| *** cnf has quit IRC | 00:03 | |
| openstackgerrit | Tim Burke proposed openstack/pyeclib: Assorted combinatoric-test fixes https://review.openstack.org/398648 | 00:04 |
|---|---|---|
| clayg | TIL copy middleware swift_source code is SSC - stands for ServerSideCopy | 00:11 |
| openstackgerrit | Timur Alperovich proposed openstack/swift: Add the container_crawler library. https://review.openstack.org/398600 | 00:25 |
| openstackgerrit | Timur Alperovich proposed openstack/swift: Script to sync Swift metadata with Elasticsearch. https://review.openstack.org/398653 | 00:25 |
| *** diogogmt has joined #openstack-swift | 00:27 | |
| timur | The metadata sync patch should show how the container crawler thing is used, but it probably makes sense to review the container crawler first. I wanted to give an example, though | 00:27 |
| *** asettle has joined #openstack-swift | 00:31 | |
| *** david-lyle_ is now known as david-lyle | 00:35 | |
| *** asettle has quit IRC | 00:35 | |
| *** Renich has joined #openstack-swift | 00:43 | |
| *** dmorita has quit IRC | 00:47 | |
| *** cnf has joined #openstack-swift | 00:49 | |
| *** tqtran has quit IRC | 01:05 | |
| clayg | do we have some huristic on EC GET that if enough nodes return 404 we return 404 rather than try more handoffs? | 01:07 |
| clayg | i think i'm seeing on 6+5 that if i get [200] * 5 and [404] * 6 responses we return 404 instead of making a 12th request maybe... | 01:10 |
| *** Renich has quit IRC | 01:24 | |
| zaitcev | guys | 01:24 |
| zaitcev | suppose you were running a cluster on EC with ISA-L. What if you just change the library to something else, would the objects be encoded exactly the same or not? | 01:24 |
| zaitcev | I mean, like jerasure | 01:25 |
| *** dmorita has joined #openstack-swift | 01:49 | |
| clayg | when is hacking 0.13 supposed to come out? I just updated flake8 and got some new (and kinda not terrible) pep8 checks.... | 01:55 |
| kota_ | zaitcev: AFAIK it's encoded as different, so we cannot decode via jerasure the object which encoded by isa-l | 01:58 |
| kota_ | back to online at my office | 01:58 |
| *** asettle has joined #openstack-swift | 02:01 | |
| zaitcev | I was afraid that would be the case, although it is a little surprising. | 02:05 |
| *** asettle has quit IRC | 02:06 | |
| openstackgerrit | zhangyanxian proposed openstack/swift: Fix typos in test_container.py https://review.openstack.org/398694 | 02:06 |
| *** klrmn has quit IRC | 02:07 | |
| openstackgerrit | zhangyanxian proposed openstack/swift: Fix typos in test_container.py https://review.openstack.org/398694 | 02:07 |
| clayg | so whatever i was eeing with the too many 404's and not enough requests must only exist on swift 2.9ish | 02:14 |
| *** siva_krish has quit IRC | 02:19 | |
| *** sgundur_ has joined #openstack-swift | 02:39 | |
| *** clu_ has quit IRC | 02:48 | |
| *** sgundur_ has quit IRC | 02:52 | |
| *** winggundamth has quit IRC | 03:00 | |
| *** Jeffrey4l has quit IRC | 03:11 | |
| *** bkopilov has quit IRC | 03:14 | |
| *** winggundamth has joined #openstack-swift | 03:25 | |
| *** klrmn has joined #openstack-swift | 03:25 | |
| *** asettle has joined #openstack-swift | 03:32 | |
| *** dmorita has quit IRC | 03:36 | |
| *** asettle has quit IRC | 03:36 | |
| *** maestropandy has joined #openstack-swift | 03:38 | |
| *** maestropandy has left #openstack-swift | 03:39 | |
| *** psachin has joined #openstack-swift | 04:12 | |
| *** deep_1 has joined #openstack-swift | 04:18 | |
| *** pdardeau has quit IRC | 04:25 | |
| *** bkopilov has joined #openstack-swift | 04:33 | |
| *** chatter has joined #openstack-swift | 04:44 | |
| chatter | Allah is doing | 04:45 |
| chatter | sun is not doing Allah is doing | 04:45 |
| chatter | moon is not doing Allah is doing | 04:45 |
| chatter | stars are not doing Allah is doing | 04:45 |
| chatter | planets are not doing Allah is doing | 04:46 |
| chatter | galaxies are not doing Allah is doing | 04:46 |
| chatter | oceans are not doing Allah is doing | 04:47 |
| chatter | mountains are not doing Allah is doing | 04:47 |
| chatter | trees are not doing Allah is doing | 04:47 |
| chatter | mom is not doing Allah is doing | 04:48 |
| chatter | dad is not doing allah is doing | 04:48 |
| chatter | boss is not doing Allah is doing | 04:48 |
| chatter | job is not doing Allah is doing | 04:49 |
| chatter | dollar is not doing Allah is doing | 04:49 |
| chatter | degree is not doing Allah is doing | 04:49 |
| chatter | medicine is not doing Allah is doing | 04:50 |
| chatter | customers are not doing allah is doing | 04:50 |
| chatter | you can not get a job without the permission of Allah | 04:50 |
| chatter | you can not get married without the permission of Allah | 04:51 |
| chatter | nobody can get angry at you without the permission of Allah | 04:51 |
| chatter | light is not doing Allah is doing | 04:52 |
| chatter | fan is not doing Allah is doing | 04:52 |
| chatter | businessess are not doing allah is doing | 04:53 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/pyeclib: Add soft warning log line when using liberasurecode <1.3.1 https://review.openstack.org/398744 | 04:53 |
| chatter | america is not doing Allah is doing | 04:53 |
| kota_ | notmyname: patch 398744 is for swift release | 04:54 |
| patchbot | https://review.openstack.org/#/c/398744/ - pyeclib - Add soft warning log line when using liberasurecod... | 04:54 |
| chatter | fire can not burn without the permission of Allah | 04:54 |
| chatter | knife can not cut without the permission of Allah | 04:55 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/pyeclib: Add soft warning log line when using liberasurecode <1.3.1 https://review.openstack.org/398744 | 04:55 |
| kota_ | just fixed small pep8 error | 04:55 |
| chatter | rulers are not doing Allah is doing | 04:56 |
| chatter | governments are not doing Allah is doing | 04:56 |
| chatter | sleep is not doing Allah is doing | 04:57 |
| chatter | hunger is not doing Allah is doing | 04:58 |
| chatter | food does not take away the hunger Allah takes away the hunger | 04:58 |
| chatter | water does not take away the thirst Allah takes away the thirst | 04:59 |
| chatter | seeing is not doing Allah is doing | 05:00 |
| chatter | hearing is not doing Allah is doing | 05:01 |
| chatter | seasons are not doing Allah is doing | 05:01 |
| chatter | weather is not doing Allah is doing | 05:01 |
| chatter | humans are not doing Allah is doing | 05:02 |
| chatter | the best amongst you are those who learn the Quran and teach it | 05:02 |
| *** asettle has joined #openstack-swift | 05:02 | |
| chatter | one letter read from book of Allah amounts to one good deed and allah multiplies one good deed ten times | 05:05 |
| chatter | hearts get rusted as does iron with water to remove rust from heart recitation of Quran and rememberance of death | 05:07 |
| *** asettle has quit IRC | 05:07 | |
| chatter | heart is likened to a mirror | 05:07 |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/swift: Enable DeprecationWarning in test environments https://review.openstack.org/381117 | 05:07 |
| *** cdelatte has quit IRC | 05:08 | |
| chatter | when a person commits one sin a black dot sustains the heart | 05:08 |
| chatter | to accept Islam say that i bear witness that there is no deity worthy of worship except Allah and Muhammad peace be upon him is his slave and messenger | 05:10 |
| chatter | read book http://www.fazaileamaal.com | 05:10 |
| jrichli | and the person to eat the last donut feels voracious ... | 05:11 |
| chatter | read book http://www.muntakhabahadith.com | 05:14 |
| chatter | need spiritual teacher visit http://www.alhaadi.org.za | 05:17 |
| *** klrmn has quit IRC | 05:20 | |
| chatter | Allah created the sky without any pillars | 05:27 |
| chatter | Allah makes the sun rise from the east and Allah makes the sun set in the west | 05:28 |
| chatter | Allah makes the day into the night and Allah makes the night into the day | 05:30 |
| chatter | Allah gives life and Allah gives death | 05:30 |
| chatter | all creation are useless,worthless,hopeless | 05:31 |
| chatter | can not do | 05:31 |
| chatter | can not benefit | 05:31 |
| chatter | can not harm | 05:31 |
| *** deep_1 has quit IRC | 05:31 | |
| chatter | Allah is the doer of each and everything | 05:32 |
| *** deep_1 has joined #openstack-swift | 05:32 | |
| chatter | when Allah wants us to stand we stand | 05:33 |
| chatter | when Allah wants us to sit we sit | 05:33 |
| chatter | i am not doing Allah is doing | 05:34 |
| chatter | you are not doing Allah is doing | 05:34 |
| chatter | atom bomb is not doing Allah is doing | 05:34 |
| chatter | rice is not doing Allah is doing | 05:34 |
| chatter | all creation get together can not create one grain of rice | 05:35 |
| chatter | all humans get together can not stop rain | 05:35 |
| chatter | all humans get together can not make anybody hungry | 05:36 |
| chatter | all humans get together can not move sun one second up or down | 05:36 |
| chatter | we can not count the hair on our head | 05:37 |
| chatter | we can not count the rain drops | 05:37 |
| chatter | we can not count the particles of sand | 05:38 |
| chatter | medicine has no power to cure | 05:38 |
| chatter | two people take same medicine one passes away and one does not | 05:41 |
| chatter | degree has no power to give job | 05:41 |
| chatter | many people have degrees but do not have jobs | 05:42 |
| chatter | sustenance does not depend on effort | 05:42 |
| chatter | one person is working very much but is earning very less | 05:43 |
| chatter | other person is working very less but is earning very much | 05:43 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/pyeclib: PyECLib 1.3.2 https://review.openstack.org/398752 | 05:48 |
| kota_ | notmyname: that's for pyeclib 1.3.2 to require the warning log | 05:49 |
| *** chatter has quit IRC | 05:49 | |
| kota_ | notmyname: perhaps, both patch 398648 and patch 393656 would be worth to be included to 1.3.2 but not strongly required | 05:53 |
| patchbot | https://review.openstack.org/#/c/398648/ - pyeclib - Assorted combinatoric-test fixes | 05:53 |
| patchbot | https://review.openstack.org/#/c/393656/ - pyeclib - Add greedy test for decode/reconstruct result solid | 05:53 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/pyeclib: Add greedy test for decode/reconstruct result solid https://review.openstack.org/393656 | 05:55 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/pyeclib: Assorted combinatoric-test fixes https://review.openstack.org/398648 | 06:02 |
| *** asettle has joined #openstack-swift | 06:03 | |
| *** asettle has quit IRC | 06:08 | |
| *** siva_krish has joined #openstack-swift | 06:10 | |
| *** siva_krish has quit IRC | 06:14 | |
| *** sanchitmalhotra has joined #openstack-swift | 06:20 | |
| *** ppai has joined #openstack-swift | 06:22 | |
| *** sanchitmalhotra has quit IRC | 06:23 | |
| *** sanchitmalhotra has joined #openstack-swift | 06:23 | |
| *** sams-gleb has joined #openstack-swift | 06:25 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/pyeclib: Assorted combinatoric-test fixes https://review.openstack.org/398648 | 06:27 |
| psachin | ping #openstack-swift, can we change timezone from UTC to local when list container ? | 06:27 |
| *** sams-gleb has quit IRC | 06:29 | |
| *** sams-gleb has joined #openstack-swift | 06:30 | |
| *** deep_1 has quit IRC | 06:33 | |
| *** sams-gleb has quit IRC | 06:34 | |
| *** dmorita has joined #openstack-swift | 06:37 | |
| kota_ | psachin: FWIW, it looks no that seems tied up to utc as hard coded | 06:37 |
| psachin | +kota, Yeah! Thanks | 06:38 |
| kota_ | psachin: :-) | 06:38 |
| *** ChubYann has quit IRC | 06:38 | |
| psachin | +kota, but it should pick up local TZ. | 06:38 |
| zaitcev | datestamp = strftime('%Y-%m-%d %H:%M:%S', utc) | 06:39 |
| psachin | +zaitcev: Hmm, I saw the code | 06:40 |
| *** dmorita has quit IRC | 06:41 | |
| *** deep_1 has joined #openstack-swift | 06:42 | |
| kota_ | hmm... I'm not sure if we should pick up the local TZ. the main reason is what is *local* | 06:42 |
| kota_ | client? proxy? object-sever? | 06:43 |
| psachin | client | 06:43 |
| kota_ | psachin: i think it's hard swift to know the client timezone, isn't it? | 06:44 |
| kota_ | swift can know the timezone swift exists but... I don't know right now how we can retrieve the timezone info from client request header... | 06:45 |
| kota_ | as common header? | 06:46 |
| psachin | +kota: A simple workaround would be to not pass utc to time.strftime(), it pick local time | 06:47 |
| psachin | let me try | 06:48 |
| patchbot | (let <variable> = <value> in <command>) -- Defines <variable> to be equal to <value> in the <command> and runs the <command>. '=' and 'in' can be omitted. | 06:48 |
| kota_ | zaitcev: where the line is from? It seems I am missing something. | 06:50 |
| zaitcev | kota_: the client. Obviosly you don't want server to return anything but UTC, because 2 clients in 2 different regions can access the same object. | 06:51 |
| kota_ | zaitcev: gotcha, I was reading upstream swift. Now heading for python-swiftclient | 06:52 |
| zaitcev | kota_: https://github.com/openstack/python-swiftclient/blob/master/swiftclient/shell.py#L500 | 06:52 |
| kota_ | zaitcev: thx | 06:53 |
| kota_ | zaitcev: sure, and exactly, we could improve the expression in the client shell | 06:57 |
| zaitcev | yeah well, it's easy to swap gmtime() to localtime(), but the problem is, we've had this default forever | 06:58 |
| zaitcev | Now you have to come up with something that makes both old users and psachin happy | 06:59 |
| zaitcev | I'm considering if we should switch -lh to useing localtime, leaving gmtime for traditional output. Obviously people are not supposed to parse output of -lh, since -l exists. | 07:00 |
| zaitcev | Therefore.... it should be safe. I think. | 07:00 |
| *** asettle has joined #openstack-swift | 07:04 | |
| psachin | +zaitcev, that does not work. | 07:06 |
| psachin | I checked | 07:06 |
| *** asettle has quit IRC | 07:08 | |
| *** sams-gleb has joined #openstack-swift | 07:16 | |
| *** pcaruana has joined #openstack-swift | 07:18 | |
| zaitcev | psachin: Indeed, for objects someone needs to run strptime() on the last_modified, then convert back with localtime(). We return ISO format there. | 07:21 |
| *** deep_1 has quit IRC | 07:50 | |
| *** deep_1 has joined #openstack-swift | 07:53 | |
| *** takashi_ has joined #openstack-swift | 07:58 | |
| *** asettle has joined #openstack-swift | 08:05 | |
| *** asettle has quit IRC | 08:09 | |
| *** rledisez has joined #openstack-swift | 08:13 | |
| *** geaaru has joined #openstack-swift | 08:27 | |
| *** jroll has quit IRC | 08:27 | |
| *** deep_1 has quit IRC | 08:36 | |
| *** jroll has joined #openstack-swift | 08:40 | |
| *** amoralej|off is now known as amoralej | 08:44 | |
| *** cbartz has joined #openstack-swift | 08:47 | |
| *** cbartz has quit IRC | 08:48 | |
| *** asettle has joined #openstack-swift | 09:06 | |
| *** takashi_ has quit IRC | 09:07 | |
| *** asettle has quit IRC | 09:10 | |
| *** oshritf has joined #openstack-swift | 09:12 | |
| *** hseipp has joined #openstack-swift | 09:13 | |
| *** deep_1 has joined #openstack-swift | 09:15 | |
| *** deep_1 has quit IRC | 09:19 | |
| *** jordanP has joined #openstack-swift | 09:26 | |
| *** sams-gleb has quit IRC | 09:27 | |
| *** sams-gleb has joined #openstack-swift | 09:28 | |
| *** sams-gleb has quit IRC | 09:32 | |
| *** npf has joined #openstack-swift | 09:33 | |
| openstackgerrit | Merged openstack/swift: Fix typos in test_container.py https://review.openstack.org/398694 | 09:45 |
| *** sams-gleb has joined #openstack-swift | 09:46 | |
| *** openstackgerrit has quit IRC | 09:48 | |
| *** openstackgerrit has joined #openstack-swift | 09:49 | |
| *** dfflanders has quit IRC | 09:51 | |
| *** nikivi has joined #openstack-swift | 09:56 | |
| openstackgerrit | Merged openstack/swift: Fix up X-Trans-Id-Extra docs https://review.openstack.org/397979 | 10:05 |
| openstackgerrit | Merged openstack/swift: Add X-Openstack-Request-Id to Access-Control-Expose-Headers https://review.openstack.org/397980 | 10:05 |
| *** asettle has joined #openstack-swift | 10:06 | |
| *** asettle has quit IRC | 10:11 | |
| *** acoles_ is now known as acoles | 10:12 | |
| *** deep_1 has joined #openstack-swift | 10:16 | |
| *** deep_1 has quit IRC | 10:21 | |
| *** asettle has joined #openstack-swift | 10:36 | |
| *** dmorita has joined #openstack-swift | 10:37 | |
| *** dmorita has quit IRC | 10:41 | |
| *** nikivi has quit IRC | 10:49 | |
| *** caiobrentano_ has joined #openstack-swift | 11:42 | |
| *** zhugaoxiao has quit IRC | 12:01 | |
| *** zhugaoxiao has joined #openstack-swift | 12:01 | |
| *** sanchitmalhotra has quit IRC | 12:18 | |
| *** bkopilov has quit IRC | 12:22 | |
| *** catintheroof has joined #openstack-swift | 12:46 | |
| *** psachin has quit IRC | 12:50 | |
| *** amoralej is now known as amoralej|lunch | 12:51 | |
| *** vint_bra has joined #openstack-swift | 12:57 | |
| *** chlong has joined #openstack-swift | 13:00 | |
| *** cbartz has joined #openstack-swift | 13:09 | |
| *** kei_yama has quit IRC | 13:11 | |
| *** deep_1 has joined #openstack-swift | 13:20 | |
| *** Jeffrey4l has joined #openstack-swift | 13:30 | |
| tdasilva | good morning | 13:32 |
| *** dmorita has joined #openstack-swift | 13:34 | |
| *** ppai has quit IRC | 13:37 | |
| *** pdardeau has joined #openstack-swift | 13:38 | |
| *** dmorita has quit IRC | 13:39 | |
| *** deep_1 has quit IRC | 13:41 | |
| *** silor has joined #openstack-swift | 13:41 | |
| *** silor1 has joined #openstack-swift | 13:46 | |
| *** silor has quit IRC | 13:47 | |
| *** silor1 is now known as silor | 13:47 | |
| *** acoles is now known as acoles_ | 13:48 | |
| *** bkopilov has joined #openstack-swift | 13:55 | |
| *** diogogmt has quit IRC | 13:58 | |
| *** hseipp has quit IRC | 14:01 | |
| *** ouchkernel has quit IRC | 14:01 | |
| tdasilva | worth reading: https://review.openstack.org/#/c/398875/1 | 14:02 |
| patchbot | patch 398875 - governance - Reference doc for new language additions | 14:02 |
| *** amoralej|lunch is now known as amoralej | 14:02 | |
| *** maestropandy has joined #openstack-swift | 14:08 | |
| *** klamath has joined #openstack-swift | 14:09 | |
| *** klamath has quit IRC | 14:09 | |
| *** maestropandy has left #openstack-swift | 14:09 | |
| *** klamath has joined #openstack-swift | 14:09 | |
| *** dstepanenko has quit IRC | 14:32 | |
| *** janonymous has quit IRC | 14:34 | |
| *** asettle__ has joined #openstack-swift | 14:36 | |
| *** asettle__ has quit IRC | 14:37 | |
| *** asettle__ has joined #openstack-swift | 14:38 | |
| *** asettle has quit IRC | 14:38 | |
| *** acoles_ is now known as acoles | 14:40 | |
| *** asettle__ is now known as asettle | 14:40 | |
| *** cdelatte has joined #openstack-swift | 14:44 | |
| *** hseipp has joined #openstack-swift | 15:00 | |
| *** sams-gleb has quit IRC | 15:02 | |
| *** hseipp has quit IRC | 15:03 | |
| *** hseipp has joined #openstack-swift | 15:03 | |
| *** sams-gleb has joined #openstack-swift | 15:03 | |
| *** sams-gleb has quit IRC | 15:04 | |
| *** ouchkernel has joined #openstack-swift | 15:06 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 15:09 |
| *** sgundur_ has joined #openstack-swift | 15:19 | |
| *** siva_krish has joined #openstack-swift | 15:20 | |
| *** sgundur_ has quit IRC | 15:23 | |
| *** sams-gleb has joined #openstack-swift | 15:28 | |
| *** oshritf has quit IRC | 15:29 | |
| *** siva_krish has quit IRC | 15:34 | |
| *** cdelatte has quit IRC | 15:35 | |
| *** dmorita has joined #openstack-swift | 15:37 | |
| *** dmorita has quit IRC | 15:42 | |
| *** diogogmt has joined #openstack-swift | 15:46 | |
| *** nikivi has joined #openstack-swift | 15:47 | |
| *** sgundur_ has joined #openstack-swift | 15:48 | |
| *** sgundur_ has quit IRC | 15:55 | |
| *** klrmn has joined #openstack-swift | 16:00 | |
| *** siva_krish has joined #openstack-swift | 16:03 | |
| *** nikivi has quit IRC | 16:06 | |
| *** janonymous has joined #openstack-swift | 16:07 | |
| *** pcaruana has quit IRC | 16:10 | |
| *** trbss has joined #openstack-swift | 16:15 | |
| trbss | Two questions 1) swift capabilities list bulk_upload, does that include extract-archive? 2) Can you perform extract-archive on temp PUT url? | 16:17 |
| *** oshritf has joined #openstack-swift | 16:30 | |
| notmyname | good morning | 16:38 |
| klamath | can someone provide info on the following cache = swift.cache under proxy.conf section [filter:authtoken]? | 16:38 |
| notmyname | trbss: 1: yes 2: yes, I think | 16:39 |
| *** sgundur_ has joined #openstack-swift | 16:39 | |
| notmyname | klamath: it allows keystone to use swift's memcache pool for tokens, thus reducing the amount of cache lookups needed. in the past, there were some optimizations in swift's memcache client code that weren't available in keystone's authtoken. IIRC, the difference is largely gone now | 16:40 |
| klamath | thanks notmyname, should the option be enabled by default? | 16:41 |
| *** oshritf has quit IRC | 16:41 | |
| *** nikivi has joined #openstack-swift | 16:43 | |
| notmyname | new governance resolution proposal about not-python https://review.openstack.org/#/c/398875/ | 16:44 |
| patchbot | patch 398875 - governance - Reference doc for new language additions | 16:44 |
| notmyname | klamath: AFAIK, yes, that should be the setting used | 16:45 |
| *** abalfour has quit IRC | 16:47 | |
| klamath | thanks, openstack-ansible is missing the option at the moment, info on that flag is nonexistent | 16:47 |
| *** abalfour has joined #openstack-swift | 16:48 | |
| notmyname | klamath: it might be worth asking the keystone team too | 16:48 |
| notmyname | klamath: and to be fair, my answers above are my understanding currently, and I'll admit to having a fair bit of ignorance around keystone current implementation of that config value | 16:49 |
| *** acoles is now known as acoles_ | 16:54 | |
| klamath | understood, we have seen decreased workload on the infra nodes with that value being added, just wanted to see if there was some issues around enabling it outside of the few uses cases ive seen. | 16:55 |
| *** geaaru has quit IRC | 16:55 | |
| *** nikivi has quit IRC | 17:00 | |
| notmyname | klamath: that's good to know. that's why it exists: to make things more efficient | 17:03 |
| rledisez | hello notmyname, starting from monday i'll get a person working full time on small file optimization, so i could really use an advice on the "acceptable way" to dev it | 17:05 |
| notmyname | rledisez: oh, cool | 17:05 |
| notmyname | rledisez: I'd like to read more over your wiki page | 17:06 |
| rledisez | notmyname: by acceptable, i mean "that has a chance to get merged upstream eventually" ;) | 17:07 |
| notmyname | :-) | 17:07 |
| notmyname | in my mind, the small file optimization would come after we get the golang object server, just because today it would involved python and we all expect that code to go away in favor of golang | 17:07 |
| rledisez | notmyname: i'll make a quick update on the page, we had a new idea yesterday with interresting results | 17:07 |
| notmyname | nice. looking forward to reading it | 17:07 |
| rledisez | sure, we are working with golang, even if we hope to deploy it on our clusters before golang is ready (cause it might take some times) | 17:08 |
| notmyname | I can't imagine what would give you that idea ;-) | 17:09 |
| rledisez | :D | 17:09 |
| *** siva_krish has left #openstack-swift | 17:20 | |
| *** hseipp has quit IRC | 17:28 | |
| *** sgundur_ has quit IRC | 17:31 | |
| *** dmorita has joined #openstack-swift | 17:35 | |
| *** dmorita has quit IRC | 17:37 | |
| *** dmorita has joined #openstack-swift | 17:37 | |
| openstackgerrit | Bryan Keller proposed openstack/swift: Correctly send 412 Precondition Failed in copy middleware https://review.openstack.org/398631 | 17:50 |
| openstackgerrit | Tim Burke proposed openstack/pyeclib: Assorted combinatoric-test fixes https://review.openstack.org/398648 | 17:58 |
| *** derekjhyang has joined #openstack-swift | 17:58 | |
| *** asettle has quit IRC | 18:00 | |
| *** rledisez has quit IRC | 18:00 | |
| *** nikivi has joined #openstack-swift | 18:01 | |
| *** nikivi has quit IRC | 18:05 | |
| *** nikivi has joined #openstack-swift | 18:07 | |
| *** silor has quit IRC | 18:09 | |
| *** tqtran has joined #openstack-swift | 18:13 | |
| *** nikivi has quit IRC | 18:26 | |
| *** jordanP has quit IRC | 18:32 | |
| *** npf has quit IRC | 18:35 | |
| *** sgundur_ has joined #openstack-swift | 18:37 | |
| clayg | whoa | 18:53 |
| notmyname | http://imgur.com/7j15tXU | 18:54 |
| *** amoralej is now known as amoralej|off | 18:55 | |
| *** diogogmt has quit IRC | 18:55 | |
| *** diogogmt has joined #openstack-swift | 18:57 | |
| *** vinsh has quit IRC | 18:57 | |
| timburke | clayg: ? | 18:57 |
| clayg | timburke: just the scrollback - not a comment on the line above it | 18:58 |
| timburke | oh, the small-files stuff? yeah, sounds cool! | 18:58 |
| notmyname | FYI if you are a core reviewer, you can `/msg chanserv op #openstack-swift <your nick>` and get ops privileges in here. that will allow you to kick people if they are abusive or spamming the channel | 18:58 |
| *** ChanServ sets mode: +o clayg | 18:59 | |
| *** notmyname was kicked by clayg (notmyname) | 18:59 | |
| clayg | nice :) | 18:59 |
| clayg | can he get back in? | 18:59 |
| timburke | lol | 18:59 |
| *** notmyname has joined #openstack-swift | 18:59 | |
| *** ChanServ sets mode: +v notmyname | 18:59 | |
| clayg | nice :D | 19:00 |
| timburke | notmyname: don't give him ideas :P | 19:00 |
| clayg | thanks notmyname ! | 19:00 |
| notmyname | also, /me can adjust channel flags and remove that ability from someone ;-) | 19:00 |
| clayg | gtk! | 19:00 |
| *** ChanServ sets mode: -o clayg | 19:01 | |
| clayg | good call | 19:01 |
| *** ChanServ sets mode: +o clayg | 19:01 | |
| clayg | cool | 19:01 |
| *** diogogmt has quit IRC | 19:02 | |
| clayg | is anyone interested to triage lp bug #1642538 - the ring loop bug? cc torgomatic cschwede | 19:31 |
| openstack | Launchpad bug 1642538 in OpenStack Object Storage (swift) "swift-ring-builder rebalance - endless loop" [Undecided,New] https://launchpad.net/bugs/1642538 | 19:31 |
| *** sgundur_ has quit IRC | 19:31 | |
| *** sgundur_ has joined #openstack-swift | 19:32 | |
| *** oshritf has joined #openstack-swift | 19:35 | |
| tdasilva | notmyname, clayg: hello! I was looking for a summary of the discussions we had in BNC re hummingbird, but all I found is this: https://etherpad.openstack.org/p/BCN-swift-working-session-5 | 19:35 |
| tdasilva | is there anything else? | 19:35 |
| notmyname | clayg might have something | 19:36 |
| *** sgundur_ has quit IRC | 19:36 | |
| *** dmorita has quit IRC | 19:36 | |
| *** ChubYann has joined #openstack-swift | 19:37 | |
| *** dmorita has joined #openstack-swift | 19:37 | |
| *** dmorita has quit IRC | 19:37 | |
| *** gyee has joined #openstack-swift | 19:38 | |
| *** dmorita has joined #openstack-swift | 19:39 | |
| *** dmorita has quit IRC | 19:41 | |
| *** dmorita has joined #openstack-swift | 19:41 | |
| *** sgundur_ has joined #openstack-swift | 19:43 | |
| *** openstackgerrit has quit IRC | 19:48 | |
| *** openstackgerrit has joined #openstack-swift | 19:49 | |
| notmyname | clayg: acoles_: kota_: tdasilva: https://gist.github.com/notmyname/5d2f8513845068e3d591214d1b93c770 | 19:49 |
| clayg | i'm guessing it's somewhere in the replica plan - a diff a unittest that hangs would probably be helpfulish | 19:49 |
| clayg | notmyname: sure | 19:49 |
| *** oshritf has quit IRC | 19:50 | |
| tdasilva | notmyname: ack, looks good | 19:50 |
| notmyname | thanks | 19:50 |
| *** arch-nemesis has joined #openstack-swift | 20:03 | |
| openstackgerrit | John Dickinson proposed openstack/swift: 2.11.0 release notes https://review.openstack.org/399215 | 20:06 |
| notmyname | when it lands, I'll ask for the release tag ^ | 20:06 |
| notmyname | doh! forgot the reno yaml version | 20:07 |
| notmyname | ugh. this is a perfect example of where different patches are grouped and don't really fit the model reno has | 20:08 |
| *** trbss has left #openstack-swift | 20:10 | |
| clayg | timburke: wow, i assumed it was falling over in planning - but no all those tiny numbers correctly come out as "Using effective overload of 0.000000 | 20:11 |
| patchbot | Error: No closing quotation | 20:11 |
| clayg | ... so it's something in placement - this is *initial* placement yeah? | 20:11 |
| timburke | clayg: yup. you can just copy/paste the commands from the bug into a terminal | 20:12 |
| timburke | might need to (temporarily?) add some more debug logging to see what's gone off the rails & why | 20:13 |
| timburke | i let it run for 6 or 7 minutes; nothin' | 20:13 |
| clayg | timburke: yeah it's very curious - the next thing that should happen is placing parts and those all log - very strange | 20:14 |
| timburke | clayg: oh... weird. you can see from the traceback it's in place_parts; i guess it never actually picks any locations, though? | 20:15 |
| clayg | tdasilva: is there anything you're looking for - we mostly only talked about the open design question on the replication-server (should it be part of the replicator daemon, the object-server, or it's own endpoint/service like rsyncd?) | 20:15 |
| *** arch-nemesis has quit IRC | 20:15 | |
| tdasilva | clayg: i guess i'm trying to figure out how could we share information on what the task items are and what people are currently working on | 20:16 |
| tdasilva | so that if someone wanted to contribute, how could they figure out what to work on | 20:16 |
| clayg | tdasilva: right | 20:17 |
| clayg | tdasilva: so a gate job that runs go test would be pretty great | 20:17 |
| notmyname | tdasilva: clayg: https://review.openstack.org/#/c/323617/ might be a starting point for that | 20:18 |
| patchbot | patch 323617 - swift (feature/hummingbird) - golang unit testing in the "normal" way | 20:18 |
| tdasilva | clayg, notmyname: mmm i'm probably missing something on what you mean. I guess from current patches, one could tell what people are working on currently, but how can someone learn what else is there to jump on? | 20:21 |
| notmyname | tdasilva: there's not a good global tracking doc right now. it's "look at patches, talk in irc" | 20:22 |
| notmyname | but getting tests running in the openstack gate would be a great help :-) | 20:22 |
| tdasilva | notmyname: lol, gotcha | 20:22 |
| clayg | tdasilva: i'm suggesting the most independent stuff to work on is all of the stuff we're going to need no matter what happens in code/implementation | 20:23 |
| openstackgerrit | John Dickinson proposed openstack/swift: 2.11.0 release notes https://review.openstack.org/399215 | 20:23 |
| clayg | so gate jobs, someway to describe what our dependencies are to packagers (I think distros like to make a some-depends-golang package that goes into a GOPATH instead of just using go get) | 20:24 |
| clayg | then when you build you building from a known version of a library | 20:24 |
| notmyname | in general, that's the stuff that -infra's been working on and the new TC governance thing is talking about (if you want a list of TODOs) | 20:25 |
| notmyname | now that I've got the release stuff done, and it seems like the XFS testing stuff is pretty close (and that's kept my head in the -infra repos), I can tackle the testing question | 20:26 |
| notmyname | clayg: tdasilva: ^ | 20:27 |
| clayg | tdasilva: another super orthogonal thing is object server compatibility - stuff like fast-post, mime put, maybe even an ec/frag disk engine - a set of (probe?) tests can drive testing both implementations | 20:27 |
| notmyname | ie getting golang tests running in the gate | 20:27 |
| tdasilva | clayg, notmyname: cool, so yeah, there's plenty to go on a TODO list, do you think it is worth throwing that on a etherpad or wiki? maybe off the ideas page? | 20:28 |
| zaitcev | gonna get obsolete super fast | 20:29 |
| clayg | tdasilva: i hear the kids like trello | 20:29 |
| notmyname | tdasilva: yeah, I need to do that. well, anyone can, but it seems like more officially my job to do that ;-) | 20:29 |
| tdasilva | oh yeah, trello | 20:29 |
| zaitcev | tdasilva: see what have you done! | 20:29 |
| tdasilva | lol | 20:30 |
| clayg | zaitcev: i'm with you brother | 20:30 |
| notmyname | zaitcev: we'll start having extra meetings to appropriately score tasks and place then on the backlog ;-) | 20:30 |
| tdasilva | lol, nothing like a standup meeting | 20:30 |
| notmyname | if one meeting is good, two must be better, right? | 20:31 |
| zaitcev | well... what's it called... "puzzles and challenges" IIRC | 20:31 |
| timburke | notmyname: might be worth letting the fix for https://bugs.launchpad.net/swift/+bug/1642538 into the release; it's fairly small. just need to write the test... | 20:32 |
| openstack | Launchpad bug 1642538 in OpenStack Object Storage (swift) "swift-ring-builder rebalance - endless loop" [Undecided,Confirmed] - Assigned to Tim Burke (1-tim-z) | 20:32 |
| timburke | (famous last words) | 20:32 |
| zaitcev | clayg: A few days ago notmyname mentioned that everyone else gave up on Proxy in Hummingbird for now, and expressed surprise that I was working on it. I didn't ask but I guessed that it's because of middleware. | 20:33 |
| notmyname | zaitcev: yeah. mostly | 20:34 |
| zaitcev | clayg: IOW, if we have to keep some kind of Python proxy _anyway_, might as well postpone | 20:34 |
| zaitcev | okay. But it's exactly the problem that Ceph RGW faced too. | 20:34 |
| notmyname | the other smaller reason for not doing any golang work is to simply focus on the backend | 20:34 |
| notmyname | zaitcev: what do you mean? | 20:34 |
| clayg | zaitcev: it's more ... it's just not a requirement - we're trying to minimize scope so we can get something done | 20:34 |
| clayg | working on the proxy is avoiding work that *has* be done | 20:34 |
| zaitcev | I see. That makes sense but I couuld not find a corner onto which to tug in the backend, and proxy seemed like separated enough from you, Sam, redbo and others. | 20:35 |
| clayg | redbo is working on container servers because you can't tell redbo what to do | 20:36 |
| clayg | sam isn't working on hummingbird | 20:36 |
| clayg | ... the rumors of the death of nadeem and dfg might be exaggerated | 20:37 |
| clayg | it's mostly you and I - and i'm tied up in hell with libec | 20:37 |
| zaitcev | As far as RGW goes, they approached the problem with a combination of rewriting for things that mush be (like Keystone client and Ceilometer), and running something known as "rgwift" as an additional proxy in front for more esoteric things. | 20:37 |
| zaitcev | A guy from Mirantis wrote that | 20:38 |
| clayg | zaitcev: ... you're trying to nerd snipe | 20:38 |
| zaitcev | clayg: I may be, but I do not understand the slang. | 20:39 |
| clayg | zaitcev: stop thinking/worrying about the proxy - it's not in scope | 20:39 |
| zaitcev | http://www.explainxkcd.com/wiki/index.php?title=356:_Nerd_Sniping | 20:39 |
| *** caiobrentano_ has quit IRC | 20:41 | |
| *** caiobrentano_ has joined #openstack-swift | 20:41 | |
| *** vinsh has joined #openstack-swift | 20:41 | |
| clayg | zaitcev: yeah that's it - you're going to get someone run over by bus - focus on what *has* to be done (hit: solving "how are we going to merge a replacement object server" is not *interesting* - it's just necessary) | 20:42 |
| zaitcev | See, I did not realize the object server even needed any changes. | 20:42 |
| zaitcev | I thought it was all in reconstructor. | 20:42 |
| clayg | timburke: good find! | 20:44 |
| notmyname | zaitcev: replication was talked about as the first step. but yeah, the object server (and it's associated daemons) are what's in scope for now | 20:45 |
| clayg | zaitcev: it doesnt have a POST verb or support MIME PUT or a diskengine for ec frags | 20:45 |
| clayg | zaitcev: i'm not interested in those things because I think replication is first | 20:45 |
| *** caiobrentano_ has quit IRC | 20:45 | |
| clayg | ... but it's not like replication doesn't *work* on hummingbird - so in theory there isn't much to do there - it's all about getting it ready to merge (refactor, test, infra) | 20:46 |
| notmyname | I need to write up, in a single place, a "here's what is in scope, here's the plan" document. I'm sorry if that hasn't been clear so far | 20:46 |
| zaitcev | MIME PUT sounds like... mundane enough job that an idiot like me can do | 20:46 |
| clayg | notmyname: the plan hasn't really changed since San Antonio | 20:46 |
| clayg | zaitcev: it would be great to have replicated encrypted data work on hummingbird object server | 20:47 |
| notmyname | clayg: I didn't think it had. | 20:47 |
| clayg | notmyname: well we sorta said "we're not going to do the object server if we don't have to" - but it turned out we sorta did/should - but other than that - not really | 20:48 |
| clayg | well... and the bit about delete code from feature-hummingbird branch - which no one has done | 20:48 |
| * notmyname goes to get lunch | 20:51 | |
| openstackgerrit | Tim Burke proposed openstack/swift: Avoid infinite loop while placing parts https://review.openstack.org/399237 | 21:07 |
| timburke | clayg: ^^ | 21:10 |
| *** ouchkernel has quit IRC | 21:18 | |
| clayg | timburke: yeah looks good | 21:19 |
| *** sgundur_ has quit IRC | 21:20 | |
| *** sams-gleb has quit IRC | 21:23 | |
| *** sams-gleb has joined #openstack-swift | 21:24 | |
| *** sams-gleb has quit IRC | 21:28 | |
| *** Jeffrey4l has quit IRC | 21:33 | |
| *** sgundur_ has joined #openstack-swift | 21:40 | |
| *** jamielennox is now known as jamielennox|away | 21:40 | |
| notmyname | if you've noticed that gerrit is nice and compact when you're logged out but it's less compact once you log in, it's because of the star images. you can add this to a local stylesheet in your browser and get more compact rows https://gist.github.com/notmyname/eeedc626c8bf49ae5d6f005a8d6778a8 | 21:45 |
| *** NKleppan has joined #openstack-swift | 21:52 | |
| mattoliverau | morning | 21:58 |
| *** lcurtis has joined #openstack-swift | 22:01 | |
| *** sgundur_ has quit IRC | 22:01 | |
| *** jamielennox|away is now known as jamielennox | 22:24 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift: breaking down tests.py into smaller pieces https://review.openstack.org/399275 | 22:24 |
| notmyname | tdasilva: yay! (however, I'm worried you'll conflict with /my/ patch ;-) | 22:30 |
| tdasilva | notmyname: lol, i bet 'mine' is easier to merge :P | 22:52 |
| notmyname | lol | 22:52 |
| tdasilva | notmyname: what's your patch again? | 22:52 |
| notmyname | actually, not sure it conflicts yet, but it's the xattrs one | 22:52 |
| notmyname | tdasilva: mattoliverau: timburke: can you take a quick look at https://review.openstack.org/#/c/399215/ please? If I can land it today, we might be able to get a release this week | 22:52 |
| tdasilva | i can add a dependency | 22:52 |
| patchbot | patch 399215 - swift - 2.11.0 release notes | 22:52 |
| mattoliverau | notmyname: will do | 22:53 |
| notmyname | tdasilva: https://review.openstack.org/#/c/336323/ but if they don't conflict, then dont' add a dependency :-) | 22:53 |
| patchbot | patch 336323 - swift - Add checksum to object extended attributes | 22:53 |
| *** sgundur_ has joined #openstack-swift | 22:54 | |
| *** sgundur_ has quit IRC | 23:02 | |
| timburke | clayg: you wrote https://github.com/openstack/swift/commit/7035639 -- why do we sort by target at https://github.com/openstack/swift/blob/2.10.0/swift/common/ring/builder.py#L808-L810 ? won't that fill up tiny places first (and be a bad thing)? or is the idea to ever-so-slightly prefer better dispersion (which i guess is a good thing)? | 23:09 |
| clayg | timburke: probably doesn't matter much - maybe some test landed like 254/256/257 and I just thought it looked better? | 23:15 |
| clayg | you could try to flip it and see what tests fail | 23:15 |
| clayg | it's all just rounding errors - there's probably some kind of bounding on the maximum number of parts that can get placed by that little loop | 23:16 |
| clayg | you definately have to place all the parts ;) | 23:16 |
| timburke | and you definitely *can't* place more than all the parts | 23:17 |
| timburke | i'm just trying to come up with what my comment should be for why, in this nicely-balanced reasonably-dispersed case, we should actually *expect* that one zone wants more parts than it has and another has more parts than it wants :-( | 23:19 |
| timburke | looking at partition counts by zone, we're perfect: 256/256/256. but region that only has one zone wants an extra part | 23:21 |
| clayg | heh, just *one* extra :D | 23:21 |
| timburke | because he's got the smaller target | 23:22 |
| clayg | timburke: does it look better if you reverse the sorting? | 23:26 |
| openstackgerrit | John Dickinson proposed openstack/swift: 2.11.0 release notes https://review.openstack.org/399215 | 23:26 |
| notmyname | mattoliverau: tdasilva: done. thanks | 23:26 |
| clayg | in all the the failing tests I'm looking it - it's more like "these two places should be get one more than this - but i have a whole one to split between the two - so i give it to the first" | 23:27 |
| notmyname | bah. forgot the reno copy of them | 23:27 |
| notmyname | gotta say, reno is just awesome | 23:27 |
| openstackgerrit | John Dickinson proposed openstack/swift: 2.11.0 release notes https://review.openstack.org/399215 | 23:28 |
| clayg | notmyname: i must say - this - https://twitter.com/notmyname/status/798988423583318018 - cracked me up | 23:28 |
| clayg | https://cdn0.vox-cdn.com/thumbor/_JI6MgeGQfe7BxV1fpL0CBpD2w0=/6x0:895x500/1600x900/cdn0.vox-cdn.com/uploads/chorus_image/image/49493993/this-is-fine.0.jpg | 23:28 |
| notmyname | :-) | 23:29 |
| *** catintheroof has quit IRC | 23:29 | |
| timburke | clayg: hey, it keeps the coffee warm | 23:29 |
| notmyname | tdasilva: I fixed the 2 things you found. one more pass at it? | 23:29 |
| pdardeau | notmyname: maybe you could sprinkle in some AI and have patchbot do that. how hard can it be? ;-) | 23:30 |
| *** catintheroof has joined #openstack-swift | 23:30 | |
| *** Jeffrey4l has joined #openstack-swift | 23:31 | |
| notmyname | pdardeau: I'm pretty sure that new features in patchbot are just further misconfigurations of whatever supybot fork I'm running ;-) | 23:31 |
| *** kei_yama has joined #openstack-swift | 23:31 | |
| openstackgerrit | Tim Burke proposed openstack/swift: Avoid infinite loop while placing parts https://review.openstack.org/399237 | 23:33 |
| *** vinsh has quit IRC | 23:33 | |
| *** vinsh has joined #openstack-swift | 23:34 | |
| *** catintheroof has quit IRC | 23:35 | |
| *** lcurtis has quit IRC | 23:46 | |
| clayg | timburke: idk, I think there's a case to be made to just reverse that sort | 23:54 |
| timburke | i don't know that there is. i think that even if we reverse the sort, we could probably come up with some topology where it breaks the other way | 23:55 |
| *** vinsh_ has joined #openstack-swift | 23:57 | |
| *** vinsh has quit IRC | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
