| notmyname | oh, and also to check with swifterdarrell that our utilization or monitoring isn't dependent on the value of the user-agent string in the logs. | 00:01 |
|---|---|---|
| Tyger | If it is, that should probably be in a test somewhere, either a unit test or one of the gate tests. | 00:02 |
| portante | notmyname: here | 00:02 |
| notmyname | portante: Tyger has proposed changing/updating some user agent strings in the logs. while that's not entirely forbidden, IMO, it's something to consider with care if it's done. mostly to check what impact it would have, if any | 00:04 |
| portante | notmyname: thanks for the summary, is there a specific gerrit review id to consider? | 00:07 |
| notmyname | I'm headed home. I'll check in this evening, and hopefully the last 2 of the 3 needed backports will have landed. if so, I'll cherry-pick them to a 2.0 branch and we'll get and rc2 tag when ttx wakes up | 00:08 |
| notmyname | portante: meet Tyger, Tyger meet portante | 00:08 |
| Tyger | portante: https://review.openstack.org/#/c/102401/ | 00:08 |
| portante | Tyger: hi | 00:08 |
| notmyname | Tyger: oh, if you're interested in logs.... | 00:09 |
| *** kevinc_ has quit IRC | 00:09 | |
| Tyger | Oh dear, that sounds ominous | 00:09 |
| notmyname | Tyger: something on my todo list that I haven't gotten to yet is to add the storage policy to the proxy access log lines. I probably won't get to it myself until next week, but if you happened to submit a patch to do it, I'd happily review it :-) | 00:10 |
| *** shri has quit IRC | 00:11 | |
| Tyger | notmyname: I'll keep it in mind... I have been meaning to investigate storage policies. | 00:11 |
| *** annegent_ has joined #openstack-swift | 00:15 | |
| *** annegent_ has quit IRC | 00:21 | |
| Tyger | portante: Any thoughts? | 00:23 |
| openstackgerrit | A change was merged to openstack/swift: Fix order dependent test in proxy.test_server https://review.openstack.org/104306 | 00:26 |
| *** dmorita has joined #openstack-swift | 00:26 | |
| peluse_ | notmyname: some strange jenkins failres on the info key patch, seemingly unlreated. Will run it again and then spend some time looking into it I guess. | 00:27 |
| * peluse_ will be in and out for a bit | 00:27 | |
| *** sungju has joined #openstack-swift | 00:29 | |
| *** mwstorer has quit IRC | 00:31 | |
| *** sungju has quit IRC | 00:33 | |
| portante | Tyger: just finished putting the kiddies to bed ... | 00:57 |
| *** nosnos has joined #openstack-swift | 01:13 | |
| *** kota_ has joined #openstack-swift | 01:18 | |
| openstackgerrit | Peter Portante proposed a change to openstack/swift: Object services user-agent string uses full name https://review.openstack.org/102401 | 01:20 |
| *** ktsuyuzaki has joined #openstack-swift | 01:20 | |
| *** nosnos has quit IRC | 01:20 | |
| *** kota_ has quit IRC | 01:22 | |
| Tyger | portante: That seems like a sensible alternative to the commit message. | 01:23 |
| *** kota_ has joined #openstack-swift | 01:23 | |
| portante | cool | 01:25 |
| *** ktsuyuzaki has quit IRC | 01:25 | |
| *** ktsuyuzaki has joined #openstack-swift | 01:25 | |
| portante | Tyger: I don't have a good sense of how this might impact existing tools that process swift log files | 01:26 |
| portante | I am guessing this change might break a few things | 01:26 |
| portante | but I could be wrong, that nobody cares about the user-agent | 01:26 |
| portante | string | 01:26 |
| Tyger | portante: That's the million dollar question isn't it. I half expected it to fail a gate check, but I have no feel for if people care about the user agent string or not. | 01:26 |
| portante | it is very helpful when debugging all the whys-and-wherefores of a problem, at least to me | 01:26 |
| *** kota_ has quit IRC | 01:27 | |
| Tyger | Yeah it was using the user agent string in the logs I noticed it to begin with. I assumed it would be object-* and couldnt find anything in the logs when I looked for that user-agent string. | 01:27 |
| *** kota_ has joined #openstack-swift | 01:32 | |
| *** ktsuyuzaki has quit IRC | 01:34 | |
| *** ktsuyuzaki has joined #openstack-swift | 01:34 | |
| openstackgerrit | Steven Lang proposed a change to openstack/swift: Object services user-agent string uses full name https://review.openstack.org/102401 | 01:36 |
| *** kota_ has quit IRC | 01:36 | |
| *** kota_ has joined #openstack-swift | 01:39 | |
| *** ktsuyuzaki has quit IRC | 01:41 | |
| *** ktsuyuzaki has joined #openstack-swift | 01:41 | |
| *** kota_ has quit IRC | 01:43 | |
| *** kota_ has joined #openstack-swift | 01:44 | |
| *** ktsuyuzaki has quit IRC | 01:46 | |
| *** ktsuyuzaki has joined #openstack-swift | 01:46 | |
| peluse_ | notmyname: saw the approval on the info patch and claygs other fix, thanks. | 01:47 |
| *** kota_ has quit IRC | 01:48 | |
| peluse_ | notmyname: if not clear from the commit message I found the bug by running a 4 node cluster with 1.13.1, doing a pile of traffic to it then upgrading nodes to v2 one by one and inspecting logs. Did that with rsync, then with ssync, then back again, then back again and then enabled a 2nd policy as default (last part between traffic generator runs though) | 01:49 |
| peluse_ | all the upgrades while the cluster was pretty active... | 01:50 |
| *** kota_ has joined #openstack-swift | 01:53 | |
| *** ktsuyuzaki has quit IRC | 01:55 | |
| *** ktsuyuzaki has joined #openstack-swift | 01:55 | |
| *** kota_ has quit IRC | 01:57 | |
| *** kota_ has joined #openstack-swift | 02:00 | |
| *** ktsuyuzaki has quit IRC | 02:02 | |
| *** ktsuyuzaki has joined #openstack-swift | 02:02 | |
| *** kota_ has quit IRC | 02:04 | |
| *** kota_ has joined #openstack-swift | 02:05 | |
| *** ktsuyuzaki has quit IRC | 02:07 | |
| *** ktsuyuzaki has joined #openstack-swift | 02:07 | |
| *** nthacker has joined #openstack-swift | 02:09 | |
| nthacker | hey guys, rather silly question here cause im new to git...what do i do with my branch once its merged | 02:09 |
| nthacker | merged to master that is | 02:09 |
| *** kota_ has quit IRC | 02:09 | |
| openstackgerrit | A change was merged to openstack/swift: Fix pop_queue for move object twice https://review.openstack.org/103731 | 02:11 |
| mattoliverau | nthacker: in your local clone? | 02:11 |
| nthacker | yes in my local clone | 02:12 |
| mattoliverau | nthacker: if its merged to master upstream then you can delete it :) | 02:12 |
| nthacker | mattoliverau: i created a branch to fix a bug. the bug was reviewed committed merged, so what do i do with my branch | 02:12 |
| nthacker | mattoliverau: im guessing deleting it, but that wont break anything right? | 02:12 |
| nthacker | mattoliverau: ok thanks! | 02:12 |
| mattoliverau | nthacker: no it wont, so long as you checkout master again and pull the latest from master before you branch out and fix another bug ;) | 02:13 |
| nthacker | mattoliverau: thanks! i should take a tutorial for git...sometime....:P | 02:14 |
| mattoliverau | nthacker: lol, your taking the git tutorial called "git by fire" :P | 02:15 |
| nthacker | mattoliverau: it does seem like that...but thankfully i broke nothing and went through the workflow | 02:16 |
| nthacker | mattoliverau: i have been wanting to poke around on the EC bits. i posted on the trello board but got no response...any ideas how to proceed there? sorry for just throwing you another question there | 02:16 |
| mattoliverau | nthacker: if you do get stuck, let me know and I'm happy to walk you through anything git related :) | 02:16 |
| mattoliverau | lol | 02:17 |
| nthacker | mattoliverau: thanks - i ll definitely take you up on that sometime( because im sure i ll break something if i keep doing this while googling it on a different tab lol) | 02:17 |
| mattoliverau | nthacker: according to the code review dashboard, this is the only EC patch out, you can take a look at it, review it, fetch the code and play, but I haven't reviewed it yet so don't know what state its in (https://review.openstack.org/#/c/103644/) | 02:20 |
| *** kota_ has joined #openstack-swift | 02:20 | |
| nthacker | mattoliverau: thanks again ! i ll take a look at this | 02:21 |
| *** ktsuyuzaki has quit IRC | 02:22 | |
| *** ktsuyuzaki has joined #openstack-swift | 02:25 | |
| *** kota_ has quit IRC | 02:27 | |
| *** kota_ has joined #openstack-swift | 02:27 | |
| hugokuo | good morning ... | 02:29 |
| *** ktsuyuzaki has quit IRC | 02:29 | |
| kota_ | hugokuo: morning | 02:30 |
| hugokuo | kota_: how's going kota_ San | 02:30 |
| kota_ | hugokuo: nice. How about you? | 02:31 |
| hugokuo | good good ~! | 02:31 |
| mattoliverau | Morning hugokuo | 02:32 |
| peluse_ | nthacker: not sure if it was super clear on Trello but the unit tests you asked about were for the external lib called PyECLib, they're not part of the swift repo (so different process and repor for review/merge) | 02:34 |
| nthacker | peluse_: ahh - ok no i didnt get that | 02:35 |
| peluse_ | the one EC related patch out there now adds a policy for EC and requires that ext library I just mentioned but it doesn't do EC yet | 02:35 |
| peluse_ | "it" meaning the Swift patch by Tushar :) | 02:35 |
| *** bkopilov has quit IRC | 02:35 | |
| peluse_ | proposals for things like the PUT/GET path, the reconstructor, Swift unit tests, etc., etc., will all depend on having a basic EC policy defined first so that one that's up there is important to get eyes on, appreciate you taking a look for sure | 02:36 |
| kota_ | mattoliverau: Hi, Matt! | 02:37 |
| zaitcev | Does anyone happen to remember what we used posix_ipc for? I remember there was some kind of shmem in it and I even found the implementation... | 02:37 |
| zaitcev | oh. It all is coming back to me | 02:38 |
| kota_ | mattoliverau: Did you comment my patch? (https://review.openstack.org/#/c/99824/) | 02:38 |
| kota_ | mattoliverau: I've fix it according to your comment and I wonder if you could review again. | 02:39 |
| mattoliverau | kota_: yes i did, and yes I will :) | 02:40 |
| mattoliverau | zaitcev: lol, sometimes it's good to just talk it out :) | 02:41 |
| kota_ | mattoriverau: Thanks! | 02:41 |
| kota_ | And if anyone here could review it, I will be happy :) | 02:43 |
| *** ktsuyuzaki has joined #openstack-swift | 02:44 | |
| *** zaitcev has quit IRC | 02:46 | |
| *** kota_ has quit IRC | 02:47 | |
| *** ktsuyuzaki has quit IRC | 02:48 | |
| *** kota_ has joined #openstack-swift | 02:48 | |
| *** ktsuyuzaki has joined #openstack-swift | 02:51 | |
| *** kota_ has quit IRC | 02:53 | |
| *** kota_ has joined #openstack-swift | 02:56 | |
| *** kota_ has quit IRC | 02:56 | |
| *** ktsuyuzaki has quit IRC | 02:57 | |
| notmyname | good evening | 03:07 |
| notmyname | looks like another 30 minutes in the gate for https://review.openstack.org/#/c/104350/ (but it's at the top, which is good) | 03:08 |
| *** nthacker has quit IRC | 03:08 | |
| peluse_ | better than a sharp stick in the eye :) | 03:10 |
| peluse_ | kota: I'll see if I can make some time tomorrow to review - is this what you brought up with clayg at the hackathon? | 03:12 |
| notmyname | peluse_: when that patch lands, I'll add it to my backport banch and push those into the proposed/2.0.0 branch so we can do the rc2 | 03:16 |
| notmyname | the 3 commits being backported are: 0f0c0e5553adeccdf6d95ee61b64ea84ffa79758 and 8a3b65107d788a8590349fc4aa02e5c8c2ad9202 and https://review.openstack.org/#/c/104350/ (in that order) | 03:16 |
| peluse_ | notmyname: sounds like a winner, now I gotta go make some dinner... | 03:19 |
| peluse_ | I'm a poet and didn't even know it | 03:19 |
| hugokuo | Anyone knows the usage of --quoted in swift-temp-url ? | 03:20 |
| notmyname | hugokuo: looking | 03:20 |
| notmyname | peluse_: winner winner chicken dinner | 03:20 |
| * hugokuo no description in swift-temp-url command help output | 03:22 | |
| notmyname | john@europa:~/Documents/swift(20bp)$ python bin/swift-temp-url GET 60 /v1/AUTH_hugo/ünicodecontainer/ob secr3t | 03:23 |
| notmyname | /v1/AUTH_hugo/ünicodecontainer/ob?temp_url_sig=1fdffcd47006e21e66e5a006e0a31ffb685abe51&temp_url_expires=1404357863 | 03:23 |
| notmyname | john@europa:~/Documents/swift(20bp)$ python bin/swift-temp-url GET 60 /v1/AUTH_hugo/ünicodecontainer/ob secr3t --quoted | 03:23 |
| notmyname | /v1/AUTH_hugo/ünicodecontainer/ob?temp_url_sig=e350bef2f5b5414f13767698993209e4fb3d93af&temp_url_expires=1404357865 | 03:23 |
| notmyname | hugokuo: so, if you end the command with --quoted, the urllib.quote() value of the path is used to compute the signature | 03:24 |
| notmyname | hugokuo: oh, wait | 03:24 |
| notmyname | hugokuo: that was wrong | 03:24 |
| notmyname | hugokuo: the opposite is true | 03:24 |
| notmyname | hugokuo: --quoted means than the URL you passed in is already quoted, and so it will first unquote it before doing the hmac | 03:25 |
| notmyname | so if you passed in "%2c" or "%20" in the url (because that's what might be given to the end user), then it will unquote that first instead of using those exact characters for the hamc | 03:26 |
| notmyname | *hmac | 03:26 |
| notmyname | hugokuo: make sense? | 03:26 |
| *** stevemac has joined #openstack-swift | 03:28 | |
| *** annegent_ has joined #openstack-swift | 03:28 | |
| hugokuo | notmyname: so if there's Chinese character in the URL. The --quoted is a must have option for swift-temp-url | 03:38 |
| notmyname | hugokuo: correct | 03:38 |
| openstackgerrit | A change was merged to openstack/swift: Fix potential missing key error in container_info https://review.openstack.org/104350 | 03:40 |
| openstackgerrit | A change was merged to openstack/swift: Small cleanup to ring serialization https://review.openstack.org/97615 | 03:40 |
| *** zhiyan_ is now known as zhiyan | 03:47 | |
| peluse_ | ^^ nice. Calling it a night... | 03:51 |
| *** nosnos has joined #openstack-swift | 03:52 | |
| *** ho has quit IRC | 03:58 | |
| *** kota_ has joined #openstack-swift | 03:59 | |
| *** bkopilov has joined #openstack-swift | 04:01 | |
| *** ktsuyuzaki has joined #openstack-swift | 04:01 | |
| *** kota_ has quit IRC | 04:04 | |
| *** annegent_ has quit IRC | 04:06 | |
| *** Kbee has joined #openstack-swift | 04:07 | |
| *** ktsuyuzaki has quit IRC | 04:08 | |
| *** annegent_ has joined #openstack-swift | 04:09 | |
| *** elambert has quit IRC | 04:09 | |
| *** ajc_ has joined #openstack-swift | 04:17 | |
| *** Tyger has quit IRC | 04:21 | |
| *** ppai has joined #openstack-swift | 04:42 | |
| *** annegent_ has quit IRC | 04:43 | |
| *** bvandenh has joined #openstack-swift | 04:44 | |
| *** annegent_ has joined #openstack-swift | 04:52 | |
| *** psharma has joined #openstack-swift | 04:54 | |
| *** stevemac has quit IRC | 04:58 | |
| notmyname | tooling is conspiring to slow down getting stuff backported. -infra has to land a patch before jenkins will pass the tests for the backported patches | 05:02 |
| *** annegent_ has quit IRC | 05:07 | |
| goodes | morning all | 05:09 |
| *** gyee has quit IRC | 05:09 | |
| goodes | question - should updating metadata (i.e. POST) on an object affect the last modified value? | 05:10 |
| mattoliverau | morning goodes | 05:28 |
| mattoliverau | notmyname: bugger :( | 05:28 |
| *** robsparker has quit IRC | 05:31 | |
| goodes | mattoliverau: heya | 05:31 |
| goodes | fundamental question - should last modified refer to the contents of the object and/or the meta data | 05:32 |
| mattoliverau | notmyname: OK, gone through all the daemons and compared what are in the sample config files and the code, there are 3 config options mssing that I have found that are missing, but nothing major. There is however a bunch of defaults in the sample configs, commented but with a value representing the default, which are different then the defaults in the code. I will assume that the code is the correct defaults in this | 05:33 |
| mattoliverau | case... now to see if the documentation differs from the truth at all. | 05:33 |
| goodes | it seems in my testing that updating the metadata also updates the last modified which could screw up any logic if someone is assuming that it refers to when the contents were last changed | 05:33 |
| goodes | or is this just a limitation of xattr that leaves us no choice? | 05:34 |
| mattoliverau | goodes: good question, it comes down to what is the "contents", is metadata as important as the object? I think is it. | 05:34 |
| goodes | mattoliverau: this could mean that middleware that updates metadata could have nasty side effects | 05:35 |
| mattoliverau | goodes: the file gets a new name (timestamp) so it isn't xattr, and I think this is required for sync | 05:35 |
| mattoliverau | goodes: true, but hopfully the middle ware only updates on a request. | 05:36 |
| mattoliverau | not all the time touching things. | 05:36 |
| goodes | or an async app that sets something like metadata to show reviewed, compliance etc could also have -ve effects | 05:36 |
| goodes | say you had a process to transcode video to different formats and you wanted to set MD to say that it has been done | 05:37 |
| goodes | unless there is an xattr limitation it would seem that it would be helpful in the POST to specify whether last-modified should be updated or not | 05:38 |
| *** robsparker has joined #openstack-swift | 05:46 | |
| goodes | mattoliverau: apparently the other half of the world is now asleep | 06:01 |
| mattoliverau | goodes: yeah, slackers :P | 06:04 |
| *** acoles has quit IRC | 06:08 | |
| *** chandan_kumar has joined #openstack-swift | 06:25 | |
| *** jamiehannaford has joined #openstack-swift | 06:33 | |
| openstackgerrit | lawrancejing proposed a change to openstack/python-swiftclient: Add CONTRIBUTING.md https://review.openstack.org/103786 | 06:55 |
| *** Anju has joined #openstack-swift | 06:59 | |
| *** Anju has quit IRC | 07:05 | |
| *** ppai has quit IRC | 07:35 | |
| mattoliverau | ok, I'm calling it a night, well from in front of the computer anyway :) Night all | 07:38 |
| mattoliverau | goodes: or should I just say night to you cause you seem to be the only one here :P | 07:38 |
| *** Midnightmyth has joined #openstack-swift | 07:41 | |
| *** nthacker_ has quit IRC | 07:44 | |
| *** mmcardle has joined #openstack-swift | 07:55 | |
| *** nacim has joined #openstack-swift | 07:58 | |
| *** nosnos has quit IRC | 08:05 | |
| *** nosnos has joined #openstack-swift | 08:05 | |
| *** mlipchuk has joined #openstack-swift | 08:07 | |
| *** foexle has joined #openstack-swift | 08:24 | |
| goodes | mattoliverau: g'nite | 08:26 |
| *** ppai has joined #openstack-swift | 08:39 | |
| *** andyandy has joined #openstack-swift | 08:58 | |
| *** dmorita has quit IRC | 09:04 | |
| *** mitz has quit IRC | 09:22 | |
| *** mitz has joined #openstack-swift | 09:26 | |
| *** Kbee has quit IRC | 10:51 | |
| openstackgerrit | A change was merged to openstack/swift: Fix exception raising in FakeConn https://review.openstack.org/104319 | 11:42 |
| *** mmcardle has quit IRC | 11:43 | |
| *** aswadr has joined #openstack-swift | 11:47 | |
| *** mkerrin has quit IRC | 11:51 | |
| *** nosnos has quit IRC | 11:55 | |
| *** nosnos_ has joined #openstack-swift | 11:58 | |
| *** mkerrin has joined #openstack-swift | 11:59 | |
| *** ppai has quit IRC | 12:02 | |
| *** bvandenh has quit IRC | 12:09 | |
| *** mmcardle has joined #openstack-swift | 12:11 | |
| *** mmcardle1 has joined #openstack-swift | 12:22 | |
| *** mmcardle has quit IRC | 12:25 | |
| *** mkollaro has joined #openstack-swift | 12:28 | |
| *** ajc_ has quit IRC | 12:30 | |
| *** fifieldt has quit IRC | 12:35 | |
| *** zigo has quit IRC | 12:37 | |
| *** zigo has joined #openstack-swift | 12:37 | |
| *** bkopilov has quit IRC | 12:41 | |
| *** zigo has quit IRC | 12:41 | |
| *** zigo has joined #openstack-swift | 12:42 | |
| *** nshaikh has joined #openstack-swift | 12:44 | |
| *** zigo has quit IRC | 12:45 | |
| *** mtreinish has quit IRC | 12:45 | |
| *** angelastreeter has joined #openstack-swift | 12:52 | |
| *** miqui has joined #openstack-swift | 12:56 | |
| zul | hi guys im looking for some guidance with respect to the storage policy stuff, should things like swift-reconciler-enqueue and swift-container-reconciler be in a seperate package and have init scripts | 12:57 |
| *** Midnightmyth has quit IRC | 13:05 | |
| *** chandan_kumar is now known as chandankumar | 13:18 | |
| *** psharma has quit IRC | 13:26 | |
| *** Kbee has joined #openstack-swift | 13:31 | |
| *** keshava_ has joined #openstack-swift | 13:33 | |
| *** keshava_ has quit IRC | 13:33 | |
| *** keshava_ has joined #openstack-swift | 13:33 | |
| *** keshava_ has quit IRC | 13:34 | |
| *** keshava_ has joined #openstack-swift | 13:34 | |
| *** keshava_ has quit IRC | 13:34 | |
| *** Kbee has quit IRC | 13:35 | |
| *** panbalag has joined #openstack-swift | 13:37 | |
| *** nosnos_ has quit IRC | 13:37 | |
| *** angelastreeter has quit IRC | 13:39 | |
| *** nosnos has joined #openstack-swift | 13:49 | |
| *** zigo has joined #openstack-swift | 13:58 | |
| *** nosnos has quit IRC | 13:59 | |
| ahale | confused by defcore emails - is swift no longer gonna be a core openstack project ? | 14:09 |
| *** bkopilov has joined #openstack-swift | 14:18 | |
| *** angelastreeter has joined #openstack-swift | 14:27 | |
| *** chandankumar has quit IRC | 14:38 | |
| *** honeybadger has joined #openstack-swift | 14:40 | |
| notmyname | ahale: I don't know what the board is going to do. the TC has expressed very strong and unified support for "openstack is the code in the integrated release". the board may or may not choose to respect that for the purposes of trademark protection | 14:49 |
| *** foexle has quit IRC | 14:51 | |
| portante | ahale, notmyname: can you point me at the specific email you are talking about? | 14:51 |
| notmyname | ahale: IMO the important things to do for swift are to make sure that as many deployers (with or without the rest of openstack)) are known and our "core capabilities" are well tested, first where it is and later in tempest (probably) | 14:51 |
| notmyname | portante: http://www.openstack.org/blog/2014/07/openstack-technical-committee-update-july-1/ and http://eavesdrop.openstack.org/meetings/tc/2014/tc.2014-07-01-20.02.log.html are good places to start | 14:53 |
| portante | notmyname: thanks | 14:54 |
| ahale | i was reading the Foundation list "Understanding DefCore" thread with the links to https://etherpad.openstack.org/p/DefCoreLighthouse.F2F | 14:55 |
| notmyname | portante: it's basically becoming obvious (IMO) that the primary reason for defcore is to figure out if someone can sell "openstack" but exclude swift from that. it has larger implications for some other projects, current and future, but that seems to be the stickiest question today (with those leading defcore seemingly trying to keep swift excluded) | 14:55 |
| notmyname | which is why it's very important, again IMO, that we as a community make swift deployments very well known | 14:56 |
| portante | notmyname: so markmc and russellb are the two leading that sentiment? | 14:57 |
| notmyname | portante: rob hirtchfeld and john mckenty are leading defcore. I had too much direct interaction with others there, so I can't comment on it. that being said, from their role on the TC, there is strong support from the TC of "openstack == code in integrated release" | 14:59 |
| portante | okay | 14:59 |
| creiht | notmyname: while I understand your point, such a strong position has other implications as well | 15:00 |
| creiht | like for example that means you have to run horizon to be called openstack | 15:00 |
| creiht | or any of the other myriad of projects | 15:00 |
| notmyname | creiht: yes, that is true. that was mentioned int he TC meeting andyou should read the transcript | 15:00 |
| creiht | heh | 15:01 |
| creiht | I'm trying to stay away from high blood pressure :) | 15:01 |
| notmyname | creiht: actually, I think you'd be pleased with what was said in the TC meeting. there was very strong unified support for swift (since that's the current lightning rod) | 15:02 |
| creiht | heh | 15:02 |
| portante | mordred seems to be leading that | 15:02 |
| notmyname | portante: he was the first to say it, but there were others too | 15:03 |
| creiht | yeah I was watching the meeting on and off yesterday | 15:03 |
| notmyname | one current issue inside of defcore is that the "core capabilities" currently listed for swift are really really wrong. and to a smaller extent there is the misperception that swift isn't pluggable/extensible so therefore can only be treated as a single unit | 15:03 |
| *** nacim has quit IRC | 15:03 | |
| notmyname | as chosen by troy toman, swift's core capabilities are: account services, container ACL, container quota, staticweb, and quotas. I proposed http://lists.openstack.org/pipermail/defcore-committee/2014-May/000173.html but I haven't seen anything come of that yet | 15:06 |
| notmyname | I think there is a json blob somewhere that maps openstack core capabilities to test names (ie the name of a function that tests it), but I think troy's 5 swift capabilities are still what's there | 15:08 |
| notmyname | so....now you know almost everything about defcore that I do. any help in navigating these waters is appreciated :-) | 15:09 |
| portante | but are the other core capabilities getting into the API features? | 15:09 |
| portante | for other projects? | 15:09 |
| *** nshaikh has quit IRC | 15:10 | |
| portante | so we need to specify that reading an object is a core capability? If that is the case, this defcore stuff sounds quite broken | 15:10 |
| ahale | its weird that the staticweb capability is considered more deployed than object-server capability on the spreadsheet, whole thing seems nuts | 15:12 |
| creiht | what is the link to the spreadsheet? | 15:12 |
| goodes | there is also a negative to this - if you wanted to do a swift only solution - you could never call that openstack | 15:13 |
| ahale | https://etherpad.openstack.org/p/DefCoreLighthouse.F2F +L76 creiht | 15:13 |
| creiht | ahale: ahh thanks | 15:13 |
| ahale | ah not L76 ChanServ | 15:13 |
| notmyname | ahale: https://wiki.openstack.org/w/images/e/e3/DefCore_Capabilities_Scoring.pdf | 15:13 |
| creiht | goodes: yeah | 15:13 |
| goodes | it seems that the whole discussion is at the wrong granularity | 15:13 |
| ahale | bah creiht not ChanServ.. L93 even | 15:14 |
| creiht | haha | 15:14 |
| notmyname | goodes: right. but is that bad? it's not "openstack" because openstack is a large umbrella of things. you may be deploying "openstack swift", but that's not "openstack". I'm ok with that | 15:14 |
| goodes | well it seems that openstack already falls into major categories i.e. computer, object, management (incl orchestration) | 15:15 |
| notmyname | I guess if I get "openstack - swift = not openstack" then I'm also ok with "swift alone = not openstack" | 15:15 |
| goodes | s/computer/compute/ | 15:15 |
| goodes | aiming for one size fits all - will infact suit no one | 15:16 |
| creiht | ahale: that spreadsheet is very confusing | 15:16 |
| portante | yeah, you can say that again | 15:16 |
| notmyname | lol | 15:16 |
| ahale | creiht: I think it defines defcore pretty well in that case | 15:16 |
| creiht | hahahaha | 15:16 |
| portante | plays well with others? | 15:16 |
| portante | takes a system view? | 15:17 |
| notmyname | there is a more current one I can't find right now where the TC has updated the scoring | 15:17 |
| * portante ... column headings | 15:17 | |
| *** honeybadger has quit IRC | 15:18 | |
| notmyname | creiht: goodes: point is, from the TC meeting, the TC isn't picking "favorites" among the integrated release. that is, the integrated release is openstack | 15:18 |
| creiht | notmyname: yeah I got that | 15:18 |
| goodes | the issue is 'integrated' - does not match the real work or make sense? | 15:18 |
| creiht | notmyname: but at the same time, you are just pushing the decision to the board | 15:18 |
| goodes | work == world | 15:19 |
| notmyname | creiht: well I'm not on the TC, so not me, per se ;-) | 15:19 |
| creiht | lol | 15:19 |
| goodes | it seems that this is more about vendors taking and selling OS solutions and not contributing back | 15:20 |
| notmyname | but I think it's slightly different that just asking the board to make the call. the TC is providing guidance. it's just that those on defcore (ie rob in the TC meeting) really didn't like the given recommendation | 15:20 |
| notmyname | goodes: yes, that is a huge part of it, I think | 15:21 |
| goodes | which if so, then this is not the right way to go about it | 15:21 |
| creiht | the funny thing is, that is already happening, and is going to happen reguardless | 15:21 |
| notmyname | and for the record, I think openstack is too young and has too little adoption today to start trying to make a shibboleth around "true openstack". now's not the time to limit deployments, IMO | 15:22 |
| creiht | heh | 15:22 |
| creiht | my argument to rob from the beginning was to get rid of the idea of core and let the markets decide | 15:22 |
| creiht | otherwise you will forever hamper innovation | 15:23 |
| goodes | most other vendors/standards committess have a 'certification' board with guidlines on how to certify and then they can apply the rule to allow someone to say 'Openstack XXXXX' - but there needs to be lewat | 15:23 |
| goodes | leway | 15:23 |
| goodes | this seems to be the case of repeating lessons that everyone else learned years ago | 15:23 |
| creiht | and prevent companies from being able to differentiate | 15:23 |
| notmyname | but the issue actually has legal concerns, since the only thing the foundation controls is the trademark. so if "openstack" is going to mean something, legally the foundation has to defend the TM, and therefore you get the questions of "well, what is openstack anyway?" | 15:24 |
| goodes | creiht: +1 | 15:24 |
| goodes | but you set guidelines with the option to review and you make companies pay to be able to use the OpenStack XXXX term | 15:25 |
| notmyname | creiht: I don't really buy the "we can't innovate if we have to use swift" argument (substitute any openstack project for "swift") | 15:25 |
| creiht | because there are artificially blessed projects | 15:25 |
| creiht | a simple example is the dashboard | 15:26 |
| goodes | or keystone | 15:26 |
| creiht | there are tons of ways to innovate/differenentiate with the dashboard | 15:26 |
| notmyname | you'll remember that when I was on the TC, I only rarely voted to add new projects :-) | 15:26 |
| creiht | sure | 15:26 |
| notmyname | (mostly in a 11-1 minority) | 15:26 |
| notmyname | ;-) | 15:26 |
| creiht | anyways... I've argued this to death before | 15:27 |
| creiht | and at this point, it is hard for me to care enough to argue any more about it | 15:27 |
| goodes | problem is that people are asking 'what does openstack mean' rather then 'what does openstack mean to the customer' - the process is too technical | 15:27 |
| notmyname | and then they win | 15:27 |
| creiht | win what? | 15:28 |
| notmyname | creiht: so ya, I think you're right that there are perhaps some openstack projects that are less suited for "innovation", but that seems different than "the code hosted at github.com/openstack/ is what openstack is" | 15:28 |
| goodes | becuase the only thing that would matter to a customer is one provider was openstack would be that they could transfer their data to another openstack provider, otherwise what does the customer care what powers the cloud solution? | 15:29 |
| notmyname | creiht: ie when one side gets tired of the argument, the other side wins. for the specific thing now, it's "does swift get tm protection from the foundation as part of openstack?" | 15:29 |
| goodes | now that apple has swift, what will happen to the name? Surely Apple checked before using the name. | 15:30 |
| ahale | what does swift get from being part of the openstack ? | 15:30 |
| notmyname | creiht: and we've seen that before with everything from bzr, launchpad, "openstack is an atomic thing rather than loosely coupled cooperating systems" etc | 15:30 |
| creiht | I guess I don't care as much about that as you do | 15:30 |
| creiht | when people say openstack, they only mean nova anyways | 15:30 |
| creiht | it has always been like that | 15:30 |
| creiht | will always be | 15:30 |
| notmyname | ahale: that's a very good question, if swift doesn't get to enjoy the foundation's tm protection. but I don't really want to explor the implications of that too much until the "if" actually happens, if it ever does | 15:32 |
| goodes | "openstack object (TM)" | 15:32 |
| creiht | goodes: yeah that was mentioned, but the board is too lazy to want to manage many trademarks | 15:32 |
| creiht | hah there you go | 15:33 |
| creiht | if it isn't worth trademarking, it isn't core :) | 15:33 |
| goodes | 3? compute (or just Openstack), object and management | 15:33 |
| creiht | goodes: oh there are like 15 now or something like that | 15:33 |
| goodes | three main pillars | 15:33 |
| creiht | true | 15:33 |
| *** acoles has joined #openstack-swift | 15:34 | |
| *** ChanServ sets mode: +v acoles | 15:34 | |
| creiht | kinda | 15:34 |
| * portante ... so much for technology | 15:35 | |
| notmyname | portante: selling it all and starting a llama farm? | 15:36 |
| * mordred will join portante in llama farming | 15:37 | |
| creiht | lol | 15:37 |
| portante | got two of us, any other takers? | 15:38 |
| notmyname | yes yes, but what does it really /mean/ to "farm"? do I have to plant my llamas int he ground or can I let them simply wander around? can we have any sort of test process around this? | 15:38 |
| * portante runs! | 15:38 | |
| creiht | notmyname: is there any other presidence for this type of thing? | 15:39 |
| notmyname | for what type of thing? | 15:39 |
| goodes | can I change the subject for a few mins? | 15:39 |
| creiht | the trademark stuff | 15:39 |
| creiht | the hadoop ecosystem is pretty huge | 15:39 |
| portante | goodes: sure | 15:39 |
| creiht | there isn't anything that says you have to have x installed to call it hadoop | 15:40 |
| notmyname | creiht: I don't know | 15:40 |
| creiht | I'm not aware of any other ecosystem that does the same | 15:40 |
| creiht | long term, people are going to care about cloud | 15:40 |
| goodes | doing a POST on a file and changing the headers updates the last modified? | 15:40 |
| creiht | openstack is just an implementation detail | 15:40 |
| goodes | is this becuase that will trigger a replicate to spread the updated metadata or is that independent? | 15:41 |
| notmyname | goodes: POST is by default implemented as a COPY | 15:41 |
| notmyname | but I don't recall without looking at the code if the "fast post" (ie not COPY) updates the last-modified | 15:42 |
| goodes | notmyname: default meaning that it can be changed or that it was a design decision | 15:42 |
| notmyname | goodes: it can be changed. but you'll lose the ability to change the content type | 15:43 |
| notmyname | there are some people working on getting the best of both worlds, but it hasn't been written yet | 15:43 |
| goodes | notmyname: I assume that would be global and not a per container setting | 15:43 |
| notmyname | goodes: correct | 15:44 |
| notmyname | goodes: int he proxy config. "post_as_copy" I think | 15:44 |
| goodes | notmyname: so there is a reference to that in the container sync documentation | 15:44 |
| notmyname | yes | 15:45 |
| goodes | Container sync will sync object POSTs only if the proxy server is set to use “object_post_as_copy = true” which is the default. So-called fast object posts, “object_post_as_copy = false” do not update the container listings and therefore can’t be detected for synchronization. | 15:45 |
| goodes | but for replication (not container sync) is this still true? | 15:46 |
| notmyname | creiht: portante: ok for other stuff, the -infra tooling is trying to get updated so that we can land those backports. then tag rc2 with those changes. unless other things have come up, we shoudl still be able to have a monday final release | 15:46 |
| *** jamiehannaford has quit IRC | 15:46 | |
| notmyname | goodes: replication is orthogonal to container sync | 15:46 |
| goodes | assuming that older implementations use container sync to copy between regions | 15:47 |
| goodes | if they no longer will rsync, ssync still replicate if object_post_as_copy = false | 15:47 |
| notmyname | I think you're confusing container sync and replication | 15:48 |
| goodes | possibly | 15:48 |
| portante | notmyname: k | 15:48 |
| peluse_ | notmyname: what's the status/plans on 2.0 at the moment? | 15:48 |
| *** jamiehannaford has joined #openstack-swift | 15:49 | |
| notmyname | peluse_: (see above) working with -infra right now. the 3 backport patches are good to go once jenkins can be convinced it's ok. | 15:49 |
| notmyname | peluse_: I think monday is still ok for final release, with (as you may have seen from dana) with some press stuff on tuesday | 15:50 |
| * peluse_ is scrolling :) | 15:50 | |
| notmyname | peluse_: about 3 lines up. pay no attention to the defcore stuff ;-) | 15:51 |
| notmyname | or you might join portante and mordred and sell it all to start a llama farm | 15:51 |
| *** gyee has joined #openstack-swift | 15:52 | |
| * peluse_ is all in when it comes to any kind of sheep thingy - aren't they close to llamas? | 15:54 | |
| creiht | heh | 15:58 |
| *** mtreinish has joined #openstack-swift | 16:00 | |
| zigo | FYI, gf-complete and jerasure have just been validated by FTP masters and just entered Debian! :) | 16:01 |
| *** mtreinish has quit IRC | 16:04 | |
| *** mtreinish has joined #openstack-swift | 16:04 | |
| notmyname | zigo: nice! | 16:05 |
| zigo | notmyname: Are you working on the erasure stuff? | 16:05 |
| *** angelastreeter has quit IRC | 16:05 | |
| zigo | I'd like to know about the status of PyECLib and separation of libec ... | 16:06 |
| zigo | I'm waiting for it to be able to package it in Debian. | 16:06 |
| *** panbalag has quit IRC | 16:07 | |
| notmyname | zigo: that's for tushar and kevin. but for people in irc, you should talk to peluse_ | 16:08 |
| zigo | Ok. | 16:08 |
| zigo | Will ping tushar (I got his email address). | 16:09 |
| *** mwstorer has joined #openstack-swift | 16:09 | |
| *** lpabon_test has joined #openstack-swift | 16:15 | |
| *** mlipchuk has quit IRC | 16:23 | |
| *** zhiyan is now known as zhiyan_ | 16:29 | |
| *** mmcardle1 has quit IRC | 16:53 | |
| notmyname | looks like about 2 hours ish for the -infra patch to land before I can recheck the backport patches | 16:55 |
| * notmyname goes offline for a bit for getting ready/commute/etc | 16:56 | |
| *** lpabon_test has quit IRC | 16:57 | |
| peluse_ | zigo: I'll get tsg in here | 17:02 |
| *** tsg has joined #openstack-swift | 17:03 | |
| tsg | zigo: ping | 17:03 |
| zigo | tsg: pong ! | 17:04 |
| tsg | zigo: this is Tushar | 17:05 |
| zigo | Oh ok, cool! | 17:06 |
| zigo | tsg: How's going the eclib and pyeclib thing? | 17:06 |
| zigo | Any progress? | 17:06 |
| tsg | zigo: progressing pretty well | 17:06 |
| zigo | :) | 17:06 |
| zigo | Cool. | 17:06 |
| tsg | the split is pyeclib:liberasurecode now | 17:06 |
| tsg | https://bitbucket.org/tsg-/liberasurecode/commits/all | 17:06 |
| tsg | we are 50% there moving stuff out of pyeclib | 17:07 |
| tsg | zigo: no more network downloads! ;) | 17:07 |
| zigo | tsg: Do you think I can already start packaging liberasurecode? | 17:07 |
| tsg | zigo: give us until Monday | 17:07 |
| zigo | Ok. | 17:07 |
| zigo | No pb. | 17:07 |
| tsg | a lot of stuff in pipeline | 17:07 |
| tsg | zigo: pyeclib remains a python api layer now | 17:08 |
| zigo | I don't want to sound like putting pressure, I just don't want to miss the November deadline to get this into Jessie! :) | 17:08 |
| zigo | It'd be just so nice to have all this in Swift for Jessie. | 17:08 |
| tsg | zigo: no worries, we have a deadline to get the first few EC Swift patches in by end of this month which depends on pyeclib :) | 17:08 |
| zigo | Awesome! :) | 17:09 |
| tsg | zigo: it might be a good idea for you to watch liberasurecode and pyeclib | 17:09 |
| tsg | so we can take in any suggestions you may have, early enough | 17:09 |
| zigo | Ok, will try. | 17:10 |
| tsg | zigo: thanks! :) | 17:10 |
| zigo | tsg: FYI, gf-complete and jerasure just entered Debian today. | 17:10 |
| tsg | zigo: awesome! thanks for getting that done - those will be "runtime" deps for liberasurecode | 17:11 |
| zigo | The only issue is that I couldn't activate SSE code, because it's compile time only. | 17:12 |
| tsg | zigo: ah, you mean any of SSE/AVX code? | 17:12 |
| zigo | Yeah, cause the Debian packages are built once, for all x86 CPUs. | 17:13 |
| tsg | zigo: is it possible to split the package into multiple subpackages - sse, sse3, avx etc variants | 17:13 |
| zigo | One way could be to have a specific .deb with the SSE activated, so that our users could choose, but it'd be best if it could be a runtime thing. | 17:13 |
| zigo | Yeah, though not sure how I'd deal with the buildd thing in Debian. | 17:14 |
| tsg | zigo: isa_l should be better that way .. it does a cpuid at runtime | 17:14 |
| zigo | What's isa_l ? | 17:14 |
| tsg | zigo: https://01.org/intel%C2%AE-storage-acceleration-library-open-source-version | 17:15 |
| tsg | provides SIMD based building blocks for erasure code (reed soloman and cauchy) | 17:16 |
| tsg | zigo: this is one of the libraries we'll add support for, in liberasurecode | 17:16 |
| tsg | zigo: let me check internally to make sure we have a go ahead to create .deb for isa_l .. will get back to you | 17:17 |
| zigo | Well, I can try to work out SSE for gf-complete & jerasure, I think. | 17:19 |
| *** Tyger has joined #openstack-swift | 17:19 | |
| zigo | Anyway, for those who deploy, it's just a mater of doing rm -rf debian/patches and rebuild, so it's pretty easy. | 17:19 |
| tsg | zigo: sounds good | 17:21 |
| zigo | tsg: Could you please add a tag to liberasurecode? | 17:21 |
| zigo | Anything, even 0.0.0.1 ... :) | 17:21 |
| *** elambert has joined #openstack-swift | 17:21 | |
| tsg | zigo: will do :) | 17:23 |
| *** aswadr has quit IRC | 17:52 | |
| *** mwstorer has quit IRC | 18:01 | |
| *** foexle has joined #openstack-swift | 18:17 | |
| notmyname | good morning again. (traffic is bad downtown today) | 18:23 |
| *** andyandy has quit IRC | 18:30 | |
| *** mwstorer has joined #openstack-swift | 18:34 | |
| Tyger | Good morning | 18:47 |
| *** foexle has quit IRC | 18:57 | |
| *** sombrafam has joined #openstack-swift | 19:00 | |
| *** sombrafam is now known as erlon | 19:00 | |
| Tyger | notmyname: While the question if changing the user agent will break any log monitoring remains unanswered, I did add an update to the logs document to patch 102401. | 19:01 |
| notmyname | thanks | 19:01 |
| *** wer has quit IRC | 19:04 | |
| *** CaioBrentano has joined #openstack-swift | 19:08 | |
| *** CaioBrentano has left #openstack-swift | 19:09 | |
| *** Tyger has quit IRC | 19:13 | |
| clayg | wow, crazy scrollback today - not sure I caught all the conversations going on | 19:20 |
| *** jamiehannaford has quit IRC | 19:20 | |
| *** zaitcev has joined #openstack-swift | 19:21 | |
| *** ChanServ sets mode: +v zaitcev | 19:21 | |
| *** robsparker has quit IRC | 19:26 | |
| *** jamiehannaford has joined #openstack-swift | 19:35 | |
| *** robsparker has joined #openstack-swift | 19:37 | |
| goodes | notmyname: so the question is, besides container sync, will setting 'object_post_as_copy = false' have any other side effects | 19:58 |
| *** Tyger has joined #openstack-swift | 20:01 | |
| *** foexle has joined #openstack-swift | 20:03 | |
| *** foexle has quit IRC | 20:10 | |
| *** foexle has joined #openstack-swift | 20:10 | |
| clayg | goodes: you don't get a new last_modified in the container listings when you do a POST, but if you HEAD the object server you get a new modified time (time of the last POST) | 20:17 |
| *** jamiehannaford has quit IRC | 20:17 | |
| goodes | clayg: the goal here is not change the last-modified when setting metadata | 20:18 |
| goodes | clayg: are you saying that is the general case? | 20:18 |
| *** jamiehannaford has joined #openstack-swift | 20:20 | |
| clayg | wut? | 20:21 |
| clayg | yeah when you post the object is "modified" | 20:21 |
| clayg | can you try setting that option on a dev setup and see if it does what you want? | 20:22 |
| goodes | clayg: testing now | 20:22 |
| *** mmcardle has joined #openstack-swift | 20:25 | |
| *** sandywalsh has quit IRC | 20:29 | |
| *** sandywalsh has joined #openstack-swift | 20:30 | |
| *** jamiehannaford has quit IRC | 20:36 | |
| *** wer has joined #openstack-swift | 20:42 | |
| *** mwstorer has quit IRC | 21:02 | |
| goodes | clayg: so weird, I set it to false, restarted everything (in vsaio) and it is still updating x-timestamp and last-modified for both head and get_object | 21:05 |
| *** foexle has quit IRC | 21:06 | |
| goodes | clayg: ah ha - get what you mean - get/head returns time stamp of .meta file and container list returns timestamp of .data file | 21:09 |
| goodes | obviously too late at night for me - time to turn in - enjoy your 4th July holidays | 21:16 |
| *** lpabon has quit IRC | 21:17 | |
| *** mwstorer has joined #openstack-swift | 21:22 | |
| notmyname | backports finally making their way through jenkins | 21:57 |
| notmyname | defcore follow up from rob: http://lists.openstack.org/pipermail/defcore-committee/2014-July/000240.html | 22:07 |
| mattoliverau | Morning | 22:10 |
| notmyname | mattoliverau: good morning to you and happy 'murica day | 22:10 |
| mattoliverau | notmyname: lol, thanks, now all I need to get my hands on are some fireworks :) | 22:13 |
| notmyname | :-) | 22:15 |
| peluse_ | morning | 22:15 |
| peluse_ | notmyname: good news on the backports! | 22:16 |
| notmyname | yup | 22:17 |
| notmyname | peluse_: oh hey! glad you are online. what's your schedule look like next week? | 22:17 |
| notmyname | mon-wed? | 22:17 |
| peluse_ | notmyname. and intersting link you posted there on defcon or defcore, whatever it is :) | 22:17 |
| peluse_ | I'm here all week like a bad lounge singer | 22:17 |
| peluse_ | well... looks like I have an all day meeting on Wed though. wheee | 22:18 |
| notmyname | peluse_: I want to have a hourish phone call with you where we talk about a spec document for the EC work. getting all the parts in swift that need to change down in one place. and maybe I'll convince torgomatic to join in too :-) | 22:18 |
| * torgomatic types faster | 22:19 | |
| peluse_ | excellent... | 22:19 |
| *** mmcardle has quit IRC | 22:19 | |
| peluse_ | pick a day and I'll let you know what times work (just not Wed) | 22:20 |
| notmyname | peluse_: tuesday | 22:20 |
| openstackgerrit | Samuel Merritt proposed a change to openstack/swift: Zero-copy object-server GET responses with splice() https://review.openstack.org/102609 | 22:21 |
| openstackgerrit | Samuel Merritt proposed a change to openstack/swift: Zero-copy object-server PUT requests with splice() https://review.openstack.org/104705 | 22:21 |
| peluse_ | what time is it in CA right now? | 22:21 |
| notmyname | peluse_: 3:22pm right now | 22:22 |
| peluse_ | 1:00 Tue work for you guys? | 22:22 |
| notmyname | CA time? | 22:22 |
| peluse_ | yes, we're the same right now | 22:22 |
| notmyname | ok. that should work for me 1pm tuesday. I'll send an invite | 22:23 |
| peluse_ | coolio | 22:23 |
| peluse_ | so on EC docs, you guys good with doing multiple commits to feature/ec this time instead of one giant one? I'll have a good start ready to post soon | 22:24 |
| notmyname | peluse_: I think we can plan this one better. I expect similar starts and missteps in the feature/ec branch as we had with SP work. so doing all of the stuff on feature/ec and then making a feature/ec-review branch may be what happens | 22:26 |
| notmyname | peluse_: either way, I see zero reason to delay or put requirements around docs for EC in the feature/ec branch :-) | 22:26 |
| peluse_ | OK, with docs specifically though I'd rather merge as we go so the feature/ec branch actually has accurate (albeit not complete) docs at any given point in time | 22:27 |
| peluse_ | cool, we were typing at the same time | 22:27 |
| notmyname | great :-) | 22:27 |
| *** elambert has quit IRC | 22:28 | |
| *** wer has quit IRC | 22:50 | |
| *** wer has joined #openstack-swift | 22:50 | |
| openstackgerrit | paul luse proposed a change to openstack/swift: Initial Erasure Code Docs https://review.openstack.org/104713 | 23:03 |
| *** tsg has quit IRC | 23:18 | |
| *** mwstorer has quit IRC | 23:23 | |
| *** mtreinish has quit IRC | 23:31 | |
| *** mtreinish has joined #openstack-swift | 23:39 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!