Tuesday, 2016-03-15

« Monday, 2016-03-14 Index Wednesday, 2016-03-16 »
*** MVenesio has joined #openstack-oslo00:14
*** MVenesio has quit IRC00:19
*** Kennan has quit IRC00:20
*** Kennan has joined #openstack-oslo00:21
*** dims has joined #openstack-oslo00:23
openstackgerritJoshua Harlow proposed openstack/tooz: Add mysql table support (if provided a database)  https://review.openstack.org/29017800:25
*** mordred has quit IRC00:31
*** dims has quit IRC00:32
*** mordred has joined #openstack-oslo00:32
*** yamamoto_ has quit IRC00:34
*** ozamiatin has quit IRC00:36
*** vilobhmm11 has quit IRC00:53
*** dims has joined #openstack-oslo00:56
*** dimtruck is now known as zz_dimtruck00:59
*** zz_dimtruck is now known as dimtruck01:12
openstackgerritJamie Lennox proposed openstack/oslo.config: Handle some native python types in config generation  https://review.openstack.org/29264701:13
*** suro-patz1 has quit IRC01:16
*** vilobhmm11 has joined #openstack-oslo01:22
*** vilobhmm111 has joined #openstack-oslo01:27
*** vilobhmm11 has quit IRC01:29
*** EinstCrazy has joined #openstack-oslo01:32
*** yamamoto_ has joined #openstack-oslo01:34
*** yamamoto_ has quit IRC01:43
*** amrith is now known as _amrith_01:49
openstackgerritJoshua Harlow proposed openstack/tooz: Add mysql table support (if provided a database)  https://review.openstack.org/29017801:57
*** vilobhmm111 has quit IRC02:00
*** salv-orlando has joined #openstack-oslo02:02
*** vilobhmm11 has joined #openstack-oslo02:03
*** browne has quit IRC02:03
*** salv-orl_ has quit IRC02:04
*** yamamoto has joined #openstack-oslo02:07
openstackgerritMerged openstack/oslo.middleware: cors: prevent WebOb setting a default Content-Type  https://review.openstack.org/29048202:09
*** MVenesio has joined #openstack-oslo02:15
*** MVenesio has quit IRC02:20
*** yamamoto has quit IRC02:27
*** dims has quit IRC02:38
*** dimtruck is now known as zz_dimtruck02:39
*** browne has joined #openstack-oslo02:43
*** EinstCrazy has quit IRC02:47
*** EinstCrazy has joined #openstack-oslo02:48
*** EinstCrazy has quit IRC02:50
*** browne has quit IRC03:00
*** yamahata has joined #openstack-oslo03:02
*** gcb has joined #openstack-oslo03:06
openstackgerritjaveme proposed openstack/oslo.messaging: Missing version parameter in can_send_version()  https://review.openstack.org/29038503:07
*** yamamoto has joined #openstack-oslo03:09
*** mtanino has quit IRC03:17
*** vilobhmm11 has quit IRC03:19
*** Kennan has quit IRC03:23
*** Kennan has joined #openstack-oslo03:26
*** links has joined #openstack-oslo03:38
*** vilobhmm11 has joined #openstack-oslo03:45
*** jdandrea_ has quit IRC03:48
*** jdandrea has joined #openstack-oslo03:51
openstackgerritJoshua Harlow proposed openstack/tooz: Add mysql table support (if provided a database)  https://review.openstack.org/29017804:00
*** MVenesio has joined #openstack-oslo04:16
*** amotoki has joined #openstack-oslo04:17
*** MVenesio has quit IRC04:20
openstackgerritjaveme proposed openstack/oslo.messaging: Missing version parameter in can_send_version()  https://review.openstack.org/29038505:04
*** EinstCrazy has joined #openstack-oslo05:07
*** EinstCrazy has quit IRC05:09
*** EinstCrazy has joined #openstack-oslo05:09
openstackgerritJoshua Harlow proposed openstack/tooz: Add mysql table support (if provided a database)  https://review.openstack.org/29017805:11
openstackgerritJoshua Harlow proposed openstack/oslo.serialization: Make msgpack registries copyable (and add __contains__)  https://review.openstack.org/28961105:17
*** amotoki has quit IRC05:21
*** amotoki has joined #openstack-oslo05:31
*** amotoki has quit IRC05:35
*** amotoki has joined #openstack-oslo05:47
*** suro-patz has joined #openstack-oslo05:49
*** zzzeek has quit IRC05:49
*** ndipanov has quit IRC05:49
*** zzzeek has joined #openstack-oslo05:50
*** salv-orl_ has joined #openstack-oslo05:54
*** amotoki has quit IRC05:55
*** salv-orlando has quit IRC05:57
*** jecarey has joined #openstack-oslo06:12
*** jecarey has quit IRC06:16
*** MVenesio has joined #openstack-oslo06:17
*** MVenesio has quit IRC06:21
*** EinstCrazy has quit IRC06:33
*** gcb has quit IRC06:45
*** gcb has joined #openstack-oslo06:57
*** suro-patz has quit IRC07:05
*** tesseract has joined #openstack-oslo07:13
*** tesseract is now known as Guest9751907:13
*** ihrachys has joined #openstack-oslo07:17
*** gcb has quit IRC07:25
*** nkrinner has joined #openstack-oslo07:28
*** fhermeni has joined #openstack-oslo07:34
*** gcb has joined #openstack-oslo07:42
*** yamahata has quit IRC07:52
*** itisha has joined #openstack-oslo07:57
*** Guest97519 has quit IRC08:00
*** mhickey has joined #openstack-oslo08:01
*** tesseract has joined #openstack-oslo08:05
*** tesseract is now known as Guest9179408:06
*** MVenesio has joined #openstack-oslo08:17
*** salv-orl_ has quit IRC08:18
*** MVenesio has quit IRC08:22
*** jpena|off is now known as jpena08:22
fhermenihi08:22
*** vilobhmm11 has quit IRC08:32
*** Guest91794 has quit IRC08:45
*** ndipanov has joined #openstack-oslo08:47
*** AJaeger has quit IRC08:54
*** AJaeger has joined #openstack-oslo08:54
*** e0ne has joined #openstack-oslo08:57
*** binarin has joined #openstack-oslo08:57
*** Guest91794 has joined #openstack-oslo09:00
*** zz_dimtruck is now known as dimtruck09:02
*** fhermeni has quit IRC09:03
*** dulek has left #openstack-oslo09:03
*** dtardivel has joined #openstack-oslo09:03
*** deray has joined #openstack-oslo09:20
*** salv-orlando has joined #openstack-oslo09:21
*** andymaier has joined #openstack-oslo09:23
*** salv-orlando has quit IRC09:29
*** salv-orlando has joined #openstack-oslo09:30
*** rcernin has joined #openstack-oslo09:31
*** lucas-dinner is now known as lucasagomes09:33
*** yassine__ has joined #openstack-oslo09:35
openstackgerritOleksii Zamiatin proposed openstack/oslo.messaging: [zmq] Reduce number of connections  https://review.openstack.org/28709409:52
*** sdague has joined #openstack-oslo10:01
*** e0ne has quit IRC10:08
*** e0ne has joined #openstack-oslo10:08
*** MVenesio has joined #openstack-oslo10:18
*** cdent has joined #openstack-oslo10:21
*** AJaeger has quit IRC10:21
*** MVenesio has quit IRC10:23
*** shardy has joined #openstack-oslo10:39
*** dimtruck is now known as zz_dimtruck10:40
*** zz_dimtruck is now known as dimtruck10:42
openstackgerritOleksii Zamiatin proposed openstack/oslo.messaging: [zmq] Refactoring consumer side  https://review.openstack.org/28304910:43
*** andymaier has quit IRC10:47
*** cdent has quit IRC10:47
*** cdent has joined #openstack-oslo10:48
*** AJaeger has joined #openstack-oslo10:57
*** yassine__ has quit IRC10:58
*** pkholkin has joined #openstack-oslo11:06
*** dbite has joined #openstack-oslo11:06
*** AJaeger_ has joined #openstack-oslo11:06
*** nkrinner_ has joined #openstack-oslo11:06
*** nkrinner_ has quit IRC11:07
*** dguitarb1 has joined #openstack-oslo11:07
*** nkrinner_ has joined #openstack-oslo11:07
*** AJaeger__ has joined #openstack-oslo11:08
*** gcb has quit IRC11:08
*** dguitarbite has quit IRC11:09
*** itisha has quit IRC11:09
*** nkrinner has quit IRC11:09
*** AJaeger has quit IRC11:10
*** AJaeger_ has quit IRC11:11
*** dbite has quit IRC11:11
*** dims has joined #openstack-oslo11:12
*** fhermeni has joined #openstack-oslo11:13
fhermenihi there11:13
-openstackstatus- NOTICE: Gerrit is going to be restarted11:14
fhermeniharlowja_at_home: ping ?11:14
*** itisha has joined #openstack-oslo11:17
openstackgerritMerged openstack/oslo.log: Explicitly exclude tests from bandit scan  https://review.openstack.org/29205711:17
openstackgerritMerged openstack/oslo.messaging: Explicitly exclude tests from bandit scan  https://review.openstack.org/29205611:17
openstackgerritMerged openstack/oslo.middleware: Disable http_proxy_to_wsgi middleware by default  https://review.openstack.org/28305211:18
*** gcb has joined #openstack-oslo11:19
*** dims has quit IRC11:20
*** dims has joined #openstack-oslo11:21
*** dims has quit IRC11:21
*** dims has joined #openstack-oslo11:21
openstackgerritOleksii Zamiatin proposed openstack/oslo.messaging: [zmq] Refactoring consumer side  https://review.openstack.org/28304911:23
*** ihrachys has quit IRC11:31
*** ihrachys has joined #openstack-oslo11:31
-openstackstatus- NOTICE: Gerrit had to be restarted because was not responsive. As a consequence, some of the test results have been lost, from 08:30 UTC to 10:30 UTC approximately. Please recheck any affected jobs by this problem.11:33
*** ihrachys has quit IRC11:36
-openstackstatus- NOTICE: Gerrit had to be restarted because was not responsive. As a consequence, some of the test results have been lost, from 09:30 UTC to 11:30 UTC approximately. Please recheck any affected jobs by this problem.11:36
flaper87harlowja_at_home: HN? really?11:45
flaper87harlowja_at_home: :P11:45
fhermeniharlowja_at_home: Can you interact a bit with the nova comments on the review ? They sound not really open to discussion to me. The TC meeting tonight will be harsh I guess11:47
*** _amrith_ is now known as amrith11:54
*** salv-orl_ has joined #openstack-oslo11:54
*** salv-orlando has quit IRC11:57
*** kgiusti has joined #openstack-oslo11:58
jaypipesfhermeni, harlowja_at_home: the resource-providers blueprint series has the most information about the plans to refactor the resource tracking and scheduler in Nova and eventually split out the scheduler. The following blueprints outline this:11:59
jaypipeshttps://review.openstack.org/#/c/225546/11:59
patchbotjaypipes: patch 225546 - nova-specs - Add concept of resource providers (MERGED)11:59
jaypipeshttps://review.openstack.org/#/c/283253/12:00
patchbotjaypipes: patch 283253 - nova-specs - Amend resource-providers database schema (MERGED)12:00
jaypipeshttps://review.openstack.org/#/c/256297/12:00
patchbotjaypipes: patch 256297 - nova-specs - resource-providers: Introduce resource classes (MERGED)12:00
jaypipeshttps://review.openstack.org/#/c/253187/12:00
patchbotjaypipes: patch 253187 - nova-specs - resource-providers: generic resource pools12:00
jaypipeshttps://review.openstack.org/#/c/260048/12:00
patchbotjaypipes: patch 260048 - nova-specs - resource-providers: Migrate compute node inventory12:00
jaypipeshttps://review.openstack.org/#/c/271779/12:00
patchbotjaypipes: patch 271779 - nova-specs - resource-providers: Move allocation fields12:00
jaypipeshttps://review.openstack.org/#/c/271823/12:00
patchbotjaypipes: patch 271823 - nova-specs - resource-providers: Scheduler filters in DB12:00
jaypipesfhermeni, harlowja_at_home: and finally, the following blueprint describes some thoughts I had about making the scheduling/placement API more generic:12:01
jaypipeshttps://review.openstack.org/#/c/183837/12:01
patchbotjaypipes: patch 183837 - nova-specs - Add generic scheduling/placement API12:01
*** yassine__ has joined #openstack-oslo12:02
*** yassine__ has quit IRC12:03
fhermenijaypipes: thanks. I will browse that. The last one especially might match a part of what we had in mind12:03
jaypipesk12:03
jaypipesI'll try to review the super scheduler BP later today.12:03
*** ihrachys has joined #openstack-oslo12:07
*** dtardivel has quit IRC12:08
*** salv-orl_ has quit IRC12:08
fhermenijaypipes: the last spec is interesting and the discussions especially. By the end, the problem being finding a concensus about a domain that is not finite. The term “distance” does not always mean the same for everyone, the same for affinity (server level, rack level, dc level, whatever level). The mix between hard and soft constraints12:10
*** shardy has quit IRC12:16
openstackgerritVipin Balachandran proposed openstack/osprofiler: Improve unit test coverage  https://review.openstack.org/29284912:18
*** MVenesio has joined #openstack-oslo12:19
*** MVenesio has quit IRC12:20
*** MVenesio has joined #openstack-oslo12:20
*** ayoung has joined #openstack-oslo12:24
*** dims has quit IRC12:25
ayoungdukhlov, sileht I'm trying to lock down the Message queueing between the controller and compute nodes in Nova.  To do this requires a regex matching the names of the queues or topics.  Do you know how I can tell what they are neamed?12:26
*** dims has joined #openstack-oslo12:26
ayoungdukhlov, sileht I made a copy of the set of queues and topics in my local install:  http://adam.younglogic.com/2016/03/what-can-talk-to-what-on-the-openstack-message-broker/12:26
ayoungdims, am I correct in understanding that the names of queues and topics used should be read out of the config file?12:29
dimsayoung : "reply_" + uuid is baked into oslo.messaging code, i think other names are picked up from process name etc i think12:31
silehtayoung, I confirm12:31
ayoungdims, ugh...that is going to be hard for ACLs12:32
*** rlrossit has joined #openstack-oslo12:32
*** edmondsw has joined #openstack-oslo12:32
ayoungdims, there is no way to write a regex that says "controler 14 can only read from x,y,z and and write to reply_<uuid> up front.  We'll have to update hte ACLs after the UUID is generated12:33
ayoungactaully, that might not be so bad...12:33
ayoungits the part where we keep anyone else from writing to it that is hard.12:33
silehtayoung, the uuid change each times the service is restarted12:33
dimsayoung : we need to see how it the ACL stuff should work with zmq / qpid+proton etc as well12:33
ayoungsileht, dims one at a time12:33
ayoungthey are not that different12:34
dims:)12:34
ayoungdims, haven't looked at zmq, but Proton is essentially the same:  regex12:34
ayoung12:34
ayoung ACL ALLOW alice CREATE QUEUE              # Grants alice permission to create all queues.12:34
ayoung      ACL DENY bob CREATE QUEUE name="myqueue"  # Denies bob permission to create a queue called "myqueue"12:34
ayoungbut I think it is a file that would need to be edited and regened12:35
ayoungsileht, which service UUID is used to generate the reply?12:35
ayounglets me make my question more focused12:35
ayounghere is what I want to do12:35
ayoungI want to make sure that only a specified hypervisor/compute node can talk on a specific channel with the controller12:36
*** yassine__ has joined #openstack-oslo12:37
ayoungwhen a controller sends a message to start a new virtual machine, where does it send it to?  And where do responses come back on?12:37
dimsdukhlov : ozamiatin_ : around? ^^12:38
ozamiatin_dims: yes12:38
dimsozamiatin_ : can you please explain the sequence of messages for a nova boot for example? (see questions from ayoung above)12:39
ayoungthanks dims12:39
ayoungdims, so rabbit has the concept of vhosts.  I wonder if we could make better use of that for isolation.  For example, createa vhost for compute node communication separate from all other, so that we don't have to worrry about, say cilometer or keystone having access to the compute node control messages12:42
*** shardy has joined #openstack-oslo12:44
silehtayoung, that should works, ceilometer is able to consume multiple vhosts, by settings multiple messaging_urls, that of course increase a lot the number of rabbit tcp connections opened by Ceilometer12:44
*** lucasagomes is now known as lucas-hungry12:45
ayoungsileht, good to know, although in this case, I would think that anything emitting to Ceilometer would be on one Vhost.  It would be the Nova control node (scheduler etc)that would have to talk to multiple...I think.12:45
ayoungWe really need a map of what is talking to what, and the expected names used12:45
silehtayoung, you can already separate rpc stuff from notification stuff12:46
ayoungsileht, excellent! link?12:46
*** shardy has quit IRC12:47
silehtayoung, you have to set transport_url into the notification group : https://github.com/openstack/oslo.messaging/blob/master/oslo_messaging/notify/notifier.py#L4012:47
*** AJaeger__ has left #openstack-oslo12:47
sileht[oslo_messaging_notifications] section, sorry12:47
*** shardy has joined #openstack-oslo12:48
silehtayoung, so you can have different vhosts or even different clusters for each part of openstack12:49
ayoungsileht, thanks...that should get me started.  I need to head in to the office.  Back on line in a bit12:49
*** gordc has joined #openstack-oslo12:50
*** amrith is now known as _amrith_12:50
ozamiatin_ayoung: dims: I think for zmq the task is a bit complicated, however we use UUID also to mark messages and theoretically can track where the message goes12:50
ozamiatin_ayoung: dims: not sure if it's possible to restrict "specified hypervisor/compute node can talk on a specific channel with the controller" currently in zmq driver12:52
*** deray has quit IRC12:53
ayoungDecided to stay here ...this is more important12:54
ayoungozamiatin_, I don't really care about ZMQ.12:56
dimsayoung : honestly, we'll need a whiteboard and a nova person :)12:56
ozamiatin_ayoung: ok12:56
ayoungozamiatin_, so long as what we are doing falls within "works for Rabbit, does not break ZMQ" we are good12:56
ayoungozamiatin_, here is the nightmare scenario: a hypervisor gets compromised. It now starts send out spurious messages, not only back to the controler, but to the other compute nodes, impersonating the controller, and takes down the whole cloud12:57
ayoungdims, I'd start with this image: http://solinea.com/wp-content/uploads/2013/06/openstack-arch-grizzly-logical-v2-1.jpg13:02
*** deray has joined #openstack-oslo13:02
openstackgerritOleksii Zamiatin proposed openstack/oslo.messaging: [zmq] Refactoring consumer side  https://review.openstack.org/28304913:03
*** yamamoto has quit IRC13:05
*** fhermeni has quit IRC13:07
*** yamahata has joined #openstack-oslo13:11
*** salv-orlando has joined #openstack-oslo13:16
*** ig0r__ has joined #openstack-oslo13:19
*** ig0r_ has quit IRC13:22
*** rlrossit has quit IRC13:25
*** ayoung has quit IRC13:26
*** dimtruck is now known as zz_dimtruck13:29
*** jpena is now known as jpena|lunch13:30
*** mriedem_away is now known as mriedem13:30
binarindims: I've looked into rabbitmq "x-expires", it works exactly as stated in documentation - documentation says "Unused means the queue has no consumers (so timer starts after creation and restarts again after last consumer disconnects), the queue has not been redeclared, and basic.get has not been invoked for a duration of at least the expiration period."13:35
binarindims: Expiry timer is not reset on sending message to the queue, and it's actually intended behaviour.13:36
*** jecarey has joined #openstack-oslo13:38
*** pradk has joined #openstack-oslo13:40
*** ayoung has joined #openstack-oslo13:40
*** jecarey has quit IRC13:41
*** jecarey has joined #openstack-oslo13:41
dimsbinarin : ack thanks for confirming, i'll have to log more information between the old and new sides of grenade to see if i can spot anything13:41
*** rlrossit has joined #openstack-oslo13:43
dimsbinarin : "rabbitmqctl report" should output all kinds of details right?13:43
*** sigmavirus24_awa is now known as sigmavirus2413:45
binarindims: Yes, it just agregates output of some other rabbitmqctl commands.13:45
dimsbinarin : is there anything else i can run in addition?13:47
*** lucas-hungry is now known as lucasagomes13:48
*** yamamoto has joined #openstack-oslo13:49
*** yamamoto_ has joined #openstack-oslo13:49
*** yamamoto has quit IRC13:53
binarindims: Do you suspect some rabbit bug related to queue expiration? Looks like it's not possible to find out exact amount of time left before queue expires, but I could write some `rabbitmqctl eval` spell that will tell whether timer is ticking and whether it has changed.13:56
dimsbinarin : thanks for that, will get some basic stuff into grenade and then we can revisit that13:57
*** Guest41345 has quit IRC14:03
ayoungsileht , dims when you said "reply_" + uuid is baked in to the code, and that the uuid changes when the server reboots...what server is that?14:11
ayoungand what is the UUID?  Is it just generated randomly each time the server boots?14:12
*** salv-orlando has quit IRC14:12
*** salv-orlando has joined #openstack-oslo14:13
*** jaosorior has joined #openstack-oslo14:15
*** boden has joined #openstack-oslo14:15
silehtayoung, exactly this is a queue with a random name created the first time a reply is needed (first time rpc.call/cast is called)14:15
ayoungsileht, to from an ACL perspective, we would have to modify the ACL for the end user after creating the queue in order to have any lock down?14:16
silehtayoung, yes14:17
ayoungsileht, is there any reason we couldn't do a pattern like <sourceuuid>-<desitinationuuid> instead of a random UUID?  It looks like we have about 300 chars to work with14:18
silehtayoung, we don't want to reuse queue when daemon are restarted, but having a better prefix can be a good enhancement14:19
ayoungsileht, because we want to make sure that old messages are discarded?14:20
silehtayoung, yes14:21
silehtI think their are other reason too, but I can't remember14:21
*** ihrachys has quit IRC14:25
*** ozamiatin has joined #openstack-oslo14:27
*** jpena|lunch is now known as jpena14:29
*** Guest91794 has quit IRC14:31
*** salv-orlando has quit IRC14:35
*** salv-orl_ has joined #openstack-oslo14:35
*** ihrachys has joined #openstack-oslo14:37
*** mtanino has joined #openstack-oslo14:41
*** yamahata has quit IRC14:44
*** zz_dimtruck is now known as dimtruck14:45
ayoungsileht, so this code here:  http://git.openstack.org/cgit/openstack/oslo.messaging/tree/oslo_messaging/_drivers/amqpdriver.py#n384  is the culprit?14:47
*** regXboi has joined #openstack-oslo14:47
ayoungand we could make that14:48
ayoungreply_q = 'reply_%s_%s' %(target.id , uuid.uuid4().hex)14:49
*** ozamiatin__ has joined #openstack-oslo14:52
*** deray has quit IRC14:54
silehtayoung, yes, but changing that need to ensure that we don't break rolling upgrade when only part of nova are updated14:56
ayoungsileht, wouldn't that change only be in effect on the calling server?14:56
ayoungthe reply name goes in the message I though?14:57
ayoungsileht, http://git.openstack.org/cgit/openstack/oslo.messaging/tree/oslo_messaging/_drivers/amqpdriver.py#n41414:57
silehtayoung, hum yep, only reply_ is hardcoded on the other side14:57
-openstackstatus- NOTICE: Launchpad OpenID SSO is currently experiencing issues preventing login. The Launchpad team is working on the issue14:58
*** ChanServ changes topic to "Launchpad OpenID SSO is currently experiencing issues preventing login. The Launchpad team is working on the issue"14:58
*** links has quit IRC14:59
*** jaosorior has quit IRC15:01
*** tesseract has joined #openstack-oslo15:06
*** tesseract is now known as Guest7146715:07
*** _amrith_ is now known as amrith15:19
*** yamahata has joined #openstack-oslo15:23
*** salv-orl_ has quit IRC15:27
*** ChanServ changes topic to "#openstack-oslo"15:31
-openstackstatus- NOTICE: Launchpad SSO is back to normal - happy hacking15:31
*** yamahata has quit IRC15:45
*** binarin has quit IRC15:46
*** amrith is now known as _amrith_15:49
*** itisha has quit IRC15:49
*** salv-orlando has joined #openstack-oslo15:57
*** nkrinner_ has quit IRC15:57
*** bogdando has quit IRC15:58
*** links has joined #openstack-oslo15:58
*** suro-patz has joined #openstack-oslo16:03
*** _amrith_ is now known as amrith16:04
*** bogdando has joined #openstack-oslo16:10
*** browne has joined #openstack-oslo16:14
*** amrith is now known as _amrith_16:19
*** permalac has joined #openstack-oslo16:22
*** permalac has quit IRC16:22
*** ozamiatin has quit IRC16:23
*** permalac has joined #openstack-oslo16:23
*** permalac has quit IRC16:23
*** permalac_ has joined #openstack-oslo16:31
*** permalac_ has quit IRC16:31
*** rcernin has quit IRC16:35
*** Guest71467 has quit IRC16:42
*** links has quit IRC16:47
*** suro-patz has quit IRC16:50
*** suro-patz has joined #openstack-oslo16:52
openstackgerritJoshua Harlow proposed openstack/taskflow: Add periodic jobboard refreshing (incase of sync issues)  https://review.openstack.org/29304116:52
dimsozamiatin_ : still around? i saw some traffic on performance channel about what you are planning for reducing # of socket connections, can you give us a summary?16:54
*** e0ne has quit IRC16:58
*** suro-patz has quit IRC17:00
dimsdmitryme : am able to capture "rabbitmq_ctl list_queues" and "rabbitmq_ctl report" http://logs.openstack.org/03/292903/1/check/gate-grenade-dsvm/cc4b94f/logs/grenade.sh.txt.gz#_2016-03-15_14_45_55_20917:13
dimsdmitryme : is there a way for me to log the params we used to create the queues?17:13
openstackgerritJoshua Harlow proposed openstack/taskflow: Add periodic jobboard refreshing (incase of sync issues)  https://review.openstack.org/29304117:14
*** openstackgerrit has quit IRC17:14
*** openstackgerrit_ is now known as openstackgerrit17:14
*** openstackgerrit has quit IRC17:15
dmitrymedims: are you interested in x-expires arg?17:15
*** openstackgerrit_ has joined #openstack-oslo17:15
dimsdmitryme : y17:15
dimsthe way things are setup in grenade, i can't instrument oslo.messaging, so i have to print what i can after the old nova is shut down and new nova starts17:15
dimsdmitryme : ^^17:15
*** openstackgerrit_ is now known as openstackgerrit17:16
dmitrymedims: then you should run something like17:16
dmitrymerabbitmqctl list_queues messages consumers arguments name17:16
*** openstackgerrit_ has joined #openstack-oslo17:16
dmitrymeif you run just ‘rabbitmqctl list_queues’, it is equivalent of ‘rabbitmqctl list_queues messages name’17:17
dimsdmitryme : what's the magic to print everything17:18
dimspossible?17:18
dmitrymeactually all options are provided in ‘man rabbitmqctl’, section list_queues17:18
dmitrymeand the list is pretty long17:18
dmitrymeI am not sure you want to get them all17:19
dimsduh. thanks :)17:19
dmitrymeit will be a mess :-)17:19
dmitryme‘arguments’ will include x-expires, if it was provided17:19
dimsgood point, will try that thanks17:19
openstackgerritVipin Balachandran proposed openstack/osprofiler: Add exception type to stop trace info  https://review.openstack.org/29305817:23
*** suro-patz has joined #openstack-oslo17:24
*** pcaruana has quit IRC17:24
*** fhermeni has joined #openstack-oslo17:26
*** fhermeni has quit IRC17:28
*** fhermeni has joined #openstack-oslo17:29
*** yassine__ has quit IRC17:30
*** pcaruana has joined #openstack-oslo17:36
*** fhermeni has left #openstack-oslo17:36
*** rlrossit has quit IRC17:40
*** mriedem1 has joined #openstack-oslo17:43
*** mriedem has quit IRC17:44
*** rlrossit_ has joined #openstack-oslo17:45
*** mriedem1 has quit IRC17:46
*** jpena is now known as jpena|off17:47
*** fhermeni has joined #openstack-oslo17:48
*** sputnik13 has joined #openstack-oslo17:48
*** fhermeni has quit IRC17:49
*** mhickey has quit IRC17:50
*** mriedem has joined #openstack-oslo17:53
*** salv-orl_ has joined #openstack-oslo17:54
openstackgerritJoshua Harlow proposed openstack/taskflow: Ensure the fetching jobs does not fetch anything when in bad state  https://review.openstack.org/29307517:54
*** salv-orlando has quit IRC17:56
*** ihrachys has quit IRC17:58
*** dims_ has joined #openstack-oslo18:03
*** dims has quit IRC18:03
*** lucasagomes is now known as lucas-dinner18:10
*** vilobhmm11 has joined #openstack-oslo18:20
harlowja_at_homedims_, what do u eventually want to do with https://etherpad.openstack.org/p/mitaka-oslo-achievements ?18:27
harlowja_at_homeML post?18:27
*** e0ne has joined #openstack-oslo18:28
*** dims_ has quit IRC18:29
harlowja_at_homecdent, chrisshoe, inc lol18:30
harlowja_at_homei would of thought they would of made hoes18:31
harlowja_at_homeoh nm, i read it right now18:31
harlowja_at_homelol18:31
cdentComfy Shoes for Comfy People™18:31
harlowja_at_homeya, i see the shoe part now18:31
harlowja_at_homejust saw hoe part, lol18:31
cdentbut yeah, I realized after the mail went out how close to hoes it was18:31
harlowja_at_homelol18:31
cdentso headdesked a little bit18:31
harlowja_at_home;)18:31
cdentoh look chris is talking about whores again18:31
harlowja_at_homelol18:32
harlowja_at_homechris and his hoes18:32
harlowja_at_homelol18:32
harlowja_at_home+2 though for the shoe issue18:34
harlowja_at_homeit is an issue18:34
cdentyeah, I think activity is not quite corrupt, but corrupted, confused by the way things work18:34
harlowja_at_homeya18:35
harlowja_at_homeprocess process and process18:36
harlowja_at_homeand shoes18:36
cdentso many shoes18:36
harlowja_at_home:-P18:36
harlowja_at_homeit reminds me of the broken windows problem18:37
harlowja_at_homeand how hard it is to get out of that problem once its begun18:37
harlowja_at_homehttps://en.wikipedia.org/wiki/Broken_windows_theory18:37
harlowja_at_hometheory not problem18:37
cdentyeah, very familiar with that theory18:37
harlowja_at_homewhere criminal vandialism == bugs, lol18:38
harlowja_at_home(and process)18:38
harlowja_at_home'The car in the Bronx was attacked within minutes of its abandonment. Zimbardo noted that the first "vandals" to arrive were a family – a father, mother and a young son – who removed the radiator and battery. '18:38
harlowja_at_homewoah18:38
harlowja_at_homewho brings there young son to do this18:38
harlowja_at_homelol18:38
cdentfamily business18:39
harlowja_at_homelol18:39
cdentchrisshoe started out stripping cars but went straight18:40
harlowja_at_homelol18:41
harlowja_at_homegood good18:42
cdentthat nearly 10 of those "lol" in the last couple minutes, you might want to take a break before you turn into a tween18:42
harlowja_at_homereminds me of SNL get straight skits18:42
harlowja_at_homelol18:42
harlowja_at_home*scared straight18:43
harlowja_at_homehttp://www.nbc.com/saturday-night-live/video/scared-straight/n12247 (and others)18:43
harlowja_at_homechrisshoe must of been in one of those sessions18:43
harlowja_at_homelol18:43
cdentheh18:44
cdentI must go put away the neighbors chickens18:44
harlowja_at_homek18:44
harlowja_at_homechrisshoe steal them again?18:44
harlowja_at_homelol18:44
cdentno that corporate behemoth Hbiranhat did18:44
harlowja_at_homelol18:46
*** _amrith_ is now known as amrith18:49
*** e0ne_ has joined #openstack-oslo18:50
*** e0ne has quit IRC18:51
*** alaski has joined #openstack-oslo18:52
alaskihello oslo-ites, I'd like to run an idea past y'all and get some feedback because I'm hitting my head against the oslo.messaging api currently18:54
alaskiI'm trying to proof out an endpoint wrapper that would receive messages immediately and queue them locally for dispatch18:55
alaskiI've been approaching this from outside of oslo.messaging, but calls and the need for an immediate reply are foiling me18:56
alaskiand afaict there's no way to explicitly handle a message reply at a later time from outside of oslo.messaging18:56
alaskiI've considered moving this functionality into oslo.messaging but that seems to require creating a new driver which means choosing the transport mechanism even though these are separate concerns18:58
cdentsileht: this ^ sounds like your beat18:58
openstackgerritJoshua Harlow proposed openstack/taskflow: Ensure the fetching jobs does not fetch anything when in bad state  https://review.openstack.org/29307518:59
harlowja_at_homealaski, hmmm18:59
harlowja_at_homelet me think if such a thing is even possible19:00
harlowja_at_homenot impossible, just didn't think we had the semantics to do that19:00
cdentalaski: I'm guessing a cast, with the caster maintaining some state (I want a response eventually) is not going to work out?19:00
harlowja_at_homesounds like a future object, lol19:00
harlowja_at_homecall_future19:00
harlowja_at_homemaybe new method named that19:00
harlowja_at_homevs call_immediate (the current stuff)19:01
alaskicdent: in theory I think it would work, where I'm stuck is how to send the response back to the right caller19:01
harlowja_at_home* https://docs.python.org/3/library/concurrent.futures.html#future-objects (future object)19:01
alaskiharlowja_at_home: right, the semantics seem to be my issue. functionally it should be doable it just doesn't currently seem to be possible19:02
cdentthe way we do messaging is so weird to me, I still haven't managed to get my head around it. I keep wanting it to be more broadcast-y19:02
alaskicdent: I've spent two days looking at this now and my head still isn't fully around it :)19:02
harlowja_at_homelol19:03
*** bnemec has quit IRC19:03
alaskihttp://paste.openstack.org/show/490585/ is basically what I'm attempting to do19:04
alaskiignoring line 30 which is there just for testing atm19:04
harlowja_at_homeright19:04
harlowja_at_homethat looks like a https://docs.python.org/3/library/concurrent.futures.html#executor-objects from my view19:05
harlowja_at_homewhich a size limit on it19:05
harlowja_at_homeexecutors also internally queue19:05
* harlowja_at_home wishes the oslo.messaging was more like an executor, lol19:06
harlowja_at_homethe stuff i have in taskflow is, perhaps can work with sileht to make the oslo.messaging stuff more like just another executor19:06
alaskiinteresting, oslo.messaging is using an executor internally. I wonder if I could make a new one that accomplishes what I want19:07
alaskithough really I want this outside of oslo.messaging, because I'm trying to have a priority queue and the priorities would come from the consuming service19:08
harlowja_at_homeright, the executor used internally though i think is just for internal dispatchinb19:09
harlowja_at_homei'd more like something like19:09
harlowja_at_home>>> remote_executor = oslo.message.get_exector(endpoints)19:09
harlowja_at_home>>> remote_executor.submit(an_endpoint.thing, args, kwargs)19:09
harlowja_at_homeor submit_cast (new method)19:10
harlowja_at_homeboth return a future object19:10
silehtharlowja_at_home, that looks so pythonic :)19:10
harlowja_at_homeso that handles the client side :-P19:10
harlowja_at_homethe worker/reciever side would need to do things differently19:10
harlowja_at_homebut ya, taaskflow has something like this in http://docs.openstack.org/developer/taskflow/workers.html#high-level-architecture19:11
harlowja_at_home*sorta like this19:11
harlowja_at_homealthough alaski i think u can just get away with a customized https://github.com/openstack/oslo.messaging/blob/master/oslo_messaging/rpc/dispatcher.py#L79 (one of these)19:11
harlowja_at_homebut sileht might say differently, lol19:12
*** dims has joined #openstack-oslo19:12
silehtharlowja_at_home, when did you plan to drop oslo.msg for taskflow ?19:12
alaskiheh, that's why I'm here :)19:12
harlowja_at_homesileht, lol, ummm19:12
harlowja_at_homeshhh19:12
harlowja_at_homelol19:12
alaskiharlowja_at_home: replacing the dispatcher does seem to be the level I'd need to integrate at right now19:12
harlowja_at_homeright19:12
harlowja_at_homereplace https://github.com/openstack/oslo.messaging/blob/master/oslo_messaging/rpc/dispatcher.py#L13519:13
harlowja_at_homewith something that takes a long time, lol19:13
*** bnemec has joined #openstack-oslo19:13
harlowja_at_homeand/or queues19:13
harlowja_at_homeperhaps queues dispatch into a executor object19:13
harlowja_at_home*submits ---> executor object (via executor submit method)19:13
harlowja_at_homemay or may not work, ha, just an idea19:13
harlowja_at_homelol19:14
dimsjust like wsgi19:14
silehtbbl19:14
*** krotscheck is now known as kro_death_by_sin19:14
alaskiI really need to look into this executor stuff, I don't really grok it yet19:14
harlowja_at_homewsgi may or may not work? :-P19:14
*** kro_death_by_sin is now known as kro_sinus_hell19:14
harlowja_at_homealaski, sure, i can help understand, its a good pattern to understand19:15
harlowja_at_homesince most lanagues are picking up that style19:15
harlowja_at_homejavascript, c++ and others (perhaps called promises)19:15
harlowja_at_homethe python version resembles the java version19:15
harlowja_at_homeanyway19:15
alaskigotcha, I understand the basic concept of promises/futures but have never looked behind the curtain19:16
harlowja_at_homek19:16
alaskiharlowja_at_home: thanks, I'll do some reading and playing around and get back with questions19:16
harlowja_at_homecool19:16
harlowja_at_homei'll bbl19:16
*** bnemec has quit IRC19:16
*** bnemec has joined #openstack-oslo19:18
*** dims has quit IRC19:18
*** e0ne_ has quit IRC19:19
*** ozamiatin__ has quit IRC19:20
*** dims has joined #openstack-oslo19:26
*** salv-orl_ has quit IRC19:43
*** dims has quit IRC19:45
*** ig0r_ has joined #openstack-oslo20:04
*** ig0r__ has quit IRC20:06
*** fhermeni has joined #openstack-oslo20:07
fhermenihi20:07
*** dims has joined #openstack-oslo20:11
*** kgiusti has left #openstack-oslo20:16
fhermeniharlowja_at_home: ping ?20:27
*** dimtruck is now known as zz_dimtruck20:30
*** ozamiatin has joined #openstack-oslo20:31
*** regXboi has quit IRC20:34
*** gordc has quit IRC20:34
*** e0ne has joined #openstack-oslo20:37
*** salv-orlando has joined #openstack-oslo20:37
*** rlrossit_ has quit IRC20:37
*** rlrossit_ has joined #openstack-oslo20:41
*** zz_dimtruck is now known as dimtruck20:44
*** ihrachys has joined #openstack-oslo20:50
*** SamYaple is now known as NotSamYaple20:52
*** NotSamYaple is now known as SamYaple20:53
*** ozamiatin has quit IRC21:01
*** gordc has joined #openstack-oslo21:03
*** sdague has quit IRC21:09
*** shardy has quit IRC21:10
*** yamahata has joined #openstack-oslo21:22
*** rlrossit_ has quit IRC21:28
*** MVenesio has quit IRC21:30
*** amrith has quit IRC21:35
*** amrith has joined #openstack-oslo21:35
*** pcaruana has quit IRC21:38
*** pradk has quit IRC21:38
*** e0ne has quit IRC21:40
*** pcaruana has joined #openstack-oslo21:52
*** mriedem has quit IRC21:55
*** yamahata has quit IRC22:00
*** harlowja has joined #openstack-oslo22:01
*** kbyrne has quit IRC22:01
*** jecarey has quit IRC22:03
*** fhermeni has quit IRC22:04
*** ozamiatin has joined #openstack-oslo22:11
*** dims_ has joined #openstack-oslo22:11
*** sigmavirus24 is now known as sigmavirus24_awa22:12
*** david-lyle has quit IRC22:13
*** dims has quit IRC22:14
*** gordc has quit IRC22:15
*** cdent has quit IRC22:15
*** dims has joined #openstack-oslo22:16
*** dims_ has quit IRC22:17
*** jecarey has joined #openstack-oslo22:17
*** dimtruck is now known as zz_dimtruck22:18
*** jecarey has quit IRC22:22
*** boden has quit IRC22:23
*** ozamiatin has quit IRC22:30
*** shakamunyi has joined #openstack-oslo22:32
*** edmondsw has quit IRC22:44
openstackgerritJoshua Harlow proposed openstack/taskflow: Ensure the fetching jobs does not fetch anything when in bad state  https://review.openstack.org/29307522:51
*** yamamoto_ has quit IRC23:05
*** dims has quit IRC23:06
openstackgerritJoshua Harlow proposed openstack/taskflow: Ensure the fetching jobs does not fetch anything when in bad state  https://review.openstack.org/29307523:29
*** MVenesio has joined #openstack-oslo23:31
*** MVenesio has quit IRC23:36
*** ihrachys has quit IRC23:45
*** harlowja has quit IRC23:47
*** salv-orlando has quit IRC23:49
*** yamamoto_ has joined #openstack-oslo23:51
« Monday, 2016-03-14 Index Wednesday, 2016-03-16 »

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