Saturday, 2015-10-24

*** diogogmt has quit IRC00:01
*** vivek-ebay has joined #openstack-lbaas00:01
*** diogogmt has joined #openstack-lbaas00:02
*** jongchoi has quit IRC00:02
*** vivek-ebay has quit IRC00:06
openstackgerritOpenStack Proposal Bot proposed openstack/octavia: Updated from global requirements
openstackgerritCarlos Garza proposed openstack/octavia: Implementing EventStreamer
*** bharathm has quit IRC01:03
*** KunalGandhi has quit IRC01:13
*** crc32 has quit IRC01:29
*** blogan_ has quit IRC01:32
*** bank has quit IRC01:41
*** alejandrito has quit IRC01:46
*** diogogmt_ has joined #openstack-lbaas01:53
*** diogogmt has quit IRC01:54
*** diogogmt_ is now known as diogogmt01:54
*** vivek-ebay has joined #openstack-lbaas02:04
*** vivek-ebay has quit IRC02:08
rm_workblogan: you there?02:28
rm_workffff us texans are f'ed02:29
rm_workflooding due to hurricane02:29
rm_workwrecking a ton of flights tomorrow02:29
sbalukoffOh geez! that sucks!02:40
rm_workwoking through options with airline rep now <_<02:41
sbalukoffGood luck!02:41
rm_worki think blogan's flight may be affected too but not sure02:41
sbalukoffjohnsom: What are the login credentials for that virtual machine?03:07
*** sbalukoff has quit IRC03:21
*** minwang2 has joined #openstack-lbaas03:29
*** sbalukoff has joined #openstack-lbaas03:31
*** bharathm has joined #openstack-lbaas03:39
*** woodster_ has quit IRC03:39
*** bharathm has quit IRC03:49
*** armax has quit IRC03:54
*** minwang2 has quit IRC04:13
*** minwang2 has joined #openstack-lbaas04:19
rm_workit was ubuntu:ubuntu04:19
*** diogogmt has quit IRC04:21
rm_workjohnsom: so it resumes, but there's no nova instances -- did you mean to set up the two lb backend nodes like were setup before?04:46
rm_workoh nm i see them04:46
rm_workthey're on admin now, instead of demo04:46
rm_worknot sure if that's what you meant to do04:46
rm_workcool, well, did the whole demo as "admin" instead of the actual demo user04:51
rm_workbut it works04:52
rm_workso that's cool :P04:52
rm_worknot sure if we want to try to fix that04:52
rm_workjohnsom: is the address in the "simulate a failover section actually right?
rm_workI feel like that's ... not related anymore04:58
*** yamamoto has quit IRC05:09
*** vivek-ebay has joined #openstack-lbaas05:13
*** minwang2 has quit IRC05:20
*** minwang2 has joined #openstack-lbaas05:24
*** jongchoi has joined #openstack-lbaas05:33
sbalukoffrm_work: Hmmm... the failover didn't work for me.05:43
sbalukoffI'm not sure what I'm doing wrong there...05:43
sbalukoffAnd yes, I ran the whole demo as the admin user as well.05:44
sbalukoffAlso, is it a known thing that a devstack doesn't survive a reboot? (ie. I'm guessing you're supposed to restack after every restart of the machine?)05:45
sbalukoffI've noticed I have to do this every time on other devstacks I've used... I'm just wondering whether I'm doing something wrong. :/05:45
sbalukoffIn any case, it's nice to have this saved VMware image--  it comes back to a "restored" state pretty quickly, and you've got everything already running, ready to do the demo.05:46
sbalukoffOther than the failover which didn't work for me... I'm feeling pretty confident about our lab. Good work there, johnsom, eh!05:46
*** minwang2 has quit IRC05:46
*** alejandrito has joined #openstack-lbaas05:48
sbalukoffAlso learned today: Do not attempt to run 12+ GB of virtual machines on a laptop with 16GB ram while also running a browser with 50 tabs open. And stuff.05:48
openstackgerritMerged openstack/neutron-lbaas: Adds a Barbican option to local.conf
johnsomSorry, took a break.  I didn't realize xgerman did demo account. I did everything as admin like I usually do.06:08
johnsom172.24.4.1:5555 is the "hack" work around for the controller IP.  It isn't the right answer, but works for now06:08
johnsomReboots always break devstack for me.  I always have to restack.  So, no answers there06:09
johnsomSurprised failover doesn't work.  Did you make sure to set the controller IP setting and restart cw and hm?06:10
johnsomCome to think about it. might not be the right answer with this image...06:13
johnsomWill have to look at that tomorrow06:14
johnsomThe static IP may have changed that.06:14
*** vivek-ebay has quit IRC06:21
*** yamamoto has joined #openstack-lbaas06:22
sbalukoffAah-- I only restarted cw, not hm.06:39
sbalukoffThat would probably be why it didn't work for me.06:39
sbalukoffSo, I think the instructions only say to restart cw. Should probably add hm in there as well.06:39
*** yamamoto has quit IRC06:49
*** bharathm has joined #openstack-lbaas06:50
*** yamamoto has joined #openstack-lbaas06:51
*** bharathm has quit IRC06:54
*** yamamoto has quit IRC07:12
*** yamamoto has joined #openstack-lbaas07:13
*** yamamoto has quit IRC07:24
*** alejandrito has quit IRC07:30
*** jongchoi has quit IRC07:39
*** jongchoi has joined #openstack-lbaas07:43
*** denghuiH has joined #openstack-lbaas08:06
*** alejandrito has joined #openstack-lbaas09:06
*** jongchoi has quit IRC09:35
*** yamamoto has joined #openstack-lbaas09:53
*** jongchoi has joined #openstack-lbaas10:43
*** alejandrito has quit IRC10:52
*** yamamoto has quit IRC11:26
*** jongchoi has quit IRC12:05
*** jongchoi has joined #openstack-lbaas12:13
*** orionx has quit IRC12:27
*** jongchoi has quit IRC12:40
*** denghuiH has quit IRC12:44
*** diogogmt has joined #openstack-lbaas12:58
*** jongchoi has joined #openstack-lbaas13:03
*** yamamoto has joined #openstack-lbaas13:08
*** yamamoto has quit IRC13:13
*** chlong has joined #openstack-lbaas13:29
*** diogogmt has quit IRC13:34
*** jongchoi has quit IRC13:46
*** yamamoto has joined #openstack-lbaas13:49
*** jongchoi has joined #openstack-lbaas13:49
*** diogogmt has joined #openstack-lbaas13:52
*** yamamoto has quit IRC13:54
*** jongchoi has quit IRC14:03
*** yamamoto has joined #openstack-lbaas14:04
*** alejandrito has joined #openstack-lbaas14:09
*** apuimedo has joined #openstack-lbaas15:05
*** alejandrito has quit IRC15:08
*** orionx has joined #openstack-lbaas15:29
*** orionx has quit IRC15:33
*** orionx has joined #openstack-lbaas15:34
*** yamamoto has quit IRC15:48
*** armax has joined #openstack-lbaas16:01
rm_worksbalukoff / sbalukoff1 / johnsom: yeah not sure if that address is right since the image IPs all changed, there's no 172.* on that box at all16:02
rm_workalso, on a reboot you should be able to just use rejoin-stack.sh16:03
rm_worknot actually restack16:03
rm_workit just starts everything again quickly16:03
rm_workhas "worked" for me16:03
rm_workyou have to manually start any nova VMs tho16:03
*** vivek-ebay has joined #openstack-lbaas16:15
*** diogogmt has quit IRC16:20
*** zigo has quit IRC16:38
*** zigo has joined #openstack-lbaas16:38
*** armax has quit IRC16:43
*** minwang2 has joined #openstack-lbaas16:44
*** yamamoto has joined #openstack-lbaas16:49
*** zigo has quit IRC16:49
*** zigo has joined #openstack-lbaas16:51
*** yamamoto has quit IRC16:54
johnsomI never get rejoin to actually work.  But maybe that is fixed now.  The processes are going but the bridge interfaces are scrambled17:19
johnsomThe address should be the gateway on the public net I think17:20
johnsomI will give it a go later and find out17:20
*** diogogmt has joined #openstack-lbaas17:29
*** diogogmt has quit IRC17:48
*** diogogmt has joined #openstack-lbaas17:50
*** minwang2 has quit IRC18:10
*** vivek-ebay has quit IRC18:11
*** vivek-ebay has joined #openstack-lbaas18:12
*** vivek-ebay has quit IRC18:17
*** johnsom_ has joined #openstack-lbaas18:27
*** johnsom has quit IRC18:29
rm_workjohnsom_: if you used HOST_IP= nothing gets scrambled because the addresses can't change :P18:44
rm_workjohnsom_: that is the primary benefit of doing it that way, I believe18:44
rm_workthe bridge interfaces for nova and neutron SHOULDN'T break, but it's possible I guess (might explain some of the wonky stuff we were seeing on German's image actually)18:45
*** apuimedo has quit IRC18:48
johnsom_It's the neutron interfaces that broke for me.  It's not that the host IP changes, it's static, it's that the interfaces don't all get put into the files for startup18:53
*** johnsom_ is now known as johnsom18:55
johnsomrm_work The public net is still so should still be correct19:02
johnsomIt is on br-ex19:03
rm_workhmm k19:03
rm_worki don't see it19:04
johnsomneutron net-list19:04
rm_worki did:19:04
johnsomWhat is your public lan19:04
rm_workifconfig | grep 17219:04
rm_workand nothing19:04
rm_workdoes it not show up as a bound interface?19:04
johnsomI have it on mine19:05
johnsomsudo ifconfig br-ex gives me the address19:05
johnsomIf you rebooted you may have lost that interface19:06
*** bank has joined #openstack-lbaas19:09
rm_worki did not...19:16
rm_workoh, no nevermind, it's showing up now wtf19:17
rm_workk, well19:17
rm_workmaybe i was just hallucinating last night19:17
rm_workor typo'd somethinf19:17
rm_workjohnsom: oh you will need to have them do19:18
johnsomOr forgot the sudo19:18
rm_worksudo chmod go+rw `tty`19:18
rm_worknah the interfaces don't change whether you sudo or not19:18
johnsomYeah, xgerman put the "script /dev/null" trick in the slides19:18
rm_workjust have to run it from /sbin/ifconfig explicitly19:18
rm_workthat is... less optimal19:18
johnsomYeah, but if you don't sudo ifconfig gives an error so the grep will come up empty19:18
rm_workthan just fixing the tty perms19:18
rm_workjohnsom: it ... doesn't?19:19
rm_workifconfig runs fine as any user19:19
rm_workit's just a debianism that it's in sbin19:19
johnsomIt's not in the path for stack user I think19:19
rm_workno actual permissions issue19:19
rm_workyou just have to run it from /sbin explicitly19:19
johnsomI usually use the chmod trick, this was the first I had heard of the script trick19:19
rm_workyeah the script trick has weird side effects tho19:19
johnsomSo, failover works.  If you don't nova delete the instance.19:20
johnsomIt looks like the failover flow reverts if the instance was deleted.  It tries to delete it and fails19:20
rm_workabout to test it19:20
rm_workjust restarted o-cw and o-hm19:20
johnsomThe steps are: update octavia.conf with controller IP/port, restart cm and hm, create a new load balancer19:21
rm_workdoing that19:21
johnsomThen stop the haproxy on the amp, wait 10 or 30 seconds, whatever is in the conf19:21
johnsomI just opened for that instance delete issue19:22
openstackLaunchpad bug 1509706 in octavia "Failover fails when instance is deleted" [Critical,New]19:22
xgermanin airport19:23
rm_workxgerman: >_>19:24
rm_workxgerman: at home waiting for rescheduled flight tomorrow at 6am >_>19:24
xgermancharging all my devices again...19:25
johnsomso, since we can't just delete the instance, we should stop the amphora-agent19:25
johnsomSince haproxy isn't running until a listener is configured19:26
rm_workhow long does it take to fail over19:26
johnsomThen you can stop the haproxy instance19:26
*** woodster_ has joined #openstack-lbaas19:26
rm_workok there it went19:26
rm_workit took a sec19:26
johnsomYeah, the default config as a long interval19:27
rm_workbut it worked19:27
rm_workcool k19:27
rm_workthis is using the UDP heartbeat stuff? :P19:27
rm_worknice to see it actually *working*19:27
rm_workok, i'll start imaging this tonight19:27
rm_workor maybe right now19:27
johnsomCool!  Thanks Adam!19:28
rm_workso many drives to do, and they are soooo slow >_>19:28
xgermannow you have more time...19:28
rm_worki still haven't looked at the troubleshooting section, hopefully someone has filled in something T_T19:28
xgermanI did19:28
rm_workwell, i WAS planning to do it on the plane :P19:28
xgermanwe are good19:28
rm_workkk cool19:28
rm_workhonestly I feel like it'd be more useful just to have a couple of us wandering around19:29
johnsomxgerman fyi, I updated the slide to include restarting the o-hm19:29
rm_workhelping people do the commands to unpack everything, get stuff up and running, etc19:29
rm_workso the speaker can just keep doing the speaker thing and not be interrupted19:29
johnsomYeah, I think we will have as many of us as people19:29
xgermanjohnsom thanks — worked for me without but who knows...19:29
rm_workheh yeah19:30
rm_workseems that way19:30
johnsomthe o-hm might only be needed to failover the failover instance19:32
johnsomOk, I am going to stop monkeying with the image, it seems to be in good shape.19:33
johnsomI wrote some slick scripts for the demo last night, so I need to make the recordings19:34
*** bank has quit IRC19:35
*** bank has joined #openstack-lbaas19:35
rm_workheh k19:59
rm_workimaging the first set19:59
rm_workthen will test one of them19:59
rm_worklooks like ~60m each to write19:59
rm_workdoing 8 at a time <_<19:59
rm_worki love tmux pane-sync20:00
*** ducttape_ has joined #openstack-lbaas20:38
*** ducttape_ has quit IRC20:42
*** armax has joined #openstack-lbaas20:47
johnsomOk, I have my demo videos and I have cut the presentation to clean up on the plane21:02
*** vivek-ebay has joined #openstack-lbaas21:14
*** armax has quit IRC21:49
*** vivek-ebay has quit IRC21:52
*** openstackgerrit has quit IRC22:16
*** openstackgerrit has joined #openstack-lbaas22:17
johnsomInteresting, NTT has a talk where they say they need UDP in LBaaS22:19
*** diogogmt has quit IRC22:21
*** ducttape_ has joined #openstack-lbaas22:29
*** ducttape_ has quit IRC22:48
*** bank has quit IRC22:50
*** bank has joined #openstack-lbaas22:54
*** bank has quit IRC23:14
*** harlowja has joined #openstack-lbaas23:26
*** harlowja has quit IRC23:26
*** openstack has joined #openstack-lbaas23:33
*** bank has joined #openstack-lbaas23:42

Generated by 2.14.0 by Marius Gedminas - find it at!