Wednesday, 2017-03-29

« Tuesday, 2017-03-28 Index Thursday, 2017-03-30 »
*** ducttape_ has joined #openstack-lbaas00:01
*** ducttape_ has quit IRC00:06
*** reedip has quit IRC00:25
*** catintheroof has quit IRC00:34
*** ducttape_ has joined #openstack-lbaas00:42
*** sanfern has quit IRC00:48
*** ducttape_ has quit IRC00:52
*** ducttape_ has joined #openstack-lbaas00:52
*** cpuga has joined #openstack-lbaas01:06
*** gongysh has joined #openstack-lbaas01:15
*** amotoki has joined #openstack-lbaas01:46
*** amotoki has quit IRC01:58
*** gongysh has quit IRC01:59
*** ducttape_ has quit IRC02:03
bbbzhaojohnsom, reedip_ , Thank you for discuss. Sorry for late..I really want to work with our team, :). I will update that spec soon.02:05
*** ducttape_ has joined #openstack-lbaas02:06
bbbzhaoreedip_, the duplicate bug https://bugs.launchpad.net/octavia/+bug/1657089 , this one is the first one RFE bug I reported, as I didn't know lb function will be move to octavia project at that time, sorry for confused.02:07
openstackLaunchpad bug 1657090 in octavia "duplicate for #1657089 [RFE]Add bandwidth_limit to vip" [Wishlist,In progress] - Assigned to zhaobo (zhaobo6)02:07
*** ducttape_ has quit IRC02:11
*** yamamoto has joined #openstack-lbaas02:29
*** reedip has joined #openstack-lbaas02:31
*** yamamoto has quit IRC02:33
*** sanfern has joined #openstack-lbaas02:45
*** gongysh has joined #openstack-lbaas02:48
*** aojea has joined #openstack-lbaas02:49
*** aojea has quit IRC02:54
*** yuanying has quit IRC02:55
*** amotoki has joined #openstack-lbaas03:00
*** ducttape_ has joined #openstack-lbaas03:07
*** links has joined #openstack-lbaas03:12
*** ducttape_ has quit IRC03:12
*** raginbajin has quit IRC03:14
*** raginbajin has joined #openstack-lbaas03:21
openstackgerritLiang Cui proposed openstack/neutron-lbaas master: add @classmethod for health_monitor  https://review.openstack.org/45115703:25
openstackgerritZhaoBo proposed openstack/octavia master: Support QoS policy apply on vip port  https://review.openstack.org/44191203:36
*** fnaval has quit IRC03:55
*** fnaval has joined #openstack-lbaas03:56
*** fnaval has quit IRC04:00
*** reedip has quit IRC04:17
*** reedip has joined #openstack-lbaas04:19
*** ducttape_ has joined #openstack-lbaas04:38
*** yuanying has joined #openstack-lbaas04:39
*** ducttape_ has quit IRC04:43
openstackgerritAnkur proposed openstack/octavia master: Octavia Feature Classification  https://review.openstack.org/45117704:51
*** gcheresh_ has joined #openstack-lbaas04:57
reedipstill working ???05:07
*** oomichi has quit IRC05:29
*** oomichi has joined #openstack-lbaas05:33
*** yamamoto has joined #openstack-lbaas05:35
*** rcernin has joined #openstack-lbaas05:39
*** yamamoto has quit IRC05:40
*** reedip has quit IRC05:54
*** fnaval has joined #openstack-lbaas05:56
*** fnaval has quit IRC06:01
*** reedip has joined #openstack-lbaas06:01
*** armax has quit IRC06:06
*** ducttape_ has joined #openstack-lbaas06:08
*** ducttape_ has quit IRC06:12
*** pcaruana has joined #openstack-lbaas06:24
*** yamamoto has joined #openstack-lbaas06:41
*** yamamoto has quit IRC06:47
*** kobis has joined #openstack-lbaas06:50
*** kobis has quit IRC06:56
*** gongysh has quit IRC07:01
*** tesseract has joined #openstack-lbaas07:14
*** aojea has joined #openstack-lbaas07:20
*** oomichi has quit IRC07:29
*** oomichi has joined #openstack-lbaas07:33
*** gongysh has joined #openstack-lbaas07:35
*** ducttape_ has joined #openstack-lbaas07:39
*** aojea_ has joined #openstack-lbaas07:41
*** kobis has joined #openstack-lbaas07:42
*** yamamoto has joined #openstack-lbaas07:43
*** ducttape_ has quit IRC07:44
*** aojea has quit IRC07:44
*** yamamoto has quit IRC07:48
*** cpuga has quit IRC07:51
*** fnaval has joined #openstack-lbaas07:57
*** fnaval has quit IRC08:01
*** krypto has joined #openstack-lbaas08:09
*** ducnc has joined #openstack-lbaas08:09
*** yuanying has quit IRC08:31
*** yamamoto has joined #openstack-lbaas08:45
*** yamamoto has quit IRC08:50
*** cody-somerville has quit IRC09:02
*** gongysh has quit IRC09:02
*** cody-somerville has joined #openstack-lbaas09:02
*** ducttape_ has joined #openstack-lbaas09:09
*** ducttape_ has quit IRC09:14
*** blogan has quit IRC09:34
*** blogan_ has joined #openstack-lbaas09:34
*** reedip has quit IRC09:34
*** reedip has joined #openstack-lbaas09:37
*** yuanying has joined #openstack-lbaas09:39
*** amotoki has quit IRC09:40
*** reedip has quit IRC09:43
*** yuanying has quit IRC09:44
*** yamamoto has joined #openstack-lbaas09:46
*** yuanying has joined #openstack-lbaas09:47
*** reedip has joined #openstack-lbaas09:48
*** yamamoto has quit IRC09:52
*** reedip has quit IRC10:01
*** krypto has quit IRC10:04
*** krypto has joined #openstack-lbaas10:04
*** amotoki has joined #openstack-lbaas10:16
*** amotoki has quit IRC10:17
*** sticker has quit IRC10:24
*** oomichi has quit IRC10:29
*** amotoki has joined #openstack-lbaas10:31
*** oomichi has joined #openstack-lbaas10:33
*** krypto has quit IRC10:35
*** krypto has joined #openstack-lbaas10:35
*** ducttape_ has joined #openstack-lbaas10:40
*** amotoki has quit IRC10:40
*** ducttape_ has quit IRC10:44
*** yamamoto has joined #openstack-lbaas10:48
nmagnezio/10:51
*** yamamoto has quit IRC10:54
*** sanfern has quit IRC10:55
*** reedip has joined #openstack-lbaas10:58
*** yamamoto has joined #openstack-lbaas11:09
*** amotoki has joined #openstack-lbaas11:11
*** yamamoto has quit IRC11:19
*** krypto has quit IRC11:30
*** krypto has joined #openstack-lbaas11:30
*** cody-somerville has quit IRC11:42
*** sanfern has joined #openstack-lbaas11:47
*** fnaval has joined #openstack-lbaas11:58
*** fnaval has quit IRC12:03
*** chlong has quit IRC12:05
*** yamamoto has joined #openstack-lbaas12:20
*** yamamoto has quit IRC12:25
*** ducttape_ has joined #openstack-lbaas12:41
*** reedip has quit IRC12:52
*** krypto has quit IRC12:59
*** krypto has joined #openstack-lbaas13:00
*** ducttape_ has quit IRC13:13
*** links has quit IRC13:20
*** yamamoto has joined #openstack-lbaas13:22
*** yamamoto has quit IRC13:28
*** catintheroof has joined #openstack-lbaas13:30
openstackgerritOpenStack Proposal Bot proposed openstack/octavia master: Updated from global requirements  https://review.openstack.org/45104613:38
*** chlong has joined #openstack-lbaas13:40
*** catintheroof has quit IRC13:41
*** krypto has quit IRC13:42
*** krypto has joined #openstack-lbaas13:42
*** sticker has joined #openstack-lbaas13:45
*** fnaval has joined #openstack-lbaas13:58
*** reedip has joined #openstack-lbaas14:00
*** ducttape_ has joined #openstack-lbaas14:02
*** armax has joined #openstack-lbaas14:11
*** armax has quit IRC14:15
*** fnaval has quit IRC14:15
*** yamamoto has joined #openstack-lbaas14:24
*** yamamoto has quit IRC14:29
reedipo/14:32
xgermano/14:36
*** kobis has quit IRC14:47
*** blogan_ is now known as blogan14:51
*** armax has joined #openstack-lbaas14:53
*** fnaval has joined #openstack-lbaas14:54
*** amotoki has quit IRC14:56
*** cpuga has joined #openstack-lbaas15:02
*** rcernin has quit IRC15:05
diltramjohnsom: ping15:06
diltramhttps://review.openstack.org/#/c/405922/54/octavia/api/v2/controllers/pool.py@8215:07
johnsomDiltram pong15:07
diltramcan you take a look into this ^^ link15:07
diltramI'm making a review of this code15:07
diltramand if I'm right those updates should be done in one session15:07
nmagnezijohnsom, someone sent me this today.. maybe we should add it to the lbaas mascot votes http://imgur.com/a/atOLO15:08
johnsomYes15:08
johnsomnmagnezi That is funny15:09
johnsomThough we already voted and I sent in our vote.15:09
johnsomdiltram So, it looks like they are sharing the passed in session.  What am I missing?15:09
diltrambut every call has own commit15:10
diltramso we can have a situation that we updated load balancer15:10
diltramand then different server will try to change listeners15:11
diltrambut they was not changed15:11
diltramso we can change their status back to active15:11
diltramdo you understand what I'm trying to tell?15:12
johnsomWell, we lock out an LB and listener in the database by putting those in pending.  I think what you are saying is this is not using a non-autocommit session, so LB could go active while the listeners are still being updated.  Correct?15:12
openstackgerritShashank Kumar Shankar proposed openstack/octavia master: Run Octavia API in a WSGI server  https://review.openstack.org/44093415:13
diltramyes15:13
diltramand then we can even change their status in different process15:13
diltramand this process will overwrite those values for ACTIVE15:13
diltrambecause it's just updating them not verifying anything15:13
johnsomWell, another process can't change it while it is not in active15:14
diltramwell it not suppose to15:14
johnsomThey are blocked if it is in a PENDING_*, but I can see the race condition here15:14
diltrambut bugs always exist :P15:14
diltramok15:15
*** cpuga has quit IRC15:15
*** cpuga has joined #openstack-lbaas15:16
johnsomWe should wrap both of those at line 119 with a non-autocommit db session, IMO15:16
diltramagree15:17
johnsomGood catch15:17
diltramthx15:17
rm_workkk15:23
rm_workyeah that was something i had to recently add15:23
rm_workso it had not really gone through as much review15:23
rm_workso we need to use lock-session there, and not commit till the end, got it15:24
johnsomIt can be a new session, I don't think it should be tied into the upper session.15:25
rm_workk15:25
rm_workso just... make a new lock_session there15:25
rm_workand that's it?15:25
*** yamamoto has joined #openstack-lbaas15:26
rm_workhmm15:27
rm_workreally that should be *anywhere* that uses _reset_lb_and_listener_statuses() right?15:28
johnsomYes,15:28
johnsomI think on the lock side we include it in the lock session there.15:29
rm_workI mean, we have the lock_session already in post(), I can reuse that right?15:29
rm_workoh, hmm maybe not15:29
johnsomNo, I don't think so15:29
rm_workmight be post-rollback for that15:29
johnsomIt should commit by then15:30
rm_workactually...15:30
rm_workk yeah i see15:30
rm_workwait15:30
johnsomI purposefully separated the handler code from the initial API code.15:30
rm_workwhat is the condition?15:30
rm_workthere's nothing after that which can fail, is there?15:31
*** yamamoto has quit IRC15:31
rm_workadding a lock session there would be ... lock_session -> do that -> commit15:31
rm_workwhich is not useful15:31
johnsomAt 119?15:31
rm_workyes?15:31
rm_worki'm trying to figure out how the pool update at 123 is relevant15:32
openstackgerritMerged openstack/octavia master: Updated from global requirements  https://review.openstack.org/45104615:32
rm_workit's the only other thing that writes to the session, right?15:32
johnsomIt would mean all of the updates to the status (LB, listener, and pool in this case) will occur atomicly as one collective.  So you don't have a situation where an LB is partially unlocked in the DB and another process grabs it and starts using it.15:32
rm_workoh, internally15:32
rm_workgot it15:32
johnsomIt's like a split-second possible race conditiion15:33
rm_workright, there's multiple *internal* things15:33
rm_workI was thinking of it as one function -> one op15:33
johnsomRight15:33
rm_workcan I just put the lock_session inside that function?15:34
rm_workso it takes care of all three places?15:34
rm_workbecause it can be a separate lock_session from anything else, right?15:34
rm_workor does the pool status update after that also matter for this15:34
rm_workah i'll just be safe and do it outside, including the pool update15:36
xgerman+1 better safe then sorry15:37
rm_workdiltram / johnsom: https://gist.github.com/rm-you/f57fc3bd85235e801025d0fa020a16ad#file-gistfile1-txt-L8-L2015:38
rm_workdoes that look reasonable?15:38
rm_workbasically cloning that across to put/delete as well15:38
diltramI know15:38
rm_workor could I do it better as a context-manager?15:38
diltramI had the same filing that we can make some decorator or anything to handle this15:39
johnsomYep, look fine to me15:39
rm_workyeah, decorator was a thought15:39
rm_workbut not sure how to make that work internally on multiple lines15:39
rm_workthat's what context-managers are for15:39
diltramagree :)15:39
diltramgo for it man :D15:40
rm_worklol15:40
rm_workk15:40
reedipjohnsom : any updates on yesterday;s discussion about the diversity tag ???15:45
reedipjust curious15:46
johnsomThey postponed the vote15:46
*** ducttape_ has quit IRC15:46
rm_workI mean it's just this, right? https://gist.github.com/rm-you/7f13ececa82d7ecf6e7c46387409dc9c15:47
*** ducttape_ has joined #openstack-lbaas15:47
rm_workjohnsom: well, that's good, i was legit worried15:47
*** ducttape_ has quit IRC15:47
johnsomI am trying to come back with some thoughts on the criteria used.  Basically it looks like 56% of the core reviews were from one company, Rackspace.  Which has the PTL and two out of four cores, so, kind of expected I would think.15:47
johnsomWe need rm_work and diltram to step up the reviews and others to step up their reviews to become core...  grin15:48
rm_worklol15:49
johnsomUsing their tool these are the results:15:50
johnsomhttps://www.irccloud.com/pastebin/kAQ0AEHz/15:50
johnsomPercentages are "from a single company"15:51
xgermana lot of this stuff is sending up a trial balloon and see who complains. Silence is consent with those guys15:52
johnsomSo the only metric that doesn't qualify is the "top core review %"15:52
xgermanyeah, we should aim for “probabtion” and step up our game15:52
johnsomI am trying to think about a way to propose different math, I think it's weighted wrong for small teams.  The intent is to have a measure of if the project will die if one company pulls out.15:54
openstackgerritAdam Harwell proposed openstack/octavia master: Install DIB from source so depends-on will work  https://review.openstack.org/44470715:54
xgermanwell, if RAX pulls out there would be some hit…15:55
johnsomYes, there would be some hit, but I don't think it would be dead15:56
*** ducttape_ has joined #openstack-lbaas15:56
rm_workit's OK, if RAX pulls out, johnsom will find a job with another company to continue being Octavia PTL, right? :P15:56
*** aojea_ has quit IRC15:56
rm_workwe could use you here15:57
johnsomHahaha, well, it's happened once before15:57
xgermanrm_work lol15:57
*** amotoki has joined #openstack-lbaas15:57
johnsomAssuming I don't throw my hands up and go bake bread or something...15:57
johnsomThough I don't see that happening any time soon, so...15:58
xgermanther eis some leadership training by a bakery15:58
*** tesseract has quit IRC15:58
rm_workjust watch a lot of Cake Boss15:58
reedipjohnsom : well , I am trying :D15:59
johnsomUgh, I'm more of a Duff fan myself15:59
xgermanhttp://lists.openstack.org/pipermail/openstack-dev/2017-March/113267.html15:59
xgermanall paid for by the foundation15:59
johnsomhttp://www.charmcitycakes.com/16:00
reedipjohnsom : whats the criteria for core :D16:00
xgerman“reviews in line with other core reviewers”16:00
johnsomreedip Good quality reviews at a quantity in line with the other cores per stackalytics16:00
xgermanbut it’s not about who is core but the # of core reviews in that metric16:01
johnsomhttp://stackalytics.com/report/contribution/octavia-group/9016:01
reedipjohnsom and whats the time duration , I mean how long would the results be checked for ? One release ?16:02
johnsomreedip Trying to game the system?16:02
johnsomWe typically look at the 30 day and 90 day charts.  Quality is a factor here.  Just +1'ing everything will not get you there....16:03
reedipjohnsom : as if others arent smart enough ?? :)16:03
xgermanafk for a bit16:03
*** amotoki has quit IRC16:03
reedipjohnsom : exactly, thats y the system cant be gamed16:03
johnsomYeah16:03
reedipso I guess the +% and Disagreement ratio may also matter16:04
reedipIIUC +% should be < 8016:04
reedipI mean its not necessary for one to just +1 everything , as you said :)16:04
openstackgerritAdam Harwell proposed openstack/octavia master: Introduce Octavia v2 API for pools  https://review.openstack.org/40592216:04
rm_workyeah, 70-80% is ok I think, though it does depend a bit on the quality of the code being pushed up during a cycle :P16:06
johnsomRight16:06
rm_workI had 90% for a bit because I was pretty much only reviewing johnsom's code... not like before when I was also reviewing a lot of blogan's stuff16:06
rm_work^_^16:06
johnsomHahaha.16:07
rm_workI wonder what mine is right now16:07
bloganim sorry16:07
reediphehehe16:08
rm_workblogan: no it was good, I needed your reviews to keep my +/- ratio in line16:08
* johnsom gives blogan a cookie16:08
bloganwell, glad i could help and ill happily accept the cookie, even though i know its one them tracking cookies16:09
bloganjokes on you though, i just stay home16:09
reedipanyways, thanks guys for the information ... will help me to get somewhere :)16:12
bloganreedip: glad i could help :)16:13
reedipblogan : :=)16:13
*** krypto has quit IRC16:14
*** krypto has joined #openstack-lbaas16:15
*** krypto has quit IRC16:15
*** krypto has joined #openstack-lbaas16:15
*** krypto has quit IRC16:15
*** krypto has joined #openstack-lbaas16:16
*** krypto has quit IRC16:16
*** krypto has joined #openstack-lbaas16:16
*** krypto has quit IRC16:16
*** krypto has joined #openstack-lbaas16:17
*** krypto has quit IRC16:17
rm_workdiltram: posting the fix for your other comment now16:20
rm_workthen should be good again16:20
rm_workif that was it :)16:20
rm_workdiltram / johnsom: up for merging that now and looking at members?16:20
johnsomI will take a quick look again16:21
rm_worksec, posting one more change as soon as tests finish running16:21
rm_workpep8 tests take longer to START than functional tests take to finish, wonder what's up with that16:21
*** yamamoto has joined #openstack-lbaas16:27
johnsomDIB is now an infra project BTW, it merged yesterday16:28
*** yamamoto has quit IRC16:33
rm_workah cool16:40
openstackgerritAdam Harwell proposed openstack/octavia master: Introduce Octavia v2 API for pools  https://review.openstack.org/40592216:41
rm_workok there we go, got distracted and tests were finished <_<16:41
rm_workoh man, I just found the Michael Johnson that is an Olympic sprinter :P16:46
rm_workI'm currently enjoying this page: https://en.wikipedia.org/wiki/Michael_Johnson16:46
johnsomYeah, had to hang up the gold shoes....16:49
* johnsom wonders why Adam is stalking him...16:51
ankur-gupta-f4I remember him. he was a boss16:51
johnsomHe was16:51
rm_workI just saw it in an article about something else16:51
rm_workand was like: lolwut16:51
johnsomCrazy fast16:52
ankur-gupta-f4johnsom: why you no like feature classification16:54
openstackgerritMerged openstack/neutron-lbaas master: Fix a releasenotes issue with rst markup.  https://review.openstack.org/44982116:54
johnsomankur-gupta-f4 It seems a bit opinionated in it's current form16:55
ankur-gupta-f4yea first iteration. but you see what its getting at right?16:55
johnsomBut I left comments for the "hard" reasons I -1'd it16:55
johnsomYeah, I see value in knowing which driver supports L7 for example16:55
ankur-gupta-f4or to your point HAproxy has no one working on it. So operators should know that hey, haproxy may not support a lot of features... etc...16:56
johnsomYeah, I think assuming that something marked "immature" is under development might not be accurate.16:57
ankur-gupta-f4but its not labeling the entire driver as immature16:57
ankur-gupta-f4its saying that this driver's support of the QoS feature is immature16:58
ankur-gupta-f4or incomplete16:58
ankur-gupta-f4or something like that16:58
ankur-gupta-f4i forget the working16:58
ankur-gupta-f4wording*16:58
openstackgerritMerged openstack/octavia master: Fix get_all method for v2 LB controller  https://review.openstack.org/44982216:59
openstackgerritMerged openstack/octavia master: Remove unused logging import  https://review.openstack.org/44945516:59
*** amotoki has joined #openstack-lbaas16:59
*** chlong has quit IRC17:03
*** amotoki has quit IRC17:04
openstackgerritMerged openstack/neutron-lbaas master: Updated from global requirements  https://review.openstack.org/45104117:05
*** links has joined #openstack-lbaas17:07
*** links has quit IRC17:19
johnsomIs anyone fixing the listener patch for the handler exception lock_session issue or should I open a bug?17:20
johnsomI guess LB PUT might be impacted as well17:21
openstackgerritMerged openstack/octavia master: change genconfig output-file path  https://review.openstack.org/44990217:28
*** yamamoto has joined #openstack-lbaas17:29
openstackgerritGerman Eichberger proposed openstack/octavia master: [WIP] Adds the ability to failover a load balancer  https://review.openstack.org/44434717:32
openstackgerritMerged openstack/octavia master: Fix some reST field lists in docstrings  https://review.openstack.org/45073617:34
*** yamamoto has quit IRC17:35
openstackgerritMerged openstack/octavia master: DIB: drop some dependencies  https://review.openstack.org/44865717:35
*** ducttape_ has quit IRC17:37
*** ducttape_ has joined #openstack-lbaas17:38
*** pcaruana has quit IRC17:52
*** chlong has joined #openstack-lbaas17:58
*** amotoki has joined #openstack-lbaas18:00
*** amotoki has quit IRC18:05
*** chlong has quit IRC18:07
*** ducttape_ has quit IRC18:07
*** ducttape_ has joined #openstack-lbaas18:09
*** ducttape_ has quit IRC18:17
*** chlong has joined #openstack-lbaas18:21
*** aojea has joined #openstack-lbaas18:24
rm_workjohnsom: you mean like the same thing i just fixed in pools?18:26
johnsomYes18:26
rm_workI looked and didn't see where it was broken18:26
rm_workwhich line in listeners?18:27
johnsomOh, ok.  I just assumed it had the same code18:27
rm_workit doesn't have the status reset thing18:27
rm_work... though I wonder if it SHOULD18:27
rm_workthat status reset function is something i wrote just for pools18:27
rm_worki'm wondering if it DOES need to do a status reset though? :/18:28
johnsomIt seems like it would have the same reset on handler failure.18:29
rm_workwell18:29
rm_workit has less objects to deal with18:29
rm_workLB would HAVE no parent object, it'd only set *itself* to ERROR18:29
rm_workListener though only sets itself to ERROR, I wonder if it needs to set LB back to ACTIVE?18:30
rm_work... or ERROR?18:30
rm_workI'm actually not sure18:30
*** yamamoto has joined #openstack-lbaas18:31
*** chlong has quit IRC18:34
openstackgerritGerman Eichberger proposed openstack/octavia master: Adds the ability to failover a load balancer  https://review.openstack.org/44434718:34
xgerman^^ johnsom, rm_work good to go. rm_work I refrained from chnagign the failover flow too much but it should failover far more rapidly now as we have an amphora prepared - not ideal but people who want HA should be running ACTIVE-PASSIVE anyway18:36
*** yamamoto has quit IRC18:36
*** ducttape_ has joined #openstack-lbaas18:37
openstackgerritNakul Dahiwade proposed openstack/octavia master: Introduce Octavia v2 API for Members  https://review.openstack.org/40593218:40
*** ducttape_ has quit IRC18:40
*** openstackstatus has joined #openstack-lbaas18:44
*** ChanServ sets mode: +v openstackstatus18:44
*** chlong has joined #openstack-lbaas18:49
rm_workwell i mean18:53
rm_workthat's also what the spares pool is for :P18:53
rm_workbut yes, it should definitely work this way18:53
xgermanit’s better than before since I will build a spare before deleting the amphora18:53
rm_workright but ... spares pool :P18:54
xgermanyep, ig we have spares pool I will refarin from that18:54
rm_worki mean yeah, same code as before, pick spares pool amp first, then try a build18:56
rm_worki was just saying, in my deploy i used spares pool, so it won't be much faster :)18:56
xgermank19:00
*** amotoki has joined #openstack-lbaas19:01
*** amotoki has quit IRC19:05
rm_workthough I guess i'll have to kill spares pool when I go active/standby19:16
xgermanno, they complement each other19:17
rm_workbecause it doesn't really work for apolocation19:17
xgermanhu?19:17
rm_workwhen you are using apolocation19:17
rm_workit sends a scheduler hint to nova19:17
rm_workbased on the existing amp19:17
rm_workright?19:17
rm_workbut if you pull from the spares pool...19:18
johnsomCorrect19:18
xgermanmmh, that should be a bug19:18
rm_workmight get an amp in the same location as your existing19:18
rm_workahhh sorry, antiaffinityu19:18
rm_workwrong term19:18
xgermanyep, got it19:18
johnsomSpares pool and the anti-affinity setting (optional) don't work well together19:18
xgermanwell, we should file a bug and fix it ;-)19:19
rm_workyeah i guess we could, uhh19:19
johnsomI don't know that nova gives us a way to fix it19:19
rm_workstore the cell info or something in the amp table19:19
rm_workand uhh19:19
rm_workhope that with enough amps, it is statistically distributed19:19
johnsomWe do store the server group already19:19
rm_workok19:19
rm_workso maybe we could do it manually...19:19
johnsomWhat if we add the instance to the server group nova migrates it?19:20
xgermanwould it do that for us?19:20
johnsomI don't think so19:20
rm_worklol19:20
johnsomThat is kind of what we would be asking for though19:21
xgermanwell, we can always interrogate the spare amps and picke the one which fits and/or build a new one19:21
johnsomHmm, maybe19:23
johnsomOh, neat, they added "soft-anti-affinity" so it doesn't just fail if there aren't enough hosts.19:24
johnsomWe should add that option19:24
xgermank19:25
*** ducttape_ has joined #openstack-lbaas19:30
xgermanjohnsom somehow that failover flow ignores the server_group if I am not blind19:31
*** yamamoto has joined #openstack-lbaas19:32
*** openstackgerrit has quit IRC19:33
*** ducttape_ has quit IRC19:33
*** chlong has quit IRC19:34
*** yamamoto has quit IRC19:38
rm_workI THOUGHT it used it...19:38
xgermanI can’t see where we bring it into the flow19:40
*** cpuga_ has joined #openstack-lbaas19:46
johnsomYeah, it might.  We may have skipped that due to the spares issue / what if it can't be filled...19:47
xgermanwell, I will see if I can straighten things19:48
*** cpuga has quit IRC19:50
rm_workanyone for +A on https://review.openstack.org/#/c/405922/ ? diltram? xgerman?19:52
rm_workdiltram: you reviewed already, right? so if I fixed your issue...19:52
diltramrm_work: moment19:54
johnsomOctavia meeting starting soon on #openstack-meeting-alt19:56
rm_workah yeah, about to be meeting time...19:56
diltramrm_work: merged19:56
rm_workdiltram: sweet! thanks19:57
rm_workMEMBERS TIME!19:57
johnsom+119:57
rm_workreally want to get done with *everything* by EOW19:57
rm_workif possible19:57
diltram+119:57
rm_workthough i haven't actually looked at HMs or L719:57
rm_work... and I may not be around Thursday/Friday actually T_T at least during normal hours19:57
diltram:P19:58
rm_workhopefully can review members after the meeting? I'll go through it again right now19:59
johnsomHM needs work last I heard19:59
rm_workk :/19:59
*** amotoki has joined #openstack-lbaas20:02
rm_workhmm i think members may need to be updated per the pools status reset stuff20:02
rm_workright?20:02
*** amotoki has quit IRC20:06
*** yamamoto has joined #openstack-lbaas20:34
*** yamamoto has quit IRC20:40
*** openstackgerrit has joined #openstack-lbaas20:41
openstackgerritMerged openstack/octavia master: Introduce Octavia v2 API for pools  https://review.openstack.org/40592220:41
*** gcheresh_ has quit IRC20:42
johnsomCores, the backport we talked about in the meeting: https://review.openstack.org/#/c/451564/20:56
*** amotoki has joined #openstack-lbaas21:02
*** amotoki has quit IRC21:07
rm_workworking on a members update21:08
*** ducttape_ has joined #openstack-lbaas21:08
*** ducttape_ has quit IRC21:21
*** aojea has quit IRC21:34
*** yamamoto has joined #openstack-lbaas21:36
rm_workalmost done (maybe)21:40
*** yamamoto has quit IRC21:42
*** cpuga has joined #openstack-lbaas21:44
*** ducttape_ has joined #openstack-lbaas21:44
*** cpuga_ has quit IRC21:46
johnsomdiltram Why is lxd gate running for stable/ocata?21:48
johnsomrm_work Can you clear your -1 on https://review.openstack.org/#/c/446959/?21:49
rm_workremoved for now21:59
rm_workbut still need to look at it further21:59
rm_workwhy didn't my -1 clear on the new patchset?21:59
johnsomDIB doesn't do stable branches so we are stuck updating this stuff21:59
johnsomNot sure on that22:00
rm_workok that's ... a lot better22:01
openstackgerritAdam Harwell proposed openstack/octavia master: Introduce Octavia v2 API for Members  https://review.openstack.org/40593222:01
rm_work^^22:01
rm_workjohnsom: yeah but doesn't dib-utils still exist in pypi?22:01
rm_workor did they DELETE it?22:01
johnsomThey don't package22:01
johnsomRight?22:02
johnsomBefore we were git clone the package in22:02
rm_workerr22:02
rm_workah22:02
rm_workhmm maybe?22:02
rm_workbut it was in our requirements, no?22:02
rm_workdib-utils22:02
rm_workand we had to switch to disk-image-builder22:02
rm_workand also, if it's cloned in, why did this patch even need to exist? https://review.openstack.org/#/c/444707/22:03
johnsomdib-utils wasn't in requirements for newton at least22:03
*** amotoki has joined #openstack-lbaas22:03
johnsomYeah , not in ocata either22:04
johnsomhttps://github.com/openstack/octavia/commit/cf9a8c8f779a3542681a4b3ea3274681cec6f68222:07
johnsomThat removed the clone22:07
*** amotoki has quit IRC22:08
*** fnaval has quit IRC22:13
rm_worklol thanks22:16
rm_worklinking me a patch *I wrote* to prove me wrong >_<22:16
rm_workthat's just cold22:16
rm_workso ... we WERE using DIB from master22:17
rm_workuntil that patch?22:17
rm_workso why such pushback against just doing the same again?22:17
rm_workre: my patch above22:17
*** yamamoto has joined #openstack-lbaas22:38
*** yamamoto has quit IRC22:42
johnsomFor me it's about trying to find some way out of this treadmill with DIB changes22:45
rm_workyeah... :/22:48
johnsomUgh, I screwed up that cherry pick anyway....  Ugh22:49
rm_worklol22:51
*** reedip has quit IRC22:51
johnsomOk, this one for newton seems to pass: https://review.openstack.org/#/c/451586/22:51
johnsomBut it didn't actually run any tests...  What is this craziness?  SKIPPED: LBaaS Extension is not enabled22:53
*** ducttape_ has quit IRC22:57
openstackgerritMichael Johnson proposed openstack/neutron-lbaas master: Imported Translations from Zanata  https://review.openstack.org/44903722:57
*** amotoki has joined #openstack-lbaas23:04
*** amotoki has quit IRC23:09
*** reedip has joined #openstack-lbaas23:26
*** reedip has quit IRC23:31
*** yamamoto has joined #openstack-lbaas23:39
*** Guest59958 has joined #openstack-lbaas23:41
*** yamamoto has quit IRC23:44
Guest59958Hello, I was wondering if I should remove the extra sections in this blueprint: https://review.openstack.org/#/c/450505/, since it appears to be failing jenkins review, or should I leave them?23:44
*** reedip has joined #openstack-lbaas23:45
johnsomGuest59958 Hi23:49
johnsomGuest59958 This should go under: https://github.com/openstack/octavia/tree/master/specs/version1 as load balancing is no longer a neutron project.23:50
johnsomThere is also a template that has to be used: https://github.com/openstack/octavia/blob/master/specs/skeleton.rst for that test to pass.23:52
johnsomLooking at the spec, there is an extra white space at the  top and the lines are too long.  The test will fail for those issues as well.23:53
*** reedip has quit IRC23:54
Guest59958johnsom: Thanks. Does that mean any code commits would also be made to the Octavia repository rather than neutron-lbaas?23:57
*** cpuga has quit IRC23:58
johnsomGuest59958 It would be best, but you can still submit them to neutron-lbaas until the Pike-2 milestone when we will feature freeze neutron-lbaas from future feature additions.  The load balancing API will be in octavia going forward.23:59
« Tuesday, 2017-03-28 Index Thursday, 2017-03-30 »

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