Wednesday, 2017-06-21

*** sshank has quit IRC00:03
*** chlong has quit IRC00:03
*** isotope has quit IRC00:18
openstackgerritAdam Harwell proposed openstack/octavia master: SSL Health Monitors didn't actually ... check very much
rm_workjohnsom: ^^ there you go :)00:20
rm_workreviewing RBAC #1 now00:20
johnsomOk, just about to a place I can look at that.00:20
*** amotoki_away is now known as amotoki00:23
*** sanfern has quit IRC01:04
*** sanfern has joined #openstack-lbaas01:23
*** JudeC has joined #openstack-lbaas01:28
*** JudeC has quit IRC01:36
*** gongysh has joined #openstack-lbaas01:38
*** KeithMnemonic1 has quit IRC01:39
*** isotope has joined #openstack-lbaas01:55
openstackgerritMichael Johnson proposed openstack/octavia master: Add RBAC enforcement to quotas v2 API
*** isotope has quit IRC02:08
*** sanfern has quit IRC02:46
*** dayou has quit IRC03:08
*** dayou has joined #openstack-lbaas03:10
*** gans has joined #openstack-lbaas03:47
*** yamamoto has joined #openstack-lbaas03:57
*** links has joined #openstack-lbaas04:07
*** armax has joined #openstack-lbaas04:28
*** cpuga has joined #openstack-lbaas04:37
*** cody-somerville has quit IRC04:37
*** cody-somerville has joined #openstack-lbaas04:37
*** armax has quit IRC04:42
*** jerrygb has joined #openstack-lbaas04:57
*** gcheresh has joined #openstack-lbaas05:00
*** jerrygb has quit IRC05:01
gansI have confirmed this bug , anyone else facing same issue?05:02
openstackLaunchpad bug 1699396 in livecd-rootfs (Ubuntu) "sha256sum mismatch in cloud images" [Undecided,Confirmed]05:02
*** links has quit IRC05:07
korean101anyone knows this ERROR? (octavia.compute.drivers.nova_driver BadRequest: Invalid key_name provided. (HTTP 400))05:15
korean101# neutron lbaas-loadbalancer-create --name test-lb demo-subnet05:16
korean101after above commands, i got ERROR05:16
*** blogan has quit IRC05:16
johnsomgans yes, the ubuntu cloud image has been broken all day.05:17
gansjohnsom, thanks for confirmation05:17
johnsomkorean101 you have a configuration issue, the keypair name you added for ssh access to the amps is not present in nova05:18
johnsomKeypair in the octavi.conf05:18
korean101/etc/octavia/.ssh/octavia_ssh_key this?05:20
johnsomIt is a nova keypair name and not a file path05:20
korean101johnsom: BTW 05:20 AM?05:20
korean101johnsom: don't sleep?05:21
johnsomYeah, it is 10:20pm here05:21
*** gongysh has quit IRC05:21
johnsomWill be soon05:21
korean101johnsom: not GMT?05:21
johnsomI am in PST05:21
korean101johnsom: OH got it05:22
korean101johnsom: /etc/octavia/.ssh/octavia_ssh_key: PEM RSA private key05:22
korean101johnsom: and keypair list OK05:22
*** links has joined #openstack-lbaas05:23
johnsomHmm, the error you shared is an error nova tells us when the keypair name in nova doesn't match what we have in octavi.conf05:24
*** armax has joined #openstack-lbaas05:24
johnsomI think the command is "nova keypair list"05:24
gansnova keypair-list05:25
gansopenstack keypair list05:25
korean101yes. hypen05:25
korean101johnsom: can you check this .conf file? (
johnsomMake sure the # amp_ssh_key_name = in octavia.conf is in that list05:27
korean101johnsom: i already succeed deploy DVR + octavia in Newton releases05:28
korean101johnsom: now i try Ocata releases. so difficult again05:28
johnsomUgh, yeah DVR still has bugs05:28
korean101johnsom: yes05:28
korean101johnsom: comment out this? amp_ssh_key_name05:29
johnsomYeah, try that and restart the processes.  I bet it will work05:29
korean101johnsom: OH!!!! got a next steps05:31
korean101johnsom: but another ERRORs05:32
korean101and in devstack not comment out amp_ssh_key_name05:32
korean101why i comment out that to deploy real machine?05:32
korean101i only reference devstack configuration05:33
johnsomYeah, octavia_ssh_key is missing from nova keypair-list05:33
korean101johnsom: so difficult for me...05:33
johnsomYou either can leave it commented out, it is only for debug, or load an ssh key into nova with that name05:34
korean101johnsom: but i already load that key name05:35
korean101johnsom: OMG05:35
johnsomWrong project maybe?05:35
korean101johnsom: same errors (ERROR octavia.compute.drivers.nova_driver BadRequest: Invalid key_name provided. (HTTP 400))05:35
korean101johnsom: not nextp steps05:36
johnsomHmm, maybe nova api log will give more information?05:36
korean101johnsom: 1minutes05:37
korean101johnsom: nothings in n-api05:39
johnsomThe 400 should be there05:40
korean101johnsom: wait johnsom05:41
korean101johnsom: please...05:41
johnsomIt is a nova error05:41
*** csomerville has joined #openstack-lbaas05:44
korean101johnsom: different ERRORs05:45
*** cody-somerville has quit IRC05:47
johnsomYeah, nova is failing.  ComputeBuildException05:47
korean101johnsom: but not key name ERROR05:48
korean101johnsom: something different05:48
johnsomI usually dig through the nova logs to see why. It is either a config problem in octavia or nova isn't setup right.  Or it i05:49
johnsomDoes not have enough resource to boot the instance05:49
korean101johnsom: but i create test-002 VM now and succeed05:49
korean101johnsom: now active running05:49
korean101new VM05:49
johnsomOk, hmm05:50
*** armax has quit IRC05:55
johnsomI need to get some sleep.  I would dig through the nova logs to see why it is failing06:04
korean101johnsom: yes thanks06:05
korean101i use ocata releases06:05
korean101johnsom: many thanks!06:05
*** JudeC has joined #openstack-lbaas06:11
*** pcaruana has joined #openstack-lbaas06:31
*** jerrygb has joined #openstack-lbaas06:36
*** kobis has joined #openstack-lbaas06:51
*** rcernin has joined #openstack-lbaas06:59
korean101johnsom: got a clues07:01
korean101johnsom: SecurityGroupNotFound: Security group dd670196-4c2d-49e0-81cd-491afa5ae056 not found (in compute node)07:01
korean101but i already create secgroup and paste in octavia.conf07:01
korean101johnsom: | dd670196-4c2d-49e0-81cd-491afa5ae056 | lb-mgmt-sec-grp       | lb-mgmt-sec-grp        |07:02
korean101johnsom: amp_secgroup_list                = dd670196-4c2d-49e0-81cd-491afa5ae05607:02
korean101johnsom: something weird07:02
*** armax has joined #openstack-lbaas07:04
korean101johnsom: !!07:07
korean101johnsom: comment out amp_secgroup_list in octavia.conf07:07
korean101johnsom: amp instance booting is good07:08
korean101johnsom: but worker can't connect to amp07:08
korean101johnsom: something wrong in octavia.conf07:08
*** cpuga has quit IRC07:08
korean101two comment out items in octavia.conf07:09
korean101johnsom: how can i do?07:09
*** tesseract has joined #openstack-lbaas07:15
*** links has quit IRC07:31
*** jerrygb has quit IRC07:34
*** links has joined #openstack-lbaas07:44
*** sticker has quit IRC08:12
*** gans819 has joined #openstack-lbaas08:18
*** gans has quit IRC08:21
rm_workkorean101: i think the user that octavia.conf has configured for nova is not the same user that you are using manually maybe?08:28
rm_workthat is what it seems like08:29
rm_workbecause your nova user (that is in octavia.conf) cannot see the key-pair or the security group08:29
nmagnezirm_work, o/08:32
*** gans819 has quit IRC08:32
rm_work /\08:32
rm_workugh lol08:32
rm_worktried to make my typo look intentional and it didn't work out T_T08:33
korean101rm_work: my normal user is demo08:33
korean101rm_work: and octavia user is also exists08:33
korean101rm_work: and i create LB via root credentials08:34
korean101rm_work: something wrong?08:34
nmagnezirm_work, yet another recheck fest
rm_workin your octavia.conf, what is under the [nova] section08:34
rm_workyou can leave out password, but what user/project08:34
korean101rm_work: but devstack's [nova] section's nothing08:36
korean101rm_work: i'll try that08:36
korean101rm_work: i added [nova] section auth information in octavia.conf08:39
*** gongysh has joined #openstack-lbaas08:39
korean101rm_work: and do not comment out (amp_ssh_key_name, amp_secgroup_list)08:39
korean101rm_work: Invalid key_name provided ERROR again...08:39
*** JudeC has quit IRC08:41
*** rcernin has quit IRC08:45
rm_workwhat user did you put08:48
rm_workah also neutron section08:48
rm_workbecause security-groups are neutron08:48
*** rcernin has joined #openstack-lbaas08:51
*** gans819 has joined #openstack-lbaas08:52
korean101rm_work: nova section's nova user08:55
rm_workkorean101: what user IS that09:00
rm_workis it the same user AND project you used to create the SSH keypair?09:00
korean101rm_work: yes. i create09:01
korean101rm_work: my own keypair09:01
rm_workkorean101: so, *as* the nova user, if you do a keypair list, the keypair shows up?09:02
rm_workusually we recommend using an "octavia-service" user or similar09:02
rm_workand give it admin in nova/neutron09:02
korean101as nova user09:04
korean101openstack keypair list09:04
korean101The request you have made requires authentication. (HTTP 401) (Request-ID: req-a67db5e3-153f-40f4-b947-baf3321cb9b5)09:04
korean101there three users have a admin role09:06
rm_workok but the *nova user* needs to have the keypair09:07
rm_workwhich i still haven't seen09:07
korean101rm_work: but nova user isn't normal user09:08
rm_workit doesn't matter09:08
rm_workkeypairs don't get shared via roles/etc like everything else09:08
korean101rm_work: also in Newton release, nova user can't seee keypairs09:08
rm_workthe *nova user* needs to actually have created the keypair09:08
rm_workthen you need to use a different user09:08
rm_workas I said, we don't recommend using some random other user09:08
rm_workwe recommend you use one "octavia-service" account09:08
*** cpuga has joined #openstack-lbaas09:09
rm_workthat account needs to own at least the nova keypair and probably also the security-group09:11
korean101rm_work: admin user can't do that?09:12
rm_workit doesn't matter what its roles are09:13
rm_worknova keypairs *are not shared* between users09:13
rm_workat all09:13
rm_workthey are only visible to the specific user that created them09:13
*** cpuga has quit IRC09:14
*** gans819 has quit IRC09:18
*** gans has joined #openstack-lbaas09:19
*** yamamoto has quit IRC09:22
openstackgerritNir Magnezi proposed openstack/octavia-dashboard master: Optimize the link address
openstackgerritNir Magnezi proposed openstack/octavia-dashboard master: Updating for octavia-dashboard
korean101rm_work: something else ERRORS09:44
korean101ERROR octavia.controller.worker.controller_worker OverQuotaClient: Quota exceeded for resources: ['security_group'].09:44
korean101rm_work: i changed octavia user09:44
korean101and rm_work and nothing amp_ssh_key_name ERROR09:44
korean101but other errors comming09:45
rm_workyeah you have a quota problem, it seems :P09:45
rm_workand errors tend to multiply09:45
rm_workbecause one thing breaks, and then everything breaks09:45
*** yamamoto has joined #openstack-lbaas09:46
openstackgerritAdam Harwell proposed openstack/octavia master: Allow using custom enum values for API fields
rm_work^^ nmagnezi there might be more to that maybe, but that seems to be the basis for what we need *I think*, if i understand the needs of the vendors correctly09:50
rm_workbut i really need feedback from Samuel / Evgeny / kobis09:51
rm_workok we can recheck/merge stuff again, supposedly the xenial images are temporarily rolled back to working versions now09:53
*** sanfern has joined #openstack-lbaas09:58
*** sanfern has quit IRC10:03
*** sanfern has joined #openstack-lbaas10:04
openstackgerritAdam Harwell proposed openstack/octavia master: SSL Health Monitors didn't actually ... check very much
openstackgerritAdam Harwell proposed openstack/python-octaviaclient master: Add TLS-HELLO option for HealthMonitors
nmagnezirm_work, aye. I'll take a look at this one :)10:08
nmagnezirm_work, see you tomorrow! (or later today for me..) :)10:09
*** yamamoto_ has joined #openstack-lbaas10:09
*** yamamoto has quit IRC10:13
rm_workyep :)10:15
openstackgerritAdam Harwell proposed openstack/octavia master: SSL Health Monitors didn't actually ... check very much
rm_work^^ missed a spot10:16
*** gongysh has quit IRC10:16
rm_worka couple of those might need more rechecks, if you would be so kind10:21
rm_workand NOW i'm off10:21
*** yamamoto_ has quit IRC10:30
*** yamamoto has joined #openstack-lbaas10:32
*** yamamoto has quit IRC10:33
*** yamamoto has joined #openstack-lbaas10:43
*** sanfern has quit IRC10:52
*** yamamoto has quit IRC10:54
*** gans has quit IRC10:54
*** sanfern has joined #openstack-lbaas10:57
*** cpuga has joined #openstack-lbaas11:11
*** cpuga has quit IRC11:15
*** gongysh has joined #openstack-lbaas11:17
*** atoth has joined #openstack-lbaas11:21
*** yamamoto has joined #openstack-lbaas11:24
*** yamamoto has quit IRC11:28
*** sanfern has quit IRC11:30
*** bzhao has quit IRC11:47
*** bzhao has joined #openstack-lbaas11:48
openstackgerritNir Magnezi proposed openstack/octavia master: Allow using custom enum values for API fields
*** yamamoto has joined #openstack-lbaas11:54
openstackgerritNir Magnezi proposed openstack/octavia master: Allow using custom enum values for API fields
*** cpuga has joined #openstack-lbaas12:45
*** sanfern has joined #openstack-lbaas12:50
*** jerrygb has joined #openstack-lbaas13:00
*** cpuga has quit IRC13:01
openstackgerritNir Magnezi proposed openstack/octavia-dashboard master: Optimize the link address
*** cpuga has joined #openstack-lbaas13:23
*** sanfern has quit IRC13:24
*** sanfern has joined #openstack-lbaas13:37
*** gongysh has quit IRC13:39
*** catintheroof has joined #openstack-lbaas13:45
*** gcheresh has quit IRC13:45
*** cpuga has quit IRC13:58
*** jerrygb has quit IRC14:04
*** chlong has joined #openstack-lbaas14:07
johnsomrm_work When you get on today, can we talk about the RBAC patch?  I don't see any major comments on it and wondered if we can defer one of those changes to the end of the chain?14:15
johnsomstderr: 'fatal: Could not read from remote repository.14:24
johnsomI think I should have just taken the day off....14:24
johnsomSo, yeah, looks like the ubuntu cloud images still don't match the sha256sums14:51
johnsomI posted to the ubuntu forums, maybe that will raise attention14:51
johnsomOur centos 7 amps are working though...  grin14:52
johnsomScore one for nmagnezi14:52
*** rcernin has quit IRC15:07
*** pcaruana has quit IRC15:12
*** kobis has quit IRC15:21
*** fnaval_ has joined #openstack-lbaas15:23
*** blogan has joined #openstack-lbaas15:32
*** sanfern has quit IRC15:48
*** sanfern has joined #openstack-lbaas15:49
*** cpuga has joined #openstack-lbaas15:59
*** cpuga_ has joined #openstack-lbaas16:01
*** tesseract has quit IRC16:03
*** cpuga has quit IRC16:04
*** yamamoto_ has joined #openstack-lbaas16:10
*** yamamoto has quit IRC16:13
*** bzhao has quit IRC16:19
*** bzhao has joined #openstack-lbaas16:20
*** sshank has joined #openstack-lbaas16:34
*** yamamoto_ has quit IRC16:47
*** yamamoto has joined #openstack-lbaas16:49
*** harlowja has joined #openstack-lbaas17:07
*** cpuga_ has quit IRC17:11
*** links has quit IRC17:15
johnsomUbuntu has fixed the issue with the cloud images17:16
*** sshank has quit IRC17:20
*** sshank has joined #openstack-lbaas17:27
*** armax has quit IRC17:29
*** cpuga has joined #openstack-lbaas17:33
*** yamamoto has quit IRC17:42
*** cody-somerville has joined #openstack-lbaas17:44
*** csomerville has quit IRC17:44
*** yamamoto has joined #openstack-lbaas17:51
*** yamamoto has quit IRC17:56
*** gcheresh has joined #openstack-lbaas18:01
*** csomerville has joined #openstack-lbaas18:02
*** cody-somerville has quit IRC18:05
*** cpuga has quit IRC18:10
*** yamamoto has joined #openstack-lbaas18:15
*** JudeC has joined #openstack-lbaas18:16
*** kbyrne has quit IRC18:22
nmagnezijohnsom, Score one for canonical :-)18:23
*** kbyrne has joined #openstack-lbaas18:23
*** cpuga has joined #openstack-lbaas18:26
openstackgerritMerged openstack/octavia master: HM Update for url_path uses incorrect validation
johnsomrm_work Did you figure out what happened with
johnsomI see you rechecked it18:30
rm_workbut, it's *better* than before18:30
rm_workat least we catch the "sometimes" one18:31
rm_workand at least it's obvious what happened18:31
rm_workso not really worse than previously18:31
rm_workand we have better debugging for catching any OTHER issues that come up, on the other patchsets that trigger them18:31
rm_workrather than rechecking that one over and over and over and hoping to randomly catch one18:31
eanderssonLets get this merged?
eanderssonSince it will cause a conflict with the general re-branding18:34
johnsomHa, I already reviewed that18:38
johnsomOk, I'm going to grab a bite to eat before the meeting.18:43
*** yamamoto has quit IRC18:44
rm_workjohnsom: k. i addressed your comments on though i don't know if you like the answers18:49
openstackgerritMerged openstack/octavia-dashboard master: Fix npm test
openstackgerritAdam Harwell proposed openstack/octavia master: Allow using custom enum values for API fields
rm_workupdated to address some things18:58
openstackgerritMerged openstack/octavia master: Add RBAC enforcement to Octavia v2 API
*** sshank has quit IRC19:00
openstackgerritAdam Harwell proposed openstack/octavia master: Octavia Feature Classification
rm_workstill trying to figure out what ^^ is exactly19:06
*** sshank has joined #openstack-lbaas19:08
*** sshank has quit IRC19:10
eanderssonrm_work, how would these features be exposed to the dashboard?19:20
eanderssonWould we add an API call to expose the features for a particular provider?19:21
johnsomrm_work It is supposed to be something like this:
johnsomIt is just docs19:26
*** jniesz has joined #openstack-lbaas19:30
openstackgerritMichael Johnson proposed openstack/octavia master: Add filtering and field selection to API
johnsomRebase so I can test, review and clean up anything the RBAC stuff broke19:37
johnsomJudeC I assume it's ok for me to fix your API patches for the RBAC changes I just made....19:42
johnsomFeel free to say otherwise19:42
JudeCYeah I don't mind at all :)19:42
*** eandersson has quit IRC19:43
JudeCI can always fix them too, still wading through this tempest stuff trying to grasp what I am doing though... lol19:43
openstackgerritMerged openstack/octavia master: Agent: swap flask responses to webob, handle 404 retries better
johnsomYeah, no problem, I can update them19:44
JudeC(wishing there were more examples of people writing tempest plugins)19:45
*** yamamoto has joined #openstack-lbaas19:45
johnsomYeah, you can ask questions in the openstack-qa channel BTW19:47
johnsomNot that I get timely replies all the time, but you can try19:47
JudeCYeah rm_work directed me there yesterday.19:47
rm_workyeah they take a while :/19:48
johnsomThey do tend to like folks that are trying to do a repo plugin, so maybe that will help you19:48
JudeCI will, I feel like I should fully understand what I am asking before posting in there. I want to make sure I at least read (most of) their documentation at least.19:48
johnsomI appreciate folks like you....19:49
JudeCThere is quite a bit...19:49
johnsomThe challenge is probably trying to figure out what is old and outdated vs. the new way...19:50
JudeCExactly... I see a lot of stuff they say not to do implemented in these other tempest plugins.19:51
*** yamamoto has quit IRC19:52
johnsomOctavia meeting starting soon on #openstack-meeting-alt19:56
*** openstackgerrit has quit IRC20:03
*** rstarmer_ has quit IRC20:07
rm_workah johnsom there was some refactoring i noticed i wanted done... so i just ... did it <_<20:10
rm_workposting shortly20:10
*** rstarmer has joined #openstack-lbaas20:10
johnsomrm_work meeting time BTW20:10
rm_workwas neck-deep in auth refactor20:11
*** rstarmer has quit IRC20:29
*** gcheresh has quit IRC20:29
*** eandersson has joined #openstack-lbaas20:31
*** ChanServ sets mode: +o johnsom20:35
*** johnsom changes topic to "Welcome to LBaaS / Octavia - Pike review priority patches at"20:36
*** rstarmer has joined #openstack-lbaas20:46
*** fnaval_ has quit IRC20:50
*** amotoki is now known as amotoki_away20:52
*** openstackgerrit has joined #openstack-lbaas20:58
openstackgerritKen Giusti proposed openstack/octavia master: Use 'get_rpc_transport' for RPC clients and servers.
johnsomnmagnezi FYI, I have not "published" the link to that API-REF yet as it is not yet finished.  L7 and quotas are still not done.21:06
openstackgerritAdam Harwell proposed openstack/octavia master: Refactor the RBAC auth enforcement a bit
rm_workjohnsom: ^^21:06
rm_workend of your chain21:06
*** fnaval has joined #openstack-lbaas21:06
johnsomCool, thanks for not making be rebase the world...  Grin21:06
rm_workyes I know how to do rebases :P21:07
rm_workerr, rather, deal with long chains21:07
openstackgerritAdam Harwell proposed openstack/octavia master: Refactor the RBAC auth enforcement a bit
johnsomDoes this mean you have reviewed the rest of the chain?21:09
rm_workor rather21:10
rm_worki decided to review it at the end21:10
rm_workbecause i got tired of looking at individual things21:10
johnsomYeah, that is cool, thanks21:11
rm_workand yeah, so much DRY21:11
rm_worksame as what I had to do at the end of the API patches21:11
johnsomYeah, well, I got into the moment of getting it done21:11
rm_workit's like, by the time you get to the end, it's super clear where the repetition is21:11
rm_workbut it's not while you're doing it from the start :P21:12
rm_workon mine i went back through and refactored everything which was an absolute PITA so I decided i wouldn't wish that on you and i'd just do it for you at the end ^_^21:12
johnsomYeah, much easier21:13
rm_worknow i just need to test in devstack21:15
rm_workwhich ... i'll get to shortly21:15
*** isotope has joined #openstack-lbaas21:16
*** chlong has quit IRC21:22
isotopeI believe this bug is effecting me, lbaas tls is not working with non-admin tenant21:33
openstackLaunchpad bug 1592612 in octavia "LBaaS TLS is not working with non-admin tenant" [High,Confirmed]21:33
isotopeis there a work around for this maybe using acl's?21:33
johnsomisotope Yes, there are two workarounds until with get this fully fixed21:34
johnsom1. Change the barbican RBAC policies to allow the octavia service account access to all of the containers in barbican.  Works, but not a great option21:34
johnsom2. Use ACLs:
johnsomWe hope to fix this soon via the Octavia v2 API, just waiting on a capability in barbican21:35
johnsomin 2. admin can (should) be the octavia service account21:36
*** cpuga has quit IRC21:43
*** catintheroof has quit IRC21:51
*** JudeC has quit IRC21:57
*** JudeC has joined #openstack-lbaas22:00
*** blogan_ has joined #openstack-lbaas22:00
*** jniesz has quit IRC22:02
isotopejohnson so admin_id is the id of the user performing the request?22:03
*** armax has joined #openstack-lbaas22:03
*** blogan has quit IRC22:03
johnsomisotope It should be the octavia service account name (or neutron service account if you are using a driver other than octavia)22:04
isotopeI tried neutron as the service account name, I'm not sure if that is correct or not22:16
johnsomYeah, it is dependent on your deployment22:20
johnsomJudeC Looking at filtering patch.22:23
*** blogan_ has quit IRC22:23
johnsomIt looks like the gt functionality isn't implemented.22:23
johnsomI am guessing the time based filtering isn't either.22:23
johnsomI think I am just going to report those as bugs and let someone add it later.22:24
JudeCohhh no I did not implement that. I was using neutron lbass docs to get feature parity.22:25
johnsomOh, well, yeah.  neutron is um....  not very up to date on the API specs22:25
johnsomI'm just going to add a bug for it and call it a day22:26
JudeC:P dang it this wouldnt have been hard to do either. Yeah if you throw in a bug I can circle back to it when I am done with this tempest stuff.22:26
johnsomIt's icing if you ask me22:26
JudeCdang I wish I would have seen this doc before, at least I am learning where to look for all of this stuff. :/22:28
johnsomAh, I thought I sent you a link.  Sorry22:29
JudeCin: isn't implemented either22:29
johnsomFeel free to ping me22:29
johnsomYeah, they don't really have  that finalized I don't think22:29
johnsomOh, in is there, it's LIKE they don't have yet22:30
openstackLaunchpad bug 1699616 in octavia "Octavia v2 API needs to support conditional filtering" [Low,Triaged]22:30
johnsomJudeC Sorry to bother you again, trying to understand the patch.22:41
johnsom Line 9822:41
johnsomWhy is this needed if we have the filtering in the query now?
JudeCThis is for field selection, so if you just want to be shown the ids you can do ?fields=id or ids and names ?fields=id&fields=name.22:44
johnsomAh, gotcha.  Hmm, seems like we can do that over in the pagination / query as well.22:45
JudeCThe way the responses work I couldn't just limit the data that was returned by the query from what I remember so I had to make the controller do the selection of the data that we want to be shown.22:45
johnsomHmm, yeah, I could see the models maybe being odd.  Going to think about it.22:46
johnsomAt least we are limiting the number of rows returned now.  That was the important one for paging22:47
JudeCI tried to do it in the query at first. IRRC I ran into an issue with the response types not being able to respond with partial data.22:47
johnsomYeah, but this current method drops them right before the response type is built22:49
johnsomThe db_to_type might have an issue with it though22:49
rm_workjohnsom / xgerman_: OK, tested the updated HM HTTPS mode, works as expected22:49
*** sshank has joined #openstack-lbaas22:50
rm_workerr sorry i still need to test the TLS-HELLO method22:50
rm_workbut only a moment22:50
JudeCjohnsom: let me pull up the code again and I can give you a more detailed answer to why I did what I did one sec.22:50
*** armax has quit IRC22:52
*** armax has joined #openstack-lbaas22:53
rm_workjohnsom: i can't seem to create a TLS-HELLO *or* a PING  healthmonitor, wtf23:00
rm_workHTTP/HTTPS/TCP work23:01
rm_workbut PING and TLS-HELLO do not23:01
johnsomIsn't PING the default if it's just check and httpchk?23:01
*** gongysh has joined #openstack-lbaas23:01
rm_workREQ: curl -g -i -X POST -H "User-Agent: osc-lib/1.6.0 keystoneauth1/2.21.0 python-requests/2.17.3 CPython/2.7.12" -H "Content-Type: application/json" -H "X-Auth-Token: {SHA1}0e8fc079e30275680844b029fedfea89604844cd" -d '{"healthmonitor": {"name": "hm2", "admin_state_up": true, "pool_id": "e758dae1-bd64-4330-a3b4-dff397d4b62b", "delay": 2, "max_retries": 2, "timeout": 5, "type":23:02
rm_work "PING"}}'23:02
rm_workRESP: [400] Date: Wed, 21 Jun 2017 23:01:31 GMT Server: WSGIServer/0.1 Python/2.7.12 Content-Length: 88 Content-Type: application/json x-openstack-request-id: req-4fe25585-4258-4264-8630-1f69ac3b325123:02
rm_workRESP BODY: {"debuginfo": null, "faultcode": "Client", "faultstring": " is not a valid option for "}23:02
rm_workliterally the same request but with a different "type" works23:02
*** gongysh has quit IRC23:03
johnsomI can try on mine, just a sec23:03
JudeCjohnsom: Ah yes because of _convert_db_to_type() I couldn't find a really elegant way to manipulate the object until it was already constructed. I am open to ideas on that though.23:04
johnsomrm_work Same here23:05
rm_workIIRC i had some ideas, need to look again at your code JudeC23:05
rm_workor rather... look for the first time T_T23:05
rm_workbeen busy23:05
rm_workugh that's a bad excuse tho23:05
rm_workjohnsom: hmm wtf23:05
JudeCnah man we all have been pretty crazy busy.23:05
rm_workjohnsom: so PING is broken this whole time lol23:06
rm_workwhy would it do that error23:06
rm_workwhat is different about that type23:06
rm_workwe don't do any further validation i thought23:06
johnsomThe API type looks right, maybe a deeper validator23:06
rm_worki am looking for one23:07
rm_workah i see where i throw that23:08
rm_workthough I don't see why it failed to render properly...23:09
johnsomIt's missing from the DB23:09
johnsommysql> select * from health_monitor_type;23:09
johnsom| name  | description |23:09
johnsom| HTTP  | NULL        |23:09
johnsom| HTTPS | NULL        |23:09
johnsom| TCP   | NULL        |23:09
johnsom3 rows in set (0.00 sec)23:09
johnsomrelational violation23:09
rm_workAHHHH I forgot about that part23:09
rm_workok whelp23:10
rm_workI can add both in this patch23:10
rm_workbut ummm23:10
rm_workthis message sucks23:10
*** armax has quit IRC23:10
rm_workwill fix23:10
johnsomAt least there is a comment acknowledging sucky-ness23:11
johnsomJudeC if we filter those in the DB call and then drop them in the object like you do, we won't have pulled that data back from the DB just to be dropped.  The type object back should just have None data in the columns that are going to be dropped by your code.23:13
johnsomDoes that make sense or am I babbling train of thought too much?23:14
JudeCNo, that makes sense.23:14
JudeCI will spin my wheels at this and see if I can come up with a fix for it tonight.23:15
johnsomSo, leaving what you have, but adding to the pagination filter block to not return the columns23:15
johnsomOk, thanks!23:15
rm_workthis falls into the category of "does anyone use this?"23:22
rm_workbecause no one has noticed that PING monitors don't work yet23:22
rm_workor else hasn't bothered to submit a bug T_T23:22
openstackgerritAdam Harwell proposed openstack/octavia master: SSL Health Monitors didn't actually ... check very much
johnsomClearly no one uses it, it NEVER was in the DB23:23
rm_worklol yep23:31
rm_workumm so23:31
rm_worknow it generates the config but23:31
rm_worki can't get it to WORK?23:31
rm_workit has servers that respond on SSL23:31
rm_workbut they show as down T_T23:31
rm_workthe config is exactly the same as before AFAICT23:31
rm_workOHHHHH nevermind23:32
rm_workthis is lulzy23:32
rm_workyeah nm the nm, i have no idea why this doesn't work23:35
rm_workyeah reverted to old code23:37
rm_workrecreated HM23:37
rm_workit still doesn't work23:37
rm_workso either something is wrong with the return that i'm getting... or ... ???23:38
*** sshank has quit IRC23:47
rm_workjohnsom: can you test that CURRENT https HM works?23:47
*** sshank has joined #openstack-lbaas23:47
rm_workdoes it really do a connect check and show members online?23:48
*** sshank has quit IRC23:48
rm_workmine is constantly showing everything offline23:48
rm_workeven with the old code23:48
johnsomOk, just a minute23:48
johnsomYeah, I see it hitting my webserver with an binary stream23:51
rm_workyeah nm i guess i was just using a bad site to test23:52
rm_worki switched to a different test site and now it works fine23:52
johnsomJust a sec, I'm going to see if I can get a tcpdump and confirm it's an SSL handshake23:52
rm_workeven though curl seemed to show a connection to both working23:52
rm_workso .. whatever, seems to be fine now23:53
rm_worknow I can test your thing23:54
*** KeithMnemonic has joined #openstack-lbaas23:54
*** isotope has quit IRC23:59

Generated by 2.15.3 by Marius Gedminas - find it at!