Friday, 2016-12-16

*** harlowja has quit IRC00:06
openstackgerritMerged openstack/octavia: Stop depending on tripleo-image-elements
johnsomSomething is still broken.  Amps are still failing over00:27
*** harlowja has joined #openstack-lbaas00:34
rm_workyeah DIB is just not respecting my settings00:43
rm_workcan't make it happen00:43
rm_workit *will not* use my repo00:43
rm_workprobably the heartbeat code has some py2 stuff00:45
johnsomYeah, I'm not sure.  It seems to be working ok locally00:45
johnsomIt has to be something the test is triggering00:46
*** zhouyaguo has quit IRC00:46
rm_workwell i still can't test locally >_<00:46
*** zhouyaguo has joined #openstack-lbaas00:49
johnsomI guess I could put an image somewhere for you00:50
rm_worki need to resolve this issue...00:53
bana_kwhy do we have default pool for the loadbalancer and what does it mean if we add one ? is it used only for l7 ?01:03
rm_workdefault pool is only for l701:03
bana_krm_work thanks !01:03
johnsomYes, a default pool is used when none of the l7 rules match01:03
bana_kgot it, thanks johnsom !01:04
*** fnaval has joined #openstack-lbaas01:05
openstackgerritMichael Johnson proposed openstack/octavia: Amphora-agent should log to a distinct location
johnsomThe health thread logging still seems to go to stdout....01:09
johnsomBut the api requests go to the new file01:09
rm_workhealth sender isn't run with gunicorn01:09
rm_workit's just a python process01:10
rm_workLOG = logging.getLogger(__name__)01:10
rm_workusing that01:10
rm_workwe split it off before we hit the gunicorn stuff01:10
johnsomSo, maybe we need another file for the health stuff.01:10
rm_workprobably wouldn't be a bad idea01:10
rm_workyou said removing your .cache helped somehow?01:11
rm_work ~/.cache/image-create/source-repositories/amphora_agent_022110a20994e6b763f32e9bbfd21a0fc6659d62/01:12
rm_workgit log there01:12
johnsomYeah, I usually end up with "missing reference" errors when stacking.  It seems to use the .cache version and not the local one01:12
rm_workshows the right commit01:12
rm_workbut it STILL isn't on the amp01:12
johnsomWant to post your devstack log?01:13
rm_worksure... minute01:13
*** fnaval has quit IRC01:15
*** fnaval has joined #openstack-lbaas01:15
*** kevo has quit IRC01:19
*** fnaval has quit IRC01:19
*** bana_k has quit IRC01:21
johnsomrm_work The git -C is failing, so devstack is setting your ref to blank01:21
rm_workgah wait01:21
rm_workOH the plugin is doing it01:21
johnsomd /opt/stack/octavia/devstack/ [m git -C /opt/stack/octavia log -1 --pretty=format:%H01:21
rm_workbecause i have it set in my bash_profile01:21
johnsomUnknown option: -C01:21
rm_worki had to change the command01:21
johnsomexport DIB_REPOREF_amphora_agent=01:21
johnsomCloning from amphora-agent cache and applying ref master01:24
johnsomNote: checking out 'd2c0c16574cd93ee21971a1b95e6cedcde78ce66'.01:24
johnsomHEAD is now at 547d03a Merge "Add support for diskimage-builder tracing"01:24
johnsomUpdating cache of /opt/stack/octavia in /opt/stack/.cache/image-create/source-repositories/amphora_agent_022110a20994e6b763f32e9bbfd21a0fc6659d62 with ref master01:25
johnsomYeah, so it's definitely getting the wrong ref01:25
*** harlowja has quit IRC01:25
xgermanare the ref’s still controlled in that element file?01:30
johnsomYes and no.01:31
johnsomThere is a default in the element, go pull master01:31
johnsomHowever, there is also an override01:31
johnsomThe two environment variables adam is trying to set.01:31
johnsomThe devstack sets those to use /opt/stack/octavia and the last commit01:32
xgermanah, that’s changed then from the last time I ran it01:33
johnsomYeah, likely01:33
rm_workok so i need to either: fix git... or01:35
rm_workremove that line from our plugin01:35
rm_workbecause i've already set it01:35
johnsomRemoving the line will get real old.01:35
johnsomWhat is up with git?  centos have some strange version?01:35
rm_workyeah it's ooooool01:35
rm_workneed to figure out how to update it01:36
rm_workbut i'm not a centos guy01:36
rm_workwhat version of git is on your box01:36
rm_workcentos is 1.801:36
rm_workmy system (osx) is 2.9.301:36
rm_worki can script this hack along with all the rest of the hacks i've scripted....01:37
rm_workPackage git- already installed and latest version01:38
johnsomgit version 1.9.1 on 14.0401:39
rm_workthere SHOULD be a newer version in some repo .....01:41
xgermanso repel and elrepo don’t have the new stuff?01:41
rm_worklooks like not?01:41
rm_worki don't even know how to tell01:41
xgermanyeah, my centos days are well in the past01:41
rm_workwhat's the equiv command to "apt-cache policy git"01:42
korean101hi guys.01:42
korean101should i only use auth_version 2?01:43
korean101not v3?01:43
korean101(in octavia)01:43
rm_worki really don't like rpm distros T_T01:43
korean101newton release01:43
xgermanrm_work: googling for fun and profit:
rm_workkorean101: it depends on your keystone setup i think01:43
xgermansee how they disable repos01:44
chenhbhi,rm_work. How is this problem about installing octavia with devstack? We noted that the error information has changed01:44
korean101rm_work: Does not Octavia force to use only auth v2?01:44
rm_workI believe we should support v2 or v301:45
korean101rm_work: many thanks01:45
johnsomkorean101 We have supported both v2 and v3.01:45
*** yamamoto has joined #openstack-lbaas01:45
xgermanyep, that is my take, too01:46
rm_workgot it via some random dude's repo01:46
rm_workscripting this...01:46
johnsomI am really suprised that centos 7 would have that old of version01:46
johnsomSomeone reported a but about that, but they wanted to change the script in a not so good way if I remember01:47
korean101and anyone knows this message? (Driver error: HTTPConnectionPool(host='', port=9876): Max retries exceeded with url: /v1/loadbalancers/ef290ebf-691c-4ffc-ba36-b4699fbe0d84)01:47
korean101i don't know why /v1/loadbalancers01:47
johnsomYes, that came in new.  It's a bug01:48
korean101johnsom: to me?01:48
korean101korean101: OMG01:48
korean101johnsom: but i succeed 'lbaas-loadbalancer-create' in (devstack stable/newton)01:49
johnsomIt snuck in during this DIB issue01:49
johnsomYes, the command succeeds, but it is telling you it is still in PENDING_CREATE.  You have to check the status to see if it finished01:50
korean101johnsom: OMG01:50
korean101johnsom: yes. i got "provisioning_status | PENDING_CREATE"01:51
johnsomThe API is asynchronus01:51
korean101johnsom: snuck -> stuck?01:51
johnsomYes, for now.  It will eventually go to ERROR01:52
rm_workok, fixed my scripts, this has cost me a day... i feel dumb for not catching it in the devstack log though01:52
rm_workHOPEFULLY this should now work01:52
*** yamamoto has quit IRC01:53
johnsomkorean101 Try checking out this version of octavia:
johnsomIt has most of the fixes and seems to be working ok.01:54
johnsomThere is still a bug we are trying to track down01:54
korean101johnsom: very appreciate01:57
johnsomSure, sorry for all this trouble.  This has been the worst two weeks for the project IMO01:57
korean101johnsom: Could you tell when the newton release will stable?01:57
korean101johnsom: Oh. thanks01:57
johnsomNot sure, I would hope this week01:58
korean101johnsom: that's christmas present for me :)01:58
*** fnaval has joined #openstack-lbaas01:58
openstackgerritMichael Johnson proposed openstack/octavia: Revert "Create Octavia api V2 directory"
johnsomLet's see if this fixes the octavia gate and the "retries /v1/loadbalancer"02:04
openstackgerritMichael Johnson proposed openstack/octavia: Revert "Create Octavia api V2 directory"
johnsomOk, time to make dinner.  Catch you folks tomorrow02:07
*** armax has quit IRC02:27
korean101johnsom: but operating_status changed ' PENDING_CREATE -> ONLINE'02:30
korean101(on devstack)02:30
*** chenhb has quit IRC02:37
*** gongysh has joined #openstack-lbaas02:37
*** chenhb has joined #openstack-lbaas02:37
*** yamamoto has joined #openstack-lbaas02:54
*** ducttape_ has joined #openstack-lbaas02:58
*** ducttape_ has quit IRC03:02
*** ducttape_ has joined #openstack-lbaas03:04
*** yamamoto has quit IRC03:05
openstackgerritMerged openstack/octavia: The amphora agent plug_vip has an indention issue
*** kevo has joined #openstack-lbaas03:19
*** fnaval has quit IRC03:21
openstackgerritMerged openstack/octavia: Fix the agent install to use pip -U
*** ducttape_ has quit IRC03:26
*** ducnc has joined #openstack-lbaas03:31
*** ducnc has quit IRC03:31
openstackgerritOpenStack Proposal Bot proposed openstack/octavia: Updated from global requirements
*** amotoki has quit IRC03:33
*** amotoki has joined #openstack-lbaas03:34
*** amotoki has quit IRC03:35
*** amotoki has joined #openstack-lbaas03:36
*** amotoki has quit IRC03:42
*** fnaval has joined #openstack-lbaas03:46
*** diogogmt has quit IRC03:54
*** catintheroof has quit IRC04:06
*** catintheroof has joined #openstack-lbaas04:06
*** catintheroof has quit IRC04:06
*** fnaval has quit IRC04:19
*** ducttape_ has joined #openstack-lbaas04:26
*** ducttape_ has quit IRC04:32
*** ducttape_ has joined #openstack-lbaas04:33
*** diogogmt has joined #openstack-lbaas04:40
*** ducttape_ has quit IRC05:00
rm_workfixed that problem. OK! I can build now! NOPE05:01
rm_worknow I'm getting this for no apparent reason: 2016-12-16 04:56:19.579 | IOError: [Errno 28] No space left on device05:02
rm_workwhen converting the image05:02
rm_workthe universe hates me05:02
rm_workit had plenty of room before ._<05:02
*** bana_k has joined #openstack-lbaas05:03
*** anilvenkata has joined #openstack-lbaas05:15
*** links has joined #openstack-lbaas05:21
openstackgerritMerged openstack/octavia: Updated from global requirements
*** sticker has quit IRC05:33
*** bana_k has quit IRC05:38
*** diogogmt has quit IRC05:39
*** ducttape_ has joined #openstack-lbaas06:00
*** ducttape_ has quit IRC06:05
*** rcernin has quit IRC06:13
*** kevo has quit IRC06:20
*** kevo has joined #openstack-lbaas06:21
*** bana_k has joined #openstack-lbaas06:24
*** rcernin has joined #openstack-lbaas06:37
*** armax has joined #openstack-lbaas06:38
*** rcernin has quit IRC06:43
*** rcernin has joined #openstack-lbaas06:54
*** eezhova has joined #openstack-lbaas06:56
*** ducttape_ has joined #openstack-lbaas07:01
*** ducttape_ has quit IRC07:07
*** gongysh has quit IRC07:22
*** eezhova has quit IRC07:45
*** pcaruana has joined #openstack-lbaas07:56
*** tesseract has joined #openstack-lbaas07:57
*** tesseract is now known as Guest3130407:58
*** yatin has quit IRC08:02
*** ihrachys has joined #openstack-lbaas08:04
*** jsheeren has joined #openstack-lbaas08:11
*** yatin has joined #openstack-lbaas08:14
*** eezhova has joined #openstack-lbaas08:20
rm_work| 54192533-0e38-4bf6-9288-07adfa96b893 | lb1  | be51d93f8690489d96a58d969a80df15 |    | ACTIVE              | octavia08:29
*** bana_k has quit IRC08:31
*** amoralej|off is now known as amoralej08:31
*** ihrachys has quit IRC08:39
*** gongysh has joined #openstack-lbaas08:49
rm_workhmm, can't get it to actually create the member nodes correctly, but whatever08:55
rm_workerr I thought subnet wasn't required for a member09:00
*** ihrachys has joined #openstack-lbaas09:00
rm_workwhoops, bedtime09:00
*** ducttape_ has joined #openstack-lbaas09:04
*** ihrachys has quit IRC09:05
*** ducttape_ has quit IRC09:09
*** ihrachys has joined #openstack-lbaas09:10
*** ducttape_ has joined #openstack-lbaas10:05
*** ducttape_ has quit IRC10:09
*** openstackgerrit has quit IRC10:18
*** yamamoto has joined #openstack-lbaas10:22
*** yamamoto has quit IRC11:03
*** ducttape_ has joined #openstack-lbaas11:05
*** ducttape_ has quit IRC11:10
*** gongysh has quit IRC11:19
*** bdeschenes has joined #openstack-lbaas11:26
*** anilvenkata has quit IRC11:31
*** yamamoto has joined #openstack-lbaas12:03
*** ducttape_ has joined #openstack-lbaas12:06
*** ducttape_ has quit IRC12:11
*** yamamoto has quit IRC12:13
*** catintheroof has joined #openstack-lbaas12:34
*** links has quit IRC12:42
*** ducttape_ has joined #openstack-lbaas13:07
*** jsheeren has quit IRC13:09
*** armax has quit IRC13:28
*** amotoki has joined #openstack-lbaas13:32
*** ducttape_ has quit IRC13:40
*** gongysh has joined #openstack-lbaas13:46
*** amoralej is now known as amoralej|lunch13:47
*** chlong has quit IRC14:00
*** jsheeren has joined #openstack-lbaas14:03
*** jsheeren has left #openstack-lbaas14:08
*** amotoki has quit IRC14:08
*** gongysh has quit IRC14:09
*** ducttape_ has joined #openstack-lbaas14:17
*** ducttape_ has quit IRC14:21
*** amotoki has joined #openstack-lbaas14:25
*** amotoki has quit IRC14:32
*** fnaval has joined #openstack-lbaas14:32
*** amoralej|lunch is now known as amoralej14:36
*** ducttape_ has joined #openstack-lbaas14:47
xgermanAdam, apparently it is now… suprised me as well15:09
*** Dave has quit IRC15:09
*** Dave has joined #openstack-lbaas15:12
*** chlong has joined #openstack-lbaas15:21
*** armax has joined #openstack-lbaas15:25
*** ihrachys has quit IRC15:42
*** kevo has quit IRC15:48
*** armax has quit IRC15:53
*** bana_k has joined #openstack-lbaas15:57
*** bana_k has quit IRC15:59
*** Guest31304 has quit IRC16:02
*** ducttape_ has quit IRC16:06
diltramso we merged that patches :)16:10
*** ducttape_ has joined #openstack-lbaas16:13
ankur-gupta-fso everything works master branch octavia & diskimage builder?16:14
johnsomWe still need to figure out the /v1/loadbalancer issue16:14
diltrambut after this my change with binding there is no problems with /v1/loadbalancers16:15
diltramthere are some other problems - timeouts?16:15
johnsomrm_work It looks like diltram -1'd the py3 string patch, can you take a look?16:15
xgermanmmh, if tye works I would keep it16:17
johnsomYeah, the comment is a bit sparse.  Maybe recommending an alternate way to do the logic?16:18
diltramxgerman: you're talking about my comment?16:18
diltramxgerman: one thing about me, I'm python maniac and I will always -1 anything what is not using proper tools in language16:18
xgermansuggests we are doing it right16:19
xgermanI doubt it will make a difference in our code but …16:20
diltrambut it's a python way to use isinstance16:20
xgermansure we can respin with tat16:21
diltramyep even in this stack overflow post is comment about ABC PEP16:21
diltramwhich tells about using isinstance16:22
*** rcernin has quit IRC16:23
xgermanok, I can respin16:23
diltramgreat :)16:23
xgermanlet me do that…16:23
*** pcaruana has quit IRC16:25
johnsomOk, thanks.  I have burned down my test environment for that since we finished our collaboration session on this yesterday.  So, I don't have a local test for that change.16:25
*** reedip_ has joined #openstack-lbaas16:28
*** chlong has quit IRC16:29
xgermanwell, I will run tox and call it a day16:33
reedip_Hi team , can someone please provide review comments for ?16:35
reedip_its about adding a timeout extension for Listeners16:35
*** kevo has joined #openstack-lbaas16:35
*** eezhova has quit IRC16:36
*** bdeschenes has quit IRC16:46
*** openstackgerrit has joined #openstack-lbaas16:47
openstackgerritGerman Eichberger proposed openstack/octavia: Fix some python3 string issues in amphora agent
xgermanok, got that working again ;-)16:48
diltramxgerman: pep8 will fail16:49
diltramthere is more than 79 chars in line16:49
xgermanok, will look into that16:51
reedip_xgerman : this is 83 chars :
reedip_pep8 will fail ( I think thats what diltram was stating)16:53
*** chlong has joined #openstack-lbaas16:53
xgermanyeah, I am on it16:53
openstackgerritGerman Eichberger proposed openstack/octavia: Fix some python3 string issues in amphora agent
xgermanlet’s hope I didn’t miss anything else17:02
*** rcernin has joined #openstack-lbaas17:04
openstackgerritLubosz Kosnik (diltram) proposed openstack/octavia: Change bind_host IP address
diltramreedip_: one thing, you not suppose to give +1 if you're seeing that tests are failing on the PS17:07
reedip_diltram : its wasnt a big issue, and it was getting fixed .... but anyways, will ensure from next time17:08
xgermannah, +1 is fine to show support… +2 definitely not17:08
diltram+1 means cores this code is ok, please review it17:09
reedip_xgerman : exactly :)17:09
diltramand no this code doesn't really works - because how do you want to merge code which fails basic gate17:09
openstackgerritLubosz Kosnik (diltram) proposed openstack/octavia: Add support for policy.json
*** anilvenkata has joined #openstack-lbaas17:12
*** bana_k has joined #openstack-lbaas17:25
*** armax has joined #openstack-lbaas17:30
*** reedip_ has quit IRC17:36
*** eezhova has joined #openstack-lbaas17:51
*** bdeschenes has joined #openstack-lbaas17:51
*** bdeschenes has quit IRC17:55
*** bdeschenes has joined #openstack-lbaas17:58
*** bdeschenes has quit IRC18:03
*** kevo has quit IRC18:23
*** kevo has joined #openstack-lbaas18:33
*** ducttape_ has quit IRC18:34
*** eezhova has quit IRC18:36
*** anilvenkata has quit IRC18:39
*** bdeschenes has joined #openstack-lbaas18:41
johnsomgreghaynes Any thoughts about tripleo-image-elements not having a stable/mitaka?  Our stable/mitaka branch is broken19:10
diltramjohnsom: all scenarios are passing, but we have some timeout issues19:25
johnsomdiltram link?19:26
rm_worki am torn between isinstance and type being fine... but i guess you're right -- though now it's annoying to read because it doesn't fit on one line, so we may as well have made it a real if/else19:26
rm_workbut whatever19:26
diltramyou can take a look into previous tests execution19:26
rm_work(just got on, my internet has been dead all morning, still is, but gave up and jumped on phone hotspot)19:27
johnsomdiltram do you have a link to the timeout issue you saw?19:27
diltramprevious test execution in this ps19:27
diltramprevious logs19:28
diltramor it was because of python3 changes?19:30
diltramI rebased that patch so maybe it fixed everything?19:30
johnsomdiltram That error does not really mean it's a timing issue.  The tests tend to report that when something goes wrong19:39
johnsomIt looks like that test run failed over the amp:19:39
diltramok, sounds good to me19:39
johnsomWell, something went wrong that caused it to failover....  It should not have failed over19:40
johnsomIt could be the python3 thing we fixed for stats, but I thought we had the dependencies right19:40
diltramSecurity Group edf57b3e-6b36-4b72-bb52-d459f1248e32 in use.19:41
diltramcleaning didn't work19:41
johnsomYeah, again, side effect of the failure / failover19:41
johnsomThe failover would have built a new one, but the test expects the old one to be there19:42
*** kevo has quit IRC19:42
johnsomWell, I guess technically the failover would have deleted the one the test created...19:42
*** ducttape_ has joined #openstack-lbaas19:42
diltramit looks like we're not removing everything properly sometimes19:42
johnsomYeah, it's clear the failover threw the test off.19:43
xgermanwrong window19:44
johnsomWow, you are applying for a loan over IRC?  ha19:44
xgermannope, archiving my mail19:44
johnsomOk, I am going to grab lunch and contemplate what I want to do about the stable branches.19:45
xgermanok, will do some reviews19:46
johnsomWe could pin them both to stable/newton tripleo-image-elements.  It works, but it is a bit odd for stable/mitaka to use stable/newton tripleo-image-elements.19:46
johnsomThey did not cut a stable/mitaka release19:46
johnsomThe one before that was icehouse, which I know won't work19:46
xgermannetwon works for me ;-19:46
*** ducttape_ has quit IRC19:47
*** ducttape_ has joined #openstack-lbaas19:47
*** amoralej is now known as amoralej|off19:48
diltramjohnsom: but it's showing that it didn't succeed with removing sec group, so problem is why it didn't remove it, and it didn't succeed because it was in use19:48
diltramor I'm missing something?19:48
xgermanlet me look19:50
xgermanyeah, it rolled back19:50
xgermanbut this is during cleanup?19:51
xgermanwell, then the system won’t care19:52
diltramin neutron there is onfo about network-vif-unplugged error19:52
xgermanyeah, but our flow will succeed I think. If you can’t delete we won’t create a new load balancer for you in the revert :-)19:54
xgermanI usually don’t read too much into those runs… the hardware they run on is weird19:54
diltrambut because of this we have failed gate tests19:56
diltramand it's sometimes occuring19:56
greghaynesjohnsom: no idea, you'll have to ask tripleo folk. I want to say they should have it19:57
diltramso personaly I would prefer to find why they exist19:57
greghaynesjohnsom: it's possible they just forgot to cut a mitaka one on that repo?19:58
xgermangreghaynes if the last one is icehouse they forgot more than mitaka ;-)19:58
diltramespecially that we can't make it voting based on this19:58
greghaynesHah, yea, I think that aligns with about when they started focusing less on t-i-e19:59
greghaynesIt looks like tripleo-heat-templates does have those stable branches20:00
greghaynesSo yea, it might just be a deal where no one asked20:00
rm_workcouldn't we just pick a good commit hash and use that?20:09
rm_workno need for a real tag technically20:09
openstackgerritLubosz Kosnik (diltram) proposed openstack/octavia: Fix some python3 string issues in amphora agent
openstackgerritLubosz Kosnik (diltram) proposed openstack/octavia: Change bind_host IP address
johnsomdiltram I am pretty sure we have not seen a failover triggered in the gate any time recently.  This is likely from the DIB issue.20:26
johnsomFailover should not be occurring during these tests.20:27
diltramjohnsom: fingers crossed20:27
rm_workugh, 10.x network doesn't work at all in my devstack >_> annoying20:28
rm_worki'm super bad at network debugging20:28
*** csomerville has quit IRC20:29
diltramrm_work: this is why they're working on this neutron tool to verify why smth is not working20:38
rm_workwell in the meantime, i can tell that it seems all the lbaas objects were created but i can't actually test to see if it works20:39
johnsomrm_work 10.x as in private?20:40
*** ducttape_ has quit IRC20:41
johnsomProbably missing a route, try from the router namespace qrouter...  ip netns exec qrouter-.... bash20:41
rm_workwhere is that from?20:44
rm_workroot on the host?20:44
rm_work*devstack host20:44
rm_workdon't appear to have any namespace20:45
xgermanon the devstack I build I needed to be in the NS to get to the (haproxy) LB20:45
rm_workthat one isn't there, but there isn't even a /etc/netns20:45
xgermanip netns list?20:45
rm_workok yeah got it, i was expecting it to be in /etc/netns20:45
xgermanthere should be a qrouter20:45
johnsomYeah, ignore /etc/netns.  Most setups don't configure that20:45
xgermanand to run something20:46
xgermanip nentns exec <namespace> <cmd>20:46
*** csomerville has joined #openstack-lbaas20:46
xgermanthis as roor and I like bash as command a lot20:46
johnsomIf that works, you are just missing a route through br-ex20:46
rm_workyep works from there20:46
rm_workneed to figure out how to add my missing route then20:47
johnsomI am still finishing lunch, but after I am done I can probably give you the command20:48
*** chlong has quit IRC20:48
rm_workmaybe I can figure it out in the meantime20:48
xgermanyeah, you can google how to add routes from namespaces20:49
johnsomNah, this is just a host route to ex-br for the subnet.  No netns needed20:50
johnsomNetns was just a test to narrow the issue20:50
rm_workgot it20:53
rm_worksudo route add -net netmask gw dev br-ex20:53
rm_workseems to work20:53
rm_worki apparently remembered enough to assemble that without any real issues, success20:54
rm_worknow to figure out how to script it20:54
rm_workk got it20:55
*** kobis has joined #openstack-lbaas20:59
*** ihrachys has joined #openstack-lbaas21:07
*** catintheroof has quit IRC21:10
*** catintheroof has joined #openstack-lbaas21:11
johnsomOk, back to figuring out what tag/commit to use for mitaka tripleo...21:11
*** catintheroof has quit IRC21:15
rm_workwoooooouhh wut21:17
rm_workthat's my mood just now21:18
*** csomerville has quit IRC21:18
*** ducttape_ has joined #openstack-lbaas21:20
xgermando I need to find cat pictures for you ;-)21:21
rm_workcurl: (35) SSL peer reports incorrect Message Authentication Code.21:22
rm_worksuper weird, never seen that before21:23
rm_workwhen curl-ing the VIP21:23
rm_work(it's an https listener)21:23
rm_workerr, https-terminated21:23
rm_workanyone familiar?21:23
xgermanwhat does Google think it is21:23
xgermanalso did you try -k21:24
rm_workyeah using -k and -v21:24
*** ihrachys has quit IRC21:24
rm_workthis is the only useful thing that comes up:
rm_workand some stackoverflow stuff about verifying system time, which i just did21:24
johnsomI have seen it before, but don't remember the issue/etc.21:24
xgermanfind cozy_bear /21:25
johnsomI don't think they even had to go that far21:25
xgermantypo s/legit/ilegit/r would have stopped that21:26
rm_worklet's see... only timezone oddity i note is that the amphora is in UTC while the host and the backend vms are in MST21:26
johnsomUgh, just finished a stack, it passed, then I realized it probably didn't do the clone since there was already a repo there.21:30
*** bdeschenes has quit IRC21:31
johnsomrm_work it looks like the devstack infrastructure for git clone won't take commit ids, just tags/branches.21:35
johnsomIt uses --branch if you pass in a "think"21:36
rm_workyeah ok, suck21:36
rm_workso we do have to bug them21:36
johnsomWell, I might be able to find a tag that makes sense21:36
johnsomThey just don't have a branch21:36
diltramok, kicked python3 to merge21:38
diltram - johnsom, rm_work may you merge this code21:38
rm_workah yep looks simple and looks passing21:39
johnsomHa, ok, well, I'm wrong.  It did check out the tag ok...  hmm, wonder what I missed trying to read the devstack code21:41
johnsomOh well, it worked, so...21:41
johnsomAre you guys cool with this fix for mitaka?21:46
johnsomCan you +1 it21:46
diltramjohnsom: for me ok, just want to see tests passing21:47
johnsomSure, NP.  Just highlighting as it is a stable and doesn't pop up here21:47
diltramjohnsom: so we still have no stable core team?21:48
johnsomYou are talking to them21:48
johnsomYeah, something else I need to figure out.21:48
johnsomI granted the neutron stable (the few left) access too as the are still supporting neutron-lbaas on the older stable releases, but after the split I am the only one on it.21:49
rm_workgrats! :P21:52
xgermanwe probably need to grow that for after the split21:52
johnsomYes, it's something we need to figure out how we want to handle21:52
johnsomI need to read up on the rules21:52
johnsomSame deal, please review the newton patch (I combined it with an outstanding patch) :
johnsomSo glad we can drop the tripleo dependency going forward...21:56
rm_workI shouldn't need "ENABLED_SERVICES+=,octavia,o-api,o-cw,o-hm" with the plugin system, right?21:57
rm_workenable_plugin already makes that happen21:57
johnsomNot 100% sure21:57
rm_workme either... and I worked on this21:57
johnsomYou would think so21:57
rm_workhow did I forget so much in 3 months21:58
xgermanyeah, I added the services since the README said so21:58
* rm_work checks21:58
xgermanI am a bit fuzzy on — do we need to make that an extension if we change the API?22:00
openstackgerritAdam Harwell proposed openstack/octavia: Amphora-agent should log to a distinct location
rm_workjust rebasing22:02
*** eezhova has joined #openstack-lbaas22:03
johnsomI thought they did add it as an extension22:06
johnsomI would need to be, yes22:06
xgermannope, I don’t see it there22:06
xgermanoh, it’s there22:06
xgermanI am blind22:07
johnsomThe cli is going to be a problem as it is frozen.  Your point about an implementation is valid.22:10
xgermanfrozen? So hows that openstack-cli stuff coming… the FWaaS people told me we culd just plug in the neutron client22:11
*** bana_k has quit IRC22:12
johnsomSo, changes to the neutron client are frozen as is the development of osc for lbaas22:12
*** kevo has joined #openstack-lbaas22:12
xgermanmmh, so how are end-users taking advantage of that change then?22:13
johnsomOSC chage?22:13
xgermanthe timeout one22:13
xgermanif we don’t do neutron nor osc22:13
diltramthey need to wait until OSC work22:13
johnsomWell, since nothing implements it (drivers) I don't think they are22:13
*** kobis has quit IRC22:14
xgermanwhat’s the point then of this change?22:14
johnsomPlus, yes, it would be API only22:14
*** kevo has quit IRC22:14
xgermanwell, API only really doesn’t benefit most of our users…22:15
johnsomYeah, I had asked that all of the timeouts be handled too, not just one at a time.  I.e. not one per extension.22:15
xgermanyeah, that would be better or we just turg that to n-lbaas22:15
*** bana_k has joined #openstack-lbaas22:19
johnsomOk, added my two pennies22:19
*** eezhova has quit IRC22:21
*** kevo has joined #openstack-lbaas22:22
xgermanTWC in Cincinnati seems to not like IPv622:27
johnsomUgh, hate that.  So far we are good on comcast here.  Though my modem can't do  the full speed.  I'm not sure I want to deal with the problems of swapping it for the 50Mbps bump.22:28
johnsomIPv6 has been good for me on comcast.  They change things everyone in a while that will hiccup it, but in general it has been stable.22:29
johnsomYou can always tunnel through HE if you need access to the ipv6 consoles in zuul22:30
openstackgerritMerged openstack/octavia: Fix some python3 string issues in amphora agent
xgermanI hard-reset my router earlier today and now I have ipV6 — weird22:39
xgermancool and now I could ssh -6 into my cloud server :-)22:42
johnsomCheck if you are filtering neighbor discovery.  Comcast made a change that would route my v6 for a couple of hours and then stop routing for my PD.  Finally figured out they turned on ND.  Changed my paranoid FW rules to all ND and fixed that.22:42
johnsomtypo, not all ND, allow ND from them22:42
xgermanmy FW is just a checkbox in my Asus Router :-)22:42
xgermanbut probably should invest in better network tech22:43
johnsomYou would shake your head at my setup....22:43
johnsomI have a bunch of subnets, SSIDs, and a pretty sizable firewall ruleset22:44
xgermanI am pretty sure mine is so broken that probably all my devices have joined botnets22:44
johnsomFor basic edge device use, you can't  beat the deal this is:
johnsomI use the POE-5 model.22:51
johnsomMy only warning is that it's a hands-on CLI type device.  The GUI is mostly useless and not novice friendly22:51
*** diltram_ has joined #openstack-lbaas22:53
*** diltram has quit IRC22:53
bana_kjohnsom: just wanted to let you know that your suggestion worked for image build error.22:54
johnsomOh good bana_k.  We are just finishing up the fixes for the branches.  I think things will be back to normal by monday22:54
bana_kjohnsom sounds great ! thanks :)22:55
johnsomI have tested master and stable/mitaka with the latest patches.  I'm testing stable/newton now22:55
diltram_johnsom: and what you're doing with this subnets and ssids?22:55
johnsomWell, I have one for "devices", one for guests, one for home office, one for the house22:56
diltram_you're using multiple ssids for HA?22:57
johnsomdevices is for those IoT devices that I don't trust, ring, echo, etc.22:58
johnsomNo, the SSIDs are Wifi ssids that map into the vlans22:58
diltram_ok, I understand22:58
diltram_what you're using to monitoring?22:59
johnsomssh - grin22:59
johnsomI mean the router and access point have some, but if I care to see what is going on I usually just hope on the devices22:59
diltram_you're not using any zabbix/riemann/etc to monitor resources inside of network?22:59
johnsomNah.  I have used them for other setups in the past, but not motivated enough just for the house.23:01
johnsomfrom .. import Lock, NeedRegenerationException23:02
johnsomImportError: cannot import name Lock23:02
johnsomSwitching a VM from mitaka to newton, kind of ugly23:03
johnsomThis is keystone code blowing up23:03
*** diltram_ has quit IRC23:07
openstackgerritMerged openstack/octavia: Change bind_host IP address
*** kevo has quit IRC23:17
*** kevo has joined #openstack-lbaas23:27
*** diltram has joined #openstack-lbaas23:32
bana_kjohnsom do we store anything in the octavia.sni table ?23:33
johnsomThat is for tls offloading.23:33
bana_kI have created a listener with default tls and sni and I dont see any data in that23:34
bana_kbut neutron.lbaas_sni has the sni info23:34
johnsomIt may not fill until you have more than one server name23:34
bana_koh I see. thanks !23:35
*** ducttape_ has quit IRC23:40
*** armax has quit IRC23:42

Generated by 2.14.0 by Marius Gedminas - find it at!