| *** omolchanov1 has quit IRC | 01:14 | |
| kota_ | good morning! | 01:15 |
|---|---|---|
| *** omolchanov has joined #openstack-swift | 01:16 | |
| mattoliverau | kota_: mornign | 01:29 |
| mattoliverau | *morning | 01:29 |
| kota_ | mattoliverau: \o/ | 01:29 |
| MooingLemur | evening :3 | 01:36 |
| kota_ | MooingLemur: :-) | 01:40 |
| *** tonanhngo_ has joined #openstack-swift | 01:48 | |
| *** tonanhngo has quit IRC | 01:50 | |
| *** sams-gleb has joined #openstack-swift | 01:53 | |
| *** chlong__ has joined #openstack-swift | 01:55 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Small fixes for ec duplciation https://review.openstack.org/438862 | 01:56 |
| *** sams-gleb has quit IRC | 01:57 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Follow up the reconstructor test coverage https://review.openstack.org/440950 | 02:03 |
| *** geaaru has quit IRC | 02:06 | |
| m_kazuhiro | tdasilva: Are you here? | 02:09 |
| kota_ | m_kazuhiro: I'm really glad you tried to poke him but I don't think he is here right now because it's sunday night for him | 02:11 |
| m_kazuhiro | kota_: OK. You are right. I will discuss with him later. | 02:15 |
| kota_ | m_kazuhiro: i think usually tdasilva is using irc bouncer to catch up the irc logs but it looks the bouncer for him is absent... | 02:16 |
| *** catintheroof has quit IRC | 02:21 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Optimize ec duplication and its md5 hashing https://review.openstack.org/421673 | 02:22 |
| *** catintheroof has joined #openstack-swift | 02:23 | |
| m_kazuhiro | kota_: Umm. So I will need to catch him directly to start discussion. | 02:23 |
| kota_ | i see | 02:23 |
| *** catintheroof has quit IRC | 02:27 | |
| *** winggundamth has joined #openstack-swift | 02:34 | |
| *** gkadam has joined #openstack-swift | 02:45 | |
| *** m_kazuhiro has quit IRC | 02:51 | |
| *** janonymous has joined #openstack-swift | 03:06 | |
| *** chatter29 has joined #openstack-swift | 03:17 | |
| *** chatter29 has quit IRC | 03:18 | |
| *** tonanhngo_ has quit IRC | 03:43 | |
| *** tonanhngo has joined #openstack-swift | 03:44 | |
| *** omolchanov has quit IRC | 03:46 | |
| *** sams-gleb has joined #openstack-swift | 03:55 | |
| *** sams-gleb has quit IRC | 03:59 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Optimize ec duplication and its md5 hashing https://review.openstack.org/421673 | 04:03 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Small fixes for ec duplciation https://review.openstack.org/438862 | 04:06 |
| *** spotz_zzz is now known as spotz | 04:13 | |
| *** adriant has quit IRC | 04:20 | |
| *** geaaru has joined #openstack-swift | 04:31 | |
| *** psachin has joined #openstack-swift | 04:49 | |
| *** klrmn has quit IRC | 05:26 | |
| *** psachin has quit IRC | 05:42 | |
| *** sams-gleb has joined #openstack-swift | 05:57 | |
| *** sams-gleb has quit IRC | 06:01 | |
| *** rcernin has joined #openstack-swift | 06:08 | |
| *** chatter29 has joined #openstack-swift | 06:21 | |
| chatter29 | hey guus | 06:21 |
| chatter29 | hey guys | 06:21 |
| chatter29 | allah is doing | 06:21 |
| chatter29 | sun is not doing allah is doing | 06:22 |
| chatter29 | 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 | 06:22 |
| *** chatter29 has quit IRC | 06:23 | |
| *** bkopilov has joined #openstack-swift | 06:24 | |
| *** geaaru has quit IRC | 06:37 | |
| *** pcaruana has joined #openstack-swift | 06:58 | |
| *** pcaruana has quit IRC | 07:05 | |
| *** pcaruana has joined #openstack-swift | 07:06 | |
| *** ChubYann has quit IRC | 07:12 | |
| *** tesseract has joined #openstack-swift | 07:13 | |
| *** SkyRocknRoll has joined #openstack-swift | 07:16 | |
| *** tonanhngo has quit IRC | 07:34 | |
| *** tonanhngo has joined #openstack-swift | 07:35 | |
| *** chlong__ has quit IRC | 07:36 | |
| *** hseipp has joined #openstack-swift | 07:41 | |
| *** sams-gleb has joined #openstack-swift | 07:59 | |
| *** sams-gleb has quit IRC | 08:02 | |
| *** sams-gleb has joined #openstack-swift | 08:02 | |
| *** geaaru has joined #openstack-swift | 08:31 | |
| *** SkyRocknRoll has quit IRC | 08:35 | |
| *** oshritf has joined #openstack-swift | 08:43 | |
| *** tone_zrt has joined #openstack-swift | 08:47 | |
| *** geaaru has quit IRC | 08:58 | |
| *** amoralej|off is now known as amoralej | 09:04 | |
| *** cbartz has joined #openstack-swift | 09:09 | |
| *** d0ugal has quit IRC | 09:14 | |
| *** d0ugal has joined #openstack-swift | 09:26 | |
| *** d0ugal has joined #openstack-swift | 09:26 | |
| *** vills_ has joined #openstack-swift | 10:00 | |
| *** openstackgerrit has quit IRC | 10:33 | |
| *** openstackgerrit has joined #openstack-swift | 10:35 | |
| openstackgerrit | Christian Schwede proposed openstack/swift master: Add support to increase object ring partition power https://review.openstack.org/337297 | 10:35 |
| *** mvk has quit IRC | 10:35 | |
| *** vills_ has quit IRC | 10:52 | |
| *** gK-1wm-su has joined #openstack-swift | 10:53 | |
| *** gK-1wm-su has quit IRC | 10:55 | |
| *** sams-gleb has quit IRC | 11:03 | |
| *** mvk has joined #openstack-swift | 11:07 | |
| *** vills has joined #openstack-swift | 11:17 | |
| *** murugesh112 has joined #openstack-swift | 11:20 | |
| murugesh112 | Hi There | 11:20 |
| *** vills has left #openstack-swift | 11:25 | |
| *** furlongm has quit IRC | 11:32 | |
| *** furlongm has joined #openstack-swift | 11:32 | |
| murugesh112 | Hi | 11:36 |
| murugesh112 | Is somebody there?? | 11:37 |
| *** sams-gleb has joined #openstack-swift | 11:52 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Generalize command test class https://review.openstack.org/441920 | 12:27 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Draft implementation of Composite Ring Zipper https://review.openstack.org/441921 | 12:27 |
| kota_ | clayg: ^^ | 12:27 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Optimize ec duplication and its md5 hashing https://review.openstack.org/421673 | 12:29 |
| *** amoralej is now known as amoralej|lunch | 12:46 | |
| *** jamielennox is now known as jamielennox|away | 12:50 | |
| *** tdasilva has joined #openstack-swift | 13:02 | |
| *** kei_yama has quit IRC | 13:10 | |
| murugesh112 | Hi | 13:20 |
| murugesh112 | I am facing issue in starting openstack-swift-proxy service | 13:21 |
| murugesh112 | can somebody help me out | 13:21 |
| *** cbartz has quit IRC | 13:38 | |
| *** klamath has joined #openstack-swift | 13:43 | |
| *** klamath has quit IRC | 13:45 | |
| *** klamath has joined #openstack-swift | 13:45 | |
| *** murugesh112 has left #openstack-swift | 13:46 | |
| *** cbartz has joined #openstack-swift | 13:48 | |
| *** tonanhngo_ has joined #openstack-swift | 13:48 | |
| *** amoralej|lunch is now known as amoralej | 13:50 | |
| *** tonanhngo has quit IRC | 13:51 | |
| *** catintheroof has joined #openstack-swift | 14:04 | |
| *** cdelatte has joined #openstack-swift | 14:31 | |
| *** sams-gleb has quit IRC | 14:41 | |
| *** sams-gleb has joined #openstack-swift | 14:42 | |
| *** sams-gleb has quit IRC | 14:46 | |
| *** gkadam has quit IRC | 14:54 | |
| *** chlong__ has joined #openstack-swift | 14:57 | |
| *** sams-gleb has joined #openstack-swift | 15:02 | |
| *** hurricanerix has quit IRC | 15:13 | |
| *** sams-gle_ has joined #openstack-swift | 15:15 | |
| *** sams-gleb has quit IRC | 15:16 | |
| *** cbartz has quit IRC | 15:17 | |
| *** david-lyle has quit IRC | 15:20 | |
| *** cbartz has joined #openstack-swift | 15:20 | |
| *** zaitcev has joined #openstack-swift | 15:20 | |
| *** ChanServ sets mode: +v zaitcev | 15:21 | |
| *** cbartz has quit IRC | 15:26 | |
| *** cbartz has joined #openstack-swift | 15:30 | |
| *** ChanServ sets mode: +v tdasilva | 15:32 | |
| *** cbartz has quit IRC | 15:41 | |
| *** cbartz has joined #openstack-swift | 15:56 | |
| *** _JZ_ has joined #openstack-swift | 16:00 | |
| *** Jeffrey4l has quit IRC | 16:02 | |
| *** chsc has joined #openstack-swift | 16:07 | |
| *** chsc has joined #openstack-swift | 16:07 | |
| *** rcernin has quit IRC | 16:08 | |
| *** SkyRocknRoll has joined #openstack-swift | 16:13 | |
| *** oshritf has quit IRC | 16:14 | |
| *** Jeffrey4l has joined #openstack-swift | 16:15 | |
| *** david-lyle has joined #openstack-swift | 16:17 | |
| *** david-lyle has quit IRC | 16:22 | |
| *** chsc has quit IRC | 16:23 | |
| notmyname | good morning | 16:40 |
| *** geaaru has joined #openstack-swift | 16:41 | |
| *** dmorita has joined #openstack-swift | 16:43 | |
| tdasilva | notmyname: o/ | 16:58 |
| *** NM has joined #openstack-swift | 17:03 | |
| *** mgagne_ has quit IRC | 17:05 | |
| *** mgagne_ has joined #openstack-swift | 17:05 | |
| clayg | kota_: you've been busy! | 17:07 |
| NM | Hi everyone! Good afternoon :) One of my proxy server is getting "STDERR: ERROR:root:Timeout getting a connection to memcached". I did some tcpdump and read some code lines and I think the problem is to use a conection from the pool. I don't see any SYN trying to open a new one. Last time we had this it was solved by a restart but I wondering if you guys think it's worth to make a deeper analysis. | 17:08 |
| *** dosaboy_ is now known as dosaboy | 17:09 | |
| *** tesseract has quit IRC | 17:10 | |
| notmyname | NM: yeah, that's one of those things that might be worth checking in to. could be as simple as raising the memcache_max_connections. could be something deeper | 17:11 |
| NM | notmyname: Thanks. I'll start increasing memcache_max_connections and see if it happens again. | 17:15 |
| timburke | good morning | 17:15 |
| *** david-lyle has joined #openstack-swift | 17:18 | |
| notmyname | NM: what's it set to now? | 17:19 |
| *** delattec has joined #openstack-swift | 17:19 | |
| notmyname | default is 2 | 17:19 |
| patchbot | Error: 'supybot.is' is not a valid configuration variable. | 17:19 |
| *** chsc has joined #openstack-swift | 17:20 | |
| *** chsc has joined #openstack-swift | 17:20 | |
| *** cdelatte has quit IRC | 17:21 | |
| acoles | timburke: morning | 17:21 |
| *** david-lyle has quit IRC | 17:24 | |
| *** vinsh_ has quit IRC | 17:25 | |
| *** rcernin has joined #openstack-swift | 17:27 | |
| *** hseipp has quit IRC | 17:27 | |
| *** david-lyle has joined #openstack-swift | 17:30 | |
| NM | notmyname: we're using the default (2) and about 24 workers. | 17:36 |
| clayg | how many memcache servers - do all of the proxies have all the memcache servers configured? single region? | 17:46 |
| clayg | NM: ^ | 17:46 |
| *** mvk has quit IRC | 17:49 | |
| NM | clayg: two servers. | 17:51 |
| *** cbartz has quit IRC | 17:58 | |
| *** delatte has joined #openstack-swift | 17:58 | |
| *** dmorita has quit IRC | 17:59 | |
| *** delattec has quit IRC | 18:00 | |
| *** dmorita has joined #openstack-swift | 18:04 | |
| *** rledisez has quit IRC | 18:05 | |
| *** rledisez has joined #openstack-swift | 18:05 | |
| *** klrmn has joined #openstack-swift | 18:06 | |
| *** chlong__ has quit IRC | 18:08 | |
| *** dmorita has quit IRC | 18:08 | |
| *** winggundamth_ has joined #openstack-swift | 18:15 | |
| *** winggundamth has quit IRC | 18:18 | |
| *** joeljwright has joined #openstack-swift | 18:23 | |
| *** ChanServ sets mode: +v joeljwright | 18:23 | |
| joeljwright | timburke: you about? | 18:29 |
| *** oshritf has joined #openstack-swift | 18:29 | |
| timburke | joeljwright: yep! | 18:29 |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: refactor some common code from crypto https://review.openstack.org/439058 | 18:30 |
| joeljwright | timburke: hello! | 18:30 |
| joeljwright | just wanted a quick chat about one of your comments on https://review.openstack.org/439058365371 | 18:30 |
| patchbot | patch 439058365371 | 18:30 |
| joeljwright | agh | 18:30 |
| joeljwright | okay, so I can't edit | 18:30 |
| *** mvk has joined #openstack-swift | 18:31 | |
| timburke | yeah! the pre/post-amble stuff | 18:31 |
| joeljwright | https://review.openstack.org/#/c/365371/ | 18:31 |
| patchbot | patch 365371 - swift - Add Preamble and Postamble to SLO and SegmentedIte... | 18:31 |
| joeljwright | it's just your comments about always calculating the slo etag/content length on GET/HEAD | 18:31 |
| joeljwright | seems strange not to just trust sysmeta if we have it… | 18:32 |
| joeljwright | but you seem to disagree | 18:32 |
| joeljwright | (otherwise we're descending through the SLO every time we get a HEAD request…) | 18:32 |
| timburke | yeah, we definitely want to trust the sysmeta if we see it... | 18:33 |
| timburke | i'm trying to remember what i said and what i was thinking... | 18:33 |
| joeljwright | :) | 18:34 |
| joeljwright | comments on patch 19 | 18:34 |
| timburke | oh, right! so, when we're doing an actual GET (so we've got the manifest loaded up *anyway*), where do we decode the pre/post-amble data? i was thinking it was in that loop... | 18:35 |
| timburke | but i also haven't actually func tested to see whether there's bad behavior, so... take everything i say with a grain of salt | 18:35 |
| joeljwright | just checking something... | 18:38 |
| timburke | the HEAD-with-sysmeta case should already be handled up in handle_slo_get_or_head... | 18:39 |
| *** spotz is now known as spotz_zzz | 18:42 | |
| joeljwright | yeah, that's true | 18:42 |
| joeljwright | but we're grabbing it out of the headers again because it's not passed as explicit args | 18:42 |
| *** spotz_zzz is now known as spotz | 18:43 | |
| timburke | yep. i think i was trying to minimize the diff a bit; probably should have worked a bit harder at not duplicating work... | 18:43 |
| joeljwright | I'm pretty sure it's okay | 18:44 |
| joeljwright | (TM) | 18:44 |
| *** winggundamth_ has quit IRC | 18:45 | |
| joeljwright | okay, I'll leave that comment for now | 18:45 |
| joeljwright | everything else you raised is addressed I think | 18:45 |
| joeljwright | there is a qn of func tests | 18:45 |
| timburke | thanks! i'll try to play around with it this week and get more comments posted | 18:46 |
| joeljwright | cheers | 18:46 |
| timburke | yeah, func tests are good...i can help write them as i go, i think | 18:46 |
| joeljwright | at least the unit tests now cover all the obvious cases | 18:47 |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Test EC chunk_transformer with larger input chunks https://review.openstack.org/442108 | 18:47 |
| *** dmorita has joined #openstack-swift | 18:48 | |
| joeljwright | thanks for taking a look Tim | 18:49 |
| *** dmorita_ has joined #openstack-swift | 18:50 | |
| *** dmorita__ has joined #openstack-swift | 18:51 | |
| *** dmorita has quit IRC | 18:52 | |
| *** joeljwright has quit IRC | 18:52 | |
| *** dmorita_ has quit IRC | 18:54 | |
| *** geaaru has quit IRC | 18:56 | |
| *** dmorita__ has quit IRC | 18:56 | |
| *** dmorita has joined #openstack-swift | 19:00 | |
| *** dmorita has quit IRC | 19:02 | |
| *** dmorita_ has joined #openstack-swift | 19:02 | |
| *** dmorita_ has quit IRC | 19:02 | |
| *** dmorita has joined #openstack-swift | 19:03 | |
| *** pcaruana has quit IRC | 19:04 | |
| *** dmorita has quit IRC | 19:04 | |
| *** dmorita_ has joined #openstack-swift | 19:05 | |
| *** dmorita has joined #openstack-swift | 19:06 | |
| *** dmorita_ has quit IRC | 19:06 | |
| *** NM has quit IRC | 19:08 | |
| *** dmorita has quit IRC | 19:10 | |
| *** dmorita has joined #openstack-swift | 19:11 | |
| *** NM has joined #openstack-swift | 19:13 | |
| *** NM has quit IRC | 19:15 | |
| *** pcaruana has joined #openstack-swift | 19:15 | |
| *** NM has joined #openstack-swift | 19:15 | |
| *** dmorita has quit IRC | 19:16 | |
| *** joeljwright has joined #openstack-swift | 19:19 | |
| *** ChanServ sets mode: +v joeljwright | 19:19 | |
| *** dmorita has joined #openstack-swift | 19:20 | |
| *** dmorita has quit IRC | 19:24 | |
| *** NM has quit IRC | 19:26 | |
| *** chlong has joined #openstack-swift | 19:31 | |
| *** oshritf has quit IRC | 19:36 | |
| *** tonanhngo has joined #openstack-swift | 19:36 | |
| *** tonanhngo_ has quit IRC | 19:39 | |
| *** ChubYann has joined #openstack-swift | 19:45 | |
| *** dmorita has joined #openstack-swift | 19:47 | |
| *** dmorita_ has joined #openstack-swift | 19:48 | |
| *** dmorita_ has quit IRC | 19:50 | |
| *** dmorita__ has joined #openstack-swift | 19:50 | |
| *** dmorita has quit IRC | 19:51 | |
| *** dmorita__ has quit IRC | 19:51 | |
| *** dmorita has joined #openstack-swift | 19:51 | |
| *** dmorita has quit IRC | 19:55 | |
| *** dmorita has joined #openstack-swift | 20:06 | |
| *** dmorita has quit IRC | 20:07 | |
| *** dmorita_ has joined #openstack-swift | 20:07 | |
| *** dmorita_ has quit IRC | 20:12 | |
| *** dmorita has joined #openstack-swift | 20:23 | |
| *** vills has joined #openstack-swift | 20:26 | |
| *** dmorita has quit IRC | 20:27 | |
| *** amoralej is now known as amoralej|off | 20:39 | |
| *** chlong has quit IRC | 20:48 | |
| *** jamielennox|away is now known as jamielennox | 21:00 | |
| *** tonanhngo has quit IRC | 21:04 | |
| *** pcaruana has quit IRC | 21:13 | |
| -openstackstatus- NOTICE: restarting gerrit to address performance problems | 21:17 | |
| *** tonanhngo has joined #openstack-swift | 21:28 | |
| *** tonanhngo has quit IRC | 21:33 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| *** Jeffrey4l has joined #openstack-swift | 21:35 | |
| *** catintheroof has quit IRC | 21:44 | |
| *** catintheroof has joined #openstack-swift | 21:45 | |
| *** catintheroof has quit IRC | 21:49 | |
| *** dmorita has joined #openstack-swift | 21:51 | |
| mattoliverau | morning | 21:53 |
| *** dmorita has quit IRC | 21:55 | |
| jrichli | timburke: I would opt for keeping "orig_req_method" always being the true original from client, and having it set any time it is involved in a request leading to internal requests. | 22:06 |
| jrichli | If I am understanding correctly, that pretty much means handling this inside of logging - like in #3 | 22:06 |
| clayg | jrichli: sure, but not subrequests - if I have a COPY which I'm implementing internally as a PUT - but first I need to make a HEAD for... reasons... that HEAD shouldn't get logged as a COPY - subrquests getting their methods changed makes no sense to me? | 22:07 |
| *** dmorita has joined #openstack-swift | 22:08 | |
| jrichli | clayg: true in logging. i am voting for logging handling it : but the value of orig_req_method meaning the client request that originated the activity. i thought that was the meaning of the header, but i could be wrong | 22:09 |
| jrichli | perhaps logging wouldnt have enough info to handle it all on its own though ... | 22:09 |
| clayg | jrichli: I think orig_req_method is in fact there mostly for logging - and I don't think we need to carry forward the original request verb through subrequests when we expect them to have different verbs and swift.source set | 22:10 |
| jrichli | i guess we cant change logging to log both orig and actual? | 22:10 |
| *** tonanhngo has joined #openstack-swift | 22:11 | |
| jrichli | i can imagine some middleware can sometimes uses orig_req_method to make different decisions. | 22:11 |
| jrichli | i know it has come up in discussions about having multiple things at once like symlinks, copy, crypto | 22:11 |
| jrichli | and maybe container-sync. i dont know a specific example now, but i will try to bring these cases to mind | 22:12 |
| timburke | clayg: so...none of the above, and go fix wsgi.py to drop orig_req_method if source is set? around https://github.com/openstack/swift/blob/2.13.0/swift/common/wsgi.py#L1115 | 22:12 |
| *** dmorita has quit IRC | 22:12 | |
| clayg | yeah... that's obviously what we want yeah? | 22:13 |
| clayg | maybe it's more subtle than I realize | 22:13 |
| jrichli | so, source is set if its an internal request, right? | 22:14 |
| clayg | I think we should log the request the client sent - I think we mostly do | 22:15 |
| clayg | homonym requests are annoying (PUT vs. X-Copy-From) | 22:15 |
| clayg | when we flat out transform the client request is annoying (post-as-copy) | 22:15 |
| *** tonanhngo has quit IRC | 22:15 | |
| timburke | or just straight-up COPY | 22:15 |
| clayg | sorry COPY vs. X-Copy-From - homonym requests are annoying! :P | 22:16 |
| clayg | then there's also like... idk staticweb sees a HEAD request makes 3 GET requests and makes up some 400ish response. | 22:17 |
| timburke | how good are we about setting swift.source during subrequests? | 22:17 |
| clayg | timburke: pretty good afaik - and anywhere we don't it's a bug | 22:18 |
| jrichli | i remembered one possible middleware later might be notifications - where clients subscribe to a particular request action. i like having both orig and actual available to middleware for choices there. | 22:18 |
| timburke | how good do we expect third-party middlewares to be about it? | 22:18 |
| clayg | i still don't see the point of setting orig_req_method on subrequests tho - that's not what it's there for - it's not like a swift.source.more_info['orig_req_method'] - it was there to fix post_as_copy - pretty sure | 22:19 |
| jrichli | oh - i like the swift.source.more_info idea ;-) | 22:20 |
| jrichli | yes, maybe i was hoping it could serve as more | 22:20 |
| clayg | jrichli: I think it'd be weird/dangerous to treat a GET from an internal middleware on equal footing as a client GET | 22:21 |
| *** dmorita has joined #openstack-swift | 22:22 | |
| *** adriant has joined #openstack-swift | 22:22 | |
| *** dmorita has quit IRC | 22:23 | |
| *** dmorita has joined #openstack-swift | 22:23 | |
| jrichli | that sounds like an argument to set orig_req_method to the client request and never set it to something else internally | 22:23 |
| clayg | jrichli: do you have a use-case where some middleware is looking at a subrequest - and needs more details about the original request that fired it off? | 22:24 |
| *** dmorita has quit IRC | 22:24 | |
| *** dmorita_ has joined #openstack-swift | 22:24 | |
| clayg | jrichli: we already *do* set orig_req_method to the client request and never set it to something else internally | 22:25 |
| *** dmorita_ has quit IRC | 22:25 | |
| *** dmorita has joined #openstack-swift | 22:26 | |
| jrichli | i thought that some of the options timburke had listed would be setting orig_req_method internally. that is why i was voting against those options. | 22:26 |
| notmyname | jrichli: we can update logging to add stuff. tack it on at the end of the log line and update docs. just can't take stuff out or change the order of stuff | 22:26 |
| *** klrmn has quit IRC | 22:26 | |
| notmyname | (i mean, we shouldn't do it often, but we've been fine with that in the past) | 22:26 |
| *** klrmn has joined #openstack-swift | 22:26 | |
| timburke | :-/ SegmentedIterable uses make_subrequest without setting a source...but that's not obviously the wrong thing to do? | 22:28 |
| timburke | versioned_writes uses get_container_info without a source and account_quotas use get_account_info without a source, but those at least fall back to *something* | 22:28 |
| *** dmorita_ has joined #openstack-swift | 22:28 | |
| timburke | and *no one* seems to use get_object_info? why do we have that? | 22:28 |
| jrichli | notmyname: ok, thanks. so could we not add orig_req_method to the logging at the end? I would assume that would help this situation where we dont know which request to put in the log. just log both. | 22:29 |
| clayg | probably david hadas | 22:29 |
| clayg | jrichli: I don't think the bug says "we don't know" - it says "we do know, and it's wrong" | 22:29 |
| *** dmorita__ has joined #openstack-swift | 22:30 | |
| *** dmorita has quit IRC | 22:30 | |
| jrichli | i read it as it doesn't match between proxy and obj server. they are both picking one, and they pick a different one. | 22:31 |
| jrichli | if they both log both, they match. | 22:31 |
| jrichli | or if proxy logs both, and obj server logs one, that one will match | 22:32 |
| clayg | jrichli: we could have all requests log all verbs - they they'd always be correct | 22:32 |
| *** dmorita_ has quit IRC | 22:33 | |
| jrichli | i think saying original and actual/internal request is still informative (beyond listing all of them) | 22:33 |
| notmyname | which bug is this? | 22:34 |
| *** dmorita__ has quit IRC | 22:34 | |
| timburke | notmyname: https://bugs.launchpad.net/swift/+bug/1657246 | 22:35 |
| openstack | Launchpad bug 1657246 in OpenStack Object Storage (swift) "Proxy logs wrong request method when validating SLO segments" [Medium,Confirmed] | 22:35 |
| notmyname | thanks | 22:35 |
| clayg | jrichli: I agree with you; I think the subtle conflict is over the what's original/actual on subrequests vs. homonym or transformed requests | 22:37 |
| clayg | jrichli: I'm trying to test a versioned stack DELETE now - where I should see a GET/PUT/DELETE all get logged - if I set "orig_request_method" to DELETE on the way in i'm pretty sure it'll look equally wonky | 22:38 |
| jrichli | if you have both logged, it tells you the system is doing a GET because of the original delete request. it is wonky. | 22:39 |
| timburke | clayg: and you shouldn't have to; copy will do it for you, *all the time*. therein lies the bug | 22:39 |
| timburke | so i think i'm partial toward a combination of http://paste.openstack.org/show/601652/ and https://gist.github.com/clayg/7e34254b9da22713d46f112f178119be -- assuming that copy uses the original user request to do the PUT half. if it's a subrequest, we might need to settle for just the first one | 22:40 |
| clayg | lol? Mar 6 17:40:53 saio proxy-server: - - 06/Mar/2017/22/40/53 DELETE /v1/AUTH_test HTTP/1.0 204 - Swift - - - - txea6aab16789541eb8ba86-0058bde575 - 0.0080 GET_ACCOUNT_INFO - | 22:41 |
| clayg | 1488840053.384970903 1488840053.392931938 - | 22:41 |
| clayg | DELETE/GET that account info! | 22:42 |
| timburke | hahaha | 22:42 |
| clayg | timburke: I love it! | 22:43 |
| clayg | jrichli: we'll have to add "annotate subrequests with all the details of the orignal request that started them" as a seperate thing on purpose | 22:43 |
| jrichli | clayg: ok. just the action would be a good start. | 22:44 |
| timburke | i'm not entirely sure how much we need that though -- if you're looking at a proxy log and see some subrequest, and you want more context, why not just grab the transaction id and look for the line that *doesn't* have a source? | 22:46 |
| jrichli | timburke: i was just thinking of middleware. like it came up when discussing notifications. but middleware isn't going to parse the logs. | 22:48 |
| clayg | timburke: we definately don't need it - I don't think a sub requests *need* to reflect anything about the original request - and even if they "could" in logs - they should not do so by mangling the VERB line in the log message | 22:54 |
| clayg | i'm surprised it took us this long to notice :'( | 22:55 |
| *** joeljwright has quit IRC | 22:55 | |
| *** sams-gle_ has quit IRC | 22:58 | |
| *** dmorita has joined #openstack-swift | 23:03 | |
| *** dmorita has quit IRC | 23:08 | |
| timburke | clayg: well, it only really affects quotas, xlo, and versioned_writes -- anything left of copy should behave like it always has | 23:10 |
| clayg | ah, interesting... | 23:12 |
| *** dmorita has joined #openstack-swift | 23:13 | |
| *** dmorita has quit IRC | 23:16 | |
| *** dmorita has joined #openstack-swift | 23:17 | |
| *** joeljwright has joined #openstack-swift | 23:18 | |
| *** ChanServ sets mode: +v joeljwright | 23:18 | |
| *** rcernin has quit IRC | 23:19 | |
| *** dmorita has quit IRC | 23:22 | |
| *** dmorita has joined #openstack-swift | 23:23 | |
| *** dmorita_ has joined #openstack-swift | 23:25 | |
| *** dmorita has quit IRC | 23:25 | |
| *** dmorita has joined #openstack-swift | 23:27 | |
| *** kei_yama has joined #openstack-swift | 23:27 | |
| *** dmorita_ has quit IRC | 23:29 | |
| *** dmorita has quit IRC | 23:32 | |
| *** klamath has quit IRC | 23:34 | |
| *** tonanhngo has joined #openstack-swift | 23:35 | |
| *** chsc has quit IRC | 23:38 | |
| *** tonanhngo has quit IRC | 23:40 | |
| *** joeljwright has quit IRC | 23:41 | |
| *** dmorita has joined #openstack-swift | 23:42 | |
| *** dmorita has quit IRC | 23:46 | |
| *** dmorita has joined #openstack-swift | 23:46 | |
| *** dmorita has quit IRC | 23:51 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Make Sphinx treat warnings as errors https://review.openstack.org/442208 | 23:56 |
| *** zaitcev has quit IRC | 23:56 | |
| *** sams-gleb has joined #openstack-swift | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!