Friday, 2018-01-26

« Thursday, 2018-01-25 Index Saturday, 2018-01-27 »
*** yamamoto has quit IRC00:00
*** yamamoto has joined #openstack-lbaas00:01
*** yamamoto has quit IRC00:01
*** yamamoto has joined #openstack-lbaas00:01
*** yamamoto has quit IRC00:06
*** yamamoto has joined #openstack-lbaas00:07
*** rstarmer has quit IRC00:15
*** armax has quit IRC00:19
*** Alex_Staf has quit IRC00:35
openstackgerritZhaoBo proposed openstack/octavia master: WIP:UDP for [2]  https://review.openstack.org/52965100:55
*** armax has joined #openstack-lbaas01:09
*** annp has joined #openstack-lbaas01:40
openstackgerritZhaoBo proposed openstack/octavia master: UDP support with netcat - DIB  https://review.openstack.org/53810501:40
*** armax has quit IRC01:53
*** dayou has quit IRC02:08
*** harlowja has quit IRC02:16
*** rstarmer has joined #openstack-lbaas02:16
*** dayou has joined #openstack-lbaas02:51
openstackgerritJeffrey Zhang proposed openstack/octavia master: Move to boot from volume rather than boot from image  https://review.openstack.org/53812403:04
*** dougwig has quit IRC03:31
*** aojea has joined #openstack-lbaas04:12
*** harlowja has joined #openstack-lbaas04:51
*** openstackstatus has quit IRC04:56
*** openstackstatus has joined #openstack-lbaas04:57
*** ChanServ sets mode: +v openstackstatus04:57
*** aojea has quit IRC05:19
*** aojea has joined #openstack-lbaas05:33
*** aojea has quit IRC05:38
*** aojea has joined #openstack-lbaas05:40
*** links has joined #openstack-lbaas05:58
*** LutzB has quit IRC06:12
*** LutzB has joined #openstack-lbaas06:17
*** gcheresh has joined #openstack-lbaas06:18
*** aojea has quit IRC06:24
*** kobis has joined #openstack-lbaas06:26
openstackgerritOpenStack Proposal Bot proposed openstack/octavia-dashboard master: Imported Translations from Zanata  https://review.openstack.org/53776506:40
*** kobis has quit IRC06:45
*** eandersson has joined #openstack-lbaas06:49
*** gcheresh has quit IRC06:51
*** kobis has joined #openstack-lbaas06:52
*** kobis has quit IRC06:53
*** kobis has joined #openstack-lbaas06:57
*** aojea_ has joined #openstack-lbaas07:00
*** kobis has quit IRC07:03
*** rstarmer has quit IRC07:20
*** harlowja has quit IRC07:23
*** gcheresh has joined #openstack-lbaas08:35
*** AlexeyAbashkin has joined #openstack-lbaas08:39
openstackgerritBar Elharar proposed openstack/octavia master: Make new-octavia-devstack.sh OS agnostic  https://review.openstack.org/53523108:55
*** Alex_Staf has joined #openstack-lbaas09:00
*** yedongcan has joined #openstack-lbaas09:02
*** aojea has joined #openstack-lbaas09:24
*** aojea_ has quit IRC09:26
*** Alex_Staf has quit IRC09:28
*** Alex_Staf has joined #openstack-lbaas09:57
*** kobis has joined #openstack-lbaas09:58
*** aojea has quit IRC10:02
*** kobis has quit IRC10:04
*** salmankhan has joined #openstack-lbaas10:11
*** aojea has joined #openstack-lbaas10:12
*** Alex_Staf has quit IRC10:16
*** aojea has quit IRC10:17
*** yamamoto has quit IRC10:30
*** rm_work has quit IRC10:42
*** rm_work has joined #openstack-lbaas10:46
cgoncalvesrm_work: I would be very interested in reviewing your patch that adds image id to the amphorae table :)10:48
*** kobis has joined #openstack-lbaas10:51
openstackgerritAkihiro Motoki proposed openstack/neutron-lbaas-dashboard master: i18n: Do not include html directives in translation strings  https://review.openstack.org/53821410:51
*** kobis has quit IRC10:53
*** yamamoto has joined #openstack-lbaas10:53
*** yedongcan has left #openstack-lbaas10:53
*** kobis has joined #openstack-lbaas10:54
*** yamamoto has quit IRC10:58
*** yamamoto has joined #openstack-lbaas11:01
*** yamamoto has quit IRC11:01
*** srihas has joined #openstack-lbaas11:20
srihashi, I am getting an error in haproxy.log -> Server ce539d6c-506a-4105-bcb8-25f7a661aec3/984b3254-fec8-40ec-933c-1d145511:22
srihas2cc635 is DOWN, reason: Layer4 connection problem, info: "No route to host", check duration: 3000ms. 0 active and 0 bac11:22
srihaskup servers left. 0 sessions active, 0 requeued, 0 remaining in queue.11:23
srihascan some one help me with the approach to solve this challenge?11:23
*** kobis has quit IRC11:30
*** salmankhan has quit IRC11:31
*** salmankhan has joined #openstack-lbaas11:35
*** Alex_Staf has joined #openstack-lbaas11:36
*** kobis has joined #openstack-lbaas11:39
*** aojea has joined #openstack-lbaas11:53
*** kobis has quit IRC11:54
*** b_bezak has joined #openstack-lbaas12:01
*** yamamoto has joined #openstack-lbaas12:02
*** kobis has joined #openstack-lbaas12:04
*** aojea has quit IRC12:09
*** aojea has joined #openstack-lbaas12:10
*** Alex_Staf has quit IRC12:11
*** yamamoto has quit IRC12:13
*** annp has quit IRC12:21
*** kobis has quit IRC12:34
*** kobis has joined #openstack-lbaas12:37
*** kobis has quit IRC12:38
*** dmellado has joined #openstack-lbaas13:58
*** aojea_ has joined #openstack-lbaas14:07
*** aojea has quit IRC14:11
*** yamamoto has joined #openstack-lbaas14:38
*** yamamoto has quit IRC14:39
*** yamamoto has joined #openstack-lbaas14:41
*** Alex_Staf has joined #openstack-lbaas14:47
*** tesseract has joined #openstack-lbaas14:49
xgerman_o/14:58
dayouo/15:09
dayoujohnsom can help, but he might be too tired to get up15:09
srihasdayou: its ok. is he in a different timezone?15:12
openstackgerritGuoqiang Ding proposed openstack/neutron-lbaas master: Choose correct log option by listener protocol  https://review.openstack.org/45361215:17
*** Alex_Staf has quit IRC15:23
dayousrihas: yes, he is pacific time if I remember correctly15:27
srihasdayou: ok :)15:29
*** fnaval has joined #openstack-lbaas15:57
*** slaweq_ has joined #openstack-lbaas16:00
*** armax has joined #openstack-lbaas16:00
johnsomsrihas Hi.  Which driver are you using? I'm guessing the old namespace haproxy driver.16:02
*** links has quit IRC16:05
*** AlexeyAbashkin has quit IRC16:06
*** salmankhan has quit IRC16:09
*** fnaval has quit IRC16:09
*** fnaval_ has joined #openstack-lbaas16:09
*** slaweq_ has quit IRC16:10
*** salmankhan has joined #openstack-lbaas16:11
*** dougwig has joined #openstack-lbaas16:14
*** atoth has joined #openstack-lbaas16:22
*** armax has quit IRC16:44
*** armax has joined #openstack-lbaas16:46
*** bbzhao has joined #openstack-lbaas17:03
*** 7JTADEBX8 has quit IRC17:05
*** harlowja has joined #openstack-lbaas17:10
*** salmankhan has quit IRC17:31
*** salmankhan has joined #openstack-lbaas17:31
*** tesseract has quit IRC17:35
*** AlexeyAbashkin has joined #openstack-lbaas17:40
*** AlexeyAbashkin has quit IRC17:45
*** slaweq_ has joined #openstack-lbaas17:49
*** b_bezak_ has joined #openstack-lbaas17:50
*** b_bezak has quit IRC17:51
*** b_bezak_ has quit IRC17:55
*** yamamoto has quit IRC17:58
*** links has joined #openstack-lbaas18:02
*** AlexeyAbashkin has joined #openstack-lbaas18:09
*** AlexeyAbashkin has quit IRC18:14
*** yamamoto has joined #openstack-lbaas18:19
*** links has quit IRC18:32
*** armax_ has joined #openstack-lbaas18:37
*** aojea_ has quit IRC18:39
*** armax has quit IRC18:40
*** armax_ is now known as armax18:40
openstackgerritMichael Johnson proposed openstack/octavia master: Add SDK list to octavia user docs  https://review.openstack.org/53832318:48
openstackgerritMichael Johnson proposed openstack/octavia master: Add SDK list to octavia user docs  https://review.openstack.org/53832318:49
*** salmankhan has quit IRC18:50
*** yamamoto has quit IRC18:52
*** atoth has quit IRC18:53
*** rstarmer has joined #openstack-lbaas19:00
*** b_bezak has joined #openstack-lbaas19:04
*** b_bezak has quit IRC19:04
*** gcheresh has quit IRC19:04
*** slaweq_ has quit IRC19:05
*** b_bezak has joined #openstack-lbaas19:06
*** yamamoto has joined #openstack-lbaas19:07
*** yamamoto has quit IRC19:12
*** harlowja has quit IRC19:14
*** links has joined #openstack-lbaas19:21
*** b_bezak has quit IRC19:25
*** b_bezak has joined #openstack-lbaas19:30
*** b_bezak has quit IRC19:30
*** b_bezak has joined #openstack-lbaas19:31
rm_workcgoncalves: i'll see if i can whip that up today -- but as johnsom said, we're in feature freeze now so probably it can't merge19:34
rm_workthough it IS just a backend thing... and it's minor... and I COULD argue that it missing that information is a *bug*... ^_^19:34
johnsomNo19:34
johnsomGrin19:34
johnsomDB schema changes are very outside the guidelines....19:35
johnsomThat said, I'm sure it will merge quickly for rocky19:35
*** b_bezak has quit IRC19:35
* rm_work shrugs19:36
rm_workworth a shot19:36
cgoncalveslool :D19:45
*** harlowja has joined #openstack-lbaas19:46
*** harlowja_ has joined #openstack-lbaas19:58
*** harlowja has quit IRC20:00
*** slaweq has joined #openstack-lbaas20:08
*** yamamoto has joined #openstack-lbaas20:15
xgerman_rm_work if you however get the info straight from nova…20:22
johnsomIt would still be an API change, so, nice try, but no20:23
*** yamamoto has quit IRC20:32
*** atoth has joined #openstack-lbaas20:32
*** links has quit IRC20:43
*** atoth has quit IRC21:08
*** aojea has joined #openstack-lbaas21:11
*** slaweq has quit IRC21:22
*** slaweq has joined #openstack-lbaas21:22
*** gcheresh has joined #openstack-lbaas21:22
*** slaweq has quit IRC21:27
*** aojea has quit IRC21:35
*** aojea has joined #openstack-lbaas21:36
*** gcheresh has quit IRC21:39
*** rstarmer has quit IRC21:42
*** aojea has quit IRC21:43
*** aojea has joined #openstack-lbaas21:44
*** slaweq has joined #openstack-lbaas21:44
*** salmankhan has joined #openstack-lbaas21:47
*** rstarmer has joined #openstack-lbaas21:49
*** salmankhan has quit IRC21:51
*** salmankhan has joined #openstack-lbaas22:06
*** aojea has quit IRC22:08
openstackgerritAdam Harwell proposed openstack/octavia-tempest-plugin master: WIP: Failover test  https://review.openstack.org/50155922:12
srihasjohnsom: LOADBALANCERV2:Haproxy:neutron_lbaas.drivers.haproxy.plugin_driver.HaproxyOnHostPluginDriver:default23:05
srihasI guess thats the old one23:05
johnsomOh, that is the even older one.23:06
srihasso the error cannot be fixed? :(23:07
johnsomWell, most of us do not use those drivers, so it is hard for us to help.23:08
srihasaha, which driver is recommended for newton ?23:08
*** slaweq has quit IRC23:09
johnsomOh, ok, they just aliased that old name to the namespace driver, so it's not as old as I thought.23:09
johnsomhttps://github.com/openstack/neutron-lbaas/blob/master/neutron_lbaas/drivers/haproxy/plugin_driver.py23:09
*** slaweq has joined #openstack-lbaas23:09
johnsomWe mostly use the Octavia driver23:10
johnsomIt brings a lot o benefits.23:10
johnsomsrihas I can *try* to help you out with that driver.23:10
srihasjohnsom: that sounds good to me :) whats step 1?23:11
johnsomOn your host with the lbaas agent, if you run "sudo ip netns" you should see a "lbaas" namespace. This is the network namespace that haproxy runs in.23:11
srihasyeah23:11
johnsomif you do "sudo ip netns exec <namespace name> ip a" it should list all of the interfaces and IP addresses configured in the namespace23:12
johnsomWhat subnet did you use for the VIP and what subnet did you use for the member?23:12
*** slaweq has quit IRC23:13
srihasI use 192.168.180.0/24 as the subnet23:16
srihasthat is where two VMs have a port each and LB has its port23:16
johnsomOk, so one armed LB, we would expect to see only one interface inside the network namespace.23:17
srihasyes23:17
srihas192.168.180.3023:18
johnsomOk, so that looks good so far.  When you curl your VIP address what do you get in a response?23:18
srihasone thing I observed is: I cant ping that LB ip from VMs23:20
johnsomYes, that is expected.  The security group on the VIP restricts that.23:20
srihasand curl from?23:20
johnsomYou should be able to curl the VIP IP and port and get a response however23:20
johnsomFrom the 192.168.180.0/24 network or through a floating IP assigned to the VIP23:21
srihascurl 192.168.180.3023:24
srihascurl: (7) Failed to connect to 192.168.180.30 port 80: No route to host23:24
srihasfrom a VM in 192 network ^23:25
johnsomHmm, ok can you paste your "netstat -rn"23:25
johnsomyou can use paste.openstack.org however, note that this is public23:25
srihasok23:26
johnsomIt will give a url you can share with us so we don't past a whole bunch of lines into IRC23:26
srihashttp://paste.openstack.org/show/654133/23:26
johnsomHa, well, that is simple23:27
johnsomInside the netns, can you do a "netstat -an | grep 80"? Also a "ps -ef| grep haproxy"23:29
srihasok23:29
johnsomWe would expect to see haproxy listening on port 8023:29
srihashttp://paste.openstack.org/show/654142/23:31
srihasyes its listening23:32
*** Bar_ has joined #openstack-lbaas23:34
openstackgerritBar Elharar proposed openstack/octavia master: Fail-proof VIP deallocation task  https://review.openstack.org/52393123:35
Bar_johnsom, do we want to kill it: https://review.openstack.org/#/c/523931 ?23:35
Bar_If so, I'll omit to try-except additions23:37
johnsomBar_ I haven't looked at that in a while.  It might be a good one to poke other cores for input.23:37
Bar_rm_work, xgerman_, would you care to vote?23:38
johnsomsrihas I see that the OS default haproxy is also running there. can we do a "netstat -anp | grep 80" (should have had you do that the first time, sorry)23:38
srihasyeah you are right23:39
srihasno need to appologise :)23:39
johnsomBar_ I am but one opinion in the octavia community... grin23:39
xgerman_added my opinion…23:44
Bar_johnsom, nonetheless, I'd rather have you on my side.23:44
*** fnaval_ has quit IRC23:44
srihasjohnsom: http://paste.openstack.org/show/654165/23:45
johnsomsrihas Ok, so far this is all looking good....23:46
srihasnice :)23:46
rm_workBar_: one hand, it will make it less likely to get stupid stuck ERROR/dead loadbalancers you can't delete23:47
rm_workon the other hand, that means we could end up with orphaned objects23:47
johnsomYou can delete them23:47
rm_work?23:47
rm_workI mean, the user could23:47
rm_workif they knew23:47
johnsomIt means you end up with orphaned VIP addresses, which is more evil IMO23:47
rm_worki mean23:48
rm_workorphaned objects like SGs or Ports in the cloud23:48
rm_worknot just DB entries23:48
johnsomYeah, but you can delete them in "ERROR"23:48
rm_workright, which then would go through and do proper cleanup right?23:48
rm_workso it's annoying a little for the user, but less orphans?23:48
johnsomRight, I don't care about orphan SGs really, I only really care about this patch as it can mean the VIP address didn't actually get freed23:48
rm_work:(23:49
rm_workyeah I guess23:49
rm_workorphan SGs are kinda... pointless23:49
rm_workthey don't really do anything23:49
rm_workor take space23:49
rm_workbut still, it feels dirty23:49
rm_worki kinda would rather have the Revert right now...23:49
johnsomMy point is, if we can't clean up/free the VIP port, that is worthy of exposing to the user in an ERROR status. As VIPs can be important to users23:49
xgerman_I have people yell at me when they see resources lay around23:50
rm_workthough since this is async on Deletes, couldn't we just... do a better job of cleaning up reliably?23:50
rm_workretrying?23:50
rm_workfix our ordering / delays?23:50
rm_worki dunno23:50
Bar_I could improve the job/flow with some retries23:50
rm_workis it possible to get in a state where there's seriously no way nova/neutron will let us clean up all our stuff?23:50
rm_workI would say we just do some retries in areas where we've not before23:51
xgerman_just had that — DB error in nova23:51
rm_workxgerman_: eugh23:51
rm_workyeah and really i HATE orphans23:51
rm_worklike23:51
rm_workREALLY hate orphans23:51
rm_work^^ please to be quoting out of context23:51
xgerman_I can see us putting orphanes into some housekeeping process and keep cleaning but…23:51
rm_workso I'd rather revert->ERROR23:51
johnsomRight, this is kind what "ERROR" is for.  It broke in a way we couldn't resolve (neutron is borked), so stop and mark in ERROR.23:52
rm_workunless we institute a whole orphan tracking process23:52
rm_worklike23:52
rm_workoctavia-orphanage23:52
johnsomThen if neutron comes back, the user can delete and it will clean up23:52
xgerman_houseeeping23:52
xgerman_rm_work, housekeeping can do that23:52
rm_workxgerman_: I would agree, but the naming is too wrong23:52
johnsomOr they can have an operator look in the logs and figure out the *right* way to clean it up23:52
rm_workhousekeepers don't deal with orphans23:52
rm_workwhat would they do with them? they just came over to clean23:52
rm_workthat's a job for an orphanage23:53
xgerman_ok, we should ERROR and have other people run orphanages23:53
rm_workI mean yeah, basically an orphanage could be passed an object system / type / ID, and be responsible for perioidically retrying a remove operation until it's successful or the object just doesn't exist for whatever reason23:53
xgerman_OpenStack Oraphanage23:54
rm_workthen an admin could check that list / get notifications, clean up the stuff, and the system would automatically remove it from tracking next time it checked23:54
xgerman_ok, brb23:54
johnsomUgh, that kind of goes against my "do no harm" mantra.  Things that magically clean up things, means we failed at our first job and will like break things in bad ways "magically".23:54
rm_workthat ... could be a nice project23:54
rm_workok, looking at registering OpenStack Orphanage project23:54
johnsomOye, so wrong23:55
rm_workAPI to give it all necessary data to track / remove objects of various types in the cloud23:55
rm_workand it can handle the tracking / retries / reporting23:55
rm_workwould actually be pretty useful23:55
rm_workIMO23:55
johnsomIS that what vitrage does? (RCA)23:55
rm_workis it?23:56
rm_workthat'd be cool23:56
johnsomhttps://governance.openstack.org/tc/reference/projects/vitrage.html23:56
srihasjohnsom: whats the next step?23:57
srihas(sorry for coming in middle)23:57
johnsomsrihas I'm thinking about that....  No, please jump in. You were here first.23:58
johnsomgrin23:58
srihas:)23:58
johnsomsrihas Let's try "ip route show table all | grep 192.168.180" inside the netns23:59
johnsomIs this a devstack?23:59
« Thursday, 2018-01-25 Index Saturday, 2018-01-27 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!