Monday, 2016-04-11

*** jerrygb has joined #openstack-lbaas00:22
*** chlong has quit IRC00:27
*** ducttape_ has joined #openstack-lbaas00:31
*** chlong has joined #openstack-lbaas00:34
*** neelashah has joined #openstack-lbaas01:21
*** ducttape_ has quit IRC01:32
*** amotoki has joined #openstack-lbaas01:38
*** bana_k has joined #openstack-lbaas01:46
*** amotoki has quit IRC02:00
*** amotoki has joined #openstack-lbaas02:10
*** jaff_cheng has joined #openstack-lbaas02:13
*** jerrygb has quit IRC02:29
*** jaff_cheng has quit IRC02:36
*** jerrygb has joined #openstack-lbaas02:42
*** yuanying has quit IRC02:46
kongMissing ports in members' network when using spared pool during failover process: <>02:47
openstackLaunchpad bug 1568653 in octavia "Missing ports in members' network when using spared pool during failover process" [Undecided,New]02:47
*** neelashah has quit IRC02:54
*** jerrygb has quit IRC03:41
kongis it the right way that I marked down the management port of the amphorae vm?03:45
*** links has joined #openstack-lbaas03:45
*** yuanying has joined #openstack-lbaas03:49
openstackgerritHenry Gessau proposed openstack/neutron-lbaas: [WIP][DNM] For Doug
*** jerrygb has joined #openstack-lbaas03:51
*** jerrygb has quit IRC03:51
*** jerrygb has joined #openstack-lbaas03:51
*** jerrygb has quit IRC03:51
*** minwang2 has joined #openstack-lbaas03:56
*** prabampm has joined #openstack-lbaas04:06
openstackLaunchpad bug 1558934 in octavia "Healthmonitor fails to spawn new amphora from spare pool" [Critical,Fix released] - Assigned to Michael Johnson (johnsom)04:11
*** prabampm has quit IRC04:12
kongbana_k: thanks! but I wonder why the team didn't backport this to stable/mitaka? I think it's critical enough04:16
kongThis bug means we can04:17
kongnot use spare pool for failover04:17
bana_kyea I am not sure about that.04:17
bana_kand actually as I read the comments on that bug I think we need the latest version of the neutron04:18
kongbana_k: yes :-( you're right04:22
*** amotoki has quit IRC04:22
*** amotoki has joined #openstack-lbaas04:24
*** jerrygb has joined #openstack-lbaas04:28
kongcore teams, can I backport to stable/mitaka? It's really important for us to deployment LBaaS in our public cloud04:30
kongsorry, have been backported, but not merged04:34
*** jerrygb has quit IRC04:37
*** jerrygb has joined #openstack-lbaas04:38
*** neelashah has joined #openstack-lbaas04:39
*** jerrygb has quit IRC04:43
*** neelashah has quit IRC04:56
*** anilvenkata has joined #openstack-lbaas05:05
*** minwang2 has quit IRC05:09
*** minwang2 has joined #openstack-lbaas05:09
*** pcaruana has joined #openstack-lbaas05:39
*** bana_k has quit IRC05:55
*** minwang2 has quit IRC06:08
*** chlong has quit IRC06:13
*** rcernin has joined #openstack-lbaas06:15
*** ChrisBenson has quit IRC06:20
*** minwang2 has joined #openstack-lbaas06:30
*** tesseract has joined #openstack-lbaas06:35
*** tesseract is now known as Guest898906:35
*** bana_k has joined #openstack-lbaas06:43
*** bana_k has quit IRC06:53
*** numan_ has joined #openstack-lbaas06:57
*** ihrachys has joined #openstack-lbaas07:03
*** Alex_Stef has joined #openstack-lbaas07:04
*** dhlorenz has joined #openstack-lbaas07:18
*** minwang2 has quit IRC07:29
*** chlong has joined #openstack-lbaas07:31
*** nmagnezi has joined #openstack-lbaas07:31
*** kobis has joined #openstack-lbaas07:48
*** yamamoto has joined #openstack-lbaas08:24
*** ArchiFleKs has quit IRC08:52
*** banszmar has quit IRC08:52
*** pcaruana has quit IRC09:00
*** banszmar has joined #openstack-lbaas09:09
kongI applied the patch in my deployment, when performing fail over, the new vm is created with the ports, but09:42
kongsorry, no but. it finally works, just to many retries09:44
openstackgerritYang Yu proposed openstack/neutron-lbaas: Fix no such option defined for service_auth
*** reedip__ has joined #openstack-lbaas11:29
*** yamamoto has quit IRC11:46
*** rtheis has joined #openstack-lbaas11:47
*** amotoki has quit IRC11:58
*** mdavidson has quit IRC12:07
*** amotoki has joined #openstack-lbaas12:09
*** yamamoto has joined #openstack-lbaas12:22
*** amotoki has quit IRC12:27
openstackgerritEvgeny Fedoruk proposed openstack/neutron-lbaas: Fixing tests to proper compare dicts.
*** mdavidson has joined #openstack-lbaas12:50
*** jschwarz has joined #openstack-lbaas12:59
openstackgerritNir Magnezi proposed openstack/neutron-lbaas: (WIP) Adds option to auto reschedule loadbalancers from dead lbaas agents
*** amotoki has joined #openstack-lbaas13:08
*** jschwarz has quit IRC13:10
*** jerrygb has joined #openstack-lbaas13:16
*** links has quit IRC13:21
*** piet has joined #openstack-lbaas13:21
*** Bjoern_ has joined #openstack-lbaas13:54
*** neelashah has joined #openstack-lbaas13:57
*** nmagnezi has quit IRC13:57
*** neelashah1 has joined #openstack-lbaas13:58
*** doug-fish has joined #openstack-lbaas14:00
*** neelashah has quit IRC14:02
*** woodster_ has joined #openstack-lbaas14:03
*** yamamoto has quit IRC14:06
*** yamamoto has joined #openstack-lbaas14:10
*** neelashah1 has quit IRC14:11
*** neelashah has joined #openstack-lbaas14:15
*** ajmiller has joined #openstack-lbaas14:20
*** kobis has quit IRC14:26
*** Alex_Stef has quit IRC14:34
*** ducttape_ has joined #openstack-lbaas14:40
*** piet has quit IRC14:49
*** yamamoto has quit IRC14:51
*** fnaval has joined #openstack-lbaas14:54
*** TrevorV has joined #openstack-lbaas14:58
*** fawadkhaliq has joined #openstack-lbaas15:01
*** ducttape_ has quit IRC15:01
*** dhlorenz_ has joined #openstack-lbaas15:02
*** ducttape_ has joined #openstack-lbaas15:02
*** anilvenkata has quit IRC15:03
*** dhlorenz has quit IRC15:05
*** numan_ has quit IRC15:07
openstackgerritFranklin Naval proposed openstack/neutron-lbaas: Barbican Scenario Test (TLS with Intermediates)
*** piet has joined #openstack-lbaas15:17
*** openstackgerrit has quit IRC15:18
*** fawadkhaliq has quit IRC15:18
*** openstackgerrit has joined #openstack-lbaas15:18
*** fawadkhaliq has joined #openstack-lbaas15:21
*** yamamoto has joined #openstack-lbaas15:21
*** matt-borland has joined #openstack-lbaas15:23
*** yamamoto has quit IRC15:27
*** ajmiller has quit IRC15:33
*** Guest8989 has quit IRC15:35
*** yamamoto has joined #openstack-lbaas15:42
*** _ducttape_ has joined #openstack-lbaas15:45
*** ducttape_ has quit IRC15:47
*** fawadkhaliq has quit IRC15:47
*** yamamoto has quit IRC15:52
*** bana_k has joined #openstack-lbaas15:55
*** johnsom_ is now known as johnsom16:00
*** armax has joined #openstack-lbaas16:03
*** neelashah has quit IRC16:05
*** neelashah has joined #openstack-lbaas16:08
*** neelashah has quit IRC16:13
*** neelashah has joined #openstack-lbaas16:15
*** fawadkhaliq has joined #openstack-lbaas16:16
*** armax has quit IRC16:16
*** fawadkhaliq has quit IRC16:16
*** rtheis_ has joined #openstack-lbaas16:18
*** rtheis has quit IRC16:19
*** fawadkhaliq has joined #openstack-lbaas16:28
*** madhu_ak has joined #openstack-lbaas16:28
*** armax has joined #openstack-lbaas16:29
*** bana_k has quit IRC16:38
*** rcernin has quit IRC16:40
*** minwang2 has joined #openstack-lbaas16:41
*** crc32 has joined #openstack-lbaas16:44
*** fawadkhaliq has quit IRC16:48
*** fawadkhaliq has joined #openstack-lbaas16:49
*** kevo has quit IRC16:50
*** yamamoto has joined #openstack-lbaas16:52
*** jerrygb has quit IRC16:53
*** jerrygb has joined #openstack-lbaas16:54
*** yamamoto has quit IRC17:00
*** fawadkhaliq has quit IRC17:06
*** fawadkhaliq has joined #openstack-lbaas17:09
*** ihrachys has quit IRC17:13
*** bana_k has joined #openstack-lbaas17:14
*** neelashah has quit IRC17:25
openstackgerritArmando Migliaccio proposed openstack/neutron-lbaas: Move away from locally installed packages
johnsomkong ping17:32
*** neelashah has joined #openstack-lbaas17:35
fnavalmadhu_ak / minwang2:  please review when you have a chance, thanks in advance!
madhu_aksure fnaval. Thanks for reiterating it17:36
fnavalmadhu_ak: yep np - it should be really ready; last iteration was just an update on the commit message17:36
*** bana_k has quit IRC17:37
*** numan_ has joined #openstack-lbaas17:40
*** jerrygb has quit IRC17:41
*** bana_k has joined #openstack-lbaas17:52
*** kevo has joined #openstack-lbaas17:53
*** neelashah has quit IRC17:56
*** amotoki has quit IRC17:59
*** alejandrito has joined #openstack-lbaas17:59
*** doug-fis_ has joined #openstack-lbaas18:01
*** fnaval_ has joined #openstack-lbaas18:02
*** doug-fi__ has joined #openstack-lbaas18:02
*** fnaval has quit IRC18:03
*** doug-f___ has joined #openstack-lbaas18:04
*** doug-fish has quit IRC18:04
*** doug-f___ is now known as doug-fish18:06
*** doug-fis_ has quit IRC18:06
*** reedip__ has quit IRC18:06
*** fnaval_ has quit IRC18:06
*** doug-fi__ has quit IRC18:07
*** neelashah has joined #openstack-lbaas18:08
*** dhlorenz_ has quit IRC18:20
*** fawadkhaliq has quit IRC18:20
*** fawadkhaliq has joined #openstack-lbaas18:21
*** crc32 has quit IRC18:21
*** cody-somerville has quit IRC18:23
*** fawadkhaliq has quit IRC18:33
*** fawadkhaliq has joined #openstack-lbaas18:34
openstackgerritMerged openstack/neutron-lbaas-dashboard: Updated from global requirements
*** piet has quit IRC18:40
*** dhlorenz has joined #openstack-lbaas18:40
*** madhu_ak has quit IRC18:53
*** doug-fish has quit IRC18:57
*** doug-fish has joined #openstack-lbaas18:58
*** bana_k has quit IRC18:58
*** doug-fis_ has joined #openstack-lbaas19:01
*** madhu_ak has joined #openstack-lbaas19:02
*** doug-fish has quit IRC19:03
*** bana_k has joined #openstack-lbaas19:03
*** doug-fis_ has quit IRC19:06
*** bana_k has quit IRC19:10
*** neelashah has quit IRC19:13
*** neelashah has joined #openstack-lbaas19:18
*** fnaval has joined #openstack-lbaas19:24
*** bana_k has joined #openstack-lbaas19:26
*** piet has joined #openstack-lbaas19:27
*** numan_ has quit IRC19:28
*** ihrachys has joined #openstack-lbaas19:36
*** mixos has joined #openstack-lbaas19:38
*** fawadkhaliq has quit IRC19:39
*** fawadkhaliq has joined #openstack-lbaas19:40
*** fnaval has quit IRC19:50
*** fnaval has joined #openstack-lbaas19:51
*** fawadkhaliq has quit IRC19:51
*** fawadkhaliq has joined #openstack-lbaas19:51
*** ajmiller has joined #openstack-lbaas19:52
*** rtheis_ has quit IRC19:53
*** fawadkhaliq has quit IRC19:54
*** fawadkhaliq has joined #openstack-lbaas19:55
*** rtheis has joined #openstack-lbaas19:55
*** openstackgerrit has quit IRC20:05
*** Bjoern_ is now known as BjoernT20:13
*** neelashah has quit IRC20:18
*** openstackgerrit has joined #openstack-lbaas20:28
*** alejandrito has quit IRC20:30
*** alejandrito has joined #openstack-lbaas20:36
*** TrevorV has quit IRC20:39
*** doug-fis_ has joined #openstack-lbaas20:40
*** openstackgerrit has quit IRC20:41
*** doug-fis_ has quit IRC20:45
*** doug-fish has joined #openstack-lbaas20:52
*** neelashah has joined #openstack-lbaas20:52
*** openstackgerrit has joined #openstack-lbaas20:55
*** neelashah has quit IRC20:56
*** doug-fish has quit IRC20:57
*** doug-fish has joined #openstack-lbaas20:57
*** neelashah has joined #openstack-lbaas21:03
kongjohnsom: pong21:07
openstackgerritmin wang proposed openstack/octavia: [WIP]: Octavia: Basic LoadBalancer Scenario Test
johnsomkong - I commented on the server group bug you posted.  Does that make sense?21:09
kongjohnsom: let me see, I just arrive in office :-)21:09
*** fawadkhaliq has quit IRC21:10
*** fawadkhaliq has joined #openstack-lbaas21:10
*** fawadkhaliq has quit IRC21:11
*** fawadkhaliq has joined #openstack-lbaas21:12
kongjohnsom: I know the usage of server group in active/standby loadbalancer implementation. What I mean is, why not we just create a 'global' server group with anti-affinity, to be used for every active/standby loadbalancer?21:12
kongI didn't see there is any difference among those server groups, except the name21:12
johnsomkong - Then nova would try to place all of the amphora on separate nodes.21:12
johnsomnova does anti-affinity for each instance inside a group, so two lbs would mean you would have to have four nodes21:13
kongjohnsom: ooh, I know what you mean21:14
kongmake sense to me now21:14
kongjohnsom: sorry for the misunderstanding21:14
johnsomOk, cool.21:14
johnsomI just wanted to make sure there wasn't another concern.21:15
kongjohnsom: yeah, we are on the same page21:15
kongjohnsom: btw, do you think it is the right way that i use 'neutron port-update' to trigger failover process?21:16
*** ajmiller has quit IRC21:16
*** ajmiller has joined #openstack-lbaas21:16
johnsomYes, that is how we force a failover, admin-state-up=False on the mgmt port21:16
kongI'm preparing for a script for our operators, aiming to upgrade the vm kernel (or patching) of the amp vms, I need to trigger failover manually21:17
johnsomYeah, it works well.  The timing is based on your health timeout settings, but works well for us21:19
johnsomThere are other ways, but they either require ssh access on the amp or DB access.  Neither of which I recommend21:20
kongjohnsom:  I just wonder why it takes so long for pluging vip in the vm?21:20
johnsomWhen we create the operator API I am sure that feature will be high on list to implement21:20
johnsomYeah, it does seem to take some time.  I noticed that myself when I was testing the namespace patch21:21
kongjohnsom: I didn't take a look at what plugging vip actually does inside the vm, can you give me some hints?21:22
johnsomWell, most of the time is on the neutron/instance side from what I saw.21:23
johnsomInside the VM, we just need the kernel to see the hot-plugged int, then we write out the interface file, update haproxy.conf, and ifup the interface.21:24
johnsomThe actions inside the vm seem to happen pretty quick21:24
*** rtheis has quit IRC21:24
kongmaybe need some debugging there :-)21:25
johnsomWith the namespaces, there is one more step of moving the interface into the namespace21:26
*** fawadkhaliq has quit IRC21:26
*** fawadkhaliq has joined #openstack-lbaas21:26
kongalthough patching will not be happened frequently, we still want the failove procedure to be fast, which is good for operators.21:26
kongyou konw, they always lost their temper when waiting :-)21:27
johnsomYes, if fast is what you need, use active/standby.  You can tune that down to a second or two21:27
kongyeah, that's our plan21:27
johnsomOur default settings for act/stndby failover time are a "generous".  If you have good hardware you can really tune those numbers down and have a very fast failover21:28
kongjohnsom: if I mark down the master amp, the slave one will take the master role. and what if the master amp comes back? Will it take the master role back?21:29
johnsomkong no, it will become backup21:29
johnsomWe don't want to failover unless we need it to21:30
kongjohnsom: yes, we just use the machanism for patching amp vms21:30
johnsomSo, in the patching case, you would down the mgmt interface on one of the amps21:31
johnsomWait for failover, then down the other21:31
kongjohnsom: ok, thanks for the confirmation.21:32
kongI think patching will be a common need for octavia operators, I can make it upstream after I finish the script and it's stable enough21:32
johnsomI'm booting an amp now, I can tell you how long the "inside vm" process takes on an amp network plug.21:32
kongjohnsom: I tested that many time yesterday :-)21:33
kongbut still glad to see your result21:33
kongjohnsom: btw, I personally really hope this patch will be merged ASAP
johnsomOk, on the amp, it's under a second.21:35
kongjohnsom: hmm....21:35
johnsomIf you include the neutron side, it's 16 seconds21:35
johnsomYeah, we have a bit of a lack of "stable" cores21:36
johnsomI want to have that one and the namespace patch merged, then I will cut another mitaka release for octavia21:37
kongjohnsom: sounds nice!21:37
johnsomBTW, those numbers above include the namespace patch21:39
kongjohnsom: cool. so what other time is spent on?21:40
kongon my lastest test, according to the log of o-cs21:41
johnsomIt's the neutron side with plugging the port.  I assume some of it is OVS time21:41
kongfrom 2016-04-11 09:48:01.621 to 2016-04-11 10:10:46.16421:41
johnsomWow, was that a failover or just a network plug?21:41
kongTask 'octavia.controller.worker.tasks.amphora_driver_tasks.AmphoraPostVIPPlug21:41
johnsomThat must be a failover with no spares pool21:42
kongjohnsom: no, just post vip plug21:42
johnsomWhat???  no21:42
kongjohnsom: I've no idea why it's so slow21:42
johnsom22 minutes?21:42
kongI install devstack using vagrant21:43
johnsomOh, ok, so that is boot time.  Yeah, so you are failing over with no spares pool, so it is waiting for the OS to finish booting.  You must not have virtualization available.  Running in virtual box?21:44
kongjohnsom: yes21:44
kongjohnsom: I will attend our standup meeting, will talk to you later21:44
*** doug-fish has quit IRC21:45
johnsomYeah, virtual box doesn't expose virtualization to the guest, so qemu drops to super slow software emulation21:45
*** matt-borland has quit IRC21:48
kongjohnsom: ok, so, it will be much faster in physical deployment, right21:55
johnsomYes.  On by lowly workstation, that step that takes you ~20 minutes, runs in about 37 seconds21:56
kongjohnsom: ok, no questions from me for now, need more digging into the source code and more testing21:57
kongjohnsom: btw, what's your time zone?21:57
johnsomWest coast US21:57
kongjohnsom: ok, I remember that :-)21:58
johnsomIn VMware you can enabled vt-x pass through.  There is also steps to enable it on kvm:
johnsomNo solution for virtual box.  It is just going to be slow for now21:58
*** piet has quit IRC21:59
kongjohnsom: never mind, deployment in virtual box is just for testing, at least , the function works well21:59
*** reedip__ has joined #openstack-lbaas22:07
*** ihrachys has quit IRC22:08
*** mixos has quit IRC22:09
*** alejandrito has quit IRC22:11
*** neelashah has quit IRC22:11
*** _ducttape_ has quit IRC22:19
*** reedip__ has quit IRC22:20
*** mixos has joined #openstack-lbaas22:31
*** mixos has quit IRC22:32
*** mixos has joined #openstack-lbaas22:34
openstackgerritmin wang proposed openstack/octavia: [WIP]: Octavia: Basic LoadBalancer Scenario Test
*** mixos has quit IRC22:34
*** mixos has joined #openstack-lbaas22:41
*** mixos has quit IRC22:44
*** fawadkhaliq has quit IRC22:51
*** fawadkhaliq has joined #openstack-lbaas22:51
*** BjoernT has quit IRC23:02
*** sbalukoff has quit IRC23:09
*** sbalukoff has joined #openstack-lbaas23:33
openstackgerritAishwarya Thangappa proposed openstack/octavia: Adds a new feature to limit the amphora build rate
*** mixos has joined #openstack-lbaas23:49

Generated by 2.14.0 by Marius Gedminas - find it at!