| *** dja has joined #openstack-swift | 00:00 | |
| notmyname | timburke: not having looked at this in a while, it does re-raise the question in my mind: why do we have *any* aws auth stuff in tempauth in the first place? | 00:06 |
|---|---|---|
| notmyname | "hysterical raisons" I'm sure | 00:07 |
| timburke | notmyname: that's my assumption. i suppose it demonstrates how other middleware authors may want to implement their middleware? | 00:08 |
| notmyname | lol, right. "example code". not at all because swift3 used to be in the repo, and when we pulled it out we didn't cleanly extract the authy bits from the auth system we don't care much about | 00:09 |
| *** _JZ_ has quit IRC | 00:11 | |
| *** dja has quit IRC | 00:12 | |
| clayg | acoles: so in ssync, self.df_mgr.yield_hashes is quite happen to have be given a list of suffixes which contain only empty hash dirs | 00:35 |
| *** dja has joined #openstack-swift | 00:35 | |
| clayg | it's not cleanup_ondisk_files (or any of the methods that sheperd around bags of filenames or ondisk_info) that detects and reaps empty hashdirs - it's _hash_suffix_dir | 00:36 |
| *** sams-gleb has joined #openstack-swift | 00:52 | |
| *** sams-gleb has quit IRC | 00:56 | |
| acoles | clayg: so it might be interesting to consider whether this code could actually move into cleanup_ondisk_files?? https://github.com/openstack/swift/blob/9be1d8ba28f9702e8e0e5024fd12f367a18506c4/swift/obj/diskfile.py#L983-L988 | 00:57 |
| *** catintheroof has joined #openstack-swift | 01:04 | |
| *** jamielennox is now known as jamielennox|away | 01:16 | |
| tdasilva | https://medium.com/@jim_dowling/reflections-on-s3s-architectural-flaws-71f14c05a5fa#.5y4ap4jup | 01:19 |
| tdasilva | interesting to start reading something that you might think might be an unbiased review only to then find out at the end that it's all a setup to introduce my project | 01:19 |
| *** catintheroof has quit IRC | 02:01 | |
| *** JimCheung has joined #openstack-swift | 02:11 | |
| clayg | acoles: *gasp* very interesting indeed! | 02:13 |
| *** klrmn has quit IRC | 02:13 | |
| clayg | tdasilva: i started to look at that the other day and didn't make it too far | 02:14 |
| kota_ | morning again at office | 02:40 |
| kota_ | it seems to sleep too much | 02:40 |
| *** tonanhngo has joined #openstack-swift | 02:42 | |
| *** tonanhngo has quit IRC | 02:46 | |
| clayg | kota_: I think this is g2g - the py35 experimental check passes -> https://review.openstack.org/#/c/438770 | 02:52 |
| patchbot | patch 438770 - pyeclib - Don't let nose play with our PYTHONPATH. | 02:52 |
| *** sams-gleb has joined #openstack-swift | 02:54 | |
| *** _JZ_ has joined #openstack-swift | 02:58 | |
| *** sams-gleb has quit IRC | 02:59 | |
| *** dmorita_ has quit IRC | 03:03 | |
| *** links has joined #openstack-swift | 03:08 | |
| kota_ | clayg: alright, it looks like you and timburke had an answer to reproduce it yeah? | 03:16 |
| kota_ | will look at | 03:17 |
| *** JimCheung has quit IRC | 03:23 | |
| *** JimCheung has joined #openstack-swift | 03:24 | |
| *** JimCheung has quit IRC | 03:25 | |
| *** JimCheung has joined #openstack-swift | 03:25 | |
| *** JimCheung has quit IRC | 03:30 | |
| *** klrmn has joined #openstack-swift | 03:50 | |
| *** rcernin has quit IRC | 03:50 | |
| *** klrmn has quit IRC | 03:52 | |
| *** dmorita has joined #openstack-swift | 04:08 | |
| *** dmorita has quit IRC | 04:10 | |
| *** dmorita has joined #openstack-swift | 04:10 | |
| *** dmorita_ has joined #openstack-swift | 04:11 | |
| *** dmorita has quit IRC | 04:11 | |
| *** dmorita has joined #openstack-swift | 04:15 | |
| *** dmorita_ has quit IRC | 04:15 | |
| *** dmorita_ has joined #openstack-swift | 04:17 | |
| *** _JZ_ has quit IRC | 04:19 | |
| *** dmorita has quit IRC | 04:19 | |
| *** dmorita_ has quit IRC | 04:21 | |
| *** dmorita has joined #openstack-swift | 04:24 | |
| *** dmorita has quit IRC | 04:27 | |
| *** psachin has joined #openstack-swift | 04:27 | |
| *** dmorita has joined #openstack-swift | 04:27 | |
| kota_ | ok, i got the same error with the way to reproduce timburke reported | 04:29 |
| *** mmotiani_ has joined #openstack-swift | 04:31 | |
| *** dmorita has quit IRC | 04:31 | |
| *** dmorita has joined #openstack-swift | 04:32 | |
| *** Jeffrey4l has joined #openstack-swift | 04:32 | |
| *** dmorita has quit IRC | 04:36 | |
| *** dmorita has joined #openstack-swift | 04:36 | |
| *** dmorita_ has joined #openstack-swift | 04:37 | |
| *** dmorita has quit IRC | 04:37 | |
| *** dmorita_ has quit IRC | 04:42 | |
| *** klrmn has joined #openstack-swift | 04:44 | |
| clayg | kota_: timburke is awesome | 04:46 |
| clayg | tdasilva: do you like diffs, follow-ups, or push-overs? | 04:46 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: refactor some common code from crypto https://review.openstack.org/440028 | 04:50 |
| *** klrmn has quit IRC | 04:53 | |
| *** sams-gleb has joined #openstack-swift | 04:57 | |
| *** dmorita has joined #openstack-swift | 04:59 | |
| *** dmorita_ has joined #openstack-swift | 05:03 | |
| *** sams-gleb has quit IRC | 05:03 | |
| *** dmorita has quit IRC | 05:04 | |
| *** dmorita_ has quit IRC | 05:07 | |
| *** dmorita has joined #openstack-swift | 05:13 | |
| *** dmorita has quit IRC | 05:13 | |
| *** dmorita has joined #openstack-swift | 05:14 | |
| *** dmorita_ has joined #openstack-swift | 05:15 | |
| *** dmorita has quit IRC | 05:15 | |
| *** dmorita_ has quit IRC | 05:18 | |
| *** adriant_ has quit IRC | 05:20 | |
| *** dja has quit IRC | 05:28 | |
| *** mmotiani_ has quit IRC | 05:54 | |
| *** dja has joined #openstack-swift | 05:56 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift master: Add container sharding to Swift containers https://review.openstack.org/423906 | 06:00 |
| *** dja has quit IRC | 06:07 | |
| *** psachin_ has joined #openstack-swift | 06:18 | |
| *** psachin_ has quit IRC | 06:19 | |
| *** dja has joined #openstack-swift | 06:26 | |
| *** JimCheung has joined #openstack-swift | 06:59 | |
| *** sams-gleb has joined #openstack-swift | 07:01 | |
| *** McMurlock1 has joined #openstack-swift | 07:03 | |
| *** ChubYann has quit IRC | 07:04 | |
| *** sams-gleb has quit IRC | 07:05 | |
| *** dmorita has joined #openstack-swift | 07:09 | |
| *** dmorita has quit IRC | 07:14 | |
| *** rcernin has joined #openstack-swift | 07:18 | |
| *** tesseract has joined #openstack-swift | 07:18 | |
| *** JimCheung has quit IRC | 07:32 | |
| *** zhugaoxiao has joined #openstack-swift | 07:34 | |
| *** hseipp has joined #openstack-swift | 07:43 | |
| *** winggundamth has joined #openstack-swift | 08:10 | |
| *** pcaruana has joined #openstack-swift | 08:13 | |
| *** sams-gleb has joined #openstack-swift | 08:16 | |
| *** bob_cheesey has joined #openstack-swift | 08:18 | |
| *** geaaru has joined #openstack-swift | 08:24 | |
| *** jordanP has joined #openstack-swift | 08:46 | |
| *** hseipp has quit IRC | 08:52 | |
| *** dja has quit IRC | 09:02 | |
| *** SkyRocknRoll has joined #openstack-swift | 09:11 | |
| *** SkyRocknRoll has quit IRC | 09:11 | |
| *** dmorita has joined #openstack-swift | 09:12 | |
| *** amoralej|off is now known as amoralej | 09:13 | |
| *** cbartz has joined #openstack-swift | 09:13 | |
| *** dmorita has quit IRC | 09:17 | |
| *** oshritf has joined #openstack-swift | 09:19 | |
| *** dja has joined #openstack-swift | 09:23 | |
| *** Jeffrey4l has quit IRC | 09:25 | |
| *** Jeffrey4l has joined #openstack-swift | 09:25 | |
| *** dja has quit IRC | 09:29 | |
| *** Jeffrey4l has quit IRC | 09:31 | |
| *** cshastri has joined #openstack-swift | 09:34 | |
| *** dmorita has joined #openstack-swift | 09:39 | |
| *** kei_yama has quit IRC | 09:39 | |
| *** bkopilov has joined #openstack-swift | 09:42 | |
| *** dmorita_ has joined #openstack-swift | 09:42 | |
| *** dmorita__ has joined #openstack-swift | 09:43 | |
| *** dmorita_ has quit IRC | 09:43 | |
| *** Jeffrey4l has joined #openstack-swift | 09:43 | |
| *** dmorita has quit IRC | 09:44 | |
| *** dmorita__ has quit IRC | 09:47 | |
| *** dmorita has joined #openstack-swift | 09:48 | |
| *** oshritf has quit IRC | 09:49 | |
| openstackgerrit | Christopher Bartz proposed openstack/python-swiftclient master: ISO 8601 timestamps for tempurl https://review.openstack.org/423377 | 09:52 |
| *** dmorita has quit IRC | 09:52 | |
| openstackgerrit | Rico Lin proposed openstack/swift master: [Fix gate]Update test requirement https://review.openstack.org/440274 | 09:55 |
| openstackgerrit | Rico Lin proposed openstack/python-swiftclient master: [Fix gate]Update test requirement https://review.openstack.org/440277 | 09:56 |
| *** jordanP has quit IRC | 09:57 | |
| openstackgerrit | Rico Lin proposed openstack/swift-bench master: [Fix gate]Update test requirement https://review.openstack.org/440282 | 10:01 |
| *** oshritf has joined #openstack-swift | 10:06 | |
| *** bob_cheesey has quit IRC | 10:25 | |
| *** sams-gle_ has joined #openstack-swift | 10:27 | |
| *** sams-gleb has quit IRC | 10:27 | |
| *** links has quit IRC | 10:31 | |
| openstackgerrit | Merged openstack/pyeclib master: Don't let nose play with our PYTHONPATH. https://review.openstack.org/438770 | 10:33 |
| *** bob_cheesey has joined #openstack-swift | 10:46 | |
| *** links has joined #openstack-swift | 10:47 | |
| *** mvk has quit IRC | 10:57 | |
| *** gcb has quit IRC | 10:57 | |
| *** bob_cheesey has quit IRC | 11:13 | |
| *** oshritf has quit IRC | 11:13 | |
| *** oshritf has joined #openstack-swift | 11:14 | |
| *** sams-gle_ has quit IRC | 11:17 | |
| *** bob_cheesey has joined #openstack-swift | 11:17 | |
| *** sams-gleb has joined #openstack-swift | 11:17 | |
| *** bob_cheesey has quit IRC | 11:18 | |
| *** oshritf has quit IRC | 11:20 | |
| *** sams-gleb has quit IRC | 11:22 | |
| *** klrmn has joined #openstack-swift | 11:22 | |
| *** klrmn1 has joined #openstack-swift | 11:25 | |
| *** klrmn has quit IRC | 11:26 | |
| *** mvk has joined #openstack-swift | 11:28 | |
| *** links has quit IRC | 11:51 | |
| *** klrmn1 has quit IRC | 11:55 | |
| kota_ | oops? patchbot is absent? | 11:58 |
| kota_ | patch 440466 | 11:59 |
| patchbot | https://review.openstack.org/#/c/440466/ - swift - Draft: TestObjController refactoring | 11:59 |
| kota_ | hello patch bot but maybe something is in trouble from gerrit notification | 11:59 |
| kota_ | acoles: patch 440466 is for the homework from Atlanta PTG. to refactor the structure of TestObjController and ECTestMixin. | 12:01 |
| patchbot | https://review.openstack.org/#/c/440466/ - swift - Draft: TestObjController refactoring | 12:01 |
| *** sams-gleb has joined #openstack-swift | 12:02 | |
| kota_ | acoles: it's still WIP because I have notes to understand the test classes relations but if you could quick check the way I'd try to go, it'll super helpful because it will be large diff anyway. | 12:03 |
| *** links has joined #openstack-swift | 12:04 | |
| *** dja has joined #openstack-swift | 12:07 | |
| *** winggundamth has quit IRC | 12:09 | |
| *** bkopilov has quit IRC | 12:10 | |
| *** gkadam has joined #openstack-swift | 12:13 | |
| *** oshritf has joined #openstack-swift | 12:26 | |
| *** oshritf has quit IRC | 12:32 | |
| *** oshritf has joined #openstack-swift | 12:33 | |
| *** oshritf has quit IRC | 12:38 | |
| *** oshritf has joined #openstack-swift | 12:38 | |
| *** gkadam has quit IRC | 12:46 | |
| *** catintheroof has joined #openstack-swift | 12:56 | |
| *** oshritf has quit IRC | 13:04 | |
| *** cshastri has quit IRC | 13:07 | |
| *** jlwhite has quit IRC | 13:15 | |
| *** jlwhite has joined #openstack-swift | 13:15 | |
| *** oshritf has joined #openstack-swift | 13:17 | |
| *** dja has quit IRC | 13:29 | |
| *** klamath has joined #openstack-swift | 13:31 | |
| *** klamath has quit IRC | 13:32 | |
| *** klamath has joined #openstack-swift | 13:32 | |
| *** chlong_ has joined #openstack-swift | 13:40 | |
| jrichli | mathiasb: thanks for setting up the devstack! | 13:45 |
| *** Jeffrey4l has quit IRC | 13:49 | |
| mathiasb | jrichli: yw! | 13:52 |
| *** mvk has quit IRC | 13:52 | |
| *** oshritf has quit IRC | 13:59 | |
| *** jith has quit IRC | 14:07 | |
| *** links has quit IRC | 14:12 | |
| *** dmorita has joined #openstack-swift | 14:23 | |
| *** dmorita has quit IRC | 14:23 | |
| *** dmorita has joined #openstack-swift | 14:24 | |
| *** Jeffrey4l has joined #openstack-swift | 14:29 | |
| *** Jeffrey4l has quit IRC | 14:35 | |
| *** _JZ_ has joined #openstack-swift | 14:44 | |
| *** _JZ_ has quit IRC | 14:44 | |
| *** gabor_antal_km has joined #openstack-swift | 14:50 | |
| *** ujjain has quit IRC | 14:50 | |
| *** gabor_antal_ has quit IRC | 14:51 | |
| *** sams-gleb has quit IRC | 14:53 | |
| *** mvk has joined #openstack-swift | 14:56 | |
| jrichli | oh - and for those who will be using devstack for testing the root secret in Barbican, I suggest that you do not reboot your devstack vm. | 15:00 |
| jrichli | unless you have already worked with devstack enough to know how to start keystone and barbican to be exactly how stack.sh sets it up. | 15:01 |
| jrichli | or unless you are ok with re-provisioning some things afterwards | 15:01 |
| *** dmorita_ has joined #openstack-swift | 15:02 | |
| *** dmorita has quit IRC | 15:04 | |
| *** dmorita_ has quit IRC | 15:07 | |
| *** _JZ_ has joined #openstack-swift | 15:09 | |
| *** dmorita has joined #openstack-swift | 15:12 | |
| *** dmorita has quit IRC | 15:16 | |
| *** sams-gleb has joined #openstack-swift | 15:17 | |
| *** hseipp has joined #openstack-swift | 15:26 | |
| *** bkopilov has joined #openstack-swift | 15:41 | |
| clayg | jrichli: there is a screenrc file that defines how services are started | 15:49 |
| clayg | jrichli: vagrant-swift-devstack has a thing so you can "ctrl-a \" kill then when you start screen all the services come up | 15:50 |
| patchbot | Error: No closing quotation | 15:50 |
| jrichli | clayg: I dabbled with screen, but gave up and found a way to restart barbican api and keystone (just apache2 restart). but now I get SSL errors. | 15:50 |
| jrichli | i dont know if a cert was lost, or i have installed some new OpenSSL version that will cause me to change config, ... still looking. | 15:51 |
| jrichli | i did make a running snapshot before i rebooted, so i can just go back. but i am still curious to find out how to recover from a devstack reboot (one with ks + bb + swift) | 15:53 |
| *** rcernin has quit IRC | 15:59 | |
| *** chlong_ has quit IRC | 16:01 | |
| *** chlong__ has joined #openstack-swift | 16:01 | |
| clayg | good thinking on the snapshot! | 16:02 |
| *** janonymous has quit IRC | 16:04 | |
| *** dmorita has joined #openstack-swift | 16:06 | |
| *** dmorita_ has joined #openstack-swift | 16:09 | |
| clayg | jrichli: does yours say SSLCertificateFile: file '/opt/stack/data/devstack-cert.pem' does not exist or is empty | 16:09 |
| *** dmorita_ has quit IRC | 16:10 | |
| jrichli | no, I get "SSLError: SSL exception connecting to https://127.0.0.1:35357: bad handshake: Error([('SSL routines', 'SSL23_GET_SERVER_HELLO', 'unknown protocol')],)" | 16:10 |
| jrichli | but nothing exists in the /opt/stack/data directory. | 16:10 |
| *** dmorita_ has joined #openstack-swift | 16:10 | |
| *** dmorita has quit IRC | 16:10 | |
| *** klrmn has joined #openstack-swift | 16:11 | |
| jrichli | I don't even know if barbican was trying to use keystone via https before when it worked. it could be that something changed that behavior - and it should just be using http | 16:11 |
| jrichli | I started barbican with /opt/stack/barbican/bin/barbican-api | 16:12 |
| jrichli | but that is where the error comes from. barbican calling out to keystone. i can try to change the config there. but it is still strange that this worked before. | 16:12 |
| *** psachin has quit IRC | 16:12 | |
| *** dmorita has joined #openstack-swift | 16:13 | |
| *** dmorita_ has quit IRC | 16:15 | |
| *** dmorita has quit IRC | 16:17 | |
| *** dmorita has joined #openstack-swift | 16:18 | |
| *** dmorita has quit IRC | 16:22 | |
| jrichli | BTW: the devstack vm I am speaking of was something I had setup last week - so this is not from using the devstack mathiasb has pointed us to | 16:24 |
| *** dmorita has joined #openstack-swift | 16:29 | |
| *** dmorita_ has joined #openstack-swift | 16:31 | |
| *** dmorita has quit IRC | 16:34 | |
| *** SkyRocknRoll has joined #openstack-swift | 16:34 | |
| *** dmorita_ has quit IRC | 16:36 | |
| timburke | good morning | 16:39 |
| *** chsc has joined #openstack-swift | 16:39 | |
| *** chsc has joined #openstack-swift | 16:39 | |
| *** pcaruana has quit IRC | 16:42 | |
| *** SkyRocknRoll has quit IRC | 16:42 | |
| *** dmorita has joined #openstack-swift | 16:45 | |
| *** dmorita_ has joined #openstack-swift | 16:46 | |
| *** dmorita has quit IRC | 16:46 | |
| *** dmorita has joined #openstack-swift | 16:50 | |
| *** dmorita_ has quit IRC | 16:51 | |
| *** dmorita has quit IRC | 16:55 | |
| *** JimCheung has joined #openstack-swift | 16:55 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Fix port ranges in install guide https://review.openstack.org/440694 | 16:57 |
| *** ChubYann has joined #openstack-swift | 16:57 | |
| *** cbartz has left #openstack-swift | 17:04 | |
| *** dmorita has joined #openstack-swift | 17:07 | |
| *** dmorita has quit IRC | 17:08 | |
| *** dmorita has joined #openstack-swift | 17:08 | |
| *** dmorita has quit IRC | 17:09 | |
| *** dmorita has joined #openstack-swift | 17:09 | |
| *** dmorita_ has joined #openstack-swift | 17:10 | |
| *** dmorita has quit IRC | 17:13 | |
| *** dmorita_ has quit IRC | 17:15 | |
| *** geaaru has quit IRC | 17:21 | |
| jrichli | FYI: I created an etherpad at https://etherpad.openstack.org/p/swift-kms-keymaster-setup | 17:29 |
| jrichli | that includes a link to mathiasb's script and has my notes from setting up things during the PTG for this testing | 17:30 |
| *** bkopilov has quit IRC | 17:31 | |
| notmyname | good morning | 17:32 |
| notmyname | jrichli: cool | 17:32 |
| notmyname | jrichli: is that linked to anywhere? like on the ideas wiki page? | 17:32 |
| *** JimCheung has quit IRC | 17:33 | |
| acoles | kota_: ack, thank | 17:33 |
| notmyname | FYI the API working group has proposed a set of guidelines for API compatibility. among other things, it defines a lot of commonly-used terms (including "compatibility"). this is a proposal we need to pay attention to | 17:43 |
| notmyname | https://review.openstack.org/#/c/421846/4/guidelines/api_compatibility.rst,unified | 17:43 |
| patchbot | patch 421846 - api-wg - Refactor and re-validate api change guidelines | 17:43 |
| *** dmorita has joined #openstack-swift | 17:46 | |
| *** dmorita has quit IRC | 17:50 | |
| *** tesseract has quit IRC | 17:52 | |
| *** hseipp has quit IRC | 17:54 | |
| *** dmorita has joined #openstack-swift | 18:01 | |
| *** McMurlock1 has quit IRC | 18:23 | |
| jrichli | notmyname: I added a link to it inside a comment on the patch, but i will take a look at the ideas page for what exists for keymaster | 18:39 |
| jrichli | and thanks for the fyi on api compatibility | 18:41 |
| *** amoralej is now known as amoralej|off | 18:44 | |
| jrichli | notmyname: just added a link to the patch and the etherpad to the existing ideas page for keymaster work | 18:52 |
| notmyname | thanks | 18:52 |
| clayg | timburke: everytime something reminds me of patch 274840 I freak out until I remember patch 311899 | 19:14 |
| patchbot | https://review.openstack.org/#/c/274840/ - swift - change default ports for servers (MERGED) | 19:14 |
| patchbot | https://review.openstack.org/#/c/311899/ - swift - Fix probe tests from commit cf48e75 (MERGED) | 19:14 |
| notmyname | FYI - S3 post-mortem https://aws.amazon.com/message/41926/ | 19:14 |
| clayg | rledisez: do you already have the patch that fixes the ssync semaphore thing so you can limit to ~4 connections per disk or whatever? | 19:21 |
| *** glange_ has joined #openstack-swift | 19:34 | |
| *** ChanServ sets mode: +v glange_ | 19:34 | |
| glange_ | https://developer.openstack.org/api-ref/object-storage/?expanded=get-object-content-and-metadata-detail <- in the Range section of object gets there are several possible formats for that header specified | 19:35 |
| glange_ | but the code only seems to accept _content_range_pattern = re.compile(r'^bytes (\d+)-(\d+)/(\d+)$') | 19:36 |
| glange_ | am I seeing that right? or am I missing something | 19:36 |
| timburke | that pattern looks more like what you'd expect in a response, as a Content-Range header | 19:37 |
| timburke | (Content-Range probably ought to be documented in the api ref...) | 19:38 |
| torgomatic | document it by reference to RFC 7233 section 4.2 | 19:39 |
| glange_ | is that documentation out of date? the one I linked? | 19:43 |
| glange_ | was it changed just to follow the RFC? | 19:43 |
| glange_ | or did it always work that way? | 19:43 |
| openstackgerrit | Jim Cheung proposed openstack/pyeclib master: Add Phazr.IO libphazr backend support https://review.openstack.org/440781 | 19:44 |
| rledisez | clayg: i uploaded this during Barcelona summit: https://review.openstack.org/#/c/390781/ | 19:45 |
| patchbot | patch 390781 - swift - WIP: Replace replication_one_per_device by custom ... | 19:45 |
| rledisez | clayg: it would need some cleanup and test, but i think this is what you’re looking for | 19:45 |
| *** tdasilva has quit IRC | 19:46 | |
| torgomatic | glange_: The docs for the Range header are correct, but redundant. Swift does what the RFC says to do for Range, bugs notwithstanding. | 19:48 |
| rledisez | clayg: i wanted to do like rsync, lockf’ing range of file instead of having multiple files, but it didn’t work, i don’t remember why, i think it’s because a lockf() is valid by process, not thread or file descriptor | 19:52 |
| clayg | rledisez: I had assumed we could leverage the semaphore stuff? | 19:53 |
| clayg | rledisez: have a locks[device].aquire() sorta thing? | 19:53 |
| *** tdasilva has joined #openstack-swift | 19:56 | |
| *** jamielennox|away is now known as jamielennox | 20:02 | |
| rledisez | clayg: if i remember well, semaphore is initialized at the very start of the object server (with the « config » object). at that point we don’t have yet the list of devices. i think this is why i didn’t choose this way | 20:06 |
| rledisez | clayg: it’s happening here: https://github.com/openstack/swift/blob/master/swift/obj/server.py#L1180-L1201 | 20:09 |
| rledisez | clayg: called from wsgi.run_wsgi() | 20:09 |
| rledisez | clayg: well, it should be do-able as this is also the place where server_per_port is used | 20:10 |
| clayg | rledisez: you could lazy load the sems anway - just create them under a global lock | 20:16 |
| clayg | rledisez: you gunna change it - or you want me to spin something | 20:16 |
| clayg | rledisez: i'm obliviuous to timezones - for all I know it's beer thirty in your neck of the woods | 20:16 |
| clayg | oh.. maybe you can't lazy create them - you *have* to make them before fork | 20:18 |
| clayg | that... should have been obvious - my bad | 20:18 |
| rledisez | exactly :/ and if they are pre-spawned, what about ring update with added/removed devices | 20:19 |
| clayg | fuuuuu | 20:19 |
| rledisez | (beer time was few hours ago, for a week day it’s more like tea ;)) | 20:19 |
| clayg | rledisez: side quesiton - are you using servers per port for your replication servers? | 20:22 |
| rledisez | clayg: yes -> https://review.openstack.org/#/c/337861/ | 20:24 |
| patchbot | patch 337861 - swift - Permit to bind object-server on replication_port | 20:24 |
| clayg | rledisez: right | 20:25 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Permit to bind object-server on replication_port https://review.openstack.org/337861 | 20:30 |
| *** oshritf has joined #openstack-swift | 20:37 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: refactor some common code from crypto https://review.openstack.org/439058 | 20:43 |
| *** oshritf has quit IRC | 20:43 | |
| *** adriant has joined #openstack-swift | 20:59 | |
| mattoliverau | Morning | 21:07 |
| *** garyj has joined #openstack-swift | 22:04 | |
| *** garyj has quit IRC | 22:10 | |
| *** dja has joined #openstack-swift | 22:12 | |
| *** sams-gleb has quit IRC | 22:40 | |
| *** catintheroof has quit IRC | 22:40 | |
| clayg | tdasilva: http://git.openstack.org/cgit/openstack-infra/project-config/tree/nodepool/scripts/configure_mirror.sh | 23:00 |
| clayg | so there's a "debian-openstack" mirror that nodepool adds when it's preparing nodes | 23:00 |
| clayg | so maybe getting a centos/epel/fedora-openstack mirror for rpm-packaging is a thing that isn't years away? | 23:01 |
| *** dja has quit IRC | 23:04 | |
| *** tdasilva has quit IRC | 23:05 | |
| *** dmorita_ has joined #openstack-swift | 23:23 | |
| *** dmorita has quit IRC | 23:23 | |
| *** tdasilva has joined #openstack-swift | 23:23 | |
| *** kei_yama has joined #openstack-swift | 23:34 | |
| *** chsc has quit IRC | 23:36 | |
| *** jamielennox is now known as jamielennox|away | 23:37 | |
| *** klamath has quit IRC | 23:37 | |
| *** dja has joined #openstack-swift | 23:37 | |
| *** dja has quit IRC | 23:41 | |
| *** _JZ_ has quit IRC | 23:44 | |
| kota_ | good morning! | 23:48 |
| notmyname | hello kota_ | 23:49 |
| kota_ | notmyname: o/ | 23:49 |
| notmyname | kota_: I just clicked +2+A ont he tempauth patch to master | 23:49 |
| kota_ | notmyname: thanks a lot! | 23:49 |
| notmyname | I'd like to see that land before doing the ocata one. does that make sense? | 23:50 |
| notmyname | maybe it's fine | 23:50 |
| notmyname | kota_: anyway, we spend a *lot* of time talking about this today | 23:50 |
| kota_ | with timburke? | 23:51 |
| notmyname | yes. and clayg | 23:51 |
| kota_ | can i get the summary? | 23:51 |
| timburke | the interactions are terrible, but we can't come up with a clean intra-middleware api in a day | 23:51 |
| notmyname | the basic summary of my current thoughts is (1) I hate that there's so much knowledge of s3 inside of auth systems (2) we're changing the interface anyway (3) can we change it to something we'll be happy to support longer term without worrying about breaking changes | 23:51 |
| notmyname | however, we have no idea what that interface would look like | 23:51 |
| notmyname | timburke mocked up a small change, but even it's not great (ie would need more), so we're looking at incremental improvements overall | 23:52 |
| kota_ | i see | 23:52 |
| notmyname | and I don't want to rush something just because it might be better. seems like a great way to make the wrong call | 23:52 |
| kota_ | +1 | 23:53 |
| notmyname | so... it's code, let's make it better where we can today and keep making it better as we go, based on what our users need | 23:53 |
| notmyname | I don't want to hold up something that's known-good and fixes a bad bug in the hope that we'll designe the perfect api in an afternoon | 23:53 |
| notmyname | so that's why I landed it | 23:53 |
| kota_ | thanks, on the (1) of the summary, I'd support that but I also have no idea because I know, we could want to one more complex auth for somewhere (i mean signature v4 of AWS) | 23:55 |
| notmyname | yep :-) | 23:55 |
| kota_ | we did it in the keystone side but tempauth still has not supported... | 23:55 |
| notmyname | it's like you were sitting in the room with us today :-) | 23:55 |
| kota_ | difficult problem for good maintainancibility :/ | 23:56 |
| kota_ | but anyway, thanks again to land the patch. :D | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!