Wednesday, 2015-01-14

*** barclaac|2 has joined #openstack-lbaas00:04
*** chlong has joined #openstack-lbaas00:05
*** dougw2 has quit IRC00:08
*** barclaac has quit IRC00:09
*** ctracey has joined #openstack-lbaas00:23
*** markmcclain has joined #openstack-lbaas00:26
*** markmcclain has quit IRC00:27
*** markmcclain1 has joined #openstack-lbaas00:27
*** markmcclain1 has quit IRC00:29
*** markmcclain has joined #openstack-lbaas00:29
*** codekobe has joined #openstack-lbaas00:33
*** markmcclain has quit IRC00:48
*** dougwig has quit IRC00:54
*** markmcclain has joined #openstack-lbaas00:57
*** sballe has joined #openstack-lbaas00:58
*** markmcclain has quit IRC00:58
*** dougwig has joined #openstack-lbaas00:58
*** markmcclain has joined #openstack-lbaas00:58
*** markmcclain has quit IRC01:00
*** sbfox has quit IRC01:04
*** xgerman has quit IRC01:11
blogandouwgig: we'd need our own octavia-db-manage wrapper which is a low priority01:19
*** mlavalle has quit IRC01:27
*** pattabi has joined #openstack-lbaas01:34
pattabihi all01:34
pattabii am getting an error as part of update_status in the (when the driver raises an exception)01:35
*** codekobe has quit IRC01:35
pattabi2015-01-13 20:19:41.097 TRACE neutron.api.v2.resource     self.db.update_status(context, db_entity.__class__._SA_MODEL,01:36
pattabi2015-01-13 20:19:41.097 TRACE neutron.api.v2.resource AttributeError: type object 'LoadBalancer' has no attribute '_SA_MODEL'01:36
pattabiany pointers ?01:36
*** codekobe has joined #openstack-lbaas01:36
*** markmcclain has joined #openstack-lbaas01:37
*** markmcclain has quit IRC01:38
*** markmcclain has joined #openstack-lbaas01:38
*** markmcclain has quit IRC01:59
rm_workdougwig: you're the one that asked for exponential :P02:22
*** sballe_ has joined #openstack-lbaas02:23
*** sballe has quit IRC02:23
*** sballe_ is now known as sballe02:23
*** markmcclain has joined #openstack-lbaas02:42
*** markmcclain has quit IRC02:42
*** markmcclain has joined #openstack-lbaas02:42
*** rm_you| has joined #openstack-lbaas02:45
*** woodster_ has quit IRC02:48
*** rm_you has quit IRC02:48
*** rm_work has quit IRC02:48
*** markmcclain has quit IRC02:49
*** markmcclain has joined #openstack-lbaas02:49
*** sballe has quit IRC02:51
*** ctracey has quit IRC02:52
*** woodster_ has joined #openstack-lbaas02:53
*** sballe_ has joined #openstack-lbaas02:53
*** rm_work has joined #openstack-lbaas02:53
*** ctracey has joined #openstack-lbaas02:58
*** codekobe has quit IRC03:08
*** codekobe has joined #openstack-lbaas03:10
*** markmcclain has quit IRC03:28
*** amotoki has joined #openstack-lbaas03:53
*** SantoshNetSacler has joined #openstack-lbaas04:01
*** markmcclain has joined #openstack-lbaas04:09
*** markmcclain has quit IRC04:10
*** markmcclain1 has joined #openstack-lbaas04:10
*** SantoshNetSacler has quit IRC04:18
*** sbalukoff has quit IRC04:23
*** sbalukoff has joined #openstack-lbaas04:34
*** hichtakk has quit IRC04:38
*** hichtakk has joined #openstack-lbaas04:39
*** hichtakk has quit IRC04:39
*** hichtakk has joined #openstack-lbaas04:39
*** hichtakk has quit IRC04:40
*** hichtakk has joined #openstack-lbaas04:41
*** hichtakk has quit IRC04:41
*** hichtakk has joined #openstack-lbaas04:42
*** hichtakk has quit IRC04:42
bloganpattabi: have you pulled the neutron-lbaas code down recently?04:42
*** hichtakk has joined #openstack-lbaas04:42
*** hichtakk has quit IRC04:42
*** hichtakk has joined #openstack-lbaas04:43
*** markmcclain1 has quit IRC04:45
*** SantoshNetScaler has joined #openstack-lbaas04:47
openstackgerritMerged stackforge/octavia: Fixed my alembic version fail
*** hichtakk has quit IRC05:23
*** hichtakk has joined #openstack-lbaas05:23
*** hichtakk has quit IRC05:23
*** hichtakk has joined #openstack-lbaas05:24
*** hichtakk has quit IRC05:24
*** hichtakk_ has joined #openstack-lbaas05:25
*** hichtak__ has joined #openstack-lbaas05:26
*** hichtakk_ has quit IRC05:26
*** hichtak__ has quit IRC05:41
*** ajmiller has quit IRC05:56
openstackgerritPattabi Ayyasami proposed openstack/neutron-lbaas: Brocade Driver for lbaas v2 data model
*** woodster_ has quit IRC06:20
*** pattabi has quit IRC06:52
openstackgerritMerged stackforge/octavia: Nova driver implementation
*** kmurari_ has joined #openstack-lbaas06:59
*** enikanorov__ has quit IRC07:08
*** apuimedo has joined #openstack-lbaas07:14
*** kongfy has joined #openstack-lbaas07:15
*** Miouge has joined #openstack-lbaas07:38
*** kongfy has quit IRC07:45
*** chlong has quit IRC07:47
*** kobis has joined #openstack-lbaas07:49
*** kmurari_ has quit IRC07:50
openstackgerritMerged stackforge/octavia: Common TLS utilies
openstackgerritMerged stackforge/octavia: Updates the controller spec to clarify API Manager
sbalukoffDon't mind me and my midnight merges...08:04
openstackgerritMerged stackforge/octavia: Removes flows from the amphora-driver-interface
*** Miouge has quit IRC08:11
*** hichtakk has joined #openstack-lbaas08:25
*** hichtakk has quit IRC08:25
*** jschwarz has joined #openstack-lbaas08:40
*** Miouge has joined #openstack-lbaas08:40
*** sbalukoff has quit IRC09:00
*** jschwarz has quit IRC09:08
*** Miouge has quit IRC09:10
*** jschwarz has joined #openstack-lbaas09:14
*** sbalukoff has joined #openstack-lbaas09:15
*** Miouge has joined #openstack-lbaas09:15
*** Miouge has quit IRC09:32
*** Miouge has joined #openstack-lbaas09:33
*** jschwarz has quit IRC09:49
*** jschwarz has joined #openstack-lbaas10:05
*** SantoshNetScaler has quit IRC10:31
*** jschwarz has quit IRC10:31
*** chlong has joined #openstack-lbaas12:06
*** enikanorov_ has joined #openstack-lbaas12:43
*** woodster_ has joined #openstack-lbaas13:00
*** chlong has quit IRC13:02
*** amotoki_ has joined #openstack-lbaas13:07
*** mestery has quit IRC13:33
dougwigsbalukoff: lol13:52
*** mestery has joined #openstack-lbaas14:02
openstackgerritMerged stackforge/octavia: Change back the amhora driver interface
*** mestery has quit IRC14:39
*** intr1nsic has joined #openstack-lbaas14:54
*** amotoki_ has quit IRC15:12
*** amotoki_ has joined #openstack-lbaas15:18
*** mestery has joined #openstack-lbaas15:21
*** Miouge has quit IRC15:22
*** apuimedo has quit IRC15:25
*** TrevorV_ has joined #openstack-lbaas15:26
*** TrevorV_ has quit IRC15:27
*** Miouge has joined #openstack-lbaas15:36
*** amotoki_ has quit IRC15:39
*** xgerman has joined #openstack-lbaas15:58
*** evgenyf has joined #openstack-lbaas16:02
*** jorgem has joined #openstack-lbaas16:24
*** sbalukoff has quit IRC16:30
*** mlavalle has joined #openstack-lbaas16:39
dougwiga2hill: morning16:39
*** evgenyf has quit IRC16:41
rm_workdougwig / sbalukoff: you guys are such pansies16:44
dougwigrm_work: what did i do this time?16:44
dougwigoh, your crazy backoff from hell?16:45
rm_workcan't even handle a 17-year backoff16:45
dougwigaka, "let's wait until technology improves to complete"16:45
dougwigflight wifi about to get turned off.16:45
rm_workrofl you're on a plane again16:46
rm_workdougwig: do you like steven's suggestion?16:46
rm_workI can do that, but *someone* wanted exponential :P16:46
openstackgerritIhar Hrachyshka proposed openstack/neutron-lbaas: Migrate to oslo.concurrency
openstackgerritIhar Hrachyshka proposed openstack/neutron-lbaas: oslo: migrate to namespace-less import paths
*** Guest8210 is now known as redrobot16:54
openstackgerritIhar Hrachyshka proposed openstack/neutron-lbaas: oslo: migrate to namespace-less import paths
openstackgerritIhar Hrachyshka proposed openstack/neutron-lbaas: Migrate to oslo.concurrency
xgermanrm_work, what did Stephen suggest this time?17:01
xgerman(still trying to wake up)17:01
rm_workjust a shorter timeout algorithm17:02
rm_workfloat((attempts + random.random()) * attempts)17:02
rm_workno one likes float(1 << attempts) + random.random() * attempts17:03
rm_workjust because it maxes at 17 years <_<17:03
rm_workif we took it down to 10 attempts it'd max at 17 minutes17:03
xgermanwhy do we need random?17:03
rm_workso they dont synchronize their timeouts17:04
rm_workhaving a backoff to avoid constantly slamming the server doesn't help if you sync up so you slam the server again the next try17:04
xgermanbackoff is usually exponential17:05
rm_workwhich is what I ahve17:05
rm_workbut dougwig and sbalukoff are pansies and can't handle the sheer power of the exponential backoff17:05
xgermanlet me chime in17:06
rm_workto be fair, 17 years is ridiculous17:06
rm_workbut i think the answer is just limiting it to 10 tries17:06
rm_worknot changing the algorithm to be non-exponential17:06
xgermanyep, that makes sense17:07
xgermanalso if you are concerned about server load you might use a cirucit breaker17:08
xgermanto short circuit requests from other threads17:08
rm_worktrying to keep this simple for now tho17:08
xgermanwell, there probbaly is some oslo library and if not there should be :-)17:09
*** sbfox has joined #openstack-lbaas17:30
dougwig10 tries would be fine.17:39
*** jorgem has quit IRC17:42
dougwigalternately, cap it at a minute at which point it will have been unserialized by the nature of the first few.17:42
*** kbyrne has quit IRC17:56
*** sbalukoff has joined #openstack-lbaas18:03
sbalukoffHowdy, y'all!18:05
*** kobis has quit IRC18:06
rm_workdougwig: yeah those were my two suggestions, I figure I'll do the 10-tries18:08
*** chlong has joined #openstack-lbaas18:10
rm_workxgerman / dougwig / sbalukoff: is 17-minute total backoffs good with you? longest single would be around 8 minutes18:34
rm_workthat seems reasonable for a network outage situation18:34
openstackgerritAdam Harwell proposed stackforge/octavia: Update certificate generator implementations
rm_worksure is a nice exponential bikeshed we've got here18:36
sbalukoffrm_work: Heh! Exactly.18:37
sbalukoffAnyway, y'all, please don't forget to fill out the weekly standup etherpad for this week:
sbalukoffI'm putting together the agenda for today's meeting. Please let me know if there's something specific y'all want to discuss as a group.18:37
sbalukoff(Otherwise, I hope to keep the meeting pretty short so we can quickly get back to writing and/or reviewing code.)18:38
sbalukoffI'm totally merging that.18:41
xgermanI didn't expect anything less from you :-)18:41
sbalukoffWhat can I say? I'm a jerk.18:43
*** apuimedo has joined #openstack-lbaas19:01
openstackgerritMerged stackforge/octavia: Update certificate generator implementations
*** jorgem has joined #openstack-lbaas19:11
blogangood afternoon gents and ladies19:14
sbalukoffGreetings, blogan!19:18
blogangreetings alien overlord!19:18
*** crc32 has joined #openstack-lbaas19:19
jorgemsbalukoff: Hey do you know if the LBaaS community was planning on any talks for the summit?19:22
sbalukoffIt's a matter of getting them in.19:23
xgermanthat's what I am hoping for19:23
sbalukoffI don't think we've formally discussed any here just yet though.19:23
jorgemGotcha, it would be nice to add that to the agenda for this week19:23
jorgemwant me to add it?19:23
sbalukoffGo for it, eh!19:23
xgermanyeah, my attempts to get ATC by reordering import statements fell flat so I need to give a talk :-)19:27
blogani feel like a flood of reviews have been merged in the last 2 weeks, its awesome!19:35
xgermanyep, way to go19:36
xgermanwe should probably aim for a panel talk Octavia - state of the union19:37
*** jorgem has quit IRC19:41
bedismorning :p19:43
*** sbfox has quit IRC19:45
bedisis there a URL for san antonio hackathon ?19:47
*** sbfox has joined #openstack-lbaas19:47
bloganwrong one19:48
*** jorgem has joined #openstack-lbaas19:50
*** ajmiller has joined #openstack-lbaas19:51
*** rm_mobile has joined #openstack-lbaas20:03
*** rm_mobile has joined #openstack-lbaas20:03
rm_workxgerman: yeah, I'm down20:20
rm_workbut only if going to Vancouver doesn't interfere with me getting selected for Tokyo20:20
*** apuimedo has quit IRC20:26
*** ptoohill has joined #openstack-lbaas20:28
*** ptoohill has quit IRC20:31
openstackgerritTrevor Vardeman proposed stackforge/octavia: haproxy reference amphora API client
bedisin which hotel are you going to go in San Antonio ?20:52
blogani have no idea which ones they are staying, just the ones listed, but you can obviously choose any that you want20:53
bloganthe ones by the airport aren't close to the tourist areas20:53
sbalukoffbedis: Not sure yet either. There's a person here who is supposed to be coordinating with me on that, but I've not yet heard back.20:53
*** markmcclain has joined #openstack-lbaas20:54
bedissbalukoff: once you know, please let me know20:57
bedisotherwise I'll choose a hilton group on20:57
bedisI'm an HHonors :)20:57
johnsomWe are at the Hilton Garden Inn.  It was on our approved list and we are HHonors too20:59
a2hillbedis, the Velencia is at the heart of downtown and is nice21:00
bedisthx, I'll check these ones21:01
bedisI'll also try to buy tickets for a game of the Spurs21:02
*** jorgem has quit IRC21:03
blogani hope they're doing better when that time comes around21:04
rm_workthere are motels like, right next to Castle, that are super cheap :P21:15
rm_workjust sayin21:15
rm_workif that had been an option in Seattle, I might have taken it21:16
*** jorgem has joined #openstack-lbaas21:17
dougwigi haven't booked a hotel yet.21:18
rm_workI thought you weren't actually going to make it?21:24
rm_workor was it that you'd miss a day or something21:24
a2hillYea, I only suggest Velencia because its downtown and its nice. But, the trip to the castle can be a pain in if staying down there, but worth it if you want to see what downtown SA has to offer21:32
a2hillCould even potentially set up a carpool. I would be willing to offer my services here as downtown is a skip and jump from my place21:33
*** TrevorV_ has joined #openstack-lbaas21:35
*** markmcclain has quit IRC21:44
*** chlong has quit IRC21:47
*** apuimedo has joined #openstack-lbaas21:53
dougwigrm_work: i'm there sunday - friday21:56
dougwigit was seattle where i couldn't make all of it21:56
xgermanyou missed out the best parts - like our world of beer outing ;-)22:02
dougwigi'm still upset that i missed out on bbq for the last excursion to texas.22:14
dougwighilton garden airport?22:16
dougwigjohnsom and hp? ^^22:17
dougwighilton garden airport?22:20
johnsomYeah, the one by the airport.  It's ~10 miles from the Rack site22:20
johnsom12828 US HIGHWAY 281 NORTH SAN ANTONIO TX 78216 US22:21
dougwigi'll book the same.22:21
dougwigi'm testing if there's a speaker limit. who all were the six? (i've got brandon, stephen, me).22:21
sbalukoffgerman and johnsom22:22
johnsomgerman and I22:22
sbalukoffAnd I think jorgem wanted to talk as well.22:22
dougwigjohnsom, what email do you use for openstack/22:23
johnsomNo, thank you...22:25
TrevorV_dougwig, can I bother you a minute22:43
dougwigTrevorV_: shoot22:44
TrevorV_I'm trying to add a package to the requirements file, and its failing22:44
TrevorV_I mean tox -r is failing22:44
TrevorV_Saying it can't install it22:44
TrevorV_Actually, nvmd22:45
TrevorV_It was a duplicate22:45
TrevorV_My fault :P22:45
dougwiguhh, happy to "help". :)22:46
*** markmcclain has joined #openstack-lbaas22:48
*** markmcclain has quit IRC22:52
*** markmcclain has joined #openstack-lbaas22:52
*** apuimedo has quit IRC22:54
*** markmcclain has quit IRC22:54
*** markmcclain1 has joined #openstack-lbaas22:54
*** markmcclain1 has quit IRC22:58
a2hillI want to :(23:09
jorgemxgerman: Hey German23:11
jorgemxgerman: Just chatted with mwang223:12
jorgemxgerman: Do you mind if I take on the API Manger blueprint? I'm trying to do something relatively easy as I'm pretty new to Python and would like to start participating in terms of code23:12
jorgemAnything else seems a little too much to start off with23:13
jorgemJava is my fortay so I just want to get my feet wet23:13
xgermanyeah, I wish we could just write the whole thing in Scala ;-)23:14
jorgemso is that okay? blogan says I can't code anymore and I want to prove him wrong!23:14
xgermansure -- our plan was to sort of fuse that with the flows I am working on but I think they will end in the deploy worker anyway23:15
jorgemxgerman: Yeah I was thinking so too. cool then. I'll start writing a spec tomorrow on this. thx!23:16
xgermanawesome. we just need to coordinate on how you fork that worker ;-)23:16
johnsomYes, that is the important hand-off23:16
jorgemsweet deal. I'll chat with you tomorrow on that23:16
jorgemboth of you23:16
jorgemSo, do I claim the blueprint on this then?23:17
jorgemjohnsom: ^^23:17
xgermanwell, I guess blogan has to do that23:17
jorgemI see your name on it ATM23:17
xgermanbut it's fins with us23:17
jorgemblogan: ^^ pleeeeease23:18
johnsomYes, grab it.  I think you need a core to change it23:18
jorgemxgerman: awesome thank you!23:18
openstackgerritTrevor Vardeman proposed stackforge/octavia: haproxy reference amphora API client
*** TrevorV_ has quit IRC23:27
*** jorgem has quit IRC23:31
rm_workwho is writing the deploy-worker code?23:45
rm_workxgerman: ^^23:45
rm_workI was kinda hoping to chat about that and check out the spec23:45
johnsomYes, I am point on deploy worker.  German is also contributing23:46
rm_workI was thinking it would probably be good to actually have the deploy-worker code handle the threading or forking internally, since it'd be required everywhere it was used, so we may as well make it inherent23:46
rm_workto reduce complexity of use / code duplication23:46
rm_workjust make it clear that there will be forking involved23:47
bloganrm_work: deploy worker is goign to be a thread right?23:47
rm_workthread/process, i don't really mind either way23:47
rm_workdoubt we'll be doing anything heavy-weight enough to run into GIL issues tho23:47
rm_workso thread is prolly fine23:48
bloganyeah so no need for a process23:48
blogananyway it sounds good in theory but i worry about thread management issues that may cause, which i do not have a concrete example of23:48
johnsomYes, we may even use some of the distributed taskflow capabilities in the future.  However, API manager will be spawning the thread23:48
bloganjohnsom: he saying he wants the api manager ot just call into the deploy_worker to spawn thread, and i assume manage it as well23:49
johnsomIf you use the notification part of oslo messaging it can handle that for us...23:49
xgermanwell, in that case what's the point of the api  manager? the deploy worker can lsiten on the queue as well?23:50
johnsomWhat other than api manager would be calling into deploy worker?23:50
johnsomYeah, I went back and forth on that.23:50
*** ajmiller has quit IRC23:50
johnsomWill API manager do work in addition to deploy worker23:51
bloganwasnt the housekeeping manager going to call deploy workers?23:51
bloganjohnsom: as far as I know it will just consume the queue and spawn deploy workers23:51
johnsomYeah, the API->database really reduced the API manager role23:52
blogani think having the queue consumption piece separate from the deploy code is good, i mean the deploy worker is just going to be a thread running inside the api manager process, so its still aprt of the api manager23:52
bloganand we may just want to rename the api manager to queue listener/consumer/receiver23:52
xgermanyep, makes sense23:53
xgermanok, so deploy worker is a library which sets up it's threadpool, eventloop, etc. and then spawns workers23:53
xgermanbased on library calls23:54
xgermanand if we need it in the housekeeper we will just import23:54
johnsomYeah, that works.23:54
xgermanawesome - seems we are ont eh same page23:57

Generated by 2.14.0 by Marius Gedminas - find it at!