*** agrebennikov has quit IRC | 00:22 | |
*** jose-phillips has quit IRC | 00:25 | |
*** itisha has quit IRC | 00:52 | |
*** pt_15 has quit IRC | 00:54 | |
*** trananhkma has joined #openstack-glance | 01:12 | |
*** smatzek has joined #openstack-glance | 01:49 | |
*** openstackgerrit has joined #openstack-glance | 02:11 | |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/glance_store: [WIP] Trying to add insecure/cafile for swift authentication https://review.openstack.org/416399 | 02:11 |
---|---|---|
*** trananhkma has quit IRC | 02:12 | |
*** smatzek has quit IRC | 02:22 | |
*** trananhkma has joined #openstack-glance | 02:54 | |
*** ducttape_ has quit IRC | 03:04 | |
*** ducttape_ has joined #openstack-glance | 03:05 | |
*** ducttape_ has quit IRC | 03:09 | |
*** prateek has quit IRC | 03:36 | |
*** prateek has joined #openstack-glance | 03:36 | |
*** mvk has quit IRC | 03:38 | |
*** prateek has quit IRC | 03:45 | |
*** jamielennox is now known as jamielennox|away | 03:51 | |
*** links has joined #openstack-glance | 03:56 | |
*** jamielennox|away is now known as jamielennox | 03:59 | |
*** ducttape_ has joined #openstack-glance | 04:02 | |
*** pdeore has joined #openstack-glance | 04:21 | |
*** ducttape_ has quit IRC | 04:25 | |
*** ducttape_ has joined #openstack-glance | 04:26 | |
*** ducttape_ has quit IRC | 04:26 | |
*** ducttape_ has joined #openstack-glance | 04:26 | |
*** bkopilov has quit IRC | 04:28 | |
*** bkopilov has joined #openstack-glance | 04:40 | |
*** nicolasbock has joined #openstack-glance | 04:42 | |
*** adisky_ has joined #openstack-glance | 04:52 | |
adisky_ | hi can anybody help me on this bug https://bugs.launchpad.net/glance/+bug/1595335?? I just want to know this is a bug or a new feature?? | 04:56 |
openstack | Launchpad bug 1595335 in Glance "Add image location fails when show_multiple_locations = false" [Undecided,Confirmed] | 04:56 |
*** ratailor has joined #openstack-glance | 05:11 | |
*** udesale has joined #openstack-glance | 05:42 | |
*** prateek has joined #openstack-glance | 05:44 | |
*** mvk has joined #openstack-glance | 05:50 | |
*** mvk has quit IRC | 06:11 | |
*** pdeore has quit IRC | 06:16 | |
*** udesale has quit IRC | 06:18 | |
*** udesale has joined #openstack-glance | 06:20 | |
*** pdeore has joined #openstack-glance | 06:27 | |
*** pcaruana has joined #openstack-glance | 06:51 | |
*** bkopilov has quit IRC | 06:52 | |
*** e0ne has joined #openstack-glance | 06:54 | |
*** mosulica has joined #openstack-glance | 07:01 | |
*** groen692 has joined #openstack-glance | 07:04 | |
*** rcernin has joined #openstack-glance | 07:15 | |
*** e0ne has quit IRC | 07:17 | |
*** tesseract has joined #openstack-glance | 07:18 | |
jokke_ | adisky_: that's not a bug ;) | 07:31 |
*** bkopilov has joined #openstack-glance | 07:33 | |
adisky_ | ok ..thats a functionality??? | 07:50 |
*** haplo37 has quit IRC | 07:57 | |
*** ezoszed has joined #openstack-glance | 07:59 | |
*** haplo37 has joined #openstack-glance | 08:06 | |
*** tshefi has joined #openstack-glance | 08:22 | |
*** zzzeek has quit IRC | 09:00 | |
*** zzzeek has joined #openstack-glance | 09:00 | |
*** pdeore__ has joined #openstack-glance | 10:05 | |
*** pdeore has quit IRC | 10:07 | |
*** mvk has joined #openstack-glance | 10:19 | |
*** pdeore__ has quit IRC | 10:21 | |
*** pdeore has joined #openstack-glance | 10:21 | |
*** e0ne has joined #openstack-glance | 10:22 | |
*** pdeore has quit IRC | 10:38 | |
*** pdeore has joined #openstack-glance | 10:39 | |
*** udesale has quit IRC | 10:53 | |
*** links has quit IRC | 10:57 | |
*** pdeore has quit IRC | 11:03 | |
*** links has joined #openstack-glance | 11:21 | |
*** ducttape_ has quit IRC | 11:25 | |
*** ducttape_ has joined #openstack-glance | 11:36 | |
*** smatzek has joined #openstack-glance | 11:40 | |
*** ducttape_ has quit IRC | 11:54 | |
*** catintheroof has joined #openstack-glance | 12:14 | |
*** cdelatte has joined #openstack-glance | 12:30 | |
*** openstackgerrit has quit IRC | 12:33 | |
*** gabor_antal has joined #openstack-glance | 12:36 | |
*** gabor_antal_ has joined #openstack-glance | 12:36 | |
*** gabor_antal_ has quit IRC | 12:38 | |
*** udesale has joined #openstack-glance | 12:55 | |
*** mosulica has quit IRC | 13:03 | |
*** itisha has joined #openstack-glance | 13:21 | |
*** ratailor has quit IRC | 13:31 | |
*** ratailor has joined #openstack-glance | 13:32 | |
*** prateek has quit IRC | 13:41 | |
*** wxy| has joined #openstack-glance | 13:43 | |
*** ratailor has quit IRC | 13:55 | |
*** agrebennikov has joined #openstack-glance | 14:10 | |
*** zul has quit IRC | 14:14 | |
*** zul has joined #openstack-glance | 14:15 | |
*** seanhandley has left #openstack-glance | 14:15 | |
*** ducttape_ has joined #openstack-glance | 14:20 | |
*** smatzek has quit IRC | 14:21 | |
*** udesale has quit IRC | 14:44 | |
agrebennikov | hi there everybody! Can somebody let me know if glance nowadays (more interested in mitaka) can terminate ssl on the api service? And also, is there a wsgi script I may use to run glance-api behind apache | 14:53 |
*** smatzek has joined #openstack-glance | 14:56 | |
*** links has quit IRC | 14:59 | |
*** udesale has joined #openstack-glance | 15:00 | |
*** mosulica has joined #openstack-glance | 15:06 | |
*** wxy|_ has joined #openstack-glance | 15:11 | |
*** wxy| has quit IRC | 15:13 | |
*** haplo37 has quit IRC | 15:13 | |
*** zzzeek has quit IRC | 15:13 | |
*** zzzeek has joined #openstack-glance | 15:14 | |
*** haplo37 has joined #openstack-glance | 15:14 | |
*** udesale has quit IRC | 15:25 | |
*** udesale has joined #openstack-glance | 15:25 | |
*** ativelkov_ has quit IRC | 15:27 | |
*** Guest66666 has quit IRC | 15:29 | |
*** Guest66666 has joined #openstack-glance | 15:29 | |
*** ativelkov has joined #openstack-glance | 15:32 | |
*** udesale has quit IRC | 15:35 | |
*** mvk has quit IRC | 15:48 | |
sigmavirus | agrebennikov: I believe Glance itself has been able to terminate TLS since liberty. That said, Apache can do that for you as well. But I odn't believe we have wsgi scripts in tree for that | 15:57 |
*** dirk has joined #openstack-glance | 15:58 | |
*** _ducttape_ has joined #openstack-glance | 15:59 | |
agrebennikov | sigmavirus, this is what I was afraid of... keeping in mind the main trend over the projects - use apache/nginx for running apis | 16:00 |
agrebennikov | sigmavirus, per ssl on python - I'm jsut struggling with it for a while (trying to set it up with glance) | 16:01 |
sigmavirus | agrebennikov: I do believe pbr has a way of generating them for us but I don't believe anyone's added those magic lines to our setup.cfg yet (nor am I certain that those magic lines would work the way we want them to) | 16:01 |
sigmavirus | agrebennikov: I know there are some config options in Glance around TLS certificates but I've never configured Glance to do TLS termination | 16:01 |
*** ChanServ sets mode: -o sigmavirus | 16:02 | |
agrebennikov | sigmavirus, correct. But the problem is - whenever I set up those 3 lines (cert, key, ca) - api just stops responding | 16:02 |
agrebennikov | even though the docs are saying - is is all I need for making it working | 16:02 |
sigmavirus | agrebennikov: nothing in logs? | 16:02 |
agrebennikov | nope | 16:02 |
agrebennikov | let me puul up the latest, sec | 16:02 |
*** ducttape_ has quit IRC | 16:03 | |
agrebennikov | since I gave up and switched to cinder :) | 16:03 |
sigmavirus | agrebennikov: have you tried using openssl s_client to see if TLS is working? | 16:03 |
agrebennikov | which works as wsgi | 16:03 |
sigmavirus | agrebennikov: also need some basic details before I can be of very much help, e.g., python version, os, etc. | 16:03 |
agrebennikov | ok, this is openstack-ansible mitaka installation | 16:04 |
agrebennikov | I'll bring the details in a bit.... have to spend some time on the meeting now. Are you around in general? | 16:04 |
*** rcernin has quit IRC | 16:09 | |
agrebennikov | so it is ubuntu, python2.7, openstack mitaka upstream | 16:10 |
agrebennikov | sigmavirus, so yeah, the request just hangs | 16:11 |
sigmavirus | agrebennikov: thta's not descriptive | 16:12 |
sigmavirus | what request hangs? | 16:12 |
agrebennikov | I understand that :) | 16:12 |
agrebennikov | if I curl to the port of glance-api - request ends up with timeout | 16:12 |
agrebennikov | since glance doesn't respond | 16:13 |
agrebennikov | this happens if I curl with http | 16:13 |
agrebennikov | if it is https - root@osa-cntl-glance-container-8dfe5d91:~# curl --insecure https://127.0.0.1:9292 | 16:13 |
agrebennikov | curl: (35) error:14094410:SSL routines:SSL3_READ_BYTES:sslv3 alert handshake failure | 16:13 |
agrebennikov | let me go ahead and try to pdb and at least see which module it hangs in | 16:15 |
sigmavirus | agrebennikov: so if you're using osa, are you not using a LB? Isn't that part of the reference architecture of that project? | 16:15 |
sigmavirus | agrebennikov: more useful would be to see what openssl s_client tells you | 16:16 |
sigmavirus | also what details do you have about the certificate you've put in place? What CN do you have in it? | 16:16 |
agrebennikov | I have an LB. But the problem is that it is not enough for the security team over here to terminate ssl on the balancer and I have to reconfigure it to terminate ssl on each service | 16:17 |
agrebennikov | cn is a wildcard, works perfectly with keystone and cinder behind apache | 16:17 |
*** dillaman has quit IRC | 16:23 | |
agrebennikov | http://paste.openstack.org/show/593879/ | 16:25 |
agrebennikov | basically this is the last pattern from pdb | 16:25 |
*** jose-phillips has joined #openstack-glance | 16:28 | |
*** dillaman has joined #openstack-glance | 16:28 | |
*** ezoszed has quit IRC | 16:29 | |
*** groen692 has quit IRC | 16:30 | |
sigmavirus | agrebennikov: don't set "ca_certs" because that means that the client is supposed to provide a certificate to authenticate with | 16:31 |
sigmavirus | In other words, only use cert_file and key_file | 16:31 |
*** _ducttape_ has quit IRC | 16:33 | |
*** ducttape_ has joined #openstack-glance | 16:33 | |
sigmavirus | rosmaita_: ping | 16:39 |
agrebennikov | sigmavirus, I have to use ca | 16:40 |
agrebennikov | and it is used across all the services :/ | 16:41 |
sigmavirus | agrebennikov: you need to use client certificate authentication? | 16:44 |
sigmavirus | agrebennikov: if that's teh case, you're not passing it appropriately from what I can tell from your curl command | 16:44 |
*** mosulica has quit IRC | 16:45 | |
*** edmondsw_ has joined #openstack-glance | 16:45 | |
*** edmondsw_ has quit IRC | 16:45 | |
sigmavirus | jokke_: ping | 16:51 |
*** TravT has joined #openstack-glance | 17:04 | |
*** wxy|_ has quit IRC | 17:14 | |
agrebennikov | sigmavirus, this is completely offtopic from what I can tell. If there is any issue with the certs themselves I'd expect to get an error. While here I definitely have a code issue since it is hanging somewhere | 17:15 |
sigmavirus | agrebennikov: when you start glance, do a list your processes. Glance should be spawning subprocesses to listen which is why you think you're seeing a hang | 17:16 |
sigmavirus | And glance just feeds the certs into the ssl module of Python | 17:16 |
*** tshefi has quit IRC | 17:18 | |
agrebennikov | sigmavirus, I'm using single process service to avoid it | 17:18 |
*** smatzek has quit IRC | 17:19 | |
sigmavirus | agrebennikov: also whether or not you're using client certificate verification is absolutely relevant | 17:19 |
sigmavirus | If your client does not provide the certificate that the server expects, you'll never complete teh TLS handshake | 17:19 |
sigmavirus | agrebennikov: if you don't know what I'm talking about, then you shouldn't be using it because ca_certs is not doing what you think it's doing | 17:20 |
*** mvk has joined #openstack-glance | 17:21 | |
*** tesseract has quit IRC | 17:21 | |
agrebennikov | sigmavirus, nevertheless this option doesn't make any difference | 17:30 |
agrebennikov | wsgi server just hangs | 17:30 |
*** e0ne has quit IRC | 17:31 | |
*** nicolasbock has quit IRC | 17:34 | |
agrebennikov | sigmavirus, but.... seems you were right about CA :) in fact the problem was in the connection between api to registry. when I removed registry_ca_file - it started to work | 17:37 |
agrebennikov | sigmavirus, thanks! :) | 17:37 |
sigmavirus | HTH | 17:38 |
*** smatzek has joined #openstack-glance | 17:46 | |
*** flwang1 has joined #openstack-glance | 17:47 | |
*** aleph1 is now known as agarner | 17:59 | |
*** dharinic is now known as dharinic|lunch | 18:16 | |
smatzek | I'm looking for some clarity on the show_multiple_locations setting. In Newton it is deprecated for removal and the release notes point at setting things in policy.json to control this behavior. Using policy.json alone to control this behavior doesn't work in Newton and it appears to not work yet in Ocata. Is removal of this property and migration to use only policy to control this still planned for Ocata? | 18:33 |
stevelle | smatzek: I would not expect actual removal to happen in Ocata | 18:36 |
flwang1 | stevelle: i think smatzek is asking if we can fix it in Ocata :) | 18:37 |
flwang1 | if it doesn't work now | 18:37 |
flwang1 | smatzek: may i know more background? | 18:37 |
flwang1 | smatzek: are you looking for using this option? | 18:37 |
flwang1 | or i should reword it: are you trying to use multi locations? | 18:38 |
stevelle | flwang1: you're probably right, I'm only answering the half I can speak to right now :) | 18:38 |
smatzek | show_multiple_locations must be set to allow Nova snapshot of a Ceph backed instance to a Ceph backed Glance. I need this functionality. | 18:38 |
flwang1 | stevelle: btw, congrats for glance core and happy new year :) | 18:38 |
stevelle | flwang1: thx and hny | 18:39 |
flwang1 | smatzek: oh, yes, if you're using ceph, you need it | 18:39 |
smatzek | so my question is, to get this functionality in Newton, and currently in Ocata I must continue to set this deprecated setting, per https://github.com/openstack/glance/blob/master/glance/api/v2/images.py#L294 | 18:39 |
smatzek | and where I'm looking for clarity is if the code above will change in Ocata to make it so that conf setting goes away and it's 100% controlled by policy. | 18:40 |
flwang1 | smatzek: hmm... that code is only using for locations update | 18:40 |
smatzek | which is used during the Nova snapshot flow | 18:41 |
flwang1 | i don't think the nova/ceph CoW functions will be impacted by that line | 18:41 |
flwang1 | oh, really? i didn't dig | 18:41 |
flwang1 | smatzek: you mean nova will call that line? | 18:42 |
smatzek | when Nova ephemeral disk and Glance backing store are both Ceph backed Nova creates a queued image, does efficient snapshot direct in Ceph and updates the image location via PATCH API to point at the new RBD object. | 18:42 |
smatzek | what's really driving this question is my proposed change to OpenStack-Ansible to set show_multiple_locations=True when Ceph is in use in this review https://review.openstack.org/#/c/413174/1 | 18:43 |
flwang1 | ah | 18:43 |
flwang1 | i see, you're talking about nova snapshot, sorry i missed it | 18:43 |
smatzek | the question is, if that property isn't going to be necessary in Ocata for the Nova Snapshot flow to work, should we be setting it in OpenStack-Ansible master/Ocata. So I'm wondering if the current Ocata behavior will change to be 100% controlled via policy. | 18:44 |
flwang1 | smatzek: we met each other btw, at rochester lab in 2013 | 18:45 |
flwang1 | smatzek: at that moment, you worked for VMcontrol team IIRC :D | 18:46 |
smatzek | yep, I worked on VMControl at that time. Now I'm working on OpenStack and Ceph cluster deployments using ansible. | 18:46 |
flwang1 | awesome | 18:47 |
flwang1 | i will look into this and will comment on your patch | 18:47 |
flwang1 | but it's not a promise we can fix it in Ocata | 18:48 |
flwang1 | i may need to talk with rosmaita_ | 18:48 |
sigmavirus | flwang1: rosmaita_'s on a much deserved vacation :) | 18:48 |
flwang1 | sigmavirus: ha, good to know, thanks | 18:49 |
sigmavirus | just so you didn't get worried about him having disappeared ;) | 18:49 |
smatzek | flaper87 may also know since his tag is in other glance comments talking about removal of that property. | 18:50 |
flwang1 | smatzek: yep, i think so, i will try and get back asap | 18:51 |
*** pcaruana has quit IRC | 18:52 | |
smatzek | I'm not really worried about if it goes away in Ocata or not. If it's staying for Ocata, fine by me, we just need to know how to set the property (or not) for Ocata. Another item of note, is that mfedosin put a similar change into DevStack here: https://review.openstack.org/#/c/279630/ | 18:53 |
flwang1 | ok, i see. no problem | 19:03 |
*** e0ne has joined #openstack-glance | 19:13 | |
*** dharinic|lunch is now known as dharinic | 19:18 | |
*** e0ne has quit IRC | 19:27 | |
*** e0ne has joined #openstack-glance | 19:29 | |
*** mosulica has joined #openstack-glance | 19:29 | |
*** flwang1 has quit IRC | 19:34 | |
*** raginbajin has quit IRC | 19:36 | |
*** lifeless has quit IRC | 19:36 | |
*** slunkad has quit IRC | 19:36 | |
*** cburgess has quit IRC | 19:36 | |
*** d34dh0r53 has quit IRC | 19:36 | |
*** dharinic has quit IRC | 19:36 | |
*** eglute has quit IRC | 19:36 | |
*** eglute has joined #openstack-glance | 19:36 | |
*** d34dh0r53 has joined #openstack-glance | 19:36 | |
*** slunkad has joined #openstack-glance | 19:36 | |
*** cburgess has joined #openstack-glance | 19:36 | |
*** lifeless has joined #openstack-glance | 19:36 | |
*** raginbajin has joined #openstack-glance | 19:38 | |
*** dharinic has joined #openstack-glance | 19:41 | |
*** mosulica has quit IRC | 19:42 | |
dharinic | sigmavirus: Would you like to have a look at this if free? It seems very close to merging. https://review.openstack.org/#/c/367528/ | 19:56 |
sigmavirus | Thanks for the reminder dharinic | 19:58 |
dharinic | Sure sigmavirus :) | 19:58 |
sigmavirus | I had tested it and it looks to be working correctly | 19:58 |
dharinic | Great. | 19:58 |
dharinic | stevelle, hemanthm https://review.openstack.org/#/c/367528/ | 20:00 |
hemanthm | dharinic: ack, added to review queue | 20:01 |
dharinic | Awesome. Thanks hemanthm | 20:02 |
*** TravT has quit IRC | 20:08 | |
stevelle | smatzek: seems like we need to fix that policy / conf option code so that the conf becomes truly optional, and policy works by-itself. | 20:09 |
smatzek | stevelle, yes that's the way I read the code as well and I'm wondering if that is still planned to make Ocata or not. Given the date and cut offs I'm guessing it won't make Ocata | 20:10 |
smatzek | this release note says it will be removed in Ocata. https://github.com/openstack/glance/blob/stable/newton/releasenotes/notes/deprecate-show-multiple-location-9890a1e961def2f6.yaml | 20:11 |
stevelle | smatzek: we don't have a patch submitted to make the above true that I have found. Further I am -1 on removing the conf option until the above is true for at least 1 full cycle, so I am against removing it in Ocata and would want to push that removal out to a later cycle TBD. | 20:12 |
smatzek | thanks | 20:13 |
stevelle | smatzek: if you want to help by contributing the needed glance patch, that could help. I am core in OSA as well and we can make sure the right thing happens there. | 20:13 |
smatzek | thanks. The OSA review link was noted above. Based on this, I think we need to get that submitted to master / backported to Newton, but we can take that up over in the openstack-ansible channel or the review itself. | 20:15 |
stevelle | smatzek: yeah, adding a comment on the review now. | 20:17 |
*** TravT has joined #openstack-glance | 20:24 | |
*** TravT has quit IRC | 20:27 | |
*** TravT has joined #openstack-glance | 20:27 | |
*** TravT_ has joined #openstack-glance | 20:31 | |
*** TravT has quit IRC | 20:31 | |
*** TravT has joined #openstack-glance | 20:39 | |
*** TravT_ has quit IRC | 20:39 | |
*** d0ugal has quit IRC | 20:39 | |
*** d0ugal has joined #openstack-glance | 20:55 | |
*** adisky_ has quit IRC | 21:09 | |
*** flwang1 has joined #openstack-glance | 21:11 | |
*** e0ne has quit IRC | 21:17 | |
*** _ducttape_ has joined #openstack-glance | 21:23 | |
*** ducttape_ has quit IRC | 21:26 | |
*** _ducttape_ has quit IRC | 21:27 | |
flwang1 | smatzek: still around? | 21:30 |
smatzek | yep | 21:30 |
flwang1 | smatzek: may i know how did you test the multi locations? | 21:30 |
flwang1 | did you just set show_multiple_locations=false and set get_image_location and delete_image_location with admin only in policy file? | 21:31 |
smatzek | set up OpenStack with Ceph backed Nova and Ceph backed Glance without the show_multiple_locations property set in glance api conf. Launch and instance from horizon. Snapshot instance from horizon. Fails with 403 from Glance inside Nova compute. | 21:32 |
flwang1 | smatzek: ok, i see, thanks | 21:32 |
smatzek | policy.json has: "context_is_admin": "role:admin", | 21:33 |
smatzek | "default": "role:admin", | 21:33 |
smatzek | and "delete_image_location": "", | 21:33 |
smatzek | "get_image_location": "", | 21:33 |
smatzek | "set_image_location": "", | 21:33 |
smatzek | so admin has policy authority to the actions | 21:33 |
flwang1 | yep, i understand. basically we need to replace the check at https://github.com/openstack/glance/blob/master/glance/api/v2/images.py#L294 with a policy check | 21:34 |
smatzek | though you could do this without Nova in the picture with Glance alone, create an image in the queued state and then use the PATCH api, http://developer.openstack.org/api-ref/image/v2/?expanded=update-an-image-detail, to add the location | 21:36 |
flwang1 | right, got it | 21:37 |
flwang1 | thanks for the clarification | 21:37 |
*** ducttape_ has joined #openstack-glance | 22:06 | |
*** smatzek has quit IRC | 22:13 | |
*** openstack has joined #openstack-glance | 22:53 | |
*** openstackgerrit has joined #openstack-glance | 23:01 | |
openstackgerrit | Dharini Chandrasekar proposed openstack/glance: Stricter checks for registry API calls https://review.openstack.org/416766 | 23:01 |
*** agrebennikov has quit IRC | 23:08 | |
*** jamielennox is now known as jamielennox|away | 23:11 | |
*** jamielennox|away is now known as jamielennox | 23:14 | |
*** ducttape_ has quit IRC | 23:16 | |
*** ducttape_ has joined #openstack-glance | 23:17 | |
*** ducttape_ has quit IRC | 23:21 | |
*** cdelatte has quit IRC | 23:23 | |
*** ducttape_ has joined #openstack-glance | 23:24 | |
*** ducttape_ has quit IRC | 23:35 | |
*** ducttape_ has joined #openstack-glance | 23:35 | |
*** ducttape_ has quit IRC | 23:40 | |
*** ducttape_ has joined #openstack-glance | 23:40 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!