| *** klrmn has joined #openstack-swift | 00:03 | |
| *** catintheroof has joined #openstack-swift | 00:08 | |
| *** arch-nemesis has quit IRC | 00:19 | |
| kota_ | good morning | 00:53 |
|---|---|---|
| kota_ | MoiingLemur, notmyname: if you are starting fromclean deployment, i don't think you need to hack the code, probably you mass up your SSDs into a region and assign other HDDs to other regions to simulate the handoff pool. | 01:14 |
| kota_ | MooingLemur:^^ (sorry, typo) | 01:15 |
| *** mmotiani_ has joined #openstack-swift | 01:17 | |
| clayg | kota_: good evening! | 01:18 |
| kota_ | clayg: welcome back :D | 01:18 |
| clayg | kota_: so... no movement on isa_l_rs_cauchy? | 01:19 |
| MooingLemur | kota_: that would require that the SSDs have a weight other than zero, correct? I suppose 0.1 would work :P | 01:19 |
| *** m_kazuhiro has joined #openstack-swift | 01:19 | |
| kota_ | clayg: i think nothing progressed after you add +2, i'm thinking of poking tsg- in this week. | 01:19 |
| kota_ | MooingLemur: ah, you doesn't want to keep anything in SSD, hmm... | 01:20 |
| MooingLemur | and even if write_affinity is set to the ssd pool, there's no guarantee that they'd appear early enough in the handoff list to be chosen. | 01:21 |
| kota_ | yeah, if no weight to the ssd handoff pool, the read would not head for them until the handoff are replicated into primaries. | 01:23 |
| kota_ | current read affinity is just sorting in primaries. | 01:23 |
| kota_ | AFAIK | 01:23 |
| kota_ | clayg: and during waiting review, I'm dougting of why i didn't make base code for isa-l and instead of most of duplicated code from isa-l-vand. | 01:25 |
| *** mmotiani_ has quit IRC | 01:28 | |
| MooingLemur | kota_: perhaps a sane way of hacking it is to have a config option for dedicated handoff zone number, configured in tandem with write_affinity in each region. All this does is simply sort the handoff list so that these appear as the first handoffs, ahead of everything else. Or perhaps 2*n replicas of them by default. | 01:29 |
| MooingLemur | for example in a 2 region (1 hdd and 1 ssd zone each) setup, region 1's proxies will see a primary list of r1z1, r2z1, r2z1, and then a handoff list of r1z2, r1z2, r1z2, followed by the rest of the normal handoffs | 01:32 |
| kota_ | MooingLemur: r1 is for ssd right? | 01:33 |
| MooingLemur | r1z1 and r2z1 for hdd, r1z2 and r2z2 for ssd | 01:33 |
| kota_ | k, zone each setup | 01:34 |
| kota_ | gotcha, it could be. | 01:34 |
| MooingLemur | but now we've got the possibility of incoherency of handoffs between regions for the read | 01:34 |
| MooingLemur | because I'd want r2z2 to be the first three handoffs on the r2 proxies | 01:35 |
| *** klrmn has quit IRC | 01:36 | |
| kota_ | it sounds to work reasonable, if i can raise up another perspective, the write affinity depends on the proxy setting, so if we have proxies for each region, one may prefer r1 but the other may prefer r2 and it may be hard to know actually the object placed which region's ssds. | 01:37 |
| kota_ | if we can assume looking up local region's ssd always, that may not be aproblem | 01:38 |
| MooingLemur | I think you're thinking about my most recent concern. request_node_count is usually set to 2 * replicas, but if you read from a different region, you don't know where the handoffs are | 01:39 |
| kota_ | MooingLemur: yeah, just thinking of you have same idea with me. | 01:39 |
| MooingLemur | That may not be a severe problem as long as a primary is available. | 01:40 |
| MooingLemur | But the idea is the handoffs will revert to primaries within a few minutes at most. | 01:40 |
| MooingLemur | so in practice it may not be a problem. | 01:40 |
| kota_ | yes, the handoff should be propageted to primaries at least in the local region. | 01:41 |
| *** clu_ has quit IRC | 01:54 | |
| *** klrmn has joined #openstack-swift | 02:07 | |
| openstackgerrit | Merged openstack/swift: Drop "--allow-insecure netifaces" from tox pip install https://review.openstack.org/401994 | 02:14 |
| *** catintheroof has quit IRC | 02:16 | |
| *** catintheroof has joined #openstack-swift | 02:16 | |
| *** catintheroof has quit IRC | 02:20 | |
| charz | morning | 02:37 |
| *** henryn has joined #openstack-swift | 02:38 | |
| *** nikivi has quit IRC | 02:40 | |
| *** bkopilov has quit IRC | 03:02 | |
| *** tqtran has quit IRC | 03:02 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/liberasurecode: Gather up common isa_l function to isa_l.c file https://review.openstack.org/403996 | 03:05 |
| kota_ | charz: morning | 03:05 |
| kota_ | clayg: if you have time, could you please check also https://review.openstack.org/403996? That may be more clean implementaion for isa-l cauchy than my older patch https://review.openstack.org/#/c/393263/ | 03:07 |
| patchbot | patch 403996 - liberasurecode - Gather up common isa_l function to isa_l.c file | 03:07 |
| patchbot | patch 393263 - liberasurecode - ISA-L Cauchy support | 03:07 |
| *** links has joined #openstack-swift | 03:10 | |
| *** links has quit IRC | 03:10 | |
| *** catintheroof has joined #openstack-swift | 03:25 | |
| charz | kota_: o/ | 03:34 |
| *** Jeffrey4l has quit IRC | 03:54 | |
| *** Jeffrey4l has joined #openstack-swift | 03:55 | |
| *** ChubYann has quit IRC | 03:58 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/liberasurecode: Gather up common isa_l function to isa_l_common.c file https://review.openstack.org/403996 | 04:07 |
| *** m_kazuhiro has quit IRC | 04:09 | |
| *** psachin has joined #openstack-swift | 04:09 | |
| *** ChubYann has joined #openstack-swift | 04:13 | |
| *** klrmn has quit IRC | 04:14 | |
| *** ppai has joined #openstack-swift | 04:15 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/pyeclib: Add soft warning log line when using liberasurecode <1.3.1 https://review.openstack.org/398744 | 04:23 |
| *** catinthe_ has joined #openstack-swift | 04:23 | |
| *** catintheroof has quit IRC | 04:26 | |
| *** catinthe_ has quit IRC | 04:26 | |
| *** catintheroof has joined #openstack-swift | 04:27 | |
| *** catintheroof has quit IRC | 04:27 | |
| *** dmorita has quit IRC | 04:34 | |
| *** bkopilov has joined #openstack-swift | 04:35 | |
| *** m_kazuhiro has joined #openstack-swift | 04:40 | |
| *** kei_yama has quit IRC | 05:18 | |
| *** kei_yama has joined #openstack-swift | 05:19 | |
| openstackgerrit | Merged openstack/pyeclib: Add soft warning log line when using liberasurecode <1.3.1 https://review.openstack.org/398744 | 05:41 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/pyeclib: PyECLib 1.3.2 https://review.openstack.org/398752 | 05:48 |
| *** Jeffrey4l has quit IRC | 06:12 | |
| *** Jeffrey4l has joined #openstack-swift | 06:12 | |
| *** McMurlock1 has joined #openstack-swift | 06:30 | |
| *** dmorita has joined #openstack-swift | 06:34 | |
| *** dmorita has quit IRC | 06:39 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:43 | |
| *** SkyRocknRoll has quit IRC | 06:50 | |
| *** sams-gleb has joined #openstack-swift | 06:54 | |
| *** sams-gleb has joined #openstack-swift | 06:55 | |
| *** tqtran has joined #openstack-swift | 07:01 | |
| *** SkyRocknRoll has joined #openstack-swift | 07:04 | |
| *** tqtran has quit IRC | 07:06 | |
| *** franklin25 has joined #openstack-swift | 07:09 | |
| notmyname | franklin25: did you have a question about cors? | 07:17 |
| franklin25 | yes, I'm having a problem uploading files to a container | 07:20 |
| notmyname | franklin25: I've always found cors to be rather ... confusing. but I'll try to help, if I can | 07:20 |
| franklin25 | Thanks! | 07:21 |
| *** pcaruana has joined #openstack-swift | 07:21 | |
| *** ChubYann has quit IRC | 07:21 | |
| notmyname | what's not working the way you expected? | 07:22 |
| franklin25 | I'm sorry though, but the developper who can generatte the problem is not yet present. and clearly skype finds it a good idea to clear chatlogs, so the error message is gone | 07:22 |
| franklin25 | I could describe it to you, but I think I'd better give you exact error messages | 07:22 |
| franklin25 | if it's allright with you I'd like to come back to this in I expect an hour or so | 07:22 |
| notmyname | franklin25: it's pretty late for me (~11:30pm). I was about to go to bed. I saw your question in #openstack and hoped I could help, but I won't be here an hour from now | 07:23 |
| franklin25 | ah, in that case I'll try to describe it best I can | 07:24 |
| franklin25 | then trying to upload a file to a container we are getting an error in the preflight about 'allowed-credentials' | 07:25 |
| franklin25 | I found the allow-credentials boolean in the keystone settings, which fixed q si,ilqr error | 07:25 |
| franklin25 | s/si,ilqr/similar/ | 07:26 |
| franklin25 | sorry, typing on a strange keyboard | 07:26 |
| notmyname | :-) | 07:26 |
| franklin25 | anyway, the first error was on a call to port 5000 which I identified as the keystone | 07:27 |
| franklin25 | now we are getting a similar error on port 8080 which I identified as the proxy | 07:27 |
| notmyname | ok | 07:27 |
| franklin25 | correct settings have eluded me till now | 07:28 |
| franklin25 | I am terribly sorry. seems like my boss needs me. This is an orphaned project -.- if you know anything, please feel free to post here I'll read it and test it. but have to go now | 07:29 |
| notmyname | franklin25: for online info, I'd point you to http://docs.openstack.org/developer/swift/cors.html | 07:29 |
| notmyname | which I'm sure you've already seen | 07:29 |
| *** takashi has joined #openstack-swift | 07:30 | |
| notmyname | franklin25: however, if there's more later, feel free to ask again in here. if I'm not around, there are people here from all sorts of timezones. perhaps someone else can help. I'll be back online tomorrow, myself | 07:30 |
| mahatic_ | notmyname: o/ hope you had good thanksgiving celebrations | 07:36 |
| notmyname | mahatic_: thanks. I did :-) | 07:36 |
| mahatic_ | nice | 07:37 |
| *** m_kazuhiro has quit IRC | 07:42 | |
| *** hseipp has joined #openstack-swift | 07:43 | |
| franklin25 | notmyname: thanks! I'll follow it up when the developper arrived | 08:06 |
| *** rledisez has joined #openstack-swift | 08:13 | |
| *** cbartz has joined #openstack-swift | 08:17 | |
| *** oshritf has joined #openstack-swift | 08:24 | |
| *** amoralej|off is now known as amoralej | 08:33 | |
| *** hseipp has quit IRC | 08:33 | |
| *** geaaru has joined #openstack-swift | 08:36 | |
| *** cbartz has left #openstack-swift | 08:38 | |
| *** rcernin has joined #openstack-swift | 08:56 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Add more test for test_byteswapped_serialization https://review.openstack.org/404096 | 09:05 |
| *** mvk has quit IRC | 09:13 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Fix non-deterministic suffix updates in hashes.pkl https://review.openstack.org/402324 | 09:18 |
| franklin25 | while using cors to upload a file to a container, the preflight spits out following error message: XMLHttpRequest cannot load http://domain:8080/v1/AUTH_0aaa3028afff4b8fbadddc0372b2e38a/container1/objectName. Response to preflight request doesn't pass access control check: Credentials flag is 'true', but the 'Access-Control-Allow-Credentials' header is ''. It must be 'true' to allow credentials. Origin 'http://192.168.42. | 09:24 |
| franklin25 | port 8080 is used by the swift proxy, but I am unable to find any documentation concerning the Access-Control-Allow-Credentials for the proxy or any other swift module. In the keystone it has been enabled | 09:25 |
| franklin25 | I have tried adding "allow_credentials = true" to the cors block in swift-proxy.conf, but this did not give any positive results after restarting the swift-proxy | 09:26 |
| *** onovy has quit IRC | 09:32 | |
| *** peterlisak has quit IRC | 09:33 | |
| *** onovy has joined #openstack-swift | 09:38 | |
| *** peterlisak has joined #openstack-swift | 09:38 | |
| *** 17WAANPQN has joined #openstack-swift | 09:43 | |
| *** mvk has joined #openstack-swift | 09:46 | |
| *** openstackgerrit has quit IRC | 09:48 | |
| *** openstackgerrit has joined #openstack-swift | 09:49 | |
| *** nexusz99 has joined #openstack-swift | 09:57 | |
| *** abalfour has quit IRC | 10:00 | |
| *** abalfour has joined #openstack-swift | 10:01 | |
| *** kei_yama has quit IRC | 10:10 | |
| *** sams-gleb has quit IRC | 10:48 | |
| *** sams-gleb has joined #openstack-swift | 10:48 | |
| *** acoles_ is now known as acoles | 10:50 | |
| *** sams-gleb has quit IRC | 10:53 | |
| *** tqtran has joined #openstack-swift | 11:03 | |
| *** sams-gleb has joined #openstack-swift | 11:05 | |
| *** takashi has quit IRC | 11:05 | |
| *** tqtran has quit IRC | 11:07 | |
| *** mvk has quit IRC | 11:08 | |
| openstackgerrit | Mathias Bjoerkqvist proposed openstack/swift: Storing encryption root secret in Barbican https://review.openstack.org/364878 | 11:16 |
| *** mvk has joined #openstack-swift | 11:20 | |
| *** psachin has quit IRC | 11:32 | |
| *** psachin has joined #openstack-swift | 11:35 | |
| *** nexusz99_ has joined #openstack-swift | 11:44 | |
| *** nexusz99 has quit IRC | 11:47 | |
| *** vint_bra has joined #openstack-swift | 12:01 | |
| *** nikivi has joined #openstack-swift | 12:12 | |
| *** bkopilov has quit IRC | 12:17 | |
| *** SkyRocknRoll has quit IRC | 12:27 | |
| *** catinthe_ has joined #openstack-swift | 12:37 | |
| *** xionchen_ has joined #openstack-swift | 12:41 | |
| *** npf has joined #openstack-swift | 12:49 | |
| *** vint_bra has quit IRC | 12:56 | |
| *** vint_bra has joined #openstack-swift | 12:58 | |
| *** vint_bra has quit IRC | 12:59 | |
| *** vint_bra has joined #openstack-swift | 13:07 | |
| *** amoralej is now known as amoralej|lunch | 13:08 | |
| *** ppai has quit IRC | 14:06 | |
| *** Worry has joined #openstack-swift | 14:09 | |
| franklin25 | I have to go. I'll probably revisit somewhere this week to pickup this problem again | 14:10 |
| *** franklin25 has quit IRC | 14:10 | |
| *** zul has quit IRC | 14:18 | |
| *** daemontool has joined #openstack-swift | 14:19 | |
| *** amoralej|lunch is now known as amoralej | 14:23 | |
| *** ChanServ sets mode: +v tdasilva | 14:28 | |
| *** bkopilov has joined #openstack-swift | 14:42 | |
| *** NeverWorry has joined #openstack-swift | 14:42 | |
| *** Worry has quit IRC | 14:42 | |
| *** sams-gleb has quit IRC | 14:47 | |
| *** sams-gleb has joined #openstack-swift | 14:47 | |
| *** sams-gleb has quit IRC | 14:52 | |
| *** NeverWorry has quit IRC | 14:54 | |
| *** acoles is now known as acoles_ | 15:04 | |
| *** tqtran has joined #openstack-swift | 15:05 | |
| *** sams-gleb has joined #openstack-swift | 15:09 | |
| *** tqtran has quit IRC | 15:09 | |
| *** Jeffrey4l has quit IRC | 15:09 | |
| *** chlong has joined #openstack-swift | 15:27 | |
| *** npf has quit IRC | 15:34 | |
| *** dmorita has joined #openstack-swift | 15:36 | |
| *** zul has joined #openstack-swift | 15:38 | |
| *** dmorita has quit IRC | 15:40 | |
| *** psachin has quit IRC | 15:40 | |
| *** klamath has joined #openstack-swift | 15:41 | |
| *** stradling has joined #openstack-swift | 15:43 | |
| *** Jeffrey4l has joined #openstack-swift | 15:44 | |
| *** stradling has quit IRC | 15:44 | |
| *** klamath has quit IRC | 15:45 | |
| *** mmotiani_ has joined #openstack-swift | 15:50 | |
| *** Jeffrey4l has quit IRC | 15:51 | |
| *** Jeffrey4l has joined #openstack-swift | 15:51 | |
| *** bkopilov_ has joined #openstack-swift | 15:56 | |
| *** bkopilov has quit IRC | 15:58 | |
| *** mmotiani_ has quit IRC | 15:58 | |
| *** madorn has quit IRC | 15:59 | |
| *** acoles_ is now known as acoles | 16:02 | |
| *** rcernin has quit IRC | 16:03 | |
| *** jlvillal has joined #openstack-swift | 16:03 | |
| jlvillal | foo,^] | 16:04 |
| patchbot | Error: Spurious "]". You may want to quote your arguments with double quotes in order to prevent extra brackets from being evaluated as nested commands. | 16:04 |
| jlvillal | notmyname: ^^^ Not sure why patchbot does that. | 16:04 |
| notmyname | jlvillal: yeah. that. if I knew exactly why, I'd make it stop. it comes down to supybot config. | 16:05 |
| notmyname | go home patchbot. you're drunk | 16:05 |
| jlvillal | notmyname: Okay thanks :) | 16:05 |
| notmyname | jlvillal: someday, I want to rewrite patchbot to be much much simpler and single-purpose | 16:06 |
| *** chlong has quit IRC | 16:07 | |
| *** StraubTW has joined #openstack-swift | 16:08 | |
| *** winggundamth_ has joined #openstack-swift | 16:15 | |
| *** winggundamth_ has quit IRC | 16:18 | |
| *** chlong has joined #openstack-swift | 16:20 | |
| *** nikivi has quit IRC | 16:20 | |
| *** mmotiani_ has joined #openstack-swift | 16:25 | |
| *** pcaruana has quit IRC | 16:36 | |
| *** mvk has quit IRC | 16:36 | |
| *** dmorita has joined #openstack-swift | 16:37 | |
| *** dmorita has quit IRC | 16:38 | |
| *** dmorita has joined #openstack-swift | 16:38 | |
| *** klrmn has joined #openstack-swift | 16:45 | |
| *** dmorita has quit IRC | 17:01 | |
| *** dmorita has joined #openstack-swift | 17:01 | |
| *** xionchen_ has quit IRC | 17:04 | |
| *** dmorita has quit IRC | 17:07 | |
| *** asettle has quit IRC | 17:08 | |
| *** chlong has quit IRC | 17:10 | |
| openstackgerrit | Drew Balfour proposed openstack/swift: add byteorder information and logic to ring files https://review.openstack.org/395237 | 17:12 |
| *** dmorita has joined #openstack-swift | 17:15 | |
| *** dmorita has quit IRC | 17:16 | |
| *** dmorita has joined #openstack-swift | 17:23 | |
| *** chlong has joined #openstack-swift | 17:23 | |
| *** rledisez has quit IRC | 17:28 | |
| *** diogogmt has joined #openstack-swift | 17:31 | |
| *** dmorita has quit IRC | 17:33 | |
| *** dmorita has joined #openstack-swift | 17:36 | |
| *** McMurlock1 has quit IRC | 17:38 | |
| *** daemontool has quit IRC | 17:39 | |
| *** mvk has joined #openstack-swift | 17:50 | |
| timburke | notmyname: franklin25: i think we'd need to add support for the Access-Control-Allow-Credentials header somewhere around https://github.com/openstack/swift/blob/2.11.0/swift/proxy/controllers/base.py#L200 | 17:55 |
| timburke | see https://www.w3.org/TR/cors/#access-control-allow-credentials-response-header -- is there any reason we'd ever want that to be anything but "true"? | 17:56 |
| *** 17WAANPQN has quit IRC | 17:57 | |
| timburke | (i'm also curious about what the client is doing that it interprets as "including credentials". i mean, sure, there's x-auth-token, but what browser knows about that being auth-related? from the description, i would expect it to only be a problem for client-side certs or Cookie/Authorization headers...) | 17:59 |
| *** asettle has joined #openstack-swift | 18:01 | |
| *** arch-nemesis has joined #openstack-swift | 18:12 | |
| *** zul has quit IRC | 18:13 | |
| zaitcev | it's for cookies | 18:20 |
| zaitcev | has no effect on Swift | 18:20 |
| *** _JZ_ has joined #openstack-swift | 18:21 | |
| *** zul has joined #openstack-swift | 18:21 | |
| zaitcev | (in case of CORS-ed resource that a browser preflights) | 18:22 |
| timburke | zaitcev: except his XMLHttpRequest seemed to want to send cookies, i guess? and failed rather than just, y'know, not sending them | 18:22 |
| zaitcev | yes, but we ignore those, don't we | 18:23 |
| zaitcev | oh | 18:23 |
| zaitcev | Oh, brother. | 18:24 |
| zaitcev | So even Vary is not enough then (well, that one is for caching proxies) | 18:24 |
| *** acoles is now known as acoles_ | 18:31 | |
| henryn | join | 18:33 |
| patchbot | Error: You don't have the admin capability. If you think that you should have this capability, be sure that you are identified before trying again. The 'whoami' command can tell you if you're identified. | 18:33 |
| henryn | I installed SAIO but now want to create a production environment for Swift. Any reco for good tutorials? The proc doc from OpenStack seems daunting to follow. | 18:36 |
| *** henryn has quit IRC | 18:43 | |
| *** dmorita has quit IRC | 18:44 | |
| *** dmorita has joined #openstack-swift | 18:44 | |
| *** dmorita has quit IRC | 18:44 | |
| *** dmorita has joined #openstack-swift | 18:44 | |
| *** henryn has joined #openstack-swift | 18:50 | |
| *** silor has joined #openstack-swift | 18:50 | |
| *** dmorita has quit IRC | 18:52 | |
| notmyname | good morning | 18:52 |
| *** dmorita has joined #openstack-swift | 18:52 | |
| *** dmorita has quit IRC | 18:52 | |
| notmyname | henryn: I'd recommend https://smile.amazon.com/OpenStack-Swift-Administering-Developing-Storage/dp/1491900822/ref=sr_1_1 | 18:53 |
| *** dmorita has joined #openstack-swift | 18:53 | |
| *** tqtran has joined #openstack-swift | 18:54 | |
| *** mmotiani_ has quit IRC | 18:55 | |
| clayg | kota_: lgtm! | 18:57 |
| *** geaaru has quit IRC | 18:59 | |
| *** clu_ has joined #openstack-swift | 19:04 | |
| henryn | notmyname: Thanks. I'm aware of the book from the SwiftStack site. Haven't ordered it yet but definitely will. | 19:10 |
| zaitcev | henryn: Joe's book is quite good, specifically for the practical hints. But you can also follow one of the guides rooted here: https://github.com/openstack/swift/blob/master/doc/source/howto_installmultinode.rst | 19:11 |
| henryn | zaitcev: Thank you, I haven't seen that yet. The abundance of information out there is overwhelming and it's tedious to sift through most of them to find the good ones. | 19:21 |
| *** amoralej is now known as amoralej|off | 19:21 | |
| openstackgerrit | Clay Gerrard proposed openstack/pyeclib: ISA-L Cauchy support https://review.openstack.org/393276 | 19:27 |
| openstackgerrit | Merged openstack/pyeclib: Add greedy test for decode/reconstruct result solid https://review.openstack.org/393656 | 19:27 |
| *** mmotiani_ has joined #openstack-swift | 19:27 | |
| openstackgerrit | Merged openstack/pyeclib: Assorted combinatoric-test fixes https://review.openstack.org/398648 | 19:28 |
| *** karenc has joined #openstack-swift | 19:31 | |
| zaitcev | henryn: Might as well select at random, I think. It all comes down to: 1) plan the nodes - storage and proxy, 2) implement the plan by building rings, modifying configs to match, and adding anything that's missing like memcached and rsync, 3) start the servers and daemons, 4) watch logs for errors, fix up | 19:35 |
| *** silor has quit IRC | 19:36 | |
| clayg | timburke: i'm looking at some unmerged cleanups and fixed for pyeclib that you worked on - i'm rebasing them on some other changed of kota's you +2'd that I merged | 19:42 |
| timburke | clayg: sounds good. i noticed the +2, followed quickly by the fact that it couldn't merge. i was going to rebase myself, but if you'd rather, that's less work for me :) | 19:46 |
| *** hseipp has joined #openstack-swift | 19:46 | |
| *** ChubYann has joined #openstack-swift | 19:50 | |
| jrichli | timburke, joeljwright, acoles, tdasilva: would you agree with my most recent post to bug 1204558 | 19:53 |
| openstack | bug 1204558 in python-swiftclient "Upload on container with write ACL reports errors" [Medium,Confirmed] https://launchpad.net/bugs/1204558 - Assigned to Peter Tran (pete-tran02) | 19:53 |
| openstackgerrit | Clay Gerrard proposed openstack/pyeclib: Fix formatted get_metadata memory leak https://review.openstack.org/385185 | 19:58 |
| tdasilva | jrichli: lol, i don't remember writing that patch | 19:59 |
| timburke | jrichli: i think the error is coming from the object HEAD around https://github.com/openstack/python-swiftclient/blob/3.2.0/swiftclient/service.py#L1846 -- we do this in an attempt to cleanup segments when overwriting. you can pass in a --leave-segments option to avoid it, and then (iirc) the upload succeeds with only a warning about being unable to create the container | 20:00 |
| openstackgerrit | Clay Gerrard proposed openstack/pyeclib: PyECLib 1.3.2 https://review.openstack.org/398752 | 20:03 |
| openstackgerrit | Clay Gerrard proposed openstack/pyeclib: Use Py_BuildValue instead of manually building dicts https://review.openstack.org/395856 | 20:03 |
| openstackgerrit | Merged openstack/pyeclib: Fix formatted get_metadata memory leak https://review.openstack.org/385185 | 20:04 |
| jrichli | timburke: yes, we have had a customer that was confused by having the warning about the 403, but not having a confirmation of the upload succeeding. | 20:07 |
| jrichli | timburke: so I was looking for confirmation that we have done all we could do, really, with it being a warning now. | 20:08 |
| jrichli | if we consider this bug resolved, it would be good to close it. the customer referenced that bug as if it was an outstanding issue to fix. | 20:11 |
| timburke | yeah, i'm not entirely happy about the LBYL approach, but the alternative seems to be to spin up a bunch of upload jobs which will all fail, attempt to create the container, then retry. so instead of one wasted request, we get like 20 :-( | 20:13 |
| timburke | i guess it's resolved? i think it's been left open because nobody likes the current situation but neither has anyone come up with workable solution | 20:13 |
| timburke | look at clayg go, merging C code left and right! | 20:21 |
| openstackgerrit | Merged openstack/liberasurecode: Fix posix_memalign handling https://review.openstack.org/340701 | 20:21 |
| clayg | timburke: no sense leaving a bunch of good code in gerrit if we're going to do a release! | 20:24 |
| clayg | also isn't the get_metadata leak thing critical? | 20:24 |
| clayg | er.. rather *wasn't* it - because the fix is merged now! pyeclib 1.3.2 is going to be the best pyeclib evar! | 20:25 |
| timburke | clayg: idk about "critical"... it came up because i was thinking about calling it rather regularly in the auditor, then kota_ pointed out why it would be a terrible idea at the time :P | 20:31 |
| *** nexusz99_ has quit IRC | 20:32 | |
| jrichli | timburke: btw, I don't think that the line you referenced is not involved in the particular test case I am looking at. I am focused on the situation in test_shell.test_upload_with_read_write_access | 20:33 |
| *** catinthe_ has quit IRC | 20:35 | |
| *** diogogmt has quit IRC | 20:40 | |
| *** diogogmt has joined #openstack-swift | 20:42 | |
| timburke | jrichli: ah, ok. the bug report seemed to be talking about how the upload will actually fail if it can't read the container | 20:44 |
| *** dmorita has quit IRC | 20:44 | |
| *** dmorita has joined #openstack-swift | 20:45 | |
| timburke | that test seems a little screwy though; it looks like we mock out two 403s (one for the container, one for the object), then only expect the message for the container to be displayed? or am i not reading that right? | 20:45 |
| jrichli | all I know is that I think I am reproducing the situation in the bug - and in the test - where I grant read and write access to a user for the objects in container, but not to container itself | 20:47 |
| jrichli | I think in the past, it was a container HEAD that would fail, but now it is the actual container create because we dont bother doing the HEAD first | 20:48 |
| timburke | right | 20:48 |
| *** dmorita has quit IRC | 20:50 | |
| jrichli | and I do only see one error. I will have to look at the test again to see this object mock. must be for something that isn't called? or its not asserted? | 20:50 |
| jrichli | i mean i only see one error when i test by hand | 20:51 |
| zaitcev | "I had a setup with < 10 parts and a dark suffix stuck " - OMG | 20:52 |
| timburke | jrichli: i think i must be mis-reading the test. the asserted behavior definitely matches what i'd expect given the test name; i must not quite be remembering what those 403s are doing for fake_http_connection | 20:52 |
| timburke | i wonder if it'd be worth adding a --skip-container-creation flag or something... | 20:53 |
| jrichli | that would at least give us a way of replying to the problem report :-) could maybe add that suggestion in the warning, too, to help out immediately | 20:55 |
| jrichli | I could take a stab at it. it's a good way to get my feet wet in swiftclient ;-) | 20:56 |
| jrichli | but then, if you know to add the flag, you'd probably know to ignore the warning | 20:57 |
| timburke | oooh... and we could add a suggestion around https://github.com/openstack/python-swiftclient/blob/3.2.0/swiftclient/shell.py#L1111 to tack on a --leave-segments on 403s... | 20:58 |
| timburke | yeah, but this gives them something immediately actionable to shut up the warning | 20:59 |
| *** hseipp has quit IRC | 21:00 | |
| jrichli | so, i didnt follow the thing about the cleaning of segments being an issue. wouldnt that only apply to SLO? I thought this was an error b/c we try to create the cont before the PUT. | 21:00 |
| jrichli | and it looks like the test uses --leave-segments, right? | 21:01 |
| jrichli | it'd be nice if there was a way to know that you have the weird acl thing going on : like a flag that says i only have acl to objs | 21:02 |
| *** sams-gleb has quit IRC | 21:02 | |
| *** sams-gleb has joined #openstack-swift | 21:03 | |
| jrichli | timburke: oh, i see now. if the test did not use that leave-segments flag, there would be more errors. and I see its purpose now. | 21:05 |
| *** sams-gleb has quit IRC | 21:08 | |
| *** dmorita has joined #openstack-swift | 21:28 | |
| mattoliverau | Morning | 21:31 |
| notmyname | mattoliverau: I guess it's the time change, but it seems like you're getting online earlier now :-) | 21:32 |
| notmyname | but yeah, we're only 5 hours apart now | 21:33 |
| *** catintheroof has joined #openstack-swift | 21:34 | |
| mattoliverau | notmyname: yeah, day light savings here. So it 8:30am atm :) | 21:35 |
| *** Jeffrey4l has quit IRC | 21:35 | |
| notmyname | I've got a few widgets on my computer for major cities near swift devs ;-) | 21:35 |
| MooingLemur | 18 hours away from me | 21:35 |
| mattoliverau | notmyname: ahh, nice | 21:35 |
| tdasilva | notmyname: heh, i do the same | 21:36 |
| mattoliverau | MooingLemur: but that also makes it 6 hours right | 21:36 |
| MooingLemur | yup | 21:36 |
| tdasilva | MooingLemur: where are you based? | 21:36 |
| MooingLemur | Phoenix | 21:36 |
| jrichli | morning, mattoliverau :-) | 21:37 |
| *** catintheroof has quit IRC | 21:38 | |
| mattoliverau | jrichli: o/ | 21:39 |
| *** mmotiani_ has quit IRC | 21:43 | |
| *** diogogmt has quit IRC | 21:46 | |
| *** Jeffrey4l has joined #openstack-swift | 21:47 | |
| *** catintheroof has joined #openstack-swift | 21:47 | |
| *** nikivi has joined #openstack-swift | 21:59 | |
| *** isotope has joined #openstack-swift | 22:01 | |
| isotope | I have swift working in my openstack | 22:03 |
| isotope | and can create containers, upload files etc... | 22:03 |
| isotope | however that only works for the admin user | 22:03 |
| isotope | What tool should I use to give other projects the ability to create containers? | 22:04 |
| isotope | I am using keystone for authentication | 22:05 |
| notmyname | isotope: you could have created the admin user and added a swift entry for it (right?) | 22:08 |
| notmyname | s/could/would | 22:08 |
| isotope | I created the swift user | 22:10 |
| isotope | and added the admin role to it | 22:10 |
| notmyname | that's the same process you'd use for the other users. create them in keystone, give them the right role so they can use swift, and that's it. the user gets a token and uses it to talk to swift | 22:11 |
| *** diogogmt has joined #openstack-swift | 22:12 | |
| *** chlong has quit IRC | 22:12 | |
| notmyname | and you can choose if you assign every keystone user to the same account in swift or if you use different accounts | 22:12 |
| isotope | @noymyname if I have an existing tenant is that the same as the "user" | 22:21 |
| notmyname | yes | 22:22 |
| isotope | hmmm, okay I see now | 22:23 |
| isotope | I have a tenant but not a role for that account | 22:23 |
| openstackgerrit | Merged openstack/swift: breaking up functional/tests.py a bit further https://review.openstack.org/401047 | 22:24 |
| notmyname | isotope: "Account" is, unfortunately, and overloaded term | 22:28 |
| notmyname | in swift, an account is where you put stuff. like a bank account. it may or may not be authorized for one or more users | 22:28 |
| isotope | yeah, I have a tenant defined and but no user in keystone associated with the account | 22:29 |
| isotope | so some how I have to create a user and associte it with both the swift user and the tenant | 22:29 |
| notmyname | wait. I thought tenant and user were the same thing | 22:30 |
| isotope | when I do a openstack tenant list | 22:30 |
| isotope | I see the project | 22:30 |
| isotope | when I do openstack user list I don't see that project | 22:30 |
| notmyname | ok. keystone tenant == swift account. | 22:31 |
| *** rcernin has joined #openstack-swift | 22:31 | |
| notmyname | the keystone user uses creds to get a token that is authorized for the tenant (swift account), based on their roles | 22:32 |
| notmyname | ...and a partridge in a pear tree | 22:32 |
| isotope | Ha | 22:32 |
| notmyname | so create a user in keystone, and add a swift endpoint to it (in its service catalog). I think that's what needs to be done | 22:33 |
| isotope | okay so I already had an keystone tenent | 22:33 |
| * notmyname is in fact setting up a new dev environment where keystone can be used | 22:34 | |
| mattoliverau | isotope: you shouhld have a swift role (like swiftoperators) which is defined in keystone and supplied to swift proxy.conf, any user you give that role will have access to swift, there account will be based of the tenent their in. | 22:37 |
| mattoliverau | really there swift account will be <reseller_prefix>_<tenent_id> which is what you should set in your swift endpoint in keystone. | 22:39 |
| mattoliverau | or more exact your swift endpoint in keystone should have internal/public urls that are in the form: 'http://1.2.3.4:8080/v1/KEY_$(tenant_id)s' where 'KEY_' is my reseller prefix defined in swift proxy config. | 22:43 |
| mattoliverau | at least that's how it's set up on my test env. noting this is running an older keystone server. | 22:43 |
| *** klrmn has quit IRC | 22:44 | |
| isotope | yup, thats how I have my endpoints set up | 22:45 |
| *** dmorita has quit IRC | 22:47 | |
| mattoliverau | isotope: nice, so if you just make users apart of the swift role (you've defined in keystone and told swift about) then they should be able to access swift. | 22:48 |
| openstackgerrit | Clay Gerrard proposed openstack/pyeclib: WIP: cleanup lint errors https://review.openstack.org/404457 | 22:48 |
| clayg | kota_: timburke: notmyname: tdasilva: how do we get a pep8 job on patch 404457 | 22:49 |
| patchbot | https://review.openstack.org/#/c/404457/ - pyeclib - WIP: cleanup lint errors | 22:49 |
| timburke | clayg: good call. presumably some changes to project-config | 22:50 |
| notmyname | I'm looking | 22:53 |
| clayg | timburke: there's a pep8 section in the tox.ini? | 22:53 |
| clayg | notmyname: timburke: my bad - yeah the job is already configured; just the tox is running the wrong args/command | 22:54 |
| clayg | i can fix that i think - question redacted | 22:54 |
| openstackgerrit | Clay Gerrard proposed openstack/pyeclib: WIP: cleanup lint errors https://review.openstack.org/404457 | 22:56 |
| *** dmorita has joined #openstack-swift | 23:00 | |
| *** dmorita has quit IRC | 23:00 | |
| *** rcernin has quit IRC | 23:01 | |
| *** diogogmt has quit IRC | 23:01 | |
| *** StraubTW has quit IRC | 23:02 | |
| *** klrmn has joined #openstack-swift | 23:02 | |
| *** dmorita has joined #openstack-swift | 23:03 | |
| *** rcernin has joined #openstack-swift | 23:03 | |
| *** catinthe_ has joined #openstack-swift | 23:03 | |
| *** rcernin has quit IRC | 23:04 | |
| *** rcernin has joined #openstack-swift | 23:04 | |
| *** catintheroof has quit IRC | 23:06 | |
| *** vint_bra has quit IRC | 23:06 | |
| *** asettle has quit IRC | 23:08 | |
| *** dmorita has quit IRC | 23:17 | |
| *** dmorita has joined #openstack-swift | 23:20 | |
| *** dmorita has quit IRC | 23:26 | |
| *** isotope has quit IRC | 23:28 | |
| *** kei_yama has joined #openstack-swift | 23:32 | |
| *** dmorita has joined #openstack-swift | 23:35 | |
| *** rcernin has quit IRC | 23:41 | |
| openstackgerrit | Clay Gerrard proposed openstack/pyeclib: Cleanup lint errors in test suite https://review.openstack.org/404457 | 23:46 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!