| *** tkay has quit IRC | 00:03 | |
| *** kyles_ne has quit IRC | 00:03 | |
| *** tkay has joined #openstack-swift | 00:04 | |
| openstackgerrit | Samuel Merritt proposed a change to openstack/swift: Test for unicode names in AccountBroker.merge_items() https://review.openstack.org/122570 | 00:07 |
|---|---|---|
| torgomatic | redbo: AccountBroker seems to not have the bug, but I wrote a test anyway ^^ | 00:07 |
| *** vu has joined #openstack-swift | 00:18 | |
| notmyname | if you `pip install matplotlib` do you get unittest errors with xprofile middleware? | 00:28 |
| *** dmorita has joined #openstack-swift | 00:32 | |
| *** occupant has quit IRC | 00:37 | |
| *** tkay has quit IRC | 00:44 | |
| *** mkollaro has quit IRC | 00:52 | |
| *** tdasilva has joined #openstack-swift | 00:53 | |
| *** insane has joined #openstack-swift | 00:57 | |
| zaitcev | Guys... The python-swiftclient 2.3.0 is considered to be a part of Juno release, right? | 00:58 |
| zaitcev | It's like "Juno RC1" or something? | 00:58 |
| insane | checking | 01:04 |
| *** tgohad has quit IRC | 01:16 | |
| *** tsg has joined #openstack-swift | 01:20 | |
| *** mwstorer has quit IRC | 01:22 | |
| redbo | torgomatic: ok, cool. | 01:25 |
| *** bill_az has quit IRC | 01:27 | |
| *** tgohad has joined #openstack-swift | 01:39 | |
| *** tsg has quit IRC | 01:42 | |
| notmyname | zaitcev: yeah, I guess | 01:48 |
| zaitcev | notmyname: thanks | 01:48 |
| notmyname | SE NOW | 01:48 |
| zaitcev | What... was that? | 01:49 |
| notmyname | all of the client projects have a rolling release cycle, but there was that mailing list post that said RELEASE NOW | 01:49 |
| notmyname | (fat fingered it) | 01:49 |
| zaitcev | oh god | 01:49 |
| zaitcev | I thought there was a nuke going of to the south-east | 01:49 |
| notmyname | anyway, it's supposed to be because of integration periods | 01:49 |
| notmyname | heh | 01:50 |
| zaitcev | okay | 01:50 |
| zaitcev | Well, it looks that Christian liked the patch that I mentioned during the IRC meeting yesterday, even though I was not sure myself that I considered all permutations. | 01:51 |
| zaitcev | But it was late for 2.3.0 anyway so its' good | 01:51 |
| zaitcev | IT'S | 01:51 |
| notmyname | :-) | 01:51 |
| *** tgohad has quit IRC | 01:56 | |
| *** stevemar has joined #openstack-swift | 02:02 | |
| *** sandywalsh has quit IRC | 02:02 | |
| stevemar | hello swift folks, looking for some info on how the tests are setup in swift... whats the deal with function/probe? | 02:03 |
| notmyname | zaitcev: since there's still several weeks before the juno release, I may look in to a 2.3.1 swiftclient release. and that would include your patch | 02:03 |
| stevemar | do the functional tests only run for the gate job: gate-swift-tox-func ? | 02:04 |
| zaitcev | notmyname: that would be great, I hope I didn't screw up in it | 02:04 |
| notmyname | stevemar: unit->tests small sections of code. functional->black box testing of API functionality. probe->white box testing like integration tests for the specific swift components (eg write data, kill a replica, make sure replication fixes it) | 02:04 |
| notmyname | stevemar: gate-swift-tox-func is just the functional tests | 02:05 |
| notmyname | stevemar: there is nothing in the CI infrastructure that does probe tests today | 02:05 |
| stevemar | notmyname, cool, and gate-swift-tox-func sets up devstack (i assume)? | 02:05 |
| notmyname | stevemar: I think so. I don't know what all services it sets up, but it only runs agains swift (with a keystone token) | 02:06 |
| stevemar | hmm, maybe not, not seeing it in the logs | 02:06 |
| notmyname | stevemar: that is, I know it definitely runs in devstack, but I don't know what else is started in devstack in that gate job | 02:06 |
| stevemar | the logs could might not be saved for the test run | 02:06 |
| stevemar | notmyname, probably just the minimal amount | 02:07 |
| stevemar | notmyname, looking @ http://logs.openstack.org/25/122225/3/check/gate-swift-tox-func/9e7503c/ might want to see the devstack jobs too, then :) | 02:07 |
| stevemar | OK cool | 02:07 |
| notmyname | and as it is such, so also as such is it unto you | 02:07 |
| notmyname | (my reply to "the logs could might not be saved..." | 02:08 |
| stevemar | ahh | 02:08 |
| stevemar | i think this is what monty was talking about in his blog post | 02:08 |
| notmyname | the swift functional tests? yes | 02:08 |
| stevemar | notmyname, since he does call out swift in that section :P | 02:08 |
| *** tsg has joined #openstack-swift | 02:08 | |
| stevemar | notmyname, i was thinking the clients would definitely benefit from this, since i doubt most of them are run against anything live | 02:09 |
| stevemar | notmyname, how do you run the tests locally? any config needed? just setup devstack and run tox -e func ? | 02:10 |
| notmyname | stevemar: swift devs use a swift-all-in-one normally. and you can either run `tox -e func` or `./.functests` | 02:11 |
| notmyname | but the functional tests actually can run against any swift endpoint. eg I can run the functests agains rackspace cloud files | 02:11 |
| notmyname | (if you want an easy way to set up a swift all in one (SAIO) look at https://github.com/swiftstack/vagrant-swift-all-in-one) | 02:11 |
| notmyname | that automates the instructions in the swift dev docs (swift.openstack.org) | 02:12 |
| notmyname | stevemar: anything else I can answer? I gotta head home from the office | 02:12 |
| stevemar | notmyname, go home! | 02:13 |
| notmyname | :-) | 02:13 |
| stevemar | notmyname, thanks for the info :) | 02:13 |
| notmyname | any time :-) | 02:14 |
| * notmyname signs off for the night | 02:14 | |
| stevemar | have fun | 02:14 |
| *** tsg has quit IRC | 02:25 | |
| *** mahatic has quit IRC | 02:37 | |
| openstackgerrit | A change was merged to openstack/swift: container.merge_items bug https://review.openstack.org/122503 | 02:42 |
| *** gyee has quit IRC | 03:25 | |
| *** haomai___ has quit IRC | 03:41 | |
| *** haomaiwang has joined #openstack-swift | 03:42 | |
| *** Kbee has joined #openstack-swift | 03:48 | |
| *** haomaiw__ has joined #openstack-swift | 03:58 | |
| *** haomaiwang has quit IRC | 04:02 | |
| *** acoles has quit IRC | 04:11 | |
| openstackgerrit | Keshava Bharadwaj proposed a change to openstack/swift: Provides proper error handling on builder unpickle https://review.openstack.org/122225 | 04:15 |
| *** elambert has joined #openstack-swift | 04:16 | |
| openstackgerrit | Rafael Rivero proposed a change to openstack/swift: Fixes several typos (Swift) https://review.openstack.org/122602 | 04:18 |
| *** haomaiw__ has quit IRC | 04:20 | |
| *** haomaiwang has joined #openstack-swift | 04:21 | |
| Kbee | Hello, How to do a recheck on a failed cluster-rox-func job ? | 04:26 |
| Kbee | ^tox | 04:27 |
| *** haomaiw__ has joined #openstack-swift | 04:30 | |
| *** haomaiwang has quit IRC | 04:33 | |
| Kbee | notmyname: torgomatic: is there anyway to do recheck on failed cluster-tox-func job ? some node failed during execution..-> One or more hosts failed while executing task 'deploy_swift' | 04:54 |
| *** zaitcev has quit IRC | 05:29 | |
| *** Kbee has quit IRC | 05:52 | |
| *** rmcall has joined #openstack-swift | 05:57 | |
| *** Kbee has joined #openstack-swift | 05:58 | |
| *** Kbee has quit IRC | 06:04 | |
| *** tkay has joined #openstack-swift | 06:09 | |
| *** tsg has joined #openstack-swift | 06:16 | |
| openstackgerrit | A change was merged to openstack/swift: Merge master to feature/ec https://review.openstack.org/122541 | 06:18 |
| *** tkay has quit IRC | 06:21 | |
| *** tsg has quit IRC | 06:25 | |
| *** k4n0 has joined #openstack-swift | 06:29 | |
| *** tsg has joined #openstack-swift | 06:34 | |
| *** Kbee has joined #openstack-swift | 06:41 | |
| *** Dafna has quit IRC | 06:43 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/swift: Imported Translations from Transifex https://review.openstack.org/122327 | 06:44 |
| *** nshaikh has joined #openstack-swift | 06:52 | |
| *** stevemar has quit IRC | 06:59 | |
| *** ppai has joined #openstack-swift | 07:00 | |
| Kbee | notmyname: Some tempest tests are failing randomly.. None related to code change... How do handle them ? | 07:01 |
| Kbee | notmyname: already performed bug rechecks twice. | 07:01 |
| *** mkollaro has joined #openstack-swift | 07:05 | |
| *** haomaiw__ has quit IRC | 07:26 | |
| *** haomaiwang has joined #openstack-swift | 07:27 | |
| *** tsg has quit IRC | 07:30 | |
| *** rmcall has quit IRC | 07:34 | |
| *** cschwede has joined #openstack-swift | 07:36 | |
| *** haomaiwang has quit IRC | 07:44 | |
| *** haomaiwang has joined #openstack-swift | 07:52 | |
| *** elambert has quit IRC | 08:05 | |
| *** jamiehannaford has joined #openstack-swift | 08:07 | |
| *** Dafna has joined #openstack-swift | 08:10 | |
| *** mkollaro has quit IRC | 08:19 | |
| *** Kbee has quit IRC | 08:32 | |
| *** Kbee has joined #openstack-swift | 08:40 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 08:51 |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 08:52 |
| *** geaaru has joined #openstack-swift | 08:53 | |
| *** Kbee has quit IRC | 09:06 | |
| *** kbee has joined #openstack-swift | 09:07 | |
| *** kbee has quit IRC | 09:22 | |
| *** kbee has joined #openstack-swift | 09:23 | |
| *** haomaiwang has quit IRC | 09:46 | |
| *** haomaiwang has joined #openstack-swift | 09:47 | |
| *** acoles_ has joined #openstack-swift | 09:47 | |
| *** ChanServ sets mode: +v acoles_ | 09:47 | |
| openstackgerrit | Lorcan Browne proposed a change to openstack/swift: Add "--no-overlap" option to swift-dispersion populate https://review.openstack.org/118411 | 09:50 |
| *** haomai___ has joined #openstack-swift | 09:51 | |
| *** aix has joined #openstack-swift | 09:54 | |
| *** haomaiwang has quit IRC | 09:54 | |
| *** cschwede has quit IRC | 10:19 | |
| *** haomai___ has quit IRC | 10:25 | |
| *** kbee has quit IRC | 11:05 | |
| *** cschwede has joined #openstack-swift | 11:10 | |
| *** dmorita has quit IRC | 11:29 | |
| *** smart_developer has left #openstack-swift | 11:44 | |
| *** smart_developer has joined #openstack-swift | 11:44 | |
| smart_developer | Has there ever been issues with one or more of Swift's logs growing at a rapid rate? | 11:46 |
| smart_developer | Possibly 100 or more GB within an hour. | 11:47 |
| *** hhuang has quit IRC | 11:51 | |
| *** insane has quit IRC | 12:14 | |
| *** nshaikh has quit IRC | 12:22 | |
| openstackgerrit | paul luse proposed a change to openstack/swift: Early framework for the EC Reconstructor https://review.openstack.org/106910 | 12:22 |
| viklund | is there anyway of having more finegrained ACLs, IE a read-write acl that can't create containers | 12:23 |
| viklund | or a list-write ACL that can list files, upload new files but can't download files. | 12:23 |
| *** tdasilva has quit IRC | 12:26 | |
| *** mkollaro has joined #openstack-swift | 12:29 | |
| *** jamiehannaford has quit IRC | 12:32 | |
| acoles_ | viklund: it is possible to set per-container ACLs: http://docs.openstack.org/developer/swift/overview_auth.html | 12:33 |
| acoles_ | viklund: the container ACLs do not support 'list-only' | 12:35 |
| openstackgerrit | Lorcan Browne proposed a change to openstack/swift: Add new features to swift-drive-audit https://review.openstack.org/122468 | 12:36 |
| viklund | acoles_: yes, I've understood that, would it be difficult to add that? | 12:37 |
| viklund | specifically, to write some auth-middleware with that | 12:37 |
| *** acoles has joined #openstack-swift | 12:38 | |
| *** ChanServ sets mode: +v acoles | 12:38 | |
| viklund | (currently I'm just hiding or showing the download button in the horizon interface, but it would be nice to have that check in swift) | 12:38 |
| *** jamiehannaford has joined #openstack-swift | 12:42 | |
| acoles_ | viklund: in principle i guess it can be done, not sure if it would be just middleware or require support in core. | 12:46 |
| viklund | ok, maybe I'll just limit it in horizon then and write a small rest-client for bulk-uploads that uses horizon | 12:47 |
| *** tdasilva has joined #openstack-swift | 12:48 | |
| viklund | Thanks! | 12:50 |
| acoles_ | viklund: if you have a use case for list-only then maybe create a swift bug report requesting that feature | 12:53 |
| viklund | I don't know how common this would be | 12:53 |
| viklund | We are building a system for managing sensitive research data (specifically human genomes) | 12:54 |
| viklund | and want to limit what is possible to take out of the system | 12:54 |
| viklund | but it would be nice if people could see what is uploaded | 12:54 |
| viklund | uploading data to the object-store (via some secure proxy) would be a simple way for a lab to get data in. | 12:58 |
| acoles | viklund: so presumably the object names alone convey useful info when listed? or do you need to let people see object metadata (but not allow content download)? | 13:02 |
| *** acoles_ has quit IRC | 13:02 | |
| acoles | viklund: just wondering if you want 'list-only' or 'head-only' (which is also not supported !) | 13:04 |
| *** cdelatte has joined #openstack-swift | 13:04 | |
| viklund | acoles: that would not matter that much actually | 13:11 |
| viklund | list-only is probably enough, but seeing size of objects would be useful | 13:12 |
| acoles | viklund: swift container listing can include object size (and timestamp, content type, hash) but i have no idea if horizon supports that | 13:14 |
| *** ppai has quit IRC | 13:23 | |
| *** peluse_ is now known as peluse | 13:30 | |
| *** ChanServ sets mode: +v peluse | 13:31 | |
| *** tsg has joined #openstack-swift | 13:39 | |
| *** k4n0 has quit IRC | 13:40 | |
| *** BAKfr has joined #openstack-swift | 13:40 | |
| *** cca has joined #openstack-swift | 13:42 | |
| *** rmcall has joined #openstack-swift | 13:43 | |
| *** cca has left #openstack-swift | 13:44 | |
| *** cca has joined #openstack-swift | 13:45 | |
| *** cca has quit IRC | 13:45 | |
| *** dmsimard_away is now known as dmsimard | 13:45 | |
| *** cca88 has joined #openstack-swift | 13:45 | |
| *** rmcall_ has joined #openstack-swift | 13:48 | |
| *** annegent_ has joined #openstack-swift | 13:48 | |
| *** rmcall has quit IRC | 13:49 | |
| *** rmcall_ is now known as rmcall | 13:49 | |
| *** mrsnivvel has quit IRC | 13:49 | |
| *** mahatic has joined #openstack-swift | 13:49 | |
| *** tsg has quit IRC | 13:55 | |
| *** elambert has joined #openstack-swift | 14:00 | |
| *** viklund_ has joined #openstack-swift | 14:00 | |
| *** openstackgerrit has quit IRC | 14:02 | |
| *** miqui_ has joined #openstack-swift | 14:03 | |
| *** stevemar has joined #openstack-swift | 14:04 | |
| *** miqui_ has quit IRC | 14:09 | |
| *** miqui_ has joined #openstack-swift | 14:09 | |
| *** tdasilva has quit IRC | 14:10 | |
| *** aix has quit IRC | 14:10 | |
| *** miqui has quit IRC | 14:10 | |
| *** viklund has quit IRC | 14:10 | |
| *** miqui_ is now known as miqui | 14:10 | |
| *** marcusvrn has quit IRC | 14:11 | |
| *** foexle has joined #openstack-swift | 14:11 | |
| *** aix has joined #openstack-swift | 14:15 | |
| *** tdasilva has joined #openstack-swift | 14:16 | |
| *** pberis1 has quit IRC | 14:43 | |
| *** cschwede has quit IRC | 14:44 | |
| *** AbyssOne has joined #openstack-swift | 14:47 | |
| *** zhiyan has quit IRC | 14:48 | |
| *** JelleB has quit IRC | 14:48 | |
| *** elambert has quit IRC | 14:48 | |
| *** serverascode has quit IRC | 14:48 | |
| *** bsdkurt has quit IRC | 14:49 | |
| *** serverascode has joined #openstack-swift | 14:49 | |
| *** zhiyan has joined #openstack-swift | 14:49 | |
| *** bsdkurt has joined #openstack-swift | 14:50 | |
| *** elambert has joined #openstack-swift | 14:51 | |
| *** mwstorer has joined #openstack-swift | 14:52 | |
| *** tongli has joined #openstack-swift | 14:56 | |
| *** pberis has joined #openstack-swift | 15:00 | |
| notmyname | good morning | 15:00 |
| *** sandywalsh_ has joined #openstack-swift | 15:02 | |
| *** elambert has quit IRC | 15:03 | |
| *** astellwag has quit IRC | 15:03 | |
| *** annegent_ has quit IRC | 15:05 | |
| notmyname | https://bugs.launchpad.net/python-swiftclient/+bug/1371650 is annoying if you are have access to containers in more than one account | 15:10 |
| *** dencaval has quit IRC | 15:13 | |
| *** openstackgerrit has joined #openstack-swift | 15:15 | |
| goodes | perhaps you need a --no-auth-check | 15:22 |
| *** IRTermite has quit IRC | 15:23 | |
| notmyname | goodes: as a new option? why should the end-user care or have to worry about that? eg it works just fine when using the API. I think the mishandling of the 403 is the bug (if not the whole HEAD request) | 15:24 |
| openstackgerrit | Christian Cachin proposed a change to openstack/swift: Updated version of swift-enc-proxy. https://review.openstack.org/122773 | 15:25 |
| acoles | notmyname: are you using it with --os-auth-token? trying to reproduce... | 15:28 |
| *** dencaval has joined #openstack-swift | 15:28 | |
| notmyname | acoles: no. | 15:29 |
| notmyname | acoles: here's the full story (and the original report) https://ask.openstack.org/en/question/47668/help-with-write-permission-in-swift/ | 15:29 |
| *** btorch_ has joined #openstack-swift | 15:30 | |
| *** glange_ has joined #openstack-swift | 15:30 | |
| *** ChanServ sets mode: +v glange_ | 15:30 | |
| *** clayg_ has joined #openstack-swift | 15:30 | |
| *** mordred_ has joined #openstack-swift | 15:31 | |
| *** a1|away_ has joined #openstack-swift | 15:31 | |
| *** clds_ has joined #openstack-swift | 15:31 | |
| *** clayg has quit IRC | 15:31 | |
| *** mordred has quit IRC | 15:31 | |
| *** glange has quit IRC | 15:31 | |
| *** clds has quit IRC | 15:31 | |
| *** btorch has quit IRC | 15:31 | |
| *** otherjon has quit IRC | 15:31 | |
| *** jeblair has quit IRC | 15:31 | |
| *** ByteSore has quit IRC | 15:31 | |
| *** mandarine has quit IRC | 15:31 | |
| *** a1|away has quit IRC | 15:31 | |
| *** redbo has quit IRC | 15:31 | |
| *** d0ugal has quit IRC | 15:31 | |
| *** tristanC has quit IRC | 15:31 | |
| *** MooingLemur has quit IRC | 15:31 | |
| *** hhuang has joined #openstack-swift | 15:31 | |
| *** mordred_ is now known as mordred | 15:31 | |
| *** MooingLe1ur has joined #openstack-swift | 15:31 | |
| *** ByteSore has joined #openstack-swift | 15:32 | |
| *** jeblair has joined #openstack-swift | 15:32 | |
| *** alpha_ori has quit IRC | 15:32 | |
| *** tanee has quit IRC | 15:32 | |
| acoles | notmyname: isn't that the intention? that --os-storage-url is to be used in conjunction with --os-auth-token? not that i was around when it was conceived but thats how i understood it | 15:32 |
| *** mandarine has joined #openstack-swift | 15:32 | |
| *** redbo has joined #openstack-swift | 15:32 | |
| *** ChanServ sets mode: +v redbo | 15:32 | |
| acoles | and its kindof implied by the usage | 15:32 |
| notmyname | acoles: maybe. it's entirely possible that I don't know the right way to use the --os-* options | 15:33 |
| *** tristanC has joined #openstack-swift | 15:33 | |
| acoles | notmyname: me too :) | 15:33 |
| notmyname | so you're saying (1) get a token and write it down with user A, then (2) send that token using the URL for user B? | 15:33 |
| *** otherjon has joined #openstack-swift | 15:33 | |
| notmyname | not sure how that's different that what I did, so I'm sure I'm missing something | 15:34 |
| *** tanee has joined #openstack-swift | 15:34 | |
| *** alpha_ori has joined #openstack-swift | 15:34 | |
| *** d0ugal has joined #openstack-swift | 15:35 | |
| *** cca88 has quit IRC | 15:35 | |
| *** d0ugal is now known as Guest50648 | 15:35 | |
| acoles | notmyname: um, my experience with it is: get a token that authenticates me as userA, use that token with a /v1/AUTH_userB/c url, if userA is in ACL for container c then request should be granted. | 15:37 |
| *** hhuang has quit IRC | 15:37 | |
| acoles | notmyname: i need to look into the HEAD you say is generated | 15:37 |
| acoles | on the account | 15:38 |
| *** tsg has joined #openstack-swift | 15:40 | |
| *** j_king_ has joined #openstack-swift | 15:44 | |
| *** JoshNang_ has joined #openstack-swift | 15:45 | |
| *** charz_ has joined #openstack-swift | 15:45 | |
| *** charz has quit IRC | 15:46 | |
| *** JoshNang has quit IRC | 15:46 | |
| *** j_king has quit IRC | 15:46 | |
| *** ekarlso has quit IRC | 15:46 | |
| *** JoshNang_ is now known as JoshNang | 15:46 | |
| *** ekarlso has joined #openstack-swift | 15:46 | |
| *** foexle has quit IRC | 15:47 | |
| *** MooingLe1ur is now known as MooingLemur | 15:47 | |
| *** BAKfr has quit IRC | 15:49 | |
| swifterdarrell | notmyname: acoles: I regularly use --os-storage-url without --os-auth-token; the behavior I've observed (and desired) was to use the auth response's X-Auth-Token but use the supplied storage URL instead of the auth response's X-Storage-Url | 15:50 |
| acoles | notmyname: is this what you see http://paste.openstack.org/show/113351/ ? those CLI commands generate no requests to swift service | 15:50 |
| acoles | swifterdarrell: ok, i didn't realise it was supposed to work that way too, makes sense though | 15:52 |
| notmyname | swifterdarrell: ya, that's what I expected | 15:53 |
| swifterdarrell | acoles: can you try w/the "-U" and "-K" before the command ("stat")? | 15:53 |
| swifterdarrell | acoles: i thought they needed to go before teh command, but it's been a while | 15:53 |
| openstackgerrit | Christian Cachin proposed a change to openstack/swift: Implement encryption proxy middleware for server-side-encryption. https://review.openstack.org/122773 | 15:54 |
| *** kyles_ne has joined #openstack-swift | 16:00 | |
| acoles | swifterdarrell: tried that, no difference, but what i did need to do is add -A (i don't often use those flags !) | 16:01 |
| acoles | that makes usage as you describe: authenticate as userA, send container request to userB account | 16:02 |
| acoles | then i see the HEAD | 16:02 |
| smart_developer | What are the .conf files that should be there in /etc/swift upon a standard install of Swift? | 16:10 |
| smart_developer | I'm think I might be missing one, but then again I might be wrong. | 16:11 |
| acoles | notmyname: can you add a trace to that bug report? so we're clear exactly what the scenario is. | 16:13 |
| acoles | notmyname: seems to be a bunch of ways to get stack traces from swiftclient playing with those options :( | 16:14 |
| notmyname | fun | 16:14 |
| notmyname | yeah, I'll try to flesh it out later | 16:14 |
| notmyname | redbo: have you looked at torgomatic's zero-copy patches? (https://review.openstack.org/#/c/102609/) you had worked on something similar a while back | 16:20 |
| *** kyles_ne has quit IRC | 16:20 | |
| *** kyles_ne has joined #openstack-swift | 16:21 | |
| *** annegent_ has joined #openstack-swift | 16:26 | |
| *** tkay has joined #openstack-swift | 16:29 | |
| *** Guest50648 is now known as d0ugal | 16:29 | |
| *** d0ugal has quit IRC | 16:30 | |
| *** d0ugal has joined #openstack-swift | 16:30 | |
| *** annegent_ has quit IRC | 16:30 | |
| redbo | notmyname: I've looked at the code, and it's neat. I intend to do some benchmarking on it soon, and a little more research on how many problems that md5 check actually catches. So far it seems like some, but not many. | 16:35 |
| *** gyee has joined #openstack-swift | 16:37 | |
| notmyname | cool, thanks. I'd really love your input on it (and if it's good I'd really like to see it land in the next 2 weeks so it can be available as part of the juno release) | 16:37 |
| *** thurloat has quit IRC | 16:43 | |
| *** sandywalsh_ has quit IRC | 16:46 | |
| *** gyee has quit IRC | 16:48 | |
| *** thurloat has joined #openstack-swift | 16:48 | |
| redbo | I guess I can just cram a configurable "use sendfile and don't md5" mode next to it later if I'm so hot on that idea. It should be able to share most of the code. | 16:51 |
| peluse | notmyname, torgomatic , tsg : feature/ec rebased as well as recon framework which I just pushed through.... | 16:54 |
| notmyname | thanks | 16:54 |
| tsg | thanks peluse! | 16:54 |
| peluse | I'm going to start tackling the recon node-node comm next, dig deeper into the various schemes we've BS'd and see what I can come up with for a proposal for next week, maybe some code by hackathon | 16:55 |
| *** stevemar has quit IRC | 17:01 | |
| *** stevemar has joined #openstack-swift | 17:02 | |
| *** robbybb has joined #openstack-swift | 17:11 | |
| robbybb | Is there any standard manifest format for storing large files (bigger than the largest supported Swift object)? | 17:11 |
| notmyname | robbybb: two of them | 17:12 |
| notmyname | robbybb: http://docs.openstack.org/developer/swift/overview_large_objects.html | 17:12 |
| notmyname | long response to openstack mailing list thread discussion sent. | 17:17 |
| *** gyee has joined #openstack-swift | 17:17 | |
| *** geaaru has quit IRC | 17:18 | |
| *** gyee has quit IRC | 17:24 | |
| *** zul has quit IRC | 17:37 | |
| *** zul has joined #openstack-swift | 17:37 | |
| clayg_ | swifterdarrell: --os-storage-url is the bomb | 17:51 |
| clayg_ | smart_developer: swift.conf [proxy|account|container|object]-server.conf should do it. | 17:52 |
| clayg_ | smart_developer: the internal client daemons (object-expirer/container-reconciler) need a seperate conf with a app:pipeline; container-sync-realms.conf can be a thing too... | 17:53 |
| *** tongli has quit IRC | 17:56 | |
| *** openstackgerrit has quit IRC | 18:01 | |
| *** openstackgerrit has joined #openstack-swift | 18:03 | |
| *** tsg has quit IRC | 18:15 | |
| *** vu has quit IRC | 18:25 | |
| *** shri has joined #openstack-swift | 18:25 | |
| *** clayg_ is now known as clayg | 18:28 | |
| *** ChanServ sets mode: +v clayg | 18:28 | |
| *** jamiehannaford has quit IRC | 18:29 | |
| *** occupant has joined #openstack-swift | 18:31 | |
| notmyname | wow. 22 PTLs to be elected. https://wiki.openstack.org/wiki/PTL_Elections_September/October_2014 | 18:49 |
| *** kyles_ne has quit IRC | 18:58 | |
| *** kyles_ne has joined #openstack-swift | 18:58 | |
| smart_developer | In Swift, does [filter:______] mean it's a middleware ? | 18:59 |
| smart_developer | Basically, is a "filter" a "middleware"? | 18:59 |
| torgomatic | smart_developer: yes, the terms are equivalent | 19:00 |
| *** kyles_ne has quit IRC | 19:03 | |
| *** aix has quit IRC | 19:06 | |
| *** LarsN has left #openstack-swift | 19:07 | |
| portante | http://lifehacker.com/373571/watch-star-wars-in-text-via-telnet | 19:12 |
| portante | amazing this site is still there and it still works | 19:13 |
| Tyger | But does Han shoot first? | 19:14 |
| notmyname | yay towel.blikenlights! | 19:14 |
| notmyname | Tyger: I've never watched it that far :-) | 19:14 |
| smart_developer | Thanks. | 19:19 |
| *** tkay has quit IRC | 19:19 | |
| smart_developer | In order to have middle, e.g., TempURL, working, do you need to do something more than just the standard install of Swift ? | 19:19 |
| smart_developer | Because I'm trying to test it out, | 19:19 |
| smart_developer | (tempURL) | 19:19 |
| smart_developer | but there is no utility/path: /opt/ss/bin/swift-temp-url | 19:19 |
| *** tkay has joined #openstack-swift | 19:20 | |
| smart_developer | I'm following https://www.swiftstack.com/docs/admin/middleware/tempurl.html instructions. | 19:20 |
| smart_developer | Under "Generating a URL" | 19:20 |
| notmyname | smart_developer: that's the path where swiftstack installs stuff. it may be different depending on how you've installed it | 19:21 |
| notmyname | smart_developer: but the swift-temp-url helper tool is in the bin directory of the source tree | 19:21 |
| notmyname | smart_developer: and to answer your other question, there is a lot of middleware included in the "standard install" of swift. tempurl is one of those. | 19:22 |
| *** vu has joined #openstack-swift | 19:23 | |
| Trixboxer | I'm writting a php web client for swift 1.4, Is there a way to allow download via the php web client ? I'm getting 401 when redirecting the user to download URL ... I think it has something to do with CORS however CORS was introduced in 1.7.5 | 19:24 |
| torgomatic | Trixboxer: generate a tempurl for the object and redirect the client there | 19:25 |
| *** kyles_ne has joined #openstack-swift | 19:29 | |
| *** ndk has joined #openstack-swift | 19:31 | |
| *** zaitcev has joined #openstack-swift | 19:31 | |
| *** ChanServ sets mode: +v zaitcev | 19:31 | |
| Trixboxer | torgomatic: thanks.. but is there way to get cors ? or i should upgrade to latest ? | 19:35 |
| Trixboxer | also is there any place where swift maintains critical updates (security/stability) against features ? | 19:36 |
| torgomatic | Trixboxer: if you want cors, you'll have to upgrade, though I don't see what cors has to do with allowing downloads directly from swift to client | 19:36 |
| torgomatic | ...I'm not sure what that last one means | 19:36 |
| torgomatic | "critical updates against features"? | 19:37 |
| *** evanjfraser has quit IRC | 19:40 | |
| *** evanjfraser has joined #openstack-swift | 19:40 | |
| smart_developer | notmyname : So basically anything that's referred to as the Swift middleware is middleware that automatically comes with your standard Swift install ? | 19:43 |
| *** kyles_ne has quit IRC | 19:44 | |
| Trixboxer | torgomatic: I mean if the visitor is browsing swift from my web php client then the web site address should be recognized by swift and let him download the file directly -- I think this is CORS and I'll have to update if I need it | 19:44 |
| *** kyles_ne has joined #openstack-swift | 19:44 | |
| Trixboxer | torgomatic: "critical updates against features" -> I mean the updates those are absolutely necessary as they solve security and stability issues | 19:44 |
| torgomatic | so client --> php app --> swift for "browsing", then client --> swift for download? | 19:45 |
| Trixboxer | yes | 19:45 |
| notmyname | smart_developer: http://docs.openstack.org/developer/swift/middleware.html | 19:45 |
| torgomatic | Trixboxer: security issues typically get backported to the latest 6-month openstack release, but no further | 19:45 |
| torgomatic | distros may do more, of course, but that's not something swift devs take care of | 19:45 |
| *** kyles_ne has quit IRC | 19:45 | |
| *** kyles_ne has joined #openstack-swift | 19:46 | |
| torgomatic | Trixboxer: I don't think this is CORS that you're looking for... I'm no CORS expert, but I think CORS is about being able to refer to Swift-hosted .js files and having Swift set appropriate headers so browsers will take them | 19:46 |
| torgomatic | I think tempurl is more likely to be the tool that you're looking for | 19:46 |
| *** vu has quit IRC | 19:48 | |
| smart_developer | Okay -> whenever I do curl -k -i https://172.20.20.11/v1/AUTH_a438ebed5eb749efa4bed51b8ccc0c92/sample_container/sample.txt?temp_url_sig=374f7b1e775cdf8ab991c9a2b1fb1e2a25e82f77&temp_url_expires=1411154923 | 19:50 |
| smart_developer | it gives me 404: Not found | 19:50 |
| smart_developer | HTTP 1.1 / 404 Not found | 19:50 |
| smart_developer | this is for tempURL | 19:50 |
| smart_developer | Any obvious errors? | 19:51 |
| *** morganfainberg is now known as CaptainMorgan | 19:52 | |
| smart_developer | I was following the SwiftStack documentation: https://www.swiftstack.com/docs/admin/middleware/tempurl.html | 19:52 |
| torgomatic | smart_developer: does that object actually exist? tempurl gives a 401 response on invalid signature, so it looks like your tempurl is working, but the object doesn't exist | 19:54 |
| notmyname | smart_developer: in that curl command, did you escape the &? (ie "\&" | 19:54 |
| clayg | notmyname: some shells are better at that than bash... | 19:54 |
| torgomatic | stick quotes around the URL; it's harder to miss a spot that way | 19:54 |
| clayg | notmyname: either way I agree with torgomatic that he wouldn't get a 404 if the sig was invalid... | 19:54 |
| notmyname | clayg: emacs doesn't count | 19:55 |
| *** stevemar has left #openstack-swift | 19:55 | |
| clayg | notmyname: idk acorwin uses some crazy shell and I'm always like "you can't do that it won't work" and he's like, "nah it'll be fine" and he's right... | 19:55 |
| Trixboxer | torgomatic: thanks, I'll go for that | 19:55 |
| smart_developer | "/&" ? | 19:56 |
| smart_developer | Where? | 19:56 |
| smart_developer | \& * | 19:56 |
| notmyname | smart_developer: nah, do what torgomatic said and put the whole thing in quotes | 19:56 |
| smart_developer | torgomatic : It's been a while since I created that particular tempURL, so it probably has expired already ? | 19:57 |
| smart_developer | That may be the 401 you're seeing. | 19:57 |
| notmyname | smart_developer: also, that's probably not what's resulting in a 404. just make sure the object still exsists | 19:57 |
| torgomatic | smart_developer: I didn't try it | 19:57 |
| clayg | peluse: you around? I suddenly realized that we don't normally trigger migrations on read's - i woulder if the logic in get_policy_stats should be to omit the container-count keys when they're not there, and then have the account utils function only add headers for the data that comes out of the function | 19:57 |
| smart_developer | Okay. | 19:57 |
| peluse | clayg, I'm here for a min - have to run into a mtg for 30 min then will be back | 19:57 |
| clayg | peluse: then we could trigger migration to add the triggers in the auditor only? | 19:58 |
| clayg | peluse: other option is trigger the migration on read and welcome ourselves to the new world of whatever i migrate whenever i need to - but then also migrate in the auditor so we don't get stuck with unmigrated containers on any more multi-policy deployments just because they don't often read all accounts? | 19:59 |
| smart_developer | notmyname torgotmatic : I can download the file, so I know it exists. | 20:00 |
| smart_developer | But I'm still getting the HTTP 1.1 / 404 Not found error. | 20:00 |
| acorwin | clayg: notmyname that's my queue to fanatically plug zsh!! http://www.zsh.org/ | 20:00 |
| *** robbybb has quit IRC | 20:00 | |
| clayg | peluse: I change the for policy in POLICES: loop in the migration into a slick sub-select, I think it's pretty groovy | 20:00 |
| peluse | clayg, I looked a bit at the auditor option and the test code got super ugly so I punted at least for now and went with what you see up there now | 20:00 |
| peluse | clayg, note that I just got back from a conf and haven't been checking a lot of things often, just realized you left some comments that I haven't read yet. Will read them here after my 1:00 mtg and ping ya back | 20:01 |
| clayg | peluse: yeah one time somewhere, not sure if it ever got pushed/merged I like imported a PreMigrationTestCase for one of the backends into a test for a background daemon - it was sorta wonky... | 20:02 |
| torgomatic | smart_developer: well, all I can tell you is that the 404 isn't coming from tempurl; there's no 404 response generated anywhere in there | 20:02 |
| * torgomatic heads to lunch | 20:02 | |
| clayg | peluse: ok no worries - i guess I'll look at tests for the auditor change and wait and think about maybe only doing migrations from there... | 20:02 |
| *** gyee has joined #openstack-swift | 20:15 | |
| smart_developer | notmyname torgomatic : Maybe the 404 issue has to do with that whenever I start with 0 containers, 0 objects | 20:16 |
| smart_developer | and then create a new container + upload a simple.txt file | 20:16 |
| -openstackstatus- NOTICE: Gerrit will be offline from 20:30 to 20:50 UTC for project renames | 20:16 | |
| *** ChanServ changes topic to "Gerrit will be offline from 20:30 to 20:50 UTC for project renames" | 20:16 | |
| smart_developer | For a long time, "swift ... stat" command only responds by saying 1 container, 0 objects. | 20:16 |
| smart_developer | And then creating another new container + uploading a new file called simple2.txt | 20:17 |
| smart_developer | yields 1 container, 0 bytes. | 20:18 |
| smart_developer | 0 objects* | 20:18 |
| smart_developer | Actually, after the first object upload (simple.txt), it yielded 0 containers, 0 objects (which was contrary to what I said a bit ago). | 20:18 |
| smart_developer | But now, after uploading the 2nd container and object, it yields 1 container, 1 object. | 20:19 |
| smart_developer | So it's hard to even discern a pattern with the info I am getting back. | 20:19 |
| smart_developer | It's not like it's consistently behind by 1, or anything like that. | 20:19 |
| smart_developer | ..... | 20:19 |
| zaitcev | Your updaters are bogging down, or maybe even not running on all nodes. | 20:20 |
| smart_developer | Now finally it (correctly) says there's 2 containers, and 2 objects, when running the "swift ... stat" command. | 20:21 |
| smart_developer | zaitcev : How long does it normally take for the correct # of containers, # of objects, info to show correctly on a small, 3-node cluster? | 20:24 |
| clayg | smart_developer: depends on the interval of the container updaters? | 20:24 |
| clayg | smart_developer: maybe 5 mins? | 20:25 |
| smart_developer | clayg : Okay, so it should be in the .conf files then, right? | 20:33 |
| smart_developer | How do you check what the default values/properties for you Swift install, are? | 20:34 |
| *** tdasilva has quit IRC | 20:35 | |
| smart_developer | (in the container-update section of container-server.conf there is nothing). | 20:35 |
| smart_developer | (the [container-update] section) | 20:35 |
| smart_developer | updater* | 20:35 |
| peluse | clayg, OK, I'm back. Yeah, agree on the auditor thing for sure and I think the idea about the change to get_policy_stats() makes sense too. Are you going to throw something up there on the audtior side? I can wait on that and then knock out the get_policy_stats() change if you want | 20:38 |
| clayg | ok, sounds reasonable - yeah I'm definately going to push up what I have going, dunno yet if i'll end up grabbing the get_policy_stats change | 20:51 |
| *** ChanServ changes topic to "Swift 2.1.0 https://launchpad.net/swift/+milestone/2.1.0 | Swift Review Dashboard: http://bit.ly/1iVBigF | Priority Reviews: https://wiki.openstack.org/wiki/Swift/PriorityReviews | Ideas: https://wiki.openstack.org/wiki/Swift/ideas" | 20:52 | |
| -openstackstatus- NOTICE: Gerrit is back online | 20:52 | |
| smart_developer | That's strange.... so I was at "swift ... stat" gave me 2 containers, 2 objects, (which is correct), but now after a while it's yielding 2 containers, only 1 object. | 20:53 |
| smart_developer | I did a "swift-init all restart" in between, but I'm not sure if this has anything to do with that ..... | 20:55 |
| *** openstackgerrit has quit IRC | 20:55 | |
| *** openstackgerrit has joined #openstack-swift | 20:57 | |
| *** gyee has quit IRC | 21:03 | |
| *** mkollaro1 has joined #openstack-swift | 21:08 | |
| smart_developer | but both objects seem to be download-able via swiftclient. | 21:09 |
| *** mkollaro has quit IRC | 21:10 | |
| *** mkollaro1 has quit IRC | 21:13 | |
| smart_developer | Now ..... container count is 2, but object count is 0. | 21:20 |
| smart_developer | Without having done anything | 21:21 |
| smart_developer | Does anyone have an example of what a formpost source code looks like? | 21:26 |
| smart_developer | Thanks. | 21:26 |
| *** CaptainMorgan is now known as morganfainberg | 21:28 | |
| *** mkollaro has joined #openstack-swift | 21:30 | |
| torgomatic | smart_developer: the formpost source is at swift/common/middleware/formpost.py | 21:42 |
| smart_developer | torgomatic : Thanks. | 21:48 |
| *** Zerpies has joined #openstack-swift | 21:48 | |
| Zerpies | Minor question - if I have a public container in swift - is there a way I can disable the file listing if I access the container URL directly in the browser? | 21:48 |
| Zerpies | s/file/object | 21:48 |
| *** mahatic has quit IRC | 21:53 | |
| smart_developer | Quick question: How do I know what the default values are for my Swift configuration? | 22:00 |
| smart_developer | I am using a modified version of Swift 1.13.1 | 22:01 |
| smart_developer | and already was told that there are differences between what I've been using, and the {proxy-server,account-server,container-server,object-server}.conf-sample files available on the web. | 22:01 |
| smart_developer | (basically the ones found here: https://github.com/openstack/swift/tree/master/etc) | 22:02 |
| smart_developer | So when certain properties/values/options aren't specified in my own .conf files, what's the most straightforward way for me to see/know what they've been set to? | 22:03 |
| smart_developer | Thanks. | 22:03 |
| *** robbybb has joined #openstack-swift | 22:09 | |
| *** rmcall has quit IRC | 22:12 | |
| *** mahatic has joined #openstack-swift | 22:12 | |
| clayg | smart_developer: normally listings aren't available on public containers (.r:*) unless you add the listing flag (.rlisting) | 22:27 |
| smart_developer | clayg : In swift/container/updater.py | 22:28 |
| smart_developer | Is this | 22:28 |
| smart_developer | self.interval = int(conf.get('interval', 300)) | 22:28 |
| smart_developer | Is the 300 in seconds, or in minutes, and where would you look? | 22:28 |
| smart_developer | That means the number it takes in order to update the container metadata, right? (for instance, the # of objects that are listed in the container)? | 22:29 |
| smart_developer | and what/where is conf.get ? | 22:29 |
| *** robbybb has quit IRC | 22:35 | |
| openstackgerrit | Florent Flament proposed a change to openstack/swift: Limit partition movement when adding a new tier https://review.openstack.org/121422 | 22:43 |
| *** dmsimard is now known as dmsimard_away | 22:50 | |
| *** robbybb has joined #openstack-swift | 22:54 | |
| *** echevemaster has joined #openstack-swift | 23:12 | |
| *** robbybb has quit IRC | 23:25 | |
| mattoliverau | smart_developer: its in seconds and I think interval is the time between runs (although I could be wrong) I'm outside on my phone so can't check the code | 23:27 |
| *** mahatic has quit IRC | 23:34 | |
| *** tkay has quit IRC | 23:37 | |
| openstackgerrit | Clay Gerrard proposed a change to openstack/swift: Add container_count to policy_stat table https://review.openstack.org/120795 | 23:42 |
| *** judd7 has joined #openstack-swift | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!