| *** david-lyle is now known as david-lyle_afk | 00:11 | |
| *** Masahiro has joined #openstack-swift | 00:19 | |
| *** tab____ has quit IRC | 00:20 | |
| *** Masahiro has quit IRC | 00:24 | |
| *** dmorita has joined #openstack-swift | 00:30 | |
| *** rmcall has quit IRC | 00:41 | |
| *** Masahiro has joined #openstack-swift | 00:45 | |
| *** gyee has quit IRC | 00:50 | |
| *** StevenK has joined #openstack-swift | 01:04 | |
| *** tellesnobrega_ has joined #openstack-swift | 01:05 | |
| *** addnull has joined #openstack-swift | 01:05 | |
| *** xianghui has joined #openstack-swift | 01:15 | |
| *** Masahiro has quit IRC | 01:31 | |
| *** xianghui has quit IRC | 01:32 | |
| *** xianghui_ has joined #openstack-swift | 01:32 | |
| *** xianghui_ is now known as xianghui | 01:33 | |
| *** Masahiro has joined #openstack-swift | 01:34 | |
| *** mitz- has quit IRC | 01:39 | |
| *** mitz_ has joined #openstack-swift | 01:41 | |
| *** oomichi_ has joined #openstack-swift | 01:57 | |
| *** haomaiwang has joined #openstack-swift | 02:14 | |
| *** Masahiro has quit IRC | 02:21 | |
| *** Masahiro has joined #openstack-swift | 02:27 | |
| *** oomichi_ has quit IRC | 02:44 | |
| *** tellesnobrega_ has quit IRC | 02:51 | |
| *** tellesnobrega_ has joined #openstack-swift | 02:53 | |
| *** sluo_wfh has joined #openstack-swift | 03:01 | |
| *** addnull has quit IRC | 03:06 | |
| *** mahatic has quit IRC | 03:07 | |
| *** rdaly2 has joined #openstack-swift | 03:12 | |
| *** jyoti-ranjan has joined #openstack-swift | 03:22 | |
| *** sluo_wfh is now known as sluo_laptop | 03:26 | |
| *** tkay has joined #openstack-swift | 03:27 | |
| *** tkay has quit IRC | 03:33 | |
| *** oomichi_ has joined #openstack-swift | 03:35 | |
| *** bill_az has quit IRC | 03:43 | |
| *** lpabon has joined #openstack-swift | 03:45 | |
| *** wangzh has joined #openstack-swift | 03:49 | |
| *** wangzh has quit IRC | 03:50 | |
| *** wangzh has joined #openstack-swift | 03:51 | |
| *** addnull has joined #openstack-swift | 04:17 | |
| *** tkay has joined #openstack-swift | 04:19 | |
| *** addnull has quit IRC | 04:20 | |
| *** addnull_ has joined #openstack-swift | 04:20 | |
| *** tkay has quit IRC | 04:20 | |
| *** stevage has joined #openstack-swift | 04:41 | |
| stevage | hello, anyone know how to make swift-ui work? I'm getting XMLHttpRequest cannot load https://keystone.rc.nectar.org.au:5000/v2.0/. No 'Access-Control-Allow-Origin' header is present on the requested resource. Origin 'https://vault.melbourne.vicnode.org.au:8888' is therefore not allowed access. | 04:41 |
|---|---|---|
| stevage | Apparently Keystone doesn't (didn't?) support CORS: http://blog.yunak.eu/2013/07/24/keystone_cors/ - so I'm wondering how do people normally make swift-ui work? | 04:41 |
| stevage | FWIW, all my data is world-readable, so I actually really want to just be able to browse files without any login anyway. | 04:44 |
| mattoliverau | stevage: sorry, I've never used swift-ui | 04:51 |
| *** fifieldt has joined #openstack-swift | 04:53 | |
| *** ppai has joined #openstack-swift | 05:01 | |
| *** wangzh has quit IRC | 05:02 | |
| *** tellesnobrega_ has quit IRC | 05:12 | |
| *** rdaly2 has quit IRC | 05:33 | |
| *** rdaly2 has joined #openstack-swift | 05:33 | |
| *** rdaly2 has quit IRC | 05:38 | |
| *** tkay has joined #openstack-swift | 05:49 | |
| *** gyee has joined #openstack-swift | 06:08 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:18 | |
| mattoliverau | well I'm going to call it a day, night swiftland | 06:26 |
| *** jyoti-ranjan has quit IRC | 06:42 | |
| *** addnull_ has quit IRC | 06:44 | |
| *** oomichi_ has quit IRC | 06:46 | |
| *** stevage has quit IRC | 06:48 | |
| *** kopparam has joined #openstack-swift | 06:57 | |
| *** kopparam has quit IRC | 07:08 | |
| *** kopparam has joined #openstack-swift | 07:08 | |
| *** tkay has quit IRC | 07:10 | |
| *** tkay has joined #openstack-swift | 07:11 | |
| *** tkay has quit IRC | 07:20 | |
| *** ho has joined #openstack-swift | 07:20 | |
| *** jyoti-ranjan has joined #openstack-swift | 07:21 | |
| openstackgerrit | Prashanth Pai proposed openstack/swift: fsync() on directories https://review.openstack.org/126923 | 07:23 |
| *** k4n0 has joined #openstack-swift | 07:25 | |
| *** addnull has joined #openstack-swift | 07:46 | |
| *** nellysmitt has joined #openstack-swift | 07:55 | |
| *** fifieldt has quit IRC | 07:56 | |
| *** gyee has quit IRC | 07:58 | |
| *** X019 has joined #openstack-swift | 07:58 | |
| *** ho has quit IRC | 08:02 | |
| *** kei_yama has joined #openstack-swift | 08:04 | |
| *** kei_yama has quit IRC | 08:05 | |
| *** ho has joined #openstack-swift | 08:07 | |
| ho | Hi, sorry to bother you. I would like to have reviews for the following patches. | 08:07 |
| ho | https://review.openstack.org/#/c/133155/ | 08:07 |
| ho | https://review.openstack.org/#/c/138342/2 | 08:08 |
| ho | https://review.openstack.org/#/c/136258/ | 08:08 |
| ho | first time to use this IRC :) | 08:08 |
| *** rledisez has joined #openstack-swift | 08:10 | |
| *** larry_ has joined #openstack-swift | 08:17 | |
| *** jistr has joined #openstack-swift | 08:17 | |
| *** addnull has quit IRC | 08:21 | |
| *** oomichi has joined #openstack-swift | 08:26 | |
| *** rledisez has quit IRC | 08:28 | |
| *** rledisez has joined #openstack-swift | 08:33 | |
| *** addnull has joined #openstack-swift | 08:37 | |
| *** ho has quit IRC | 08:38 | |
| *** oomichi has quit IRC | 08:40 | |
| *** nellysmitt has quit IRC | 08:41 | |
| *** jordanP has joined #openstack-swift | 08:59 | |
| *** jordanP has joined #openstack-swift | 08:59 | |
| *** ppai has quit IRC | 09:03 | |
| *** ppai has joined #openstack-swift | 09:17 | |
| *** larry_ has left #openstack-swift | 09:19 | |
| *** kopparam_ has joined #openstack-swift | 09:24 | |
| *** kopparam has quit IRC | 09:27 | |
| *** nellysmitt has joined #openstack-swift | 09:29 | |
| *** Masahiro has quit IRC | 10:10 | |
| *** tellesnobrega_ has joined #openstack-swift | 10:10 | |
| *** Masahiro has joined #openstack-swift | 10:11 | |
| *** aix has joined #openstack-swift | 10:14 | |
| openstackgerrit | Xiang Hui proposed openstack/swift: Fix getaddrinfo if dnspython is installed. https://review.openstack.org/116618 | 10:17 |
| *** tellesnobrega_ has quit IRC | 10:23 | |
| *** kopparam has joined #openstack-swift | 10:29 | |
| *** dmorita has quit IRC | 10:29 | |
| *** koppara__ has joined #openstack-swift | 10:30 | |
| *** tellesnobrega_ has joined #openstack-swift | 10:31 | |
| *** kopparam_ has quit IRC | 10:32 | |
| *** Masahiro has quit IRC | 10:33 | |
| *** kopparam has quit IRC | 10:33 | |
| *** tellesnobrega_ has quit IRC | 10:43 | |
| *** haomaiwang has quit IRC | 10:57 | |
| *** NellyK has joined #openstack-swift | 11:01 | |
| *** NellyK is now known as nellysmi_ | 11:02 | |
| *** nellysmitt has quit IRC | 11:04 | |
| *** rebelshrug has quit IRC | 11:04 | |
| *** nellysmi_ has quit IRC | 11:25 | |
| *** addnull has quit IRC | 11:26 | |
| *** Masahiro has joined #openstack-swift | 11:34 | |
| *** acoles_away is now known as acoles | 11:35 | |
| *** Masahiro has quit IRC | 11:38 | |
| *** lpabon has joined #openstack-swift | 12:04 | |
| *** nellysmitt has joined #openstack-swift | 12:04 | |
| *** david-lyle_afk is now known as david-lyle | 12:09 | |
| *** aix has quit IRC | 12:09 | |
| *** SkyRocknRoll has quit IRC | 12:16 | |
| *** david-lyle is now known as david-lyle_afk | 12:17 | |
| *** david-lyle_afk is now known as david-lyle | 12:17 | |
| *** jistr has quit IRC | 12:20 | |
| *** david-lyle is now known as david-lyle_afk | 12:23 | |
| *** david-lyle_afk is now known as david-lyle | 12:23 | |
| *** nshaikh has joined #openstack-swift | 12:23 | |
| *** david-lyle is now known as david-lyle_afk | 12:31 | |
| *** david-lyle_afk is now known as david-lyle | 12:31 | |
| *** addnull has joined #openstack-swift | 12:37 | |
| *** jistr has joined #openstack-swift | 12:40 | |
| *** addnull has quit IRC | 12:42 | |
| *** addnull has joined #openstack-swift | 12:44 | |
| *** koppara__ has quit IRC | 12:46 | |
| *** Masahiro has joined #openstack-swift | 12:46 | |
| *** Masahiro has quit IRC | 12:51 | |
| *** david-lyle is now known as david-lyle_afk | 13:00 | |
| *** aix has joined #openstack-swift | 13:04 | |
| *** tdasilva has joined #openstack-swift | 13:05 | |
| *** tdasilva has quit IRC | 13:08 | |
| *** jyoti-ranjan has quit IRC | 13:15 | |
| *** miurahr has joined #openstack-swift | 13:24 | |
| *** miurahr has left #openstack-swift | 13:25 | |
| *** zul has quit IRC | 13:30 | |
| *** ppai has quit IRC | 13:31 | |
| *** zul has joined #openstack-swift | 13:34 | |
| *** ppai has joined #openstack-swift | 13:45 | |
| *** kopparam has joined #openstack-swift | 13:49 | |
| *** ppai has quit IRC | 13:50 | |
| *** bill_az has joined #openstack-swift | 13:59 | |
| *** ppai has joined #openstack-swift | 14:05 | |
| *** rdaly2 has joined #openstack-swift | 14:11 | |
| openstackgerrit | Daniel Wakefield proposed openstack/python-swiftclient: Change tests to use new CaptureOutput class. https://review.openstack.org/140690 | 14:13 |
| *** rdaly2 has quit IRC | 14:16 | |
| *** annegent_ has joined #openstack-swift | 14:20 | |
| *** addnull has quit IRC | 14:22 | |
| *** ppai has quit IRC | 14:27 | |
| *** Masahiro has joined #openstack-swift | 14:35 | |
| *** rdaly2 has joined #openstack-swift | 14:38 | |
| *** Masahiro has quit IRC | 14:39 | |
| *** rdaly2_ has joined #openstack-swift | 14:40 | |
| *** tdasilva has joined #openstack-swift | 14:40 | |
| *** ppai has joined #openstack-swift | 14:40 | |
| *** rdaly2_ has quit IRC | 14:41 | |
| *** rdaly2 has quit IRC | 14:42 | |
| *** ppai has quit IRC | 14:47 | |
| *** miurahr has joined #openstack-swift | 14:49 | |
| *** dmsimard_away is now known as dmsimard | 14:51 | |
| *** addnull has joined #openstack-swift | 15:03 | |
| *** miurahr has quit IRC | 15:04 | |
| *** bill_az has quit IRC | 15:22 | |
| *** kopparam has quit IRC | 15:36 | |
| *** nshaikh has quit IRC | 15:43 | |
| *** dmsimard is now known as dmsimard_away | 15:45 | |
| *** dmsimard_away is now known as dmsimard | 15:48 | |
| *** lcurtis has joined #openstack-swift | 16:03 | |
| *** jistr has quit IRC | 16:10 | |
| *** Nadeem_ has joined #openstack-swift | 16:12 | |
| *** Nadeem_ has quit IRC | 16:12 | |
| *** Nadeem_ has joined #openstack-swift | 16:13 | |
| *** X019 has quit IRC | 16:24 | |
| *** Masahiro has joined #openstack-swift | 16:24 | |
| *** panbalag has joined #openstack-swift | 16:27 | |
| *** Masahiro has quit IRC | 16:28 | |
| *** annegent_ has quit IRC | 16:29 | |
| *** kopparam has joined #openstack-swift | 16:29 | |
| *** AbyssOne is now known as a1|away | 16:34 | |
| *** X019 has joined #openstack-swift | 16:37 | |
| *** linuxgeek_ has joined #openstack-swift | 16:45 | |
| *** zhiyan has quit IRC | 16:46 | |
| *** serverascode___ has quit IRC | 16:47 | |
| *** ipolyzos has quit IRC | 16:47 | |
| *** nottrobin_ has quit IRC | 16:48 | |
| *** goodes has quit IRC | 16:48 | |
| linuxgeek_ | hi, swift is configured as a backend for glance. uploaded a 15gb image file and when launching an instance using this image, it failed IOError: [Errno 32] Corrupt image download. Checksum was 17a79c79bb35822cef49239c5ca26545 expected 4f7d0c57c74527872afb4cc7cc5a61af | 16:48 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Warn if multiple replicas are stored within same region/zone/node https://review.openstack.org/140478 | 16:49 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Warn if multiple replicas are stored within same region/zone/node https://review.openstack.org/140478 | 16:51 |
| *** Vijay has joined #openstack-swift | 16:53 | |
| *** rledisez has quit IRC | 17:03 | |
| *** david-lyle_afk is now known as david-lyle | 17:06 | |
| *** dmsimard is now known as dmsimard_away | 17:09 | |
| *** k4n0 has quit IRC | 17:12 | |
| *** nellysmitt has quit IRC | 17:25 | |
| *** Vijay has quit IRC | 17:37 | |
| *** aix has quit IRC | 17:45 | |
| notmyname | good morning | 17:45 |
| *** gyee has joined #openstack-swift | 17:46 | |
| *** annegent_ has joined #openstack-swift | 17:47 | |
| *** nottrobin_ has joined #openstack-swift | 17:57 | |
| *** zaitcev has joined #openstack-swift | 17:58 | |
| *** ChanServ sets mode: +v zaitcev | 17:58 | |
| * notmyname preemptively apologizes for leave/join irc messages that are soon to happen | 18:01 | |
| *** annegent_ has quit IRC | 18:03 | |
| *** serverascode___ has joined #openstack-swift | 18:03 | |
| *** ChanServ sets mode: +v cschwede | 18:04 | |
| *** zhiyan has joined #openstack-swift | 18:09 | |
| *** acorwin has quit IRC | 18:10 | |
| *** ctennis has quit IRC | 18:10 | |
| *** otherjon has quit IRC | 18:10 | |
| *** zackmdavis has quit IRC | 18:10 | |
| *** bobby2 has quit IRC | 18:10 | |
| *** joearnold has quit IRC | 18:10 | |
| *** anderstj has quit IRC | 18:10 | |
| *** hugokuo has quit IRC | 18:10 | |
| *** alpha_ori has quit IRC | 18:10 | |
| *** charz has quit IRC | 18:10 | |
| *** chrisnelson has quit IRC | 18:10 | |
| *** amandap has quit IRC | 18:10 | |
| *** mlanner has quit IRC | 18:10 | |
| *** swifterdarrell has quit IRC | 18:10 | |
| *** goodes has joined #openstack-swift | 18:11 | |
| *** Masahiro has joined #openstack-swift | 18:13 | |
| *** kopparam has quit IRC | 18:16 | |
| *** ipolyzos has joined #openstack-swift | 18:17 | |
| *** jordanP has quit IRC | 18:17 | |
| *** Masahiro has quit IRC | 18:17 | |
| *** acorwin has joined #openstack-swift | 18:18 | |
| *** amandap has joined #openstack-swift | 18:19 | |
| *** anderstj has joined #openstack-swift | 18:19 | |
| *** nellysmitt has joined #openstack-swift | 18:19 | |
| *** bobby2 has joined #openstack-swift | 18:20 | |
| *** charz has joined #openstack-swift | 18:20 | |
| *** nellysmitt has quit IRC | 18:20 | |
| *** chrisnelson has joined #openstack-swift | 18:21 | |
| *** ctennis has joined #openstack-swift | 18:21 | |
| *** swifterdarrell has joined #openstack-swift | 18:22 | |
| *** ChanServ sets mode: +v swifterdarrell | 18:22 | |
| *** hugokuo has joined #openstack-swift | 18:24 | |
| *** joearnold has joined #openstack-swift | 18:24 | |
| *** mlanner has joined #openstack-swift | 18:25 | |
| *** alpha_ori has joined #openstack-swift | 18:27 | |
| *** otherjon has joined #openstack-swift | 18:27 | |
| *** zackmdavis has joined #openstack-swift | 18:29 | |
| *** annegent_ has joined #openstack-swift | 18:34 | |
| *** cutforth has joined #openstack-swift | 18:35 | |
| *** annegent_ has quit IRC | 18:39 | |
| *** ctennis has quit IRC | 18:39 | |
| *** ctennis has joined #openstack-swift | 18:39 | |
| *** flwang has quit IRC | 18:40 | |
| notmyname | swift team meeting in #openstack-meeting in 15 minutes | 18:45 |
| *** gvernik has joined #openstack-swift | 18:47 | |
| *** addnull has quit IRC | 18:49 | |
| quack_quack_ | i'm trying to enable proxy-logging. Here's my pipeline: http://pastebin.com/ZnEMNrYY | 18:52 |
| *** flwang has joined #openstack-swift | 18:52 | |
| quack_quack_ | and i've also added this to my proxy-conf: http://pastebin.com/VSSRVK8C | 18:53 |
| notmyname | quack_quack_: bulk seems like a weird place | 18:53 |
| quack_quack_ | bulk was inserted there by default. I'm using packstack/RDO | 18:53 |
| notmyname | https://github.com/openstack/swift/blob/master/etc/proxy-server.conf-sample#L79 | 18:54 |
| notmyname | quack_quack_: and proxy-logging isn't there by default in RDO? | 18:54 |
| quack_quack_ | nope | 18:54 |
| quack_quack_ | it's driving me crazy | 18:54 |
| notmyname | hmm.. | 18:54 |
| notmyname | zaitcev: wat? ^^ | 18:55 |
| quack_quack_ | i'm trying to figure out what's going on with s3_token | 18:55 |
| quack_quack_ | notmyname: this is my default pipeline - http://pastebin.com/EUeqb31N | 18:57 |
| notmyname | quack_quack_: I'd recommend looking at the sample proxy config I pasted above | 18:57 |
| *** kota_ has joined #openstack-swift | 18:58 | |
| quack_quack_ | notmyname: would you recommend just using the sample instead of what I have now? | 18:58 |
| quack_quack_ | f it. i'm just gonna give it a try. | 18:59 |
| notmyname | heh | 18:59 |
| notmyname | swift team meeting time in #openstack-meeting | 18:59 |
| notmyname | wheeee | 19:00 |
| quack_quack_ | ok i restarted openstack-swift-proxy.service | 19:01 |
| quack_quack_ | no errors | 19:01 |
| *** annegent_ has joined #openstack-swift | 19:02 | |
| notmyname | dfg: around? | 19:04 |
| *** jistr has joined #openstack-swift | 19:05 | |
| *** tsg has joined #openstack-swift | 19:09 | |
| *** kopparam has joined #openstack-swift | 19:17 | |
| *** X019 has quit IRC | 19:25 | |
| *** annegent_ has quit IRC | 19:26 | |
| *** annegent_ has joined #openstack-swift | 19:27 | |
| *** jistr has quit IRC | 19:30 | |
| *** kota_ has quit IRC | 19:42 | |
| *** acoles is now known as acoles_away | 19:43 | |
| mattoliverau | ok, now I'm confused what to do (post meeting), damn summer means the sun is up... I guess I'll go make breakfast.. bbl | 19:44 |
| zaitcev | quack_quack_: That pipeline and the filter look about right, so what is the actual problem? | 19:44 |
| *** tab____ has joined #openstack-swift | 19:48 | |
| *** tsg has quit IRC | 19:49 | |
| *** kopparam has quit IRC | 19:51 | |
| *** Masahiro has joined #openstack-swift | 20:01 | |
| notmyname | 2.2.1 stuff ordered on https://wiki.openstack.org/wiki/Swift/PriorityReviews now | 20:01 |
| *** Masahiro has quit IRC | 20:06 | |
| *** cutforth has quit IRC | 20:07 | |
| *** annegent_ has quit IRC | 20:09 | |
| *** echevemaster has joined #openstack-swift | 20:12 | |
| tdasilva | clayg: around? | 20:26 |
| clayg | tdasilva: not really - getting on a call | 20:26 |
| tdasilva | clayg: ok, no problem.. | 20:27 |
| *** annegent_ has joined #openstack-swift | 20:39 | |
| *** annegent_ has quit IRC | 20:44 | |
| quack_quack_ | zaitcev: I'm gonna check this. I have a feeling it's actually working and i'm just not enabling debug logging. | 20:48 |
| *** gvernik has quit IRC | 20:48 | |
| *** kopparam has joined #openstack-swift | 20:49 | |
| zaitcev | quack_quack_: I have to admit that when I hit issue with s3_token, I usually end resorting to fp=open("/tmp/dump",'a'); fp.write("debug info"); fp.close() all over that thing. | 20:50 |
| quack_quack_ | cool thanks zaitcev | 20:51 |
| zaitcev | For a long time the logging was FUBAR in both s3_token and switf3 because they used LOG, but it's completely mute when such middleware is loaded into Swift. It must use Swift's own swift.utils.get_logger(). | 20:51 |
| quack_quack_ | zaitcev: while i've got you | 20:51 |
| quack_quack_ | I've been trying to figure this out for a couple days | 20:51 |
| quack_quack_ | but the POST /v2.0/s3tokens api call | 20:52 |
| notmyname | zaitcev: quack_quack_: check out q https://pypi.python.org/pypi/q/2.4 | 20:52 |
| torgomatic | fwiw the "q" module is basically the Cadillac of debugging-by-dumping-to-a-file | 20:52 |
| quack_quack_ | I don't know if I'm not looking hard enoguh | 20:52 |
| quack_quack_ | but I can't find any documentatoin on this | 20:52 |
| quack_quack_ | so as I understand the code | 20:52 |
| quack_quack_ | in order to check my S3 credential | 20:52 |
| quack_quack_ | I need to pass | 20:52 |
| torgomatic | import q; q.q(thingy, other_thingy, blah) --> dumps to /tmp/q with pretty-printing and line numbers and stuff | 20:52 |
| quack_quack_ | access key ID | 20:52 |
| quack_quack_ | a token which is an HMAC hash of 'something', usually a canonical request | 20:53 |
| quack_quack_ | and a signature calculated from that token | 20:53 |
| quack_quack_ | ? | 20:53 |
| quack_quack_ | thanks notmyname and torgomatic | 20:53 |
| zaitcev | I don't know much about the split between Keystone and Swift either, in regards to S3. I only implemented S3 for another project. As you observed S3 uses HMAC and there's a definiton of it in Amazon's devel guide. | 20:55 |
| zaitcev | So | 20:55 |
| quack_quack_ | right | 20:55 |
| quack_quack_ | i'm getting a 401 error when I pass it a garbage token and signature | 20:55 |
| zaitcev | Logically, all the things except the plaintext need to be shipped to Keystone, which then finalizes HMAC comparison | 20:55 |
| quack_quack_ | i *think* that's the expected out put if you pass garbage. | 20:55 |
| quack_quack_ | for a little while, i though I needed ot pass a valid admin token or something | 20:55 |
| zaitcev | Man, I completely forgot what I did to make it work. I remember some tinkering... | 20:57 |
| zaitcev | And I had to have tenant:account thing as Amazon username or something like that. | 20:59 |
| quack_quack_ | that's something I don't understand at all. it seems like you can do tenant:account or ec2 access key ID | 21:00 |
| quack_quack_ | i think you only use ec2 access key ID if you have keystone enabled | 21:00 |
| quack_quack_ | it's like, we need some documentation or something :) | 21:00 |
| zaitcev | hmm... the only note I saved was to run "keystone-manage credentials add zaitcev EC2 zaitcev keypass admten" | 21:02 |
| zaitcev | admten being the password's plaintext | 21:02 |
| zaitcev | But you're obviously past that stage. | 21:03 |
| zaitcev | so in the end captured headers look like this (has to have double colon): Authorization: AWS zaitcev:admin:yyiMmswAXi6C6dsNKiaksjooxHI= | 21:03 |
| zaitcev | apparently at the time I thought this information was the key to make swift3 work | 21:05 |
| *** annegent_ has joined #openstack-swift | 21:05 | |
| quack_quack_ | did you have keystone in the pipeline? | 21:06 |
| quack_quack_ | or keystoneauth or whatever | 21:06 |
| zaitcev | I don't think it's necessary if normal Swift apps aren't used, but I did. | 21:08 |
| zaitcev | pipeline = healthcheck cache s3token authtoken swift3 keystone proxy-logging dlo proxy-server | 21:09 |
| quack_quack_ | ok thanks | 21:10 |
| notmyname | zaitcev: what version of swift is in rdo? | 21:10 |
| zaitcev | notmyname: sec... Lon (other guy) built 2.2.0 but I dunno if it made out to repos | 21:11 |
| mattoliverau | morning again | 21:11 |
| notmyname | zaitcev: just curious since that looks like a pretty old pipeline construction (position of dlo, but on the other hand it's here) and only one proxy-logging | 21:12 |
| zaitcev | notmyname: it's 1.13.1 right now | 21:12 |
| notmyname | also no catch_errors or gatekeeper | 21:12 |
| notmyname | zaitcev: ok, thanks | 21:12 |
| *** kopparam has quit IRC | 21:21 | |
| *** tdasilva has quit IRC | 21:21 | |
| *** X019 has joined #openstack-swift | 21:23 | |
| *** JonB has joined #openstack-swift | 21:30 | |
| JonB | How does erasure encoding change the number of zones required in a region? | 21:33 |
| notmyname | JonB: it doesn't. zones are a way to get better distribution across failure domains | 21:35 |
| JonB | notmyname: So, 5 zones if still good enough? | 21:36 |
| JonB | 2 parts of the object will not land on the same server/disk | 21:36 |
| JonB | notmyname: *5 zones is still good enough? | 21:37 |
| notmyname | JonB: you've piqued my curiosity | 21:38 |
| notmyname | JonB: but in answer to your question, 5 is fine. so is 50. so is 1 | 21:38 |
| JonB | If I have 14 parts (of a single object) and 5 zones then 3 parts could land on the same disk, right? | 21:38 |
| notmyname | only if you have one drive per zone | 21:39 |
| mattoliverau | JonB: on the same zone, no necessarily the same disk. | 21:39 |
| mattoliverau | what notmyname said | 21:39 |
| notmyname | what mattoliverau said | 21:39 |
| notmyname | JonB: swift arranges data across drives>disks>zones>regions. it will distribute the data as widely as possible across those failure domains | 21:40 |
| notmyname | JonB: swift uses what you have and doesn't require that you have a certain number (beyond "you should have more than 3 drives in your cluster") | 21:41 |
| JonB | notmyname: Yes, all these same rules will still apply to EC, correct? | 21:43 |
| JonB | except for min of 14 drives, then? | 21:44 |
| notmyname | JonB: yup. | 21:45 |
| JonB | Thanks! | 21:46 |
| notmyname | JonB: what's your swift cluster look like today? | 21:46 |
| JonB | Only test ones, 5 zones, 2 nodes per zones, 12 disks | 21:46 |
| JonB | I am looking at 1/2 PB nodes I am wondering about replication times going that dense. Have you had experience with that? | 21:48 |
| JonB | notmyname: 72x8TB | 21:49 |
| *** lpabon has quit IRC | 21:49 | |
| notmyname | JonB: ya, that's pretty dense. you're right to worry about replication. also your ingest and draining will take a long time. I'd only think that nodes that big would make sense if they are a small fraction of the overall cluster size | 21:50 |
| *** Masahiro has joined #openstack-swift | 21:50 | |
| notmyname | but still you have the replication issues | 21:50 |
| notmyname | JonB: have you looked at some of the 1U 12 drive chassis? they have some nice density while giving you a pretty small failure domain | 21:51 |
| JonB | Not yet, I did not pick out the nodes. 12 drive / 1 U is a bit less dense per U vs 72 disks/4U. I really need to test that a bit if we go that way. | 21:55 |
| *** Masahiro has quit IRC | 21:55 | |
| JonB | notmyname: Would a 2 or 5 node PACO be enough to test a hardware setup. Note: I am not looking at putting object nodes that dense. | 21:57 |
| notmyname | JonB: it really depends on what you're targeting as your prod deployment and your expected workloads | 21:58 |
| JonB | notemyname: One last question :) All the logic/workload for EC is on the proxy nodes, right? and thanks! | 21:59 |
| notmyname | JonB: no, not all of it. the read/write path encoding is on the proxy | 22:01 |
| notmyname | JonB: but reconstructing missing fragments (eg in the case of a drive failure) will happen on the object nodes | 22:01 |
| notmyname | JonB: but also remember that you can designate a different set of hardware for your ec storage policy (ie different SKU that has more CPU) | 22:02 |
| JonB | notemyname: Yeah, I am seeing that... OK, thanks, that makes sense. | 22:03 |
| JonB | notmyname: * | 22:03 |
| *** miurahr has joined #openstack-swift | 22:04 | |
| *** annegent_ has quit IRC | 22:05 | |
| quack_quack_ | torgomatic: omg thanks so much for the tip on q | 22:06 |
| quack_quack_ | q, where have you been all my life | 22:07 |
| torgomatic | quack_quack_: hehe :) | 22:07 |
| torgomatic | I only learned about it from a PyCon lightning talk maybe... 2 years ago? 3? Sometime in Santa Clara, that's for sure | 22:07 |
| *** charz has quit IRC | 22:09 | |
| *** mattoliverau has quit IRC | 22:09 | |
| *** serverascode___ has quit IRC | 22:09 | |
| *** mattoliverau has joined #openstack-swift | 22:10 | |
| *** zigo has quit IRC | 22:11 | |
| *** zigo has joined #openstack-swift | 22:12 | |
| *** lpabon has joined #openstack-swift | 22:12 | |
| *** charz has joined #openstack-swift | 22:12 | |
| *** serverascode___ has joined #openstack-swift | 22:12 | |
| *** lpabon has quit IRC | 22:21 | |
| *** JonB has quit IRC | 22:43 | |
| *** geaaru has quit IRC | 22:48 | |
| *** lcurtis has quit IRC | 22:51 | |
| *** mattoliverau has quit IRC | 22:53 | |
| *** mattoliverau has joined #openstack-swift | 22:53 | |
| *** lcurtis has joined #openstack-swift | 23:06 | |
| *** kopparam has joined #openstack-swift | 23:07 | |
| *** annegent_ has joined #openstack-swift | 23:11 | |
| clayg | cschwede: torgomatic: I think the heart of the issue is when a "teir" has an aggregate weight that is greater than 1/replicanth of the total percentage of weight across the cluster | 23:16 |
| clayg | cschwede: torgomatic: so if you have three servers and three replicas and one server has 1% more capacity but you still want as-unique-as-possible data placement you need to either inrecase the weights in the other servers, or decrease the weights on the one | 23:17 |
| *** Nadeem_ has quit IRC | 23:18 | |
| clayg | cschwede: torgomatic: if you have 5 zones and three replicas it's not a problem unless one zone has more then 1/replicanth of the total weight - then again I think the other zones won't be able to take the 2/replicanths of the weight and that over-weighted zone will end up doing double duty for some part | 23:19 |
| clayg | same thing with 2 regions and 3 replicas, if you want to make sure that the dr region has 1 copy of every part it has to have > 1/replicanth of the total weight | 23:19 |
| clayg | torgomatic: I see you're back at your desk - you buying any of that before I go start ripping up swifterdarrell's all_replicas_of_a_part_in_a_tier change? | 23:24 |
| torgomatic | clayg: that all sounds about right | 23:24 |
| clayg | torgomatic: did you have something else going on where there was a device with zero weight and parts assigned that wouldn't come off? | 23:25 |
| torgomatic | clayg: yeah, I've been trying to get a test written for a fix I have... the fix works on ctennis's builder file that he posted, but getting it to happen again is stupidly difficult | 23:26 |
| torgomatic | in a unit test, that is | 23:26 |
| clayg | torgomatic: k, so but unrelated? in that builder none of the tiers have an aggregate weight > 1/replicanth | 23:27 |
| torgomatic | clayg: no, but in that builder zone 3 has exactly 1/3 the weight | 23:28 |
| clayg | hey man - I didn't say >= i sayd >! | 23:28 |
| clayg | add another zone, it's easy | 23:28 |
| torgomatic | and most of the zero-weight devices are zone 3, so clearly zone 3 used to be more than 1/3, but stuff's just not evacuated because of this bug | 23:29 |
| *** miurahr has quit IRC | 23:29 | |
| clayg | yuk | 23:29 |
| torgomatic | yup | 23:31 |
| *** tab____ has quit IRC | 23:37 | |
| *** kopparam has quit IRC | 23:39 | |
| *** Masahiro has joined #openstack-swift | 23:39 | |
| *** Masahiro has quit IRC | 23:43 | |
| *** lcurtis has quit IRC | 23:45 | |
| *** annegent_ has quit IRC | 23:51 | |
| *** tellesnobrega_ has joined #openstack-swift | 23:55 | |
| *** lcurtis has joined #openstack-swift | 23:57 | |
| *** tellesnobrega_ has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!