Wednesday, 2014-03-05

« Tuesday, 2014-03-04 Index Thursday, 2014-03-06 »
*** chandan_kumar has joined #openstack-climate03:42
*** chandan_kumar has quit IRC03:46
*** chandankumar_ has quit IRC05:28
*** chandan_kumar has joined #openstack-climate05:38
*** saju_m has joined #openstack-climate07:02
*** bauzas has joined #openstack-climate07:54
*** Nikolay_1t has joined #openstack-climate08:06
*** Nikolay_St has quit IRC08:11
*** openstackgerrit has quit IRC08:11
*** saju_m has quit IRC08:11
*** bauzas has quit IRC08:11
*** chandan_kumar has quit IRC08:11
*** YorikSar has quit IRC08:11
*** scroiset has quit IRC08:11
*** SergeyLukjanov has quit IRC08:11
*** DinaBelova has quit IRC08:11
*** f_rossigneux has quit IRC08:11
*** openstackgerrit has joined #openstack-climate09:01
*** bauzas has joined #openstack-climate09:01
*** saju_m has joined #openstack-climate09:01
*** chandan_kumar has joined #openstack-climate09:01
*** YorikSar has joined #openstack-climate09:01
*** SergeyLukjanov has joined #openstack-climate09:01
*** DinaBelova has joined #openstack-climate09:01
*** scroiset has joined #openstack-climate09:01
*** f_rossigneux has joined #openstack-climate09:01
*** bauzas has quit IRC09:14
*** saju_m has quit IRC09:19
*** bauzas has joined #openstack-climate09:27
*** saju_m has joined #openstack-climate09:35
bauzasDinaBelova: something weird happened with https://bugs.launchpad.net/climate/+bug/128817911:14
bauzasfor some reason, an Ubuntu guy triaged this bug11:14
openstackgerritSylvain Bauza proposed a change to stackforge/climate: Fix Empty Context stack issue  https://review.openstack.org/7816411:17
*** chandan_kumar has quit IRC11:25
DinaBelovahm, maybe that's some pecan/wsme error?11:27
DinaBelovamm11:27
DinaBelovaI don't know really :D11:27
bauzasnah, I think the guy was drunk11:33
DinaBelova:D11:35
Nikolay_1tbauzas: :D11:37
*** chandan_kumar has joined #openstack-climate12:11
*** casanch1 has joined #openstack-climate12:58
*** cmart has joined #openstack-climate13:20
bauzasYorikSar: you there, about discussing https://bugs.launchpad.net/climate/+bug/128817913:30
bauzasYorikSar: ?13:30
casanch1bauzas: hi13:44
bauzascasanch1: hi13:44
casanch1please, whenever you have some time, could you take a look at the notifications review https://review.openstack.org/#/c/75189/13:45
casanch1I think it'll be great to have your input13:45
bauzascasanch1: I'll do my best :)13:46
casanch1thanks13:46
YorikSarbauzas: Hi13:56
YorikSarSo... What about it?13:56
bauzasYorikSar: I just placed some comments within the bug13:56
bauzasYorikSar: feel free to read these13:56
bauzasYorikSar: I also gave you the stacktraces13:56
YorikSarbauzas: Ok, sure.13:57
bauzasYorikSar: I'm currently digging into Pecan to see with after() is called without before()13:57
YorikSarbauzas: In fact I find it really strange that after() can be called w/o before().13:57
bauzasYorikSar: yup, sounds like a Pecan bug13:57
YorikSarbauzas: Oh, no.13:57
YorikSarbauzas: It isn't.13:57
bauzasthat's another thread then ?13:57
DinaBelovabauzas, btw - to get rid of duplicate option error test your patch with previous oslo.messaging13:57
DinaBelovaor update openstack common13:58
DinaBelovaby using appropriate patchg13:58
bauzasDinaBelova: don't worry, I just went through the merged patch13:58
YorikSarbauzas: If before() wouldn't have been called we'd get another error - no context object would be available.13:58
bauzasit just says that the stack is empty13:58
YorikSarbauzas: So it means that after() is called twice.13:59
bauzaswell, I can put some logging in there13:59
bauzaswe'll see13:59
bauzashold on, instrumenting this13:59
YorikSarbauzas: This is some Pecan's bullshitting, actually... I guess we can add a flag that will prevent second __exit__ from happening.13:59
bauzaswell, sounds ok to me14:00
bauzaswe put some flag in the state dict14:00
bauzasand if the flag is there, run __exit__()14:00
bauzasand of course, consume the flag14:00
bauzasok, will do this14:00
YorikSarbauzas: Wait...14:01
YorikSarbauzas: We should make sure we don't do smth nondeterministic here...14:01
bauzasdon't you think it would be good to replace the IndexError exception by a more readable one14:01
bauzasmmm14:01
YorikSarbauzas: I've just thought of a possible reason for this, actually.14:01
YorikSarbauzas: It might be because of on_error called before after()14:02
bauzason_error() is never called14:02
bauzasWSME is catching the exceptions14:02
bauzasoh14:02
bauzasok, lemme instrument this14:02
YorikSarbauzas: You should check it.14:02
YorikSarbauzas: If that's the case, you should store error in the request object and pass it to __exit__14:03
bauzashold 5 mins14:04
bauzasYorikSar: you got it14:06
bauzas2014-03-05 15:06:28.012 3229 DEBUG climate.api.v2.hooks [-] ***** CTX ERROR ***** on_error /home/bauzas/climate/climate/api/v2/hooks.py:5814:06
bauzas2014-03-05 15:06:28.016 3229 DEBUG climate.api.v2.hooks [-] ***** CTX AFTER ***** after /home/bauzas/climate/climate/api/v2/hooks.py:5414:06
bauzasoh dear...14:06
bauzasthat's not going to on_error()14:07
bauzasthat's going to on_error()14:07
bauzasthat would be worth totally removing it14:07
bauzasas WSME is catching all exceptions14:07
bauzasYorikSar: what do you think about it?14:08
YorikSarbauzas: We should definitelly handle on_error. Because WSME can fail. Or smth else can fail before it.14:09
YorikSarbauzas: But if WSME handled the error, we don't need to pass it into Context.14:09
bauzasYorikSar: so, what would be the best option ? storing the error within the state dict ?14:09
YorikSarbauzas: In fact, since after() is called anyway,  you should just remove on_error and it'll be fine.14:10
bauzas(15:07:16) bauzas: that's going to on_error()14:10
bauzas(15:07:39) bauzas: that would be worth totally removing it14:10
bauzas:)14:10
bauzasok, will remove the on_error() hook14:10
bauzasafter() is called anyway14:11
bauzaswill put a comment for clarity14:11
YorikSarbauzas: Ok.14:11
bauzaswelcome in my Pecan/WSME buggy world14:11
bauzasstill have one last question tho14:12
bauzasI personnally think it would be good to raise a more understandable exception than IndexError for the __exit__() clause14:12
bauzasI can propose ContextException('stack is empty')14:12
bauzasYorikSar: do you agree ?14:13
YorikSarbauzas: I want to avoid Climate-specific exceptions in contexts...14:13
bauzaswhy ? that's climate.context :)14:14
YorikSarbauzas: You can add descriptive assert before that line, actually.14:14
bauzasunless you want to put it in oslo :)14:14
YorikSarbauzas: Because I want it to become oslo.context probably one day ;)14:14
bauzas:)14:14
*** saju_m has quit IRC14:15
YorikSarbauzas: You can add smth like this:14:15
bauzaswell, ok for a simple assert then14:15
YorikSarassert self._context_stack, "context stack is already empty"14:15
bauzasok14:18
bauzaswill amend my commit14:18
bauzasthanks14:18
*** chandan_kumar has quit IRC15:51
*** chandan_kumar has joined #openstack-climate15:57
openstackgerritSylvain Bauza proposed a change to stackforge/climate: Port to Pecan/WSME for API v2  https://review.openstack.org/7101116:28
openstackgerritCristian A Sanchez proposed a change to stackforge/climate: Sends notifications at lease events  https://review.openstack.org/7518916:47
openstackgerritSylvain Bauza proposed a change to stackforge/climate: Add extensions for API v2 controllers  https://review.openstack.org/7830917:11
casanch1guys, is it possible to start climate-m with remote debugging options like the rest of the openstack components?17:12
bauzascasanch1: not yet17:13
bauzasyou have to manually inject a pdb.set_trace()17:14
casanch1oh ok17:14
bauzascasanch1: I'm personnally using ST3 for texting17:14
casanch1i'm using pycharm but I cannot do remote debug17:14
bauzascasanch1: I dislike Pydev or PyCharm17:14
bauzascasanch1: too way huge for me17:14
bauzascasanch1: I have to go17:15
casanch1bauzas: ok, thanks17:15
*** pcargnel has joined #openstack-climate17:15
*** chandan_kumar has quit IRC17:16
*** bauzas has quit IRC17:19
*** openstackgerrit has quit IRC17:32
*** openstackgerrit has joined #openstack-climate18:26
*** bauzas has joined #openstack-climate18:55
*** openstackgerrit has quit IRC19:02
*** casanch1_ has joined #openstack-climate20:05
*** casanch1 has quit IRC20:06
*** casanch1_ is now known as casanch120:07
*** casanch1 has quit IRC20:11
*** casanch1 has joined #openstack-climate20:33
*** cmart has quit IRC20:34
*** casanch1 has quit IRC21:10
*** bauzas has quit IRC21:27
*** pcargnel has quit IRC23:02
*** openstack has joined #openstack-climate23:03
-dickson.freenode.net- [freenode-info] channel trolls and no channel staff around to help? please check with freenode support: http://freenode.net/faq.shtml#gettinghelp23:03
« Tuesday, 2014-03-04 Index Thursday, 2014-03-06 »

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