14:16
vkmcle poptárt14:16
flaper87you're missing an è at the end14:16
vkmclol14:16
flaper87I don't know french but I'm sure you're missing it14:16
flaper87:D14:16
vkmcI know a little of French14:17
vkmcand yeah, it is missing14:17
vkmcFrench people love letters with accents that are not really pronounced14:17
*** adrian_otto has joined #openstack-zaqar14:18
kgriffsI have a feeling that french pop-tarts are about as similar to American pop-tarts as italian pizza is to Pizza Hut14:19
flaper87LOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOL14:19
flaper87I went to Pizza Hut once - that was back in Venezuela. I died a little that day14:20
flaper87https://pbs.twimg.com/media/BxBEIA7CQAAKbd0.jpg:large14:21
earnThishey guys, can anyone provide a little more detail into whats required to resolve this bug https://bugs.launchpad.net/zaqar/+bug/1236540 ? Possibly kgriffs ?14:22
kgriffsearnThis: looking14:23
kgriffsearnThis: OK, I haven't checked this for a while - wpf's comment is interesting. I think what I was looking for was a comprehensive audit of everywhere we return HTTP 400 and making sure we are using standard classes.14:28
kgriffsFalcon 0.2 will have these classes: http://paste.openstack.org/show/108301/14:28
kgriffshowever, Zaqar is currently pinned to <0.2, so for now we would need to use Zaqar error classes (fix up existing ones and/or add new ones as necessary)14:29
*** cpallares has joined #openstack-zaqar14:33
earnThiskgriffs: alright, thanks14:34
*** alcabrera is now known as alcabrera|afk14:42
kgriffssure thing14:42
*** sriram has quit IRC14:47
*** sriram1 has joined #openstack-zaqar14:47
flaper87meeting in 1014:47
openstackgerritA change was merged to openstack/zaqar: Updated from global requirements  https://review.openstack.org/11915114:49
openstackgerritA change was merged to openstack/zaqar: Remove recursive imports  https://review.openstack.org/11975414:52
flaper87vkmc: ^14:53
flaper87vkmc: you need to remove the dependency on the other patch14:54
vkmcflaper87, yup, updating it14:54
vkmcthanks14:54
flaper87malini|afk: meeting15:01
flaper87everyone meeeeting15:01
cpallareswhich channel is the meeting?15:03
cpallaresflaper87: ^^15:04
flaper87#openstack-meeting-alt-315:05
*** alcabrera|afk is now known as alcabrera15:15
openstackgerritA change was merged to openstack/zaqar: Add ZAQAR_TEST_EVERYTHING to enable all tests  https://review.openstack.org/11955515:26
openstackgerritA change was merged to openstack/zaqar: Add a dox.yml config file  https://review.openstack.org/11974015:40
*** adrian_otto has quit IRC15:41
flaper87vkmc: minutes ? :D :D :D15:44
vkmcflaper87, of course, it's my thing :p15:44
cpallaresflaper87: You gave me the wrong channel, just saying...15:44
vkmccpallares, just in case -alt isn't enough... there are three versions of it in flaper87's head :p15:44
flaper87vkmc: LOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOL15:46
flaper87I'm sorry15:46
flaper87:(15:46
flaper87erm, cpallares ^15:46
* flaper87 is not sure what's wrong with him15:46
* flaper87 just knows there are *many* things wrong with him15:46
* vkmc wonders what happened with openstack-meeting-215:47
cpallaresvkmc: lol15:50
cpallaresflaper87: Have you seen the hobbit 3 trailers?15:51
flaper87cpallares: yup yup yup15:52
flaper87:D :D :D :D15:52
vkmcnewbie question: is there a practical way to remove a dependency in git?15:57
flaper87vkmc: if you have a 2 patches dependency chain15:58
flaper87you can do something like15:58
flaper87$ git rebase -i HEAD~215:58
flaper87That should give you a list of patches to rebase on15:58
flaper87from that list, remove the line of the commit you don't want (yes just remove it)15:59
vkmcyup15:59
flaper87and then you're done15:59
vkmccool15:59
vkmc:D15:59
openstackgerritKurt Griffiths proposed a change to openstack/zaqar: In the Redis driver, only read the msg body when needed  https://review.openstack.org/11915315:59
kgriffsthar she blows!15:59
kgriffslook lively now!15:59
vkmcnothing in the log... awesome!15:59
kgriffsarrr15:59
flaper87btw, I found some other use-cases from our old notifications etherpad15:59
flaper87https://etherpad.openstack.org/p/zaqar-integrated-projects-use-cases15:59
kgriffsbrb15:59
openstackgerritVictoria Martínez de la Cruz proposed a change to openstack/zaqar: Adds the API reference to the devs guide  https://review.openstack.org/11639015:59
kgriffsflaper87: did prashanth give an update on redis sentinel support?16:05
kgriffsI just saw this16:05
kgriffshttps://github.com/andymccurdy/redis-py#sentinel-support16:05
kgriffsI think we actually really need this in the RC16:06
kgriffsotherwise, we can't say redis driver is HA16:06
flaper87not that I know :/16:06
kgriffscrappies16:07
kgriffsalcabrera, vkmc: have you heard from prashanth16:07
alcabrerakgriffs: no. I vaguely remember her was interviewing for a thing.16:07
alcabreraperhaps that time is now?16:07
kgriffswe need a patch submitted TODAY for sentinel support16:07
kgriffsi had asked prashanth to work on that16:08
kgriffs(early last week)16:08
kgriffsguess I should have given a deadline.16:08
* kgriffs forgot16:08
vkmckgriffs, I didnt!16:09
kgriffsanybody bored with nothing to do today? ;)16:09
* flaper87 is about to get into yet-another-meeting16:11
flaper87kgriffs: what's that thing about?16:11
flaper87lets see if I can help16:11
kgriffsflaper87: master-slave support for Redis16:11
kgriffsso you can have HA16:12
kgriffsrequires a little bit of coding due to the design of redis-py16:12
kgriffshttps://github.com/andymccurdy/redis-py#sentinel-support16:12
flaper87mmh, damn. then I think I won't be able to get it done today16:12
kgriffsI don't feel good about saying redis is "done" until we can say we have this16:12
kgriffswell, if anyone sees prashanth ask him if he has started on this16:13
*** sgotliv has quit IRC16:13
kgriffsi may have some time later today to do it16:13
* kgriffs crosses fingers16:13
flaper87kgriffs: ok. Please let me know, otherwise I'll take a look at it tomorrow16:14
kgriffskk16:14
*** alcabrera is now known as alcabrera|afk16:38
*** reed has joined #openstack-zaqar16:41
kgriffsflaper87: ping16:42
*** mpanetta has quit IRC16:50
openstackgerritVictoria Martínez de la Cruz proposed a change to openstack/zaqar: Adds the API reference to the devs guide  https://review.openstack.org/11639016:50
vkmcoff to lunch, brb16:55
openstackgerritKurt Griffiths proposed a change to openstack/zaqar: Fix: Redis driver does not fall back to default port  https://review.openstack.org/11982016:55
openstackgerritKurt Griffiths proposed a change to openstack/zaqar: Add missing flavors_controller method to Redis driver  https://review.openstack.org/11982116:56
openstackgerritKurt Griffiths proposed a change to openstack/zaqar: In storage driver base, raise instance of NotImplementedError  https://review.openstack.org/11982317:00
*** adrian_otto has joined #openstack-zaqar17:03
*** sriram1 is now known as sriram17:35
*** alcabrera|afk is now known as alcabrera17:53
*** sgotliv has joined #openstack-zaqar18:31
*** openstackgerrit has quit IRC18:51
*** earnThis has quit IRC18:58
*** openstackgerrit has joined #openstack-zaqar19:07
*** jchai is now known as jchai_afk19:08
*** itisit has quit IRC19:09
*** echevemaster has joined #openstack-zaqar19:51
*** sgotliv has quit IRC19:57
*** sgotliv has joined #openstack-zaqar19:58
kgriffspeeps, can I get some reviews? Also, let me know if y'all have some patches you would like me to focus on. https://review.openstack.org/#/q/status:open+project:openstack/zaqar,n,z20:00
vkmcsure thing20:00
kgriffsvkmc: btw, do you know why we don't have redis in our requirements.txt?20:04
kgriffswe have pymongo20:04
kgriffsredis is in test-requirements.txt20:04
vkmcw00t20:04
vkmcit should be in requirements.txt20:04
kgriffsok, that's what I was thinking20:04
kgriffsvkmc: would you mind adding a patch for that real quick?20:04
vkmckgriffs, sure thing20:04
vkmcI should stop saying 'sure thing'20:05
kgriffsmove redis req from test-requirements.txt to requirements.txt20:05
kgriffsvkmc: heh20:05
kgriffsgets a person in trouble, that's for sure. :D20:05
kgriffsanyway, thanks!20:05
kgriffsbrb20:08
vkmcnp!20:08
*** jchai_afk is now known as jchai20:11
*** arunkant has quit IRC20:17
*** flwang1 has joined #openstack-zaqar20:19
*** echevemaster has quit IRC20:20
openstackgerritKurt Griffiths proposed a change to openstack/zaqar: Fix zaqar-bench hanging after resetting queues  https://review.openstack.org/11990020:23
flaper87kgriffs: probably because redis is not our default storage20:23
*** flwang1 has quit IRC20:24
flaper87TBH, I think pymongo shouldn't be in requirements either20:24
flaper87We recommend mongodb as the production storage but I'm more of the idea we shouldn't have a *default* one20:24
kgriffsyeah20:24
kgriffs+1 for no "default"20:25
vkmchmm.. good thinking20:25
vkmcso I remove pymongo20:25
vkmcinstead of adding redis20:25
flaper87vkmc: +120:25
flaper87;)20:25
kgriffsas for requirements.txt - we would need to document that extra pip install step if we remove pymongo from requirements.txt20:25
flaper87If we move it to test-requirements, that should just work. All our test environements install things from there20:25
kgriffswhat about operators though? they aren't going to use test-requirements.txt20:26
flaper87sounds good to me. Probably it's worth to add it to the user-guide20:26
flaper87they're going to use packages and that's normally left to the distro20:26
kgriffson the other hand, we can recommend hiredis as well even though it isn't in global-reqs. :D20:26
flaper87for example, in fedora, we'll probably just add pymongo as a dependency20:26
flaper87but that's just because it's cheap20:26
flaper87indeed20:26
flaper87Distros do different things anyway20:27
flaper87there are many packages with dependencies that are not a strong requirement for the project20:27
flaper87that depends on customers, distro goals and other things20:27
flaper87In other words, requirements.txt doesn't affect operators, distros do20:27
kgriffsflaper87: speaking of distros... what is the most recent version of Redis can we depend on?20:29
* kgriffs switches context20:29
flaper87ah I meant to look that up and forgot20:30
* flaper87 tries to remember the fucking command20:30
flaper87kgriffs: it looks like it is 2.620:33
kgriffsif we can gate on 2.6, that would super rad20:35
*** alcabrera is now known as alcabrera|afk20:35
kgriffsbecause that is when lua support landed20:35
vkmckgriffs, how can I reproduce this? https://review.openstack.org/#/c/119900/121:08
vkmcjust run the bench?21:08
kgriffsyeah21:09
kgriffszaqar-bench -v21:09
kgriffssee if it happens ony your box21:09
kgriffsI should actually mark this as WIP until we figure out root cause21:09
vkmck21:09
kgriffsI see the problem on devian wheezy21:09
*** sriram has quit IRC21:09
kgriffsdebian21:09
kgriffsheh21:09
vkmcoh ok21:11
vkmctesting on Ubuntu Precise21:12
vkmchey btw, I pinged you last week to ask you how did you set up the environment to test https://bugs.launchpad.net/zaqar/+bug/136134621:13
kgriffsoh, it is all debian wheezy21:14
kgriffsuWSGI + gevent21:14
vkmcI'm just too rookie with uWSGI so I'm afraid I didn't set it up correctly21:14
kgriffsconfig: http://paste.openstack.org/show/100592/21:14
kgriffsapp.py: http://paste.openstack.org/show/100593/21:14
kgriffsvkmc: nah, it is more likely I made a mistake in my config.21:14
kgriffsI'm OK at system administration, but no expert21:15
kgriffspython 2.7.321:15
vkmcand then simply... /usr/bin/uwsgi --ini uwsgi.ini21:16
vkmcright?21:16
kgriffsyeah, although I think you don't need to explicitly put --ini21:18
kgriffsuwsgi uwsgi.ini21:18
vkmcok, I'll try again21:18
vkmcand after that I killed mongo21:18
*** jchai has quit IRC21:20
kgriffskill the primary21:20
vkmcbut I didn't see anything in sterr21:20
kgriffsoh snap... i think we have a bug in python-zaqarclient21:21
* kgriffs is multitasking21:21
vkmcoh noes21:21
kgriffshttps://gist.github.com/anonymous/3940b7e6f34d58ede01021:22
vkmchmm, how did you reproduce that?21:23
kgriffsI just ran zaqar-bench with a consumer worker enabled21:23
vkmcbtw, I could check this https://review.openstack.org/#/c/119900/121:24
vkmcand it's happening21:24
vkmcwithout the fix it hangs, with the fix it succesfully finish21:24
kgriffsso strange21:24
vkmcyeah21:24
vkmcit's new though21:24
kgriffsI suspect something strange with gevent monkey patching21:24
kgriffsjust a hunch21:24
vkmcthat error of zaqarclient is raising on the bench tool in master?21:26
kgriffsno, it is the one with that patch above applied21:26
kgriffs11990021:26
vkmck21:27
vkmclet me run it21:27
vkmchttp://paste.openstack.org/show/108481/21:30
vkmcok21:30
vkmcnow we can panic as a team21:30
vkmcD:21:30
kgriffscrud21:30
vkmcdid u find it already?21:30
kgriffsno21:30
vkmcoh21:30
vkmcunderstood that it was some crud operation21:30
vkmcnow, thanks to urban dictionary, I know what you truly meant21:31
kgriffsi started looking but the problem isn't obvious21:31
vkmcit's not21:33
vkmcit's something with gevent though21:34
vkmc'greenlet' is a library gevent uses IIRC21:35
*** flaper87 is now known as flaper87|afk21:38
vkmckgriffs, about this change https://review.openstack.org/#/c/11982321:51
vkmcwe should change this in other modules as well21:51
vkmcto keep consistency21:51
vkmcshould we pospone it?21:51
*** sgotliv has quit IRC21:52
kgriffsoh, I must have missed it. Are there other places too?21:52
vkmcyup... for instance in transport/base.py21:53
*** alcabrera|afk is now known as alcabrera21:53
vkmchttps://github.com/openstack/zaqar/blob/master/zaqar/queues/transport/base.py21:54
vkmcwell it's ONE appeareance21:54
vkmcbut I saw it in other places as well21:54
kgriffsah, good catch21:57
kgriffsmy grep must be less educated than I realized21:57
kgriffsyeah, just drop a -1 with a comment and I'll circle back on it21:57
vkmcconsider it DONE22:01
*** earnThis has joined #openstack-zaqar22:02
kgriffsrock on22:03
*** earnThis has left #openstack-zaqar22:03
vkmcbrb22:10
kgriffsvkmc: can you work on fixing that claim_id bug in the client?22:14
kgriffsthe message __init__ does not take a "claim_id" param22:14
kgriffsbut one is being passed22:14
vkmckgriffs, will do22:15
kgriffsthanks!22:15
vkmcnp22:15
vkmc:)22:15
kgriffsI'm fixing a bug in the redis MessageEnvelope patch atm22:15
* kgriffs thinks zaqar-bench is a pretty good smoke test!22:16
kgriffsvkmc: you're a lifesaver!22:16
openstackgerritKurt Griffiths proposed a change to openstack/zaqar: In the Redis driver, only read the msg body when needed  https://review.openstack.org/11915322:35
openstackgerritKurt Griffiths proposed a change to openstack/zaqar: Fix: Redis driver does not fall back to default port  https://review.openstack.org/11982022:35
openstackgerritKurt Griffiths proposed a change to openstack/zaqar: Add missing flavors_controller method to Redis driver  https://review.openstack.org/11982122:35
*** alcabrera is now known as alcabrera|afk22:39
kgriffsflaper87|afk: if you see this, please work on implementing support for Redis sentinel. I am fixing a claims bug in the driver that is blocking the benchmark, so won't have time.22:44
kgriffsflwang: or you could do it if you have a few minutes?22:45
kgriffsthe Redis driver needs support for master-slave: https://github.com/andymccurdy/redis-py#sentinel-support22:45
kgriffsit should be optional, i guess (needs to be enabled in zaqar.conf by the operator)22:46
flwangkgriffs: I will take a look at it today, I assume we just need add a config entry for that, is it?22:46
kgriffsslightly more complex than that, but not much22:46
kgriffsyou also need to look where StrictRedis would normally be instantiated in the redis driver and instead create a Sentinel instance22:47
kgriffsand then rturn sentinel.discover_master22:47
kgriffsfinally, you will need to simulate a master failover and see if redis-py automatically recovers, or we have to add some retry logic ourselves22:48
kgriffsflwang: if you could do that, it would be super crazy awesome. I'm up against the wall with so many things on my plate right now22:48
kgriffswe need in in time for RC22:48
flwangkgriffs: ok, I will work on that. so the deadline for this is RC date, right?22:49
kgriffsyeah, but the sooner the better so we can have some time to smoke test22:49
flwangkgriffs: in other words, I don't think I can complete it TODAY :)22:49
flwangkgriffs: yep, I see22:49
flwangis there a bug or something to track this?22:49
kgriffsflwang: mmm, hold on sec22:50
kgriffsnow there is22:52
kgriffshttps://bugs.launchpad.net/zaqar/+bug/136702022:52
flwangawesome22:52
kgriffsthanks man. you're a lifesaver22:53
flwangkgriffs: you can create new life, God22:55
*** kgriffs is now known as kgriffs|afk23:20
« Sunday, 2014-09-07 Index Tuesday, 2014-09-09 »

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