| *** two_tired2 has joined #openstack-swift | 01:02 | |
| *** psachin has joined #openstack-swift | 02:39 | |
| *** itlinux has joined #openstack-swift | 03:56 | |
| *** pcaruana has joined #openstack-swift | 04:14 | |
| *** pcaruana has quit IRC | 04:38 | |
| *** viks__ has joined #openstack-swift | 05:00 | |
| *** two_tired2 has quit IRC | 05:05 | |
| *** e0ne has joined #openstack-swift | 05:08 | |
| openstackgerrit | Sam Morrison proposed openstack/swift master: s3 secret caching https://review.openstack.org/603529 | 05:13 |
|---|---|---|
| *** mikecmpbll has quit IRC | 05:20 | |
| *** pcaruana has joined #openstack-swift | 05:43 | |
| *** e0ne has quit IRC | 05:49 | |
| *** jistr has quit IRC | 05:55 | |
| *** jistr has joined #openstack-swift | 05:56 | |
| *** silor has joined #openstack-swift | 05:57 | |
| *** silor1 has joined #openstack-swift | 06:27 | |
| *** silor has quit IRC | 06:31 | |
| *** silor1 has quit IRC | 06:33 | |
| *** silor has joined #openstack-swift | 06:34 | |
| *** silor1 has joined #openstack-swift | 06:36 | |
| *** silor has quit IRC | 06:40 | |
| *** silor1 has quit IRC | 06:40 | |
| *** hseipp has joined #openstack-swift | 06:45 | |
| *** rcernin has quit IRC | 07:02 | |
| *** stelucz has joined #openstack-swift | 07:15 | |
| *** psachin has quit IRC | 07:21 | |
| *** psachin has joined #openstack-swift | 07:27 | |
| *** rcernin has joined #openstack-swift | 07:56 | |
| *** mikecmpbll has joined #openstack-swift | 07:58 | |
| *** e0ne has joined #openstack-swift | 08:10 | |
| *** gkadam has joined #openstack-swift | 08:16 | |
| *** e0ne has quit IRC | 08:20 | |
| *** jistr has quit IRC | 08:30 | |
| *** jistr has joined #openstack-swift | 08:31 | |
| *** e0ne has joined #openstack-swift | 09:04 | |
| *** rcernin has quit IRC | 09:16 | |
| *** e0ne has quit IRC | 10:27 | |
| *** silor has joined #openstack-swift | 10:29 | |
| *** stelucz has quit IRC | 10:41 | |
| *** mahatic has quit IRC | 10:46 | |
| *** mahatic has joined #openstack-swift | 10:56 | |
| *** ChanServ sets mode: +v mahatic | 10:56 | |
| *** psachin has quit IRC | 11:13 | |
| *** pcaruana has quit IRC | 11:15 | |
| *** psachin has joined #openstack-swift | 11:27 | |
| *** e0ne has joined #openstack-swift | 11:53 | |
| *** guimaluf has quit IRC | 12:21 | |
| *** guimaluf has joined #openstack-swift | 12:26 | |
| *** psachin has quit IRC | 12:41 | |
| *** two_tired2 has joined #openstack-swift | 13:37 | |
| *** silor has quit IRC | 13:39 | |
| *** two_tired2 has quit IRC | 14:13 | |
| openstackgerrit | Merged openstack/swift master: Clean up comment about 404s leaking out during COPYs https://review.openstack.org/535981 | 14:39 |
| openstackgerrit | Merged openstack/swift master: Allow kmip_keymaster to be configured in proxy-server.conf https://review.openstack.org/604937 | 14:39 |
| *** pcaruana has joined #openstack-swift | 15:01 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: update 1space url in associated projects https://review.openstack.org/605467 | 15:55 |
| *** gkadam has quit IRC | 16:25 | |
| *** mikecmpbll has quit IRC | 16:33 | |
| *** e0ne has quit IRC | 16:43 | |
| notmyname | good morning | 16:44 |
| *** hseipp has quit IRC | 16:55 | |
| timburke | so, can we follow through on the "In the future, the ability to use pickle serialization will be removed." threat from https://github.com/openstack/swift/commit/e1ff51c ? | 17:01 |
| timburke | 'cause i want to be able to reason more easily about the types we get when loading from memcache | 17:02 |
| notmyname | timburke: I don't know. that was only 6 years ago... ;-) | 17:03 |
| notmyname | timburke: or to say the same thing in different words, "yes! kill it with fire!" | 17:04 |
| notmyname | as long as we leave the ability to read the pickled data, and just remove the ability to write it, we should be ok. maybe we'd need a read-and-type-normalize thing so we always get consistent types in memcache | 17:05 |
| timburke | so we've got a flag to go switch from "read and write pickle" to "read and write json, but be willing to read pickle" to "only read and write json" -- and i want just that last one | 17:08 |
| timburke | it doesn't count as killing it with fire if we still have the config option, even if we remove one of the possible values :-P | 17:08 |
| notmyname | timburke: yeah, I mean remove the config option, but I personally have data in a swift cluster that I wrote circa 2012 or before that I'd still like to have access to ;-) | 17:10 |
| timburke | notmyname: the question is, do you still have data in your memcache that you wrote circa 2012 that you'd still like access to? | 17:11 |
| notmyname | very unlikely :-) | 17:12 |
| notmyname | oh, wait. did I misread/misremember? I was thinking about the metadata serialization format | 17:12 |
| notmyname | if it's just memcache, yeah, about 48 hours should be plenty | 17:12 |
| *** silor has joined #openstack-swift | 18:20 | |
| *** pcaruana has quit IRC | 18:24 | |
| *** silor has quit IRC | 18:33 | |
| *** silor has joined #openstack-swift | 19:01 | |
| *** silor1 has joined #openstack-swift | 19:05 | |
| *** silor has quit IRC | 19:08 | |
| *** silor1 is now known as silor | 19:08 | |
| *** e0ne has joined #openstack-swift | 19:15 | |
| *** e0ne has quit IRC | 19:30 | |
| *** silor has quit IRC | 19:44 | |
| *** mikecmpbll has joined #openstack-swift | 19:58 | |
| clayg | zaitcev: timburke: bumped into this (again) while lamenting how crappy py2's super method is -> http://python-future.org/faq.html#what-is-the-relationship-between-future-and-six | 20:33 |
| zaitcev | clayg: interesting, considering that bytes is a built-in type in py27 (as an alias to str). They probably ask to have every b'' literal using future.bytes('foo'). | 20:36 |
| clayg | i'm not sure about that | 20:38 |
| timburke | i keep going back and forth on six. sometimes i like it, other times i just want to embrace the fact that writing py2/3 compliant programs amounts to a comparatively simple form of https://en.wikipedia.org/wiki/Polyglot_(computing) | 20:42 |
| zaitcev | I think we're not limited by six in our project. | 20:44 |
| zaitcev | The problem I'm seeing is mostly the Unicode and tracking of what's encoded and what's not, what library uses what arguments, and what's the on-disk formats are. That's what makes Swift difficult to port. An the sheer size. | 20:45 |
| zaitcev | So, swapping six for a more awesome package isn't going to help us much. Maybe fractionally. But it's only IMHO. | 20:46 |
| timburke | agreed. and the idiocy that is WSGI on py3 | 20:46 |
| zaitcev | Yes, PEP-333/-3333 is somewhat problematic. | 20:47 |
| timburke | i think i've said this before, but i don't think i've ever been so concerned about the types of variables before when writing python | 20:50 |
| kota_ | good morning | 20:55 |
| mattoliverau | Morning | 20:58 |
| kota_ | mattoliverau: o/ | 20:58 |
| *** m_kazuhiro has joined #openstack-swift | 20:59 | |
| notmyname | swift team meeting time | 21:00 |
| *** m_kazuhiro has quit IRC | 21:40 | |
| notmyname | my "any filesystem as long as it's xfs" was a play on the henry ford quote of "any color you want as long as it's black" | 21:43 |
| notmyname | and that made me look up if he actually said that. turns out, he did! but it's a more interesting story | 21:43 |
| notmyname | http://oplaunch.com/blog/2015/04/30/the-truth-about-any-color-so-long-as-it-is-black/ | 21:44 |
| notmyname | there are some really interesting comments on sales and marketing vs engineering in there | 21:44 |
| notmyname | https://wiki.openstack.org/wiki/Swift/PriorityReviews updated | 22:18 |
| DHE | I looked at ZFS as an alternative that wouldn't require an auditor (or a much simpler one). I think I mentioned this before. Unfortunately it doesn't scale well in certain scenarios which is pretty much where Swift likes to sit... | 22:31 |
| notmyname | DHE: yeah, that sounds similar to what I've heard from other people too | 22:34 |
| DHE | ZFS makes a lot of design decisions that are good for data reliability but bad for overhead. Making a pool per-disk in a 60-disk chassis also makes me feel highly uncomfortable. :) | 22:35 |
| *** rcernin has joined #openstack-swift | 22:35 | |
| notmyname | I had a home zfs file server for years and years. it's a good system. just, like you said, sits in a different place than swift does | 22:36 |
| DHE | I'm looking at a design that would replace multiple ZFS boxes with a swift cluster. So I was looking at ZFS initially and then that idea just went up in smoke as I learned more. | 22:37 |
| *** rcernin has quit IRC | 22:53 | |
| *** rcernin has joined #openstack-swift | 22:55 | |
| mattoliverau | morning | 23:17 |
| *** jlvillal has joined #openstack-swift | 23:24 | |
| *** rcernin_ has joined #openstack-swift | 23:41 | |
| *** rcernin has quit IRC | 23:43 | |
| *** openstackgerrit has quit IRC | 23:49 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!