Thursday, 2014-01-16

« Wednesday, 2014-01-15 Index Friday, 2014-01-17 »
*** SergeyLukjanov_ is now known as SergeyLukjanov04:59
*** SergeyLukjanov is now known as SergeyLukjanov_06:22
*** chandankumar has joined #openstack-climate06:23
*** ppetit has quit IRC08:39
*** ppetit has joined #openstack-climate08:57
openstackgerritNikolay Starodubtsev proposed a change to stackforge/climate: Add utils to work with openstack components  https://review.openstack.org/5229609:31
openstackgerritNikolay Starodubtsev proposed a change to stackforge/climate: Implement keystone trust support  https://review.openstack.org/4800209:32
openstackgerritNikolay Starodubtsev proposed a change to stackforge/climate: Implement basic plugin for VM management  https://review.openstack.org/4914209:33
*** DinaBelova has joined #openstack-climate09:42
*** DinaBelova has quit IRC09:59
*** DinaBelova has joined #openstack-climate09:59
*** bauzas has joined #openstack-climate10:18
*** SergeyLukjanov_ is now known as SergeyLukjanov10:25
bauzasfolks, could I please have a status for https://blueprints.launchpad.net/climate/+spec/basic-vm-plugin ?10:25
bauzasI saw recent changes on the reviews10:26
bauzasNikolay_St, DinaBelova ?10:30
*** SergeyLukjanov is now known as SergeyLukjanov_10:35
*** SergeyLukjanov_ is now known as SergeyLukjanov10:51
DinaBelovabauzas, hello11:20
bauzasDinaBelova: o/11:20
DinaBelovaIt's now almost done - Nick needs to remove some debug lines from this change :)11:21
DinaBelovaand add unit tests11:21
DinaBelovaand that's it11:21
DinaBelovait's working and tested for now with change for climate-nova repo11:21
DinaBelovabauzas, btw, now I'm going to test your policy patch11:26
DinaBelova:)11:26
bauzasplease notice I gave a -111:27
bauzasfor utils11:27
bauzasthere was a typo11:27
bauzasclient_version vs. version11:27
bauzaswas unclear11:27
DinaBelovaI asked Nick to fix that11:27
bauzassmall fix but necessary :(11:27
bauzasok11:28
DinaBelovaMmm..11:28
DinaBelovaI've got an idea11:28
bauzasonce it's done, I'll test11:28
bauzassure, go for it11:28
DinaBelovaI can't find Nick now, but I may fix this small thing myself :)11:28
DinaBelovaand you'll be possible to review it :)11:28
bauzaswell, I think I'll test it even without the change11:29
DinaBelova:D11:29
bauzasthat's only a cosmetic but necessary change11:29
bauzasso I can test it11:29
bauzasno need to hurry for it11:30
bauzasesp. if you're already in a rush :)11:30
bauzasDina, please check out http://status.openstack.org/zuul for rechecks11:30
bauzasclimate is not there11:31
bauzaseven in the gate...11:31
DinaBelovayes, I saw that11:32
DinaBelovathat's why I asked reverify11:32
DinaBelovabut it seems not helping11:32
DinaBelova:(11:32
openstackgerritDina Belova proposed a change to stackforge/climate: Add utils to work with openstack components  https://review.openstack.org/5229611:32
DinaBelovacosmetic change delivered :)11:33
bauzascoolio11:35
bauzaswell, I'm wondering if 'reverify no bug' has been get rid of11:36
bauzasI remember a thread on the ML11:36
bauzasdo a recheck instead11:37
DinaBelovammm11:37
DinaBelovaas I remember it was about rechecking without any bug number11:37
bauzasor ask infra guys :)11:37
bauzasnope11:37
bauzasrecheck no bug still exists :)11:38
DinaBelovabauzas, I'll find that ML conversation :)11:38
DinaBelovaseems like I missed that :)11:38
DinaBelovathanks11:38
bauzasgotcha11:38
bauzashold on11:38
bauzaswill give you the link11:38
DinaBelova:D11:38
DinaBelovamm11:38
DinaBelovaas for Zuul11:38
DinaBelovaQueue lengths: 481 events, 0 results.11:38
bauzashttp://lists.openstack.org/pipermail/openstack-dev/2013-December/021649.html11:39
DinaBelovait seems that queue is really huge11:39
DinaBelovareleases are close11:39
bauzasyup11:39
DinaBelovawe have there 3 changes in check pipeline11:39
DinaBelovaQueue lengths: 483 events, 46 results - it's not going to become better :D11:41
DinaBelovathanks for link11:41
bauzasno pb11:42
bauzasDinaBelova: btw, CFP just opened11:42
bauzasfor the summit11:42
DinaBelovayeeep :)11:44
DinaBelovadeadline Feb 14, as I see :)11:44
openstackgerritFrançois Rossigneux proposed a change to stackforge/climate: Physical host reservation plugin (lease update)  https://review.openstack.org/6712111:48
openstackgerritFrançois Rossigneux proposed a change to stackforge/climate: Physical host reservation plugin  https://review.openstack.org/5428511:48
openstackgerritFrançois Rossigneux proposed a change to stackforge/climate: Physical host reservation (manager and base plugin)  https://review.openstack.org/6470611:48
openstackgerritFrançois Rossigneux proposed a change to stackforge/climate: Physical host reservation (DB related things)  https://review.openstack.org/6470511:48
bauzasDinaBelova: so that means we do have 4 weeks for producing an abstract :)11:55
bauzasfair enough :)11:55
DinaBelova:D11:55
DinaBelovaf_rossigneux, you here?11:56
DinaBelovaf_rossigneux, I'm about https://review.openstack.org/#/c/67121/1/climate/manager/service.py11:56
DinaBelovathat's great job, btw11:56
DinaBelovaf_rossigneux, but speaking about update method there is one moment11:57
bauzasDinaBelova: indeed11:57
bauzasI just gave a +2 on it btw.11:57
bauzasf_rossigneux should port your changes inside his commit11:57
bauzaswithout doing a cherry-pick11:57
DinaBelovabauzas, yes, exactly11:57
bauzasI forgot to tell this to f_rossigneux11:57
bauzasf_rossigneux, we are talking about a small PR11:58
DinaBelovahttps://review.openstack.org/#/c/63773/211:58
DinaBelovathis one11:58
bauzasexactly, was googling it :)11:58
bauzasbtw., f_rossigneux has rights for +2/A :)11:58
DinaBelova:D11:58
DinaBelovayeees11:58
DinaBelova:)11:58
bauzasthat would be even the simpliest solution :D11:59
bauzasno need to backport a change by hand :)11:59
DinaBelova*smiling and waiting for +2/A*11:59
bauzasor I can corrupt scroiset11:59
DinaBelova:)11:59
bauzashold on11:59
DinaBelovaas you wish :)11:59
bauzasscroiset is in front of me while f_rossigneux is 80km away :)12:00
DinaBelova:D12:00
DinaBelovaso you, Swann and Patrick are in Grenoble?12:01
DinaBelovaam I right?12:01
DinaBelovawhere is Francois? :)12:01
scroisetDinaBelova: Lyon12:02
DinaBelovacool!12:02
DinaBelovaI love photos of your places :)12:02
DinaBelovaFrance is beautiful :)12:02
bauzasif you have budget, we could do a mid-Juno sprint :D12:03
DinaBelova:D I believe I'll had no free budget because of my autumn wedding :)12:04
DinaBelovabut who knows :D12:04
scroisetjust approuved https://review.openstack.org/#/c/63773/212:05
DinaBelovacoolio :)12:05
DinaBelovalet's wait 12 hours for our small poor patches :)12:05
bauzasoh congrats ;)12:05
scroisetcongrat! where will be you wedding .. Moscou ?12:05
DinaBelovaI believe in my city - Saratov :)12:05
DinaBelovaand then some trip :) Maybe in France, btw, we have not decided yet )12:06
bauzas=)12:06
openstackgerritA change was merged to stackforge/python-climateclient: Initial Climate client implementation  https://review.openstack.org/4776112:10
DinaBelovawow :)12:11
DinaBelovaIt happened :)12:11
DinaBelovabauzas, am I right that by default any admin or user from the same tenant may, for example, get list of leases?12:16
DinaBelovathat might be created before by other user from that project?12:16
DinaBelovabauzas, I'm just confused by my little test. I've created lease for VM using extension as admin, then I changed my user to non-admin one and tried to list leases12:20
DinaBelovathat was successful12:20
DinaBelovais that normal with default policy?12:21
*** DinaBelova has quit IRC12:23
*** DinaBelova has joined #openstack-climate12:25
DinaBelovabauzas, you there?12:26
*** DinaBelova has quit IRC12:38
*** DinaBelova has joined #openstack-climate12:42
DinaBelovabauzas, sorry, my laptop decided to take a nap12:45
DinaBelova:)12:45
openstackgerritFrançois Rossigneux proposed a change to stackforge/climate: Physical host reservation plugin (lease update)  https://review.openstack.org/6712112:46
bauzassorry, was afk12:49
f_rossigneuxI am here DinaBelova.12:49
bauzasDinaBelova: so you test policies ?12:50
bauzas(13:20:48) DinaBelova: bauzas, I'm just confused by my little test. I've created lease for VM using extension as admin, then I changed my user to non-admin one and tried to list leases12:50
DinaBelovabauzas,yes12:50
DinaBelovaf_rossigneux, never mind12:50
bauzashttps://review.openstack.org/#/c/57200/27/etc/policy.json12:51
bauzaslook at it12:51
bauzas10    "climate:leases:create": "rule:admin_or_owner",12:51
bauzas9    "climate:leases:get": "rule:admin_or_owner",12:51
bauzas4    "admin_or_owner":  "rule:admin or tenant_id:%(tenant_id)s",12:51
bauzasso, yes, that's normal your test works12:52
DinaBelovayes, I see that12:52
DinaBelovammm12:52
DinaBelovabut I also changed tenant12:52
bauzaspolicies won't check if you have rights to see other tenants12:52
bauzasthat's another concern :)12:52
bauzasthat has to be managed on the DB side12:53
bauzasor by using tenanted objects12:53
DinaBelovaok, got it12:53
bauzas(thanks to the conductor)12:53
bauzasplease note that default is !12:54
bauzaswhich means we deny queries12:54
bauzasunless we explicitely give the right12:55
bauzasyou can test changing policy.json12:55
bauzason the fly12:55
bauzasthere is no need of reloading the API12:56
DinaBelovayes, I know that :)12:56
bauzasbut maybe I misunderstood nova code...12:57
bauzasthere is not a lot of documentation about that :(12:58
DinaBelovaI just think that user not from tenant should not see leases created by other tenant12:58
DinaBelovaI mean that's ok with listing leases - that will be done by DB12:58
DinaBelovaautomaticly12:58
DinaBelovaas got showing one lease - that's done by lease ID12:59
DinaBelovaand it other user will somehow have it, it would be not correct letting him see it12:59
DinaBelovaas it seems to me12:59
DinaBelovaalthough, as after DB fixing https://review.openstack.org/#/c/59109/ - other user won't see any wrong leases, so that's ok13:00
DinaBelovaso that's ok :)13:01
DinaBelovaf_rossigneux, I just wanted to say that your commit https://review.openstack.org/#/c/67121/2 should use https://review.openstack.org/#/c/63773/ change. It'll be merged asap after Zuul will do that :)13:02
DinaBelovaf_rossigneux, it's about date format for updating coming from cleint13:02
DinaBelovaf_rossigneux after it'll be merged, you will need to rebase and fix small conflicts13:03
bauzasDinaBelova: found some ops docs about custom json13:05
bauzashttp://docs.openstack.org/trunk/openstack-ops/content/customize_auth.html13:05
bauzas[1] Shows a rule which evaluates successfully if the current user is an administrator or the owner of the resource specified in the request (tenant identifier is equal).13:05
bauzasthat doesn't help much13:05
DinaBelovayeeeeep13:06
DinaBelova...13:06
DinaBelovabecause now there is no such checking as I see13:06
DinaBelovabauzas, if this tenant stuff will be processed by DB, that rule might be some how incorrect13:07
DinaBelovaI mean it has no real value13:07
bauzasyep13:07
bauzasthe main issue is that I really have no idea on where the real check is done :)13:07
DinaBelovaeven with other tenant now it will work13:07
f_rossigneuxDinaBelova, I fix the dateformat in my patch too. I think we can abandon the "Fix date processing while lease updating" change.13:07
DinaBelovahm13:07
bauzasGeneric rules: compare an attribute in the resource with an attribute extracted from the user's security credentials and evaluates successfully if the comparison is successful. For instance "tenant_id:%(tenant_id)s" is successful if the tenant identifier in the resource is equal to the tenant identifier of the user submitting the request.13:08
bauzashttp://docs.openstack.org/trunk/openstack-ops/content/projects_users.html13:08
DinaBelovaf_rossigneux, 1/ I see no '%Y-%m-%dT%H:%M:%S.%f' format in your patch set13:09
DinaBelova2/ common rule is to use earlier CR13:09
DinaBelova:)13:09
f_rossigneuxThe user uses a date format like "2014-01-15 21:36". Do you agree ?13:10
DinaBelovaf_rossigneux, from client we are using --prolong-for option13:10
DinaBelovathat goes to climate and calculates real update value13:10
DinaBelovaand it'll be in  '%Y-%m-%dT%H:%M:%S.%f' format - because that's datetime storing in DB13:11
DinaBelovafor creation we really use "2014-01-15 21:36" - because it's more comfortable for user13:11
DinaBelovauser does not say now lease update <id> --end_date="2014-01-15 21:36"13:12
DinaBelovahe says now13:12
DinaBelovalease update <id> --prolong-for 5m13:12
DinaBelovahttps://github.com/stackforge/python-climateclient/blob/master/climateclient/v1/leases.py#L5813:14
DinaBelovawe may really abandon my CR, but only in case you or me will add the same as you did in https://review.openstack.org/#/c/67121/2/climate/manager/service.py line 222 to client13:15
DinaBelovabauzas, scroiset, how are you about it?13:15
f_rossigneuxOk. I am just wondering if its a good idea to keep a very high precision (milliseconds...). In my patchset I am truncating all dates the keep only the minutes (no seconds and milliseconds), because it causes problems...13:15
DinaBelovaf_rossigneux, yes13:16
DinaBelovathat's why I love your idea to use your code13:16
DinaBelovabut that required some client changes13:16
DinaBelovaso ->>> we may really abandon my CR, but only in case you or me will add the same as you did in https://review.openstack.org/#/c/67121/2/climate/manager/service.py line 222 to client13:16
bauzaswhich CR do you want to abandon ?13:16
DinaBelovahttps://review.openstack.org/#/c/63773/13:16
DinaBelovathis format is not so nice really13:17
DinaBelovamilisecs are too much :)13:17
DinaBelovaso I propose to abandon it13:17
DinaBelovaand use Francois's13:17
DinaBelovabut he or I should add that nice format to request made by client13:18
bauzasI'm only passing HH:MM in the client13:18
DinaBelovaммм13:18
DinaBelovammm13:18
bauzassee https://review.openstack.org/#/c/66810/2/climateclient/v1/shell_commands/leases.py13:19
bauzasL12013:19
bauzasI didn't noticed the .f stuff13:19
DinaBelovabauzas, you are rigth using that13:20
DinaBelovawhat's wrong?13:20
bauzasoh ok13:20
bauzasI dunno13:20
bauzas:)13:20
DinaBelova:)13:20
DinaBelovaok so13:20
bauzaswell, I'm a bit lost with your discussions13:20
bauzas:)13:20
DinaBelova1/ I'll abandon my fix patch13:20
bauzas(to be frank, I'm on the phone at the same time... )13:20
DinaBelova2/ f_rossigneux will add new patch to client13:21
DinaBelovaeveryone's ok?13:21
DinaBelovaadd + or smth :)13:21
f_rossigneuxOk for me.13:21
scroisetok as you prefere13:21
DinaBelovabauzas?13:21
bauzasok13:22
bauzas+113:22
DinaBelovai think he's ok :)13:22
DinaBelovayeep13:22
DinaBelovaf_rossigneux, will you provide this date format to client update lease method asap? That's needed to be merged for this release too13:23
f_rossigneuxI will do this immediately.13:23
DinaBelovaf_rossigneux, thank you13:23
DinaBelova:)13:23
DinaBelovabauzas, as for policies, still I have no clear view of this tenant request filtering13:25
bauzasI'm trying to get the final word on how's done in Nova13:25
bauzasor Neutron13:26
DinaBelovaok13:26
bauzasbecause Ceilometer doesn't go clear this way13:26
bauzaswhat I can propose is to land the patch but raise a bug in parallel13:26
bauzasif there is not more to do, then we will close it as invalid13:26
DinaBelova+113:27
bauzasI would propose the prio as High13:27
bauzasor Criticial13:27
DinaBelovahigh, I suppose13:27
DinaBelovaClimate will work without this change :)13:27
bauzasalthough I'm not sure it's good to leave Critical bugs behind a release... :)13:27
bauzasCritical are showstoppers :)13:28
bauzashttp://zqfan.github.io/openstack/2013/03/28/openstack-essex-nova-policy/13:28
DinaBelovaI think this bug should not even become a part of this release13:28
bauzasmaybe I could plea for help13:28
bauzasbecause I have to dig in the Nova/Neutron codes...13:28
bauzashard moment :)13:28
DinaBelovabecause it won't be researched anyway :)13:29
DinaBelovayep13:29
DinaBelovayou just won't get anought time before release13:29
bauzasthis blogpost is quite valuable13:29
bauzasbut sounds like the check is done on the wsgi side :)13:29
bauzaswhich means this is not done on the policies level13:29
DinaBelovatesting shows it does not :(13:30
bauzasit does not what ? :D13:30
DinaBelovaI may look on lease created by other user13:30
bauzasNova filters per tenant13:30
bauzasClimate does not13:30
DinaBelovaand it seems like that's it..13:30
DinaBelovayes13:30
bauzasthese two assertions don't say that policies are responsible for checking tenancy :)13:31
DinaBelovamaybe we have to wit until your DB fix will be merged13:31
bauzasI mean, the policy code :)13:31
DinaBelova:D13:31
DinaBelovaand see how that will look like then :)13:31
DinaBelovabauzas, agreed? :)13:31
bauzasand as per I saw how it's done in Nova, that's fairly close on what we do with Climate13:31
DinaBelova+113:31
DinaBelovai also researched that13:32
bauzasso that means we have to do 2 things for filtering per tenat13:32
bauzas1. filter on the DB side13:32
bauzas2. filter on the API side13:33
DinaBelovait seems that 1. will be enough for correct working13:33
bauzasyep, but not enough for security reasons :)13:33
DinaBelovabecause DB won't give results13:33
DinaBelovaI mean that will be ok for 0.113:33
DinaBelovabauzas, also here https://review.openstack.org/#/c/66563/ - I've written one comment :)13:34
bauzasyup, saw it13:34
bauzasthat's in my own Zuul pipe13:34
bauzasI also have a 24h delay :D13:34
DinaBelova:D13:34
DinaBelovabauzas, just interesting, where are you testing your PR use case? I mean what kind of lab? I mean you are reserving full hosts, so I think you need at least not the smallest one :)13:39
bauzasyou want me to be frank ? :D13:39
bauzason my laptop, running 2 VMs13:40
bauzasone as devstack ctrl13:40
bauzasone as devstack-compute13:40
bauzasthat's enough for testing all combinations :)13:40
bauzasf_rossigneux was responsible for electing the hosts13:41
bauzasmaybe he ran more hosts13:41
bauzasthe good news is that once we tag 0.1, I'll deploy Climate on our integ lav13:41
bauzaslab13:41
bauzasaround 20 machines13:41
bauzasbig babies13:42
DinaBelova:D13:42
bauzaswell, a git blame around extensions_authorizer in Nova gives alaski as committer13:50
bauzasI'll try to ask him directly13:50
bauzasI'm looking at the gerrit reviews13:51
bauzashttps://bugs.launchpad.net/nova/+bug/92796613:52
*** DinaBelova has quit IRC13:55
*** DinaBelova has joined #openstack-climate14:00
bauzas(15:08:22) flaper87: bauzas: got your question. So, that's still a DB role14:11
bauzashurrah!14:11
DinaBelovavah14:12
DinaBelovaokay :)14:12
DinaBelovaI need to have some dinner :)14:12
DinaBelovaso will be awk14:12
bauzasafk you mean ? :D14:14
bauzashave a good dinner14:14
*** DinaBelova has quit IRC14:23
*** SergeyLukjanov is now known as SergeyLukjanov_a14:35
openstackgerritFrançois Rossigneux proposed a change to stackforge/python-climateclient: Fix date format in update lease  https://review.openstack.org/6717314:38
*** SergeyLukjanov_a is now known as SergeyLukjanov15:02
openstackgerritFrançois Rossigneux proposed a change to stackforge/python-climateclient: Fix date format in update lease  https://review.openstack.org/6717315:41
openstackgerritFrançois Rossigneux proposed a change to stackforge/climate: Physical host reservation plugin (lease update)  https://review.openstack.org/6712115:47
*** SergeyLukjanov is now known as SergeyLukjanov_16:24
*** SergeyLukjanov_ is now known as SergeyLukjanov16:30
openstackgerritSylvain Bauza proposed a change to stackforge/python-climateclient: Implement support for provisioning hosts to ClimateClient  https://review.openstack.org/6656316:34
*** ppetit has quit IRC16:35
*** SergeyLukjanov is now known as SergeyLukjanov_17:01
*** SergeyLukjanov_ is now known as SergeyLukjanov17:19
*** bauzas has quit IRC17:25
*** chandankumar has quit IRC17:41
*** SergeyLukjanov is now known as SergeyLukjanov_17:44
openstackgerritA change was merged to stackforge/climate: Policy management for Climate  https://review.openstack.org/5720018:01
openstackgerritA change was merged to stackforge/python-climateclient: Fix date format in update lease  https://review.openstack.org/6717318:14
*** openstackgerrit has quit IRC18:52
*** openstackgerrit has joined #openstack-climate18:52
*** SergeyLukjanov_ is now known as SergeyLukjanov19:03
*** Alan_ has joined #openstack-climate19:58
*** Alan_ is now known as Guest1627519:59
*** Guest16275 has left #openstack-climate19:59
*** bauzas has joined #openstack-climate20:11
*** SergeyLukjanov is now known as SergeyLukjanov_a22:09
*** SergeyLukjanov_a is now known as SergeyLukjanov_22:10
*** bauzas has quit IRC22:39
« Wednesday, 2014-01-15 Index Friday, 2014-01-17 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!