Thursday, 2017-02-16

*** gongysh has joined #openstack-lbaas00:30
*** chlong has joined #openstack-lbaas00:32
*** m-greene has quit IRC00:32
*** Swami_ has joined #openstack-lbaas00:33
*** m-greene has joined #openstack-lbaas00:34
*** Swami_ has quit IRC00:38
*** m-greene has quit IRC00:53
johnsomrm_work I don't know man.  It seems to me like sqlite is just ignoring "rollback"01:03
*** chlong has quit IRC01:08
johnsomIf I force it to use mysql instead of sqlite the test passes01:22
johnsomIt rolls back like it should01:22
rm_workerr i do remember something about sqlite transactions being ...01:30
johnsomYeah, and throw oslo.db and sqlalchemy in the mix and you get mud01:31
johnsomWe get this fun too:     sqlalchemy.orm.exc.FlushError: New instance <LoadBalancer at 0x7f910c94b750> with identity key (<class 'octavia.db.models.LoadBalancer'>, ('27a4271c-516a-4e8b-afba-31555f4123d3',)) conflicts with persistent instance <LoadBalancer at 0x7f910c9f2110>01:37
johnsomOk, so that one is easy to fix.  It's just sqlalchemy not cleaning up sessions correctly.01:46
johnsomNow, back to sqlite not honoring rollbacks01:46
*** gongysh has quit IRC02:09
johnsom sqlalchemy.exc.OperationalError: (sqlite3.OperationalError) cannot rollback - no transaction is active [SQL: u'ROLLBACK;']02:21
johnsomYeah, what a mess.  Ok, I think I need to look at this tomorrow.02:21
*** sanfern has joined #openstack-lbaas02:52
*** amotoki has joined #openstack-lbaas02:53
*** gongysh has joined #openstack-lbaas03:06
*** fnaval has joined #openstack-lbaas03:17
*** m-greene has joined #openstack-lbaas03:25
*** joehuang has joined #openstack-lbaas03:27
joehuanghello, I am joehuang working on how to use Octavia together with tricircle to realize adding members from multi-region to a load balancer.03:30
joehuangone question is why o-hm0 port was added to br-int, and how it works for heath monitor, and have to do it in plugin.sh03:32
*** amotoki has quit IRC03:41
*** armax has quit IRC03:59
*** links has joined #openstack-lbaas04:02
*** yamamoto has joined #openstack-lbaas04:10
*** ipsecguy_ has quit IRC04:25
*** ipsecguy has joined #openstack-lbaas04:26
*** amotoki has joined #openstack-lbaas04:29
*** amotoki has quit IRC04:40
*** jerrygb has quit IRC04:41
*** amotoki has joined #openstack-lbaas04:48
*** dileepr has quit IRC04:49
*** gongysh has quit IRC05:13
*** gongysh has joined #openstack-lbaas05:19
*** csomerville has quit IRC05:20
*** csomerville has joined #openstack-lbaas05:20
*** fnaval has quit IRC05:36
*** joehuang has left #openstack-lbaas06:22
*** kobis has joined #openstack-lbaas06:36
*** rcernin has joined #openstack-lbaas07:03
*** tesseract has joined #openstack-lbaas07:13
*** yamamoto has quit IRC07:27
*** eezhova has joined #openstack-lbaas07:29
*** amotoki_ has joined #openstack-lbaas07:29
*** amotoki has quit IRC07:32
*** gcheresh has joined #openstack-lbaas07:48
*** jsheeren has joined #openstack-lbaas07:52
*** cody-somerville has joined #openstack-lbaas08:02
*** yamamoto has joined #openstack-lbaas08:03
*** csomerville has quit IRC08:05
*** kobis has quit IRC08:06
*** Alex_Staf has joined #openstack-lbaas08:13
*** pcaruana has joined #openstack-lbaas08:14
*** pcaruana has quit IRC08:20
*** pcaruana has joined #openstack-lbaas08:22
*** kobis has joined #openstack-lbaas08:25
korean101rm_work: hi rm!08:30
korean101rm_work: i'm again to see you...08:30
korean101rm_work: i already success deploy octavia with your helps08:31
korean101rm_work: but recently i try again, i got trouble08:31
korean101rm_work: type this command (neutron security-group-create lbaas) and i got this message (WARNING octavia.amphorae.drivers.haproxy.rest_api_driver [-] Could not connect to instance. Retrying.) endless08:32
*** yamamoto has quit IRC08:32
*** yamamoto has joined #openstack-lbaas08:33
korean101rm_work: i log into amphora instance and doesn't exist any haproxy netns08:33
korean101rm_work: like this (root@amphora-a7bbc2fe-6d63-418a-9f38-960dc6d5e288:~# ip netns)08:34
korean101rm_work: login via qdhcp namespace (on network node)08:34
korean101rm_work: this is my os-collect-config.log in amphora VM (
*** yamamoto has quit IRC08:37
*** eezhova has quit IRC08:39
*** eezhova has joined #openstack-lbaas09:05
*** yamamoto has joined #openstack-lbaas09:17
*** yamamoto has quit IRC09:19
*** yamamoto has joined #openstack-lbaas09:20
*** yamamoto has quit IRC09:25
*** openstackgerrit has quit IRC09:32
*** amotoki_ has quit IRC10:01
*** amotoki has joined #openstack-lbaas10:13
*** amotoki has quit IRC10:20
*** yamamoto has joined #openstack-lbaas10:22
*** amotoki has joined #openstack-lbaas10:35
*** yamamoto has quit IRC10:40
*** amotoki has quit IRC10:42
*** sanfern has quit IRC11:19
*** yamamoto has joined #openstack-lbaas11:38
*** yamamoto has quit IRC11:47
*** reedip has quit IRC11:54
*** reedip_ has joined #openstack-lbaas11:59
*** sanfern has joined #openstack-lbaas12:09
*** catintheroof has joined #openstack-lbaas12:14
*** yamamoto has joined #openstack-lbaas12:25
*** gongysh has quit IRC12:31
*** jerrygb has joined #openstack-lbaas12:42
*** jerrygb has quit IRC12:42
*** jerrygb has joined #openstack-lbaas12:43
*** chlong has joined #openstack-lbaas12:56
*** Alex_Staf has quit IRC13:00
*** RichB has joined #openstack-lbaas13:03
*** RichB has quit IRC13:15
*** catintheroof has quit IRC13:20
*** Alex_Staf has joined #openstack-lbaas13:24
nmagnezisindhu, ping re:
nmagnezisindhu, is it somehow ready for review? (i see it is in merge conflict)13:25
nmagnezindahiwade, ^^13:25
sindhunmagnezi: hi, I'm not working on that. It's been completely taken over by ndahiwade13:28
nmagnezigot it. thanks sindhu :)13:29
*** links has quit IRC13:39
*** yamamoto has quit IRC13:43
*** yamamoto has joined #openstack-lbaas13:49
*** yamamoto has quit IRC13:53
*** gongysh has joined #openstack-lbaas13:55
*** diltram_ has joined #openstack-lbaas14:01
*** diltram has quit IRC14:03
*** kberger has quit IRC14:06
*** kberger has joined #openstack-lbaas14:06
*** nmagnezi has quit IRC14:20
*** matt-borland has joined #openstack-lbaas14:21
*** nmagnezi has joined #openstack-lbaas14:23
*** yamamoto has joined #openstack-lbaas14:50
*** fnaval has joined #openstack-lbaas14:53
*** yamamoto has quit IRC14:57
*** m-greene has quit IRC15:00
*** chlong has quit IRC15:05
ndahiwadeHi nmagnezi , It is not ready yet, I have replied to your comment on gerrit just now.15:37
ndahiwadeThe review order for the API patches would be: Loadbalancer--> Listener -->Pool --> Members, L7Policy --> L7Rule.15:37
*** gongysh has quit IRC15:43
*** faizy has joined #openstack-lbaas15:45
*** gcheresh has quit IRC15:55
*** jsheeren has quit IRC16:02
*** rcernin has quit IRC16:03
johnsomkorean101 Still around?16:04
johnsomThe step where you are getting continual retries is the step where the netns is created, so since it did not complete I would expect the netns to not exist.16:05
johnsomFrom your description your controller worker process is unable to reach the amphora.  This is a problem with the network setup for the lb-mgmt-net16:06
*** Alex_Staf has quit IRC16:09
*** pcaruana has quit IRC16:11
*** tesseract has quit IRC16:14
*** faizy has quit IRC16:15
*** diltram_ is now known as diltram16:24
johnsomMorning diltram_  How was Houston?16:24
*** openstackgerrit has joined #openstack-lbaas16:26
openstackgerritNakul Dahiwade proposed openstack/octavia master: Introduce Octavia v2 API for Members
*** eezhova has quit IRC16:35
*** eezhova has joined #openstack-lbaas16:38
*** Alex_Staf has joined #openstack-lbaas16:41
*** kobis has quit IRC16:42
*** m-greene has joined #openstack-lbaas16:52
openstackgerritNakul Dahiwade proposed openstack/octavia master: Align Octavia API to n-lbaasv2 for L7Policy
*** m-greene has quit IRC16:59
*** chlong has joined #openstack-lbaas16:59
*** Alex_Staf has quit IRC17:08
openstackgerritMerged openstack/octavia master: Fix the service type for octavia
*** eezhova has quit IRC17:31
*** kobis has joined #openstack-lbaas17:32
diltramjohnsom: it was OK17:36
diltram6 hours both ways of driving so a little tired17:36
diltrambut my in-laws are here so it's easier for us :)17:37
openstackgerritShashank Kumar Shankar proposed openstack/octavia master: Align Octavia API to n-lbaasv2 for L7Rules
*** sanfern has quit IRC17:47
*** sanfern has joined #openstack-lbaas17:52
*** SumitNaiksatam has joined #openstack-lbaas17:55
*** m-greene has joined #openstack-lbaas17:55
openstackgerritMichael Johnson proposed openstack/octavia master: Fix quota lock on single_create with auth enabled
*** armax has joined #openstack-lbaas18:02
*** sanfern has quit IRC18:10
*** sanfern has joined #openstack-lbaas18:11
*** Swami_ has joined #openstack-lbaas18:26
*** eezhova has joined #openstack-lbaas18:27
*** Swami has joined #openstack-lbaas18:36
*** Swami_ has quit IRC18:38
*** SumitNaiksatam has quit IRC18:43
*** Swami has quit IRC18:43
*** ipsecguy_ has joined #openstack-lbaas19:07
*** ipsecguy has quit IRC19:09
*** SumitNaiksatam has joined #openstack-lbaas19:10
*** jidar has joined #openstack-lbaas19:28
*** jidar has left #openstack-lbaas19:29
*** gcheresh has joined #openstack-lbaas19:32
*** chlong has quit IRC19:56
*** mdurrant has joined #openstack-lbaas20:00
*** cody-somerville has quit IRC20:01
*** csomerville has joined #openstack-lbaas20:02
*** jsheeren has joined #openstack-lbaas20:07
*** jsheeren has left #openstack-lbaas20:09
johnsomBummer, something other than just the osic cloud changed...  I'm getting the same strange kernel issue on RAX cloud20:13
openstackgerritNakul Dahiwade proposed openstack/octavia master: Align Octavia API to n-lbaasv2 for L7Policy
rm_workjohnsom: is that a kernel issue? or a connection issue to the mountpoint?20:19
rm_workVFS: Cannot open root device "LABEL=cloudimg-rootfs" or unknown-block(0,0): error -6'20:20
johnsomI am guessing it is init_ramfs not mounting20:20
johnsomI think it's another DIB issue....20:21
rm_workgreghaynes: o/20:22
johnsomI need to go eat something, then I will be back and dig into this...  Sigh.20:23
johnsomMaybe it's related to this?
greghaynesI'm currently at lunch but I'll read some backscroll in a bit20:25
johnsomgreghaynes Hi, thanks for joining.
johnsomWe are seeing kernel panics20:25
johnsomKernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)'20:25
johnsomStarted this week, maybe yesterday-ish20:26
greghaynesNice, that looks like a bad FS or something similar20:26
rm_workyeah wondering if maybe it's a DIB change or something?20:27
rm_worksee you after lunch20:27
*** kobis has quit IRC20:35
*** kberger has quit IRC20:39
*** armax has quit IRC20:41
*** armax has joined #openstack-lbaas20:41
*** gcheresh has quit IRC20:42
openstackgerritSindhu Devale proposed openstack/octavia master: Octavia v2 API for listeners
*** armax has quit IRC20:44
*** ducttape_ has quit IRC20:51
*** ducttape_ has joined #openstack-lbaas20:51
*** aleph1 is now known as agarner_away20:54
blogansshank: PoolResponse should have a list of loadbalancers, sorry20:56
blogansshank: and a list of listeners20:56
bloganbut i think it'll onlybe a list of one item20:56
sshankI was comparing the response now.20:56
rm_workomg seriously we ended up with load-balancing21:02
rm_workare you kidding me21:02
rm_workbeen trying to slowly make everything consistent, from loadbalancer and load_balancer, and now we have load-balancing as our service name >_<21:03
* rm_work cries21:03
bloganill push for Load-Balancing next21:04
bloganto be fair though, load balancer is two words (even though balancer is not a recognized word)21:05
rm_workright wut21:05
bloganjust call it LB21:05
bloganit can either be load balancer or linux bridge, there won't be any confusion21:06
bloganor lima beans21:06
rm_workwhat dictionary does pycharm use21:06
rm_workthat it can never figure out "balancer"21:06
blogani have yet to find a legit dictionary that actually defines balancer21:07
bloganlol i just need to find one21:07
blogani've always used google okay, and its never in the first page21:07
bloganjohnsom: nlbaas defined the status DISABELD21:09
bloganbut your table does not (and octavia did not either)21:10
bloganso then if we keep octavia v2 the same as nlbaas, i don't think OFFLINE means disabled, DISABLED does21:11
*** yamamoto has joined #openstack-lbaas21:17
johnsomrm_work It is a "load balancing" service that has "load balancers"21:18
johnsomThat was my thought at least21:18
xgermannetwork as a SERVICE THAT NETWORKS21:18
xgermansorrym accidentially hit the all-caps21:19
*** yamamoto has quit IRC21:19
rm_workjohnsom: yeah it's the hyphen i'm sad about21:19
johnsomblogan That is unfortunate.  So operating status has "DISABLED" or provisioning?  (I need to restack at the moment)21:20
bloganoperating status21:20
xgermanrm_work that was strongly suggested on the mailing list21:21
johnsomrm_work I originally had it without the hyphen21:21
*** armax has joined #openstack-lbaas21:22
rm_workyeah i know21:22
rm_workand I +2'd that21:22
xgermanyeah, where would openstack be if everybody made up their own service name spelling,,,21:22
rm_workI could propose a followup patch to change it again21:23
rm_workbut probably no one would care enough21:23
rm_workso i guess we're stuck21:23
johnsomRight here, where we are....21:23
johnsomrm_work I think it might go back to how Doug put us in for a project:
johnsomAlso, it's used more than just there now, so...  You would need to turn a larger ship21:24
rm_workugh, fine21:29
rm_worki'll never have to touch it21:29
rm_workso whatever21:29
*** armax has quit IRC21:30
johnsomSo, I guess I'm going to be looking into the gate issue now21:31
johnsomThe OpenStack gremlins do not want me to test the load balancer API or continue on the api-ref...21:32
rm_workthat generally seems to be the case21:33
xgermanI am a bit under the weather so slow id good ;-)21:35
*** jerrygb has quit IRC21:41
*** ducttape_ has quit IRC21:42
johnsomblogan So what is the plan with "DISABLED"?  Change octavia code and DB and put converter in v1?21:42
xgermanSince V1 is going away we can change DB21:43
greghaynesrm_work: hey, so its unlikely taht a dib change caused your issues - we havent merged anything since over a week and last release was longer ago than that21:44
johnsomI would lean that way, it's a bit of work though21:44
greghaynesrm_work: did you all add anything / change images size, or anything of that sort?21:44
rm_workthis is the base cirros image right?21:45
greghaynesoh, it is?21:45
rm_workso octavia code doesn't touch that, it's base devstack21:45
johnsomNo, it's the DIB image21:45
rm_workoh, it is?21:45
rm_workwhy does it say cirros21:45
johnsomUnless cirros is using an 16.04 kernel: SSH_AUTH_SOCK=/tmp/ssh-qgce8DIEfJ/agent.299221:45
rm_work2017-02-16 19:57:05.737594 | 2017-02-16 19:57:05.737 |     b'tempest.lib.exceptions.SSHTimeout: Connection to the via SSH timed out.'21:45
rm_work2017-02-16 19:57:05.739117 | 2017-02-16 19:57:05.738 |     b'User: cirros, Password: None'21:45
johnsomLinux version 4.4.0-62-generic (buildd@lcy01-30) (gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.4) )21:46
johnsomSorry, wrong paste21:46
rm_worki am pretty sure all images are borked21:46
rm_worknot just ours21:46
greghaynesah, yea21:46
greghaynesif its cirros then definitely not a dib thing :p21:46
greghaynesid ask the infra folks if this is a known deal21:46
rm_worki thought we built the cirros image too, no?21:47
rm_workjust download it and upload it straight?21:47
rm_workhmm k21:47
greghayneser, right, we dont make that21:47
rm_workyeah johnsom, do you think the cirros issue is unrelated?21:47
greghaynesrandom thought - its possible youre running out of disk space or memory and the block device isnt getting properly created21:47
rm_workit's possible we have two issues21:47
rm_workso is this the infra disk issue?21:47
johnsomMaybe, but it's happened over two different clouds now21:48
greghaynesno idea, I havent really been around openstack CI all week so I dont know the state of things21:48
johnsomI do see that devstack is loading cirros 0.3.5 which last time I knew we were using 0.3.421:49
bloganjohnsom: we cna just dynamically change it from what the db has21:49
bloganjohnsom: so we don't have to do some migrations21:49
bloganor change v121:49
johnsomYeah, just kicking the can though...21:50
*** ducttape_ has joined #openstack-lbaas21:50
johnsomrm_work greghaynes I think we have a winner:
greghaynesmakes sense21:51
*** SumitNaiksatam has quit IRC21:52
bloganjohnsom: well we're already dynamically change enabled to admin_state_up teh same way...we decided to do that for the same db migrations and no changes to v121:52
*** ankur-gupta-f1 has quit IRC21:52
bloganand i thought v1 was going to hang around for a while...21:53
johnsomAnyone know what IRC channel the devstack folks are in?21:56
johnsomRaising the revert question there....21:58
*** beardedeagle has joined #openstack-lbaas21:58
*** dileepr has joined #openstack-lbaas21:59
*** ducttape_ has quit IRC22:03
*** chlong has joined #openstack-lbaas22:05
*** ducttape_ has joined #openstack-lbaas22:10
openstackgerritShashank Kumar Shankar proposed openstack/octavia master: Introduce Octavia v2 API for pools
*** chlong has quit IRC22:11
*** armax has joined #openstack-lbaas22:27
*** yamamoto has joined #openstack-lbaas22:35
*** kobis has joined #openstack-lbaas22:37
*** kobis has quit IRC22:39
*** matt-borland has quit IRC22:42
*** catintheroof has joined #openstack-lbaas23:02
*** eezhova has quit IRC23:10
johnsomSo summary, they changed out the cirros version and changed the file format, which caused devstack to decide to just take the first image it sees in the glance list which happened to be our amp image23:14
johnsomIt sounds like we need to wait on this:
johnsomI may try a depends on to test the theory23:14
johnsomSecond question is why did our amp image not boot, but that I suspect is just the flavor size they use for the cirros in tempest23:15
*** beardedeagle has quit IRC23:20
openstackgerritMichael Johnson proposed openstack/octavia master: Fix quota lock on single_create with auth enabled
openstackgerritSindhu Devale proposed openstack/octavia master: DB migration to add ID column to HM
rm_workprobably yeah23:49
*** armax has quit IRC23:58

Generated by 2.14.0 by Marius Gedminas - find it at!