Tuesday, 2017-08-01

*** noslzzp has joined #openstack-barbican00:19
*** liujiong has joined #openstack-barbican01:08
*** randomhack has joined #openstack-barbican01:17
*** randomhack has quit IRC01:21
*** rmascena has quit IRC01:58
*** https_GK1wmSU has joined #openstack-barbican02:10
*** https_GK1wmSU has left #openstack-barbican02:11
dave-mcc_johnsom rm_work can you confirm that barbican client 4.5.1 fixes the issues you had?  i'll update global requirements for pike once we know that all is well again.02:29
*** liujiong has quit IRC02:50
*** liujiong has joined #openstack-barbican02:51
*** raddaoui has quit IRC02:54
openstackgerritOpenStack Proposal Bot proposed openstack/python-barbicanclient master: Updated from global requirements  https://review.openstack.org/48810903:03
*** dave-mcc_ has quit IRC03:48
*** noslzzp has quit IRC04:02
rm_worktonyb: ^^ we will test it with octavia at least, and tell you it is OK for us... the other project that had obvious problems was Heat, IIRC05:11
rm_workbut i think the usage pattern is the same05:12
tonybrm_work: Thanks.05:12
tonybrm_work: ISTR that the glance/nova etc testign we did didn't show it up so I'm trying to get ahead of the game05:13
tonybrm_work: I assume it's just a matter of a no-op change in octavia / heat that Depends-On the u-c bump right?05:13
rm_workwe could do that, yes05:14
rm_workI was literally just going to test it locally and give you a thumbs up, but your idea is better :P05:14
rm_worki figured since I raised the issue originally, i would have some sort of extra sign-off credit maybe ^_^05:15
rm_workumm... though interestingly, 4.5.1 doesn't seem to have solved it locally... not sure why actually05:16
*** randomhack has joined #openstack-barbican05:17
tonybrm_work: Oh phooey :(05:18
*** randomhack has quit IRC05:22
rm_workyeah this sucks05:36
rm_workyeah crap, we were using BOTH import methods I mentioned05:41
rm_workso in the fix patch: https://review.openstack.org/#/c/487721/05:42
rm_workI commented twice that this only fixed half of the problem05:42
rm_workand then somehow let myself be convinced that because the specific example I used wasn't actually doing it that way, that it was somehow still OK to not fix that too05:42
rm_workbut WE are using it that was as well <_<05:43
rm_workso, I was correct05:43
rm_work"Note that they used to be available from within the client as well"05:43
rm_worksecret = barbican_client.secrets.create(...)05:43
rm_workThat's not using the client object... That's using the imported module:05:43
rm_workfrom barbicanclient import client as barbican_client05:43
rm_workTHAT client doesn't have the imports anymore for the objects...05:43
tonybrm_work: time for 4.5.2?05:51
rm_workshould have stuck to my guns >_>05:54
rm_workbecause even if castellan didn't use it that way, SOMEONE could have05:56
rm_work(we did)05:56
*** jaosorior has quit IRC06:05
tonybrm_work: :(06:06
*** jaosorior has joined #openstack-barbican06:24
liujiongtonyb rm_work: sad to hear that06:24
rm_workbasically, see my original comments06:25
liujiongrm_work: ok, now I know folks import barbicanclient.secrets or barbicanclient.client.secrets06:27
*** pcaruana has joined #openstack-barbican07:06
openstackgerritJeremy Liu proposed openstack/python-barbicanclient master: Support import modules from barbicanclient.client module  https://review.openstack.org/48951807:50
liujiongrm_work: I send a new patch https://review.openstack.org/#/c/489518/, can you confirm whether that's proper fix for octavia? thanks07:54
*** randomhack has joined #openstack-barbican08:17
*** randomhack has quit IRC08:21
*** sapcc-bot2 has joined #openstack-barbican08:49
*** sapcc-bot has quit IRC08:53
*** salmankhan has joined #openstack-barbican08:58
rm_workliujiong: i will have to do it tomorrow, sorry T_T09:03
rm_workin the morning, for me (about 8 hours)09:03
liujiongrm_work: ok, thank you!09:04
*** liujiong has quit IRC10:12
*** tpatzig_ has joined #openstack-barbican11:29
*** seife_ has joined #openstack-barbican11:29
*** carthaca_ has joined #openstack-barbican11:29
*** mkoderer_ has joined #openstack-barbican11:29
*** seife_ has quit IRC11:31
*** tpatzig_ has quit IRC11:31
*** mkoderer_ has quit IRC11:31
*** carthaca_ has quit IRC11:31
*** raildo has joined #openstack-barbican12:10
*** catintheroof has joined #openstack-barbican12:20
*** dave-mccowan has joined #openstack-barbican12:54
openstackgerritDavanum Srinivas (dims) proposed openstack/castellan master: Use Stevedore for better extensions  https://review.openstack.org/48401513:04
*** alee has quit IRC13:19
dave-mccowanjohnsom rm_work ping13:38
*** noslzzp has joined #openstack-barbican13:47
*** liujiong has joined #openstack-barbican13:48
*** liujiong has quit IRC14:15
*** alee has joined #openstack-barbican14:16
aleedave-mccowan, ping14:31
dave-mccowanalee pong14:32
aleedave-mccowan, note email from Luke Hinds upstream14:32
aleefrom security team14:32
alee(not sure if you responded yet)14:32
dave-mccowanalee yep.  he pinged me earlier today too.14:32
aleeI had thought you all had decided on first two days security and seconf two days barbican14:32
aleebut it definitely makes sense for us to share etherpads etc.14:33
aleedave-mccowan, anyways will let you respond14:33
dave-mccowanyep, i replied and said we should share and co-ordinate.14:34
aleedave-mccowan, cool14:35
dave-mccowanalee i release barbican client 4.5.1 with the fix for object entry points.  do you know anyone in heat or lbaas who is around today to confirm that everything is working for them now?14:36
dave-mccowanalee i want to confirm before getting feature-freeze exception to update global requirements14:37
aleedave-mccowan, dont know anyone on the top of my head14:37
aleedave-mccowan, maybe rm_work14:38
*** pcaruana has quit IRC15:22
*** zz_dimtruck has quit IRC15:33
openstackgerritDavanum Srinivas (dims) proposed openstack/castellan master: [WIP] Vault based key manager  https://review.openstack.org/48308015:37
*** pcaruana has joined #openstack-barbican15:40
*** rmascena has joined #openstack-barbican15:40
*** andreas_s has quit IRC15:41
*** raildo has quit IRC15:42
johnsomdave-mccowan Hi15:48
johnsomI am just coming off some vacation time, so might not have the answers you are looking for, but I can at least ask15:48
dave-mccowanjohnsom can you check if barbican client 4.5.1 is working for you now?15:49
dave-mccowanjohnsom i need to request a feature-freeze-exemption to update global requirements, so i want to make sure everything is working before i do.15:49
johnsomI see there is a patch up with g-r changes, but there is a lot of red on it.  It kind of looks  like an infra blip though, rechecking.  Though if the UC change hasn't happened yet, this must be unrelated.15:52
johnsomI will probably need to pull this down and test local.15:52
dave-mccowanjohnsom i just noticed jeremy put up another patch for this.  maybe it's not fixed yet.  i'll check with him tonight.15:56
johnsomOk, we will work to verify today.  Where would you like results?  Here?  Is there bug?16:05
openstackLaunchpad bug 1706841 in python-barbicanclient "Client refactor broke all object entrypoints" [High,In progress] - Assigned to Jeremy Liu (liujiong)16:31
dave-mccowanjohnsom ^^ and ping me here too please16:32
johnsomSure, NP16:32
rm_workdave-mccowan: you were off last night but16:41
rm_worki verified it is NOT16:41
rm_workand there is a new patch up16:41
rm_workjohnsom / dave-mccowan i'm tracking this16:42
dave-mccowanrm_work thanks.  do you know if the new patch works?16:42
rm_workdave-mccowan: if you go to eavesdrop for last night you can see our conversation16:42
rm_workit was about ... midnight16:42
rm_workno, i need to test it this morning16:42
rm_workbut probably16:42
dave-mccowandims ^^ fyi16:42
dave-mccowanrm_work thanks!16:43
dimsthanks dave-mccowan. yes rm_work, i saw your chat with tonyb16:43
rm_worktesting now16:43
*** randomhack has joined #openstack-barbican16:50
rm_workok almost there16:51
rm_worki see one last bug >_<16:51
rm_worktracking it down16:51
rm_workappears some function signature changed16:51
rm_workahh this one is maybe my fault16:51
rm_workI'm using an evil "isinstance" check16:52
rm_workhmm or maybe it is catching an actual bug, one sec16:53
*** pcaruana has quit IRC16:58
*** randomhack has quit IRC16:58
mtreinishdave-mccowan: did you see my response to your review on: https://review.openstack.org/#/c/476996/17:02
openstackgerritMatthew Treinish proposed openstack/barbican master: Use devstack functions for deploying barbican-svc  https://review.openstack.org/47699617:02
rm_workyeah this is an interesting one17:15
rm_worktrying to figure out how to solve this testing issue without having the new stuff17:15
rm_workone sec, let me merge a patch17:21
*** randomhack has joined #openstack-barbican17:21
rm_workdave-mccowan: so here's the issue17:21
rm_workdave-mccowan: i was doing an isinstance on the returned client object in one test17:21
rm_workbut it has moved to be a v1.client.Client17:21
rm_workso, do we consider that if someone did that, they deserve to be broken?17:22
rm_workI am willing to patch it in Octavia17:22
rm_workbut there's technically no guarantee others didn't do this17:22
rm_workBUT... <_< well17:22
rm_workyou can't catch EVERY way someone used your stuff17:22
*** salmankhan has quit IRC17:23
rm_workI've changed it to be more pythonic and just check for the things we expect exist in the object, rather than an "evil" isinstance()17:23
dave-mccowanrm_work :-(  i hate breaking anyone, especially so close to release time.17:23
rm_workI am not sure how you fix an isinstance to pass17:24
rm_workis there a trick that allows you do pass that?17:24
rm_workI know that it is considered a very evil check17:24
rm_workbasically the `barbicanclient.v1.client.Client` object needs to pass an isinstance() check that is looking for `barbicanclient.client.Client`17:25
*** randomhack has quit IRC17:26
rm_workbut, that would technically be lieing in the case of wanting to know if you got the right thing17:26
rm_workbecause then the real `barbicanclient.client.Client` would also pass that test, but the test would be wrong17:26
rm_workI think maybe it's better to break a test that does that because it is a really bad test17:26
dave-mccowanrm_work i'll defer to you; if you're ok with patching to lbaas then we should be good to go.18:19
rm_workworking on getting it merged18:19
rm_workyou can merge that last fix18:19
rm_workand prepare the new versiob18:19
dimsrm_work : this one to be sure :) https://review.openstack.org/#/c/489518/18:23
rm_workyou can cut the new release once that merges and it will be ok18:41
rm_workwe'll still need to merge OUR fix but we can do that today18:41
openstackgerritMerged openstack/python-barbicanclient master: Support import modules from barbicanclient.client module  https://review.openstack.org/48951818:47
*** salmankhan has joined #openstack-barbican19:01
*** randomhack has joined #openstack-barbican19:02
*** peter-hamilton has joined #openstack-barbican19:33
*** peter-hamilton has quit IRC19:33
*** peter-hamilton has joined #openstack-barbican19:34
*** randomhack has quit IRC19:47
*** salmankhan has quit IRC20:29
*** dave-mccowan has quit IRC21:27
*** alee has quit IRC21:52
*** rmascena has quit IRC22:00
*** catintheroof has quit IRC22:23
*** dave-mccowan has joined #openstack-barbican22:43
*** alee has joined #openstack-barbican22:50
*** pbourke has quit IRC22:53
*** pbourke has joined #openstack-barbican22:55
-openstackstatus- NOTICE: osic nodes have been removed from nodepool due to a problem with the mirror host beginning around 22:20 UTC. please recheck any jobs with failures installing packages.23:46

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