| *** tonanhngo has quit IRC | 00:18 | |
| *** tonanhngo has joined #openstack-swift | 00:18 | |
| *** tonanhngo has quit IRC | 00:18 | |
| *** tovin07_ has joined #openstack-swift | 00:50 | |
| clayg | rledisez: I think it would be really helpful to me if I could see your handoff/sync_only and override_devices plumbing... you just did everything from the commandline right? you only have a single .conf for all reconstructors? | 00:55 |
|---|---|---|
| *** pdardeau has joined #openstack-swift | 00:58 | |
| kota_ | good morning | 01:01 |
| kota_ | clayg: patch 459023 ? | 01:02 |
| patchbot | https://review.openstack.org/#/c/459023/ - liberasurecode - Consistently use zlib for crc32 | 01:02 |
| clayg | do you have any idea about that change? tdasilva was asking and timburke is out? | 01:03 |
| kota_ | looking | 01:03 |
| clayg | timburke says "it's really bad but it should be impossible to cause a problem" "tdasilva says - yeah i think this would break me" | 01:04 |
| kota_ | i'm not entierly sure but the second topic on the commit message seemed to effect timburke's activity IIRC | 01:05 |
| kota_ | saying "load timing of zlib" | 01:05 |
| kota_ | and the order of load timing (strictly which lib is used, zlib or hard corded in the liberasurecode) affects to the result of crc checksum, maybe? | 01:07 |
| kota_ | and he hit the different crc with same data in... maybe his mac os | 01:07 |
| * kota_ is moving to bug report description to understand what is the bad problem | 01:09 | |
| * kota_ is on https://bugs.launchpad.net/liberasurecode/+bug/1666320 | 01:09 | |
| openstack | Launchpad bug 1666320 in liberasurecode "crc32 implementation is schizophrenic" [Undecided,In progress] - Assigned to Tim Burke (1-tim-z) | 01:09 |
| kota_ | probably the bad case is as Bad news: Despite comments declaring that these are using the same polynomials [3], they're not! The first implements CRC32C [4], while the table-based implementation seems to be based off of zlib [5]. | 01:10 |
| openstackgerrit | liuyamin proposed openstack/swift master: Fix the reST field raises in docstrings https://review.openstack.org/451143 | 01:11 |
| *** itlinux has joined #openstack-swift | 01:13 | |
| tdasilva | kota_: clayg I mean, i'm not even sure i was testing things correctly really, but basically I put an object into a ec storage policy and then pointed timburke's script at one of the fragments and got a mismatch | 01:14 |
| tdasilva | from what I understood, timburke expected that never to happen | 01:14 |
| clayg | yeah same here - it WOMM then I tried it on another node after you called shenanigans | 01:15 |
| kota_ | tdasilva: check_crc.py in lp? | 01:15 |
| clayg | yeah, so we disproved that - so hopefully he can come back with some better plan? | 01:15 |
| tdasilva | kota_: yeah | 01:15 |
| clayg | kota_: https://launchpadlibrarian.net/307221898/check_crc.py | 01:15 |
| kota_ | alright, I will look at the patch with priority | 01:16 |
| clayg | ok, so that settles that - it's every bit as scary as it looks | 01:16 |
| clayg | kota_: I don't think that patch can merge as is - let's wait for more info from timburke next week | 01:16 |
| clayg | tdasilva: you're wanting to do a pyeclib/libec release - yeah? | 01:16 |
| tdasilva | clayg: kota_ mean while, i'd like to do a release of libec, but maybe we can wait until next week when timburke is back | 01:16 |
| clayg | tdasilva: don't hold up for this patch - it's not ready | 01:16 |
| tdasilva | clayg: yeah! | 01:17 |
| kota_ | i think, what I can tell right now is we should add zlib in dindep.txt if we want to use it always. | 01:17 |
| kota_ | timburke: ok | 01:17 |
| tdasilva | kota_: but i did check that my system has zlib installed, so I think it's not a matter of zlib being on the system, but whether it is getting loaded or not??? | 01:18 |
| clayg | kota_: timburke was trying to explain that somehow that if you start not-using zlib then it dynamically changes to start using zlib all hell breaks loose? It's possible the nodes tdasilva and I found where check_crc returns "mismatch" are at dire risk of one misplaced package install quarantineing the entire datadir | 01:19 |
| clayg | tdasilva: good point - i'm getting the mismatch on a node with zlib and zlib-dev installed | 01:19 |
| clayg | tdasilva: so maybe it *is* somehow build time? i wonder why timburke says it's run time... | 01:20 |
| tdasilva | maybe zaitcev has some clues on this??? ^^^ | 01:21 |
| kota_ | clayg: IIRC, we could have a chance to create the situation in running time | 01:21 |
| kota_ | maybe... pre-loading for .so files... | 01:21 |
| tdasilva | meanwhile i'm going to update https://review.openstack.org/#/c/471084/ and hopefully we can get that merged ?? | 01:21 |
| patchbot | patch 471084 - liberasurecode - Release 1.5.0 | 01:21 |
| kota_ | I forget how to set but basically it can be by sort of environment variable | 01:22 |
| kota_ | it looks like https://stackoverflow.com/questions/426230/what-is-the-ld-preload-trick | 01:22 |
| clayg | oh right, not if it's installed, more like when it's loaded | 01:22 |
| kota_ | something like `LD_PRELOAD=/path/to/my/zlib.so ` or liberasurecode or both... | 01:23 |
| kota_ | I'll try it | 01:23 |
| *** JimCheung has quit IRC | 01:24 | |
| *** JimCheung has joined #openstack-swift | 01:25 | |
| clayg | LD_PRELOAD=/usr/local/lib/liberasurecode.so.1.4.0 ./check_crc.py /srv/node3/sdb3/objects-1/589/c2b/93464008f434853ad9b74c8c5381ac2b/1498598728.24878#2#d.data | 01:25 |
| clayg | ^ says mismatch on a node that previously said OK | 01:25 |
| kota_ | oh yeah | 01:26 |
| clayg | but that's bullshit... shouldn't the check script not be effected by runtime bullshit in the same way!? | 01:26 |
| clayg | like it doesn't even load pyeclib? | 01:26 |
| kota_ | i think, if NOT preload liberasurecode will use zlib because python requires zlib.so loaded | 01:26 |
| clayg | but where does check_crc load liberasurecode? some throwing the damn .so into the mix replace something in zlib's brain!? | 01:28 |
| * kota_ is looking at check_crc.py | 01:28 | |
| *** cshastri has joined #openstack-swift | 01:29 | |
| kota_ | ah, it doesn't use liberasurecode, ok. | 01:29 |
| * kota_ is thinking | 01:29 | |
| *** JimCheung has quit IRC | 01:29 | |
| kota_ | hmm... no idea, right now. need deep dive into the change and the problem | 01:32 |
| *** davidchu has joined #openstack-swift | 01:36 | |
| *** davidchu has left #openstack-swift | 01:37 | |
| clayg | i hate that bug report | 01:44 |
| kota_ | :/ | 01:44 |
| *** gyee has quit IRC | 01:53 | |
| kota_ | LD_PRELOAD="/usr/local/lib/liberasurecode.so /usr/lib/x86_64-linux-gnu/libz.so" python check_crc.py 1478688450.50122#2#d.data -> mismatch | 02:01 |
| *** itlinux has quit IRC | 02:01 | |
| kota_ | LD_PRELOAD="/usr/lib/x86_64-linux-gnu/libz.so /usr/local/lib/liberasurecode.so" python check_crc.py 1478688450.50122#2#d.data -> match | 02:01 |
| kota_ | so it looks like libz should be loaded before liberasurecode.so | 02:02 |
| * kota_ is going to dive into liberasurecode liking | 02:02 | |
| kota_ | linking | 02:02 |
| *** mat128 has quit IRC | 02:27 | |
| kota_ | clayg: it looks like symbol search problem, if i understand correctly | 02:31 |
| clayg | symbol search problem!? Can we use ML to train it to better recognize the symbols? | 02:32 |
| kota_ | clayg: maybe zlib and liberasurecode have conflict function name as crc32 and then, when calling crc3 | 02:32 |
| clayg | symbolism is hard for computers IME | 02:32 |
| kota_ | crc32 in liberasurecod.so is used | 02:33 |
| kota_ | LD_DEBUG=symbols,bindings python check_crc.py 1478688450.50122#2#d.data | 02:33 |
| clayg | so liberasurecode monkey patches zlib - LAME | 02:33 |
| kota_ | this find zlib crc32 symbol | 02:33 |
| kota_ | but | 02:33 |
| kota_ | LD_PRELOAD="/usr/local/lib/liberasurecode.so" LD_DEBUG=symbols,bindings python check_crc.py 1478688450.50122#2#d.data | 02:33 |
| kota_ | shows only liberasurecode's crc32 | 02:34 |
| *** htruta` has quit IRC | 02:34 | |
| *** raginbajin has quit IRC | 02:34 | |
| * kota_ is trying to rename the liberasurecode's crc32... | 02:34 | |
| *** htruta has joined #openstack-swift | 02:35 | |
| kota_ | clayg: https://gist.github.com/bloodeagle40234/011251d8377d6a53c6dffab203400b55 | 02:39 |
| *** aselius has quit IRC | 02:40 | |
| *** tonyb_ is now known as tonyb | 02:40 | |
| kota_ | clayg: it's interesting. that small change save the preloaded liberasurecode problem. | 02:40 |
| *** raginbajin has joined #openstack-swift | 02:40 | |
| *** lucasxu has joined #openstack-swift | 02:40 | |
| * kota_ means LD_PRELOAD="/usr/local/lib/liberasurecode.so" LD_DEBUG=symbols,bindings python check_crc.py 1478688450.50122#2#d.data becomes OK | 02:40 | |
| clayg | in that it always uses the zlib crc32 implementation - which I think is what timburkes change accomplishes | 02:41 |
| kota_ | yeah | 02:41 |
| clayg | but... if we ship that - and new libec encounters some frag written down the crc32_l it will check against zlib crc32 and conclude things don't match | 02:42 |
| kota_ | good point | 02:42 |
| kota_ | if the libec's crc32 was used in some ec fragments, they could be handled as broken :/ | 02:43 |
| kota_ | that sounds problematci | 02:44 |
| kota_ | tic | 02:44 |
| kota_ | but... no way to find the used lib for crc32? | 02:45 |
| kota_ | a couple of ways to handle the case... | 02:48 |
| kota_ | 1. ignore, if mismatch crc found but the frags created by older libec version | 02:49 |
| kota_ | 2. keepking current crc32 imple with renaming and then, try both impl against to check crc | 02:50 |
| kota_ | against to a fragment | 02:50 |
| clayg | I'd love to compare a hexdump of my bad frags with a hexdump from tdasilva's bad frags | 02:50 |
| clayg | I can't find anywhere in this metadata where the crc32_l checksum is written down - and yet i'm not quarantining frags left and right - so I don't really know what's going on? | 02:51 |
| clayg | isn't there a pyeclib thing to get the metadata - does it check the checksum? | 02:52 |
| kota_ | clayg: in usual case, python requires zlib at first so the crc32 symbol should be find at zlib IIRC | 02:52 |
| kota_ | if you're using pyeclib | 02:52 |
| kota_ | I'm not sure if we can find the case liberasurecode loaded first | 02:53 |
| clayg | yeah except I have a frag on a node in my lab - and it definitely doesn't have the zlib crc written down where the check_crc script is expecting it | 02:53 |
| kota_ | timburke may know though | 02:53 |
| clayg | timburke said nope, and me and tdasilva both found existence proofs | 02:53 |
| kota_ | hmm... interesting | 02:55 |
| * kota_ goes to lunch | 02:55 | |
| *** lucasxu has quit IRC | 02:56 | |
| *** JimCheung has joined #openstack-swift | 02:57 | |
| *** lucasxu has joined #openstack-swift | 03:07 | |
| *** gkadam has joined #openstack-swift | 03:18 | |
| mattoliverau | Hey all, fyi, I'm currently driving between beach house and Melbourne.. about 1/3 - 1/2 way.. don't think I'll make the 007 meeting, but I I'll try :) | 03:19 |
| kota_ | mattoliverau: enjoy! | 03:19 |
| mattoliverau | (noting it's an 8 - 9 hour drive without breaks) | 03:20 |
| kota_ | wow | 03:20 |
| clayg | is that tonite? I wonder if I'll be up... | 03:28 |
| *** JimCheung has quit IRC | 03:33 | |
| *** itlinux has joined #openstack-swift | 03:38 | |
| *** noark9 has joined #openstack-swift | 03:45 | |
| *** links has joined #openstack-swift | 04:00 | |
| *** lucasxu has quit IRC | 04:06 | |
| *** JimCheung has joined #openstack-swift | 04:07 | |
| *** psachin has joined #openstack-swift | 04:09 | |
| *** klrmn has quit IRC | 04:09 | |
| *** JimCheung has quit IRC | 04:11 | |
| *** qwertyco has joined #openstack-swift | 04:47 | |
| *** itlinux has quit IRC | 05:15 | |
| *** ChubYann has quit IRC | 05:42 | |
| *** pdardeau has quit IRC | 05:49 | |
| *** pdardeau has joined #openstack-swift | 05:59 | |
| mahatic_ | clayg: yup, in an hour from now | 06:00 |
| *** rcernin has joined #openstack-swift | 06:02 | |
| *** cschwede has joined #openstack-swift | 06:12 | |
| *** ChanServ sets mode: +v cschwede | 06:12 | |
| *** skudlik has joined #openstack-swift | 06:15 | |
| openstackgerrit | Van Hung Pham proposed openstack/swift master: Switch from oslosphinx to openstackdocstheme https://review.openstack.org/478402 | 06:39 |
| *** hseipp has joined #openstack-swift | 06:52 | |
| cschwede | Hi! Note: Swift meeting starting in 4 minutes in #openstack-swift | 06:56 |
| acoles | good morning | 06:57 |
| mathiasb | good morning! | 07:01 |
| cschwede | hmm, is the meetbot on #openstack-meetings dead? | 07:01 |
| mahatic_ | cschwede: fwiw, I don't see your nick on #openstack-meeting | 07:01 |
| tovin07_ | cschwede, #openstack-meeting not #openstack-meetings | 07:02 |
| tovin07_ | without `s` :D | 07:02 |
| acoles | I'm in #openstack-meeting and see no activity, did we start yet? | 07:03 |
| mahatic_ | acoles: it started now | 07:03 |
| acoles | mahatic_: k :) | 07:03 |
| cschwede | bummer, yes. i have a new bouncer, epic fail :/ | 07:03 |
| *** m_kazuhiro has joined #openstack-swift | 07:04 | |
| *** tesseract has joined #openstack-swift | 07:08 | |
| *** oshritf_ has joined #openstack-swift | 07:41 | |
| rledisez | clayg: here is my reconstructor revert/sync only patch + init script + relevant conf https://gist.github.com/rledisez/f905dcca6be565da04ee12efdae8ad52 | 07:42 |
| *** m_kazuhiro_ has joined #openstack-swift | 07:45 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: WIP: Add multiple worker processes strategy to reconstructor https://review.openstack.org/478416 | 07:47 |
| *** m_kazuhiro has quit IRC | 07:48 | |
| clayg | rledisez: ^ I think this patch could potentially simply some of that then | 07:49 |
| clayg | rledisez: i'd love some feedback before getting too deep into testing | 07:50 |
| clayg | I think after merging https://review.openstack.org/#/c/428408/ | 07:51 |
| patchbot | patch 428408 - swift - Don't rehash primaries in reconstructor handoffs_o... | 07:51 |
| clayg | it will be easier to add the -r and -s options - the sync_group_size and revert_group_size is going to be a little different ... I expect to just set workers = X in the .conf | 07:52 |
| *** gkadam is now known as gkadam-afk | 07:53 | |
| *** cbartz has joined #openstack-swift | 07:59 | |
| acoles | cschwede: thank you for chairing the meeting | 08:00 |
| clayg | cschwede: did you name the next chair? | 08:00 |
| mahatic_ | cschwede: thank you | 08:00 |
| rledisez | cschwede: good job! thx | 08:01 |
| cschwede | no problem, you're welcome! surprised by how fast time went by | 08:01 |
| clayg | 👉 👃 | 08:01 |
| cschwede | clayg: no, i didn't :/ | 08:01 |
| acoles | cschwede: I just realised that clayg precedes you alphabetically so he should have chaired today!! :P | 08:01 |
| cschwede | clayg: are you volunteering? ;) | 08:01 |
| clayg | https://en.wikipedia.org/wiki/Nose_goes | 08:01 |
| acoles | hehee | 08:01 |
| mahatic_ | hehe | 08:03 |
| * acoles away for breakfast | 08:03 | |
| mahatic_ | acoles: enjoy | 08:03 |
| *** m_kazuhiro_ has quit IRC | 08:06 | |
| *** joeljwright has joined #openstack-swift | 08:23 | |
| *** ChanServ sets mode: +v joeljwright | 08:23 | |
| openstackgerrit | Van Hung Pham proposed openstack/swift master: Switch from oslosphinx to openstackdocstheme https://review.openstack.org/478402 | 09:08 |
| *** pcaruana has joined #openstack-swift | 09:28 | |
| acoles | rledisez: I'm just working on some tests for patch 472659 | 09:30 |
| patchbot | https://review.openstack.org/#/c/472659/ - swift - Allow to rebuild a fragment of an expired object | 09:30 |
| rledisez | acoles: thx! you also mentioned a probe test that could be easily modified? | 09:39 |
| *** mvk has quit IRC | 09:46 | |
| *** yiorgos_272 has joined #openstack-swift | 09:52 | |
| acoles | rledisez: yes we can probably adapt this https://review.openstack.org/#/c/460073/ | 09:53 |
| patchbot | patch 460073 - swift - Add probe test for ssync of unexpired metadata to ... | 09:53 |
| openstackgerrit | Van Hung Pham proposed openstack/swift master: Switch from oslosphinx to openstackdocstheme https://review.openstack.org/478402 | 09:53 |
| *** yiorgos_272 has left #openstack-swift | 09:54 | |
| *** rcernin has quit IRC | 09:55 | |
| *** mat128 has joined #openstack-swift | 09:58 | |
| *** gkadam-afk is now known as gkadam | 09:59 | |
| *** tovin07_ has quit IRC | 10:02 | |
| *** rcernin has joined #openstack-swift | 10:03 | |
| *** rcernin has quit IRC | 10:09 | |
| *** rcernin has joined #openstack-swift | 10:11 | |
| *** mvk has joined #openstack-swift | 10:13 | |
| *** joeljwright has quit IRC | 10:13 | |
| *** noark9 has quit IRC | 10:49 | |
| *** qwertyco has quit IRC | 10:51 | |
| *** amit213 has quit IRC | 11:02 | |
| *** pdardeau has quit IRC | 11:05 | |
| *** cshastri has quit IRC | 11:10 | |
| *** pdardeau has joined #openstack-swift | 11:15 | |
| *** pcaruana has quit IRC | 12:02 | |
| *** pcaruana has joined #openstack-swift | 12:02 | |
| *** geaaru has joined #openstack-swift | 12:03 | |
| notmyname | good morning | 12:11 |
| notmyname | cschwede: thanks for running the meeting. I'll read over the logs later | 12:13 |
| cschwede | notmyname: no problem, you're welcome! next time i need to shrink the agenda a bit... run out of time | 12:15 |
| notmyname | heh | 12:15 |
| notmyname | I have *no* idea what you mean ;-) | 12:15 |
| *** chlong_ has quit IRC | 12:15 | |
| cschwede | notmyname: i'm sure that never happened to you ;) | 12:16 |
| *** pcaruana has quit IRC | 12:16 | |
| cschwede | notmyname: biggest topics: i set up an etherpad for queens ptg (https://etherpad.openstack.org/p/swift-ptg-queens) based on our pike etherpad, hope that's ok. also two action items for myself until next week | 12:17 |
| notmyname | wonderful | 12:17 |
| notmyname | thank you | 12:17 |
| *** noark9 has joined #openstack-swift | 12:19 | |
| notmyname | I'm at a "partner training" today and tomorrow in boston. I'm not sure yet how busy it will be (wrt being online and following up on swift things) | 12:19 |
| *** pcaruana has joined #openstack-swift | 12:23 | |
| *** pcaruana has quit IRC | 12:23 | |
| *** pcaruana has joined #openstack-swift | 12:23 | |
| *** pcaruana has quit IRC | 12:29 | |
| *** pcaruana has joined #openstack-swift | 12:37 | |
| *** pcaruana has quit IRC | 12:43 | |
| *** pcaruana has joined #openstack-swift | 12:44 | |
| *** MVenesio has joined #openstack-swift | 12:44 | |
| *** gkadam has quit IRC | 12:51 | |
| *** hseipp has quit IRC | 12:53 | |
| *** pcaruana has quit IRC | 12:53 | |
| *** saint_ has joined #openstack-swift | 13:01 | |
| *** hseipp has joined #openstack-swift | 13:04 | |
| *** pcaruana has joined #openstack-swift | 13:05 | |
| *** saint__ has joined #openstack-swift | 13:12 | |
| *** pcaruana has quit IRC | 13:13 | |
| *** lucasxu has joined #openstack-swift | 13:14 | |
| *** pcaruana has joined #openstack-swift | 13:14 | |
| *** links has quit IRC | 13:16 | |
| *** saint_ has quit IRC | 13:16 | |
| *** saint__ has quit IRC | 13:25 | |
| *** klamath has joined #openstack-swift | 13:26 | |
| *** klamath has quit IRC | 13:26 | |
| *** klamath has joined #openstack-swift | 13:27 | |
| *** saint_ has joined #openstack-swift | 13:28 | |
| *** links has joined #openstack-swift | 13:32 | |
| *** chlong_ has joined #openstack-swift | 13:34 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Allow to rebuild a fragment of an expired object https://review.openstack.org/472659 | 13:34 |
| openstackgerrit | Alistair Coles proposed openstack/swift master: Allow to rebuild a fragment of an expired object https://review.openstack.org/472659 | 13:38 |
| acoles | rledisez: ^^ added unit and probe tests | 13:38 |
| *** zhurong_ has joined #openstack-swift | 13:38 | |
| *** pcaruana has quit IRC | 13:40 | |
| *** pcaruana has joined #openstack-swift | 13:46 | |
| *** links has quit IRC | 13:56 | |
| *** saint_ has quit IRC | 14:03 | |
| *** saint_ has joined #openstack-swift | 14:04 | |
| *** jaosorior has quit IRC | 14:05 | |
| *** jaosorior has joined #openstack-swift | 14:08 | |
| *** ouchkernel has joined #openstack-swift | 14:08 | |
| tdasilva | interesting discussion this morning about bug triaging. One point to make is that maybe we need to change our processes a bit (or maybe I need to change my own process) where we (I) focus more on the patches in gerrit on what TODO and instead we should change focus to bugs (launchpad or future storyboard) | 14:08 |
| tdasilva | For example, we have the Swift Priortity patches on the wiki, but maybe the priority list should be the list of bugs and how they get prioritized there. | 14:09 |
| *** ouchkernel has quit IRC | 14:09 | |
| *** ouchkernel has joined #openstack-swift | 14:10 | |
| *** saint_ has quit IRC | 14:32 | |
| acoles | tdasilva: good point - it's too easy to not see the gaps (bugs that are not being worked on) | 14:48 |
| tdasilva | acoles: yeah, i think if we do a good job of focusing on the bugs we can make the list there be a good point for knowing what to review, what to fix and also the priorities | 14:56 |
| *** vinsh has joined #openstack-swift | 14:56 | |
| *** hseipp has quit IRC | 15:00 | |
| *** hseipp has joined #openstack-swift | 15:00 | |
| *** zhurong_ has quit IRC | 15:03 | |
| *** geaaru has quit IRC | 15:06 | |
| *** rcernin has quit IRC | 15:06 | |
| *** links has joined #openstack-swift | 15:07 | |
| *** klamath has quit IRC | 15:07 | |
| *** gyee has joined #openstack-swift | 15:07 | |
| *** klamath has joined #openstack-swift | 15:07 | |
| *** hseipp has quit IRC | 15:15 | |
| *** hseipp has joined #openstack-swift | 15:16 | |
| *** geaaru has joined #openstack-swift | 15:17 | |
| *** klamath has quit IRC | 15:18 | |
| *** klamath has joined #openstack-swift | 15:19 | |
| *** aselius has joined #openstack-swift | 15:29 | |
| *** ouchkern_ has joined #openstack-swift | 15:29 | |
| *** ouchkernel has quit IRC | 15:30 | |
| *** skudlik has quit IRC | 15:32 | |
| *** ouchkern_ has quit IRC | 15:34 | |
| *** links has quit IRC | 15:36 | |
| *** geaaru has quit IRC | 15:38 | |
| *** vinsh has quit IRC | 15:42 | |
| *** klrmn has joined #openstack-swift | 15:42 | |
| *** ouchkernel has joined #openstack-swift | 15:50 | |
| *** ouchkern_ has joined #openstack-swift | 15:55 | |
| *** ouchkernel has quit IRC | 15:58 | |
| *** Sukhdev_ has joined #openstack-swift | 15:58 | |
| *** ouchkern_ has quit IRC | 15:59 | |
| *** itlinux has joined #openstack-swift | 16:01 | |
| *** klrmn has quit IRC | 16:08 | |
| *** ouchkernel has joined #openstack-swift | 16:09 | |
| *** psachin has quit IRC | 16:10 | |
| *** ouchkern_ has joined #openstack-swift | 16:14 | |
| *** ouchkernel has quit IRC | 16:16 | |
| *** cschwede has quit IRC | 16:21 | |
| *** klamath has quit IRC | 16:24 | |
| *** Sukhdev_ has quit IRC | 16:25 | |
| *** cbartz has quit IRC | 16:25 | |
| *** ouchkernel has joined #openstack-swift | 16:26 | |
| *** ouchkern_ has quit IRC | 16:27 | |
| *** ouchkernel has quit IRC | 16:28 | |
| *** ouchkernel has joined #openstack-swift | 16:28 | |
| *** ouchkern_ has joined #openstack-swift | 16:32 | |
| *** ouchkernel has quit IRC | 16:35 | |
| *** JimCheung has joined #openstack-swift | 16:39 | |
| *** noark9 has quit IRC | 16:49 | |
| *** oshritf_ has quit IRC | 16:50 | |
| clayg | @acoles whoot! | 16:50 |
| clayg | acoles: no new version of patch 477000 today? | 16:50 |
| patchbot | https://review.openstack.org/#/c/477000/ - swift - WIP: Ring rebalance respects co-builders' last_par... | 16:50 |
| clayg | kota_: ^ did you look at this? torgomatic ^ | 16:51 |
| clayg | tdasilva: what's going on with cschwede 's bounder situation? he's never around async anymore? | 16:52 |
| zaitcev | wow, this still exists - http://etherpad.openstack.org/swift3 | 16:55 |
| zaitcev | 2011 | 16:56 |
| acoles | clayg: not yet but there are hours left ;) (you told me to write some tests for rledisez so I did that) | 16:58 |
| *** ouchkern_ has quit IRC | 16:58 | |
| *** ujjain has quit IRC | 16:59 | |
| *** ujjain has joined #openstack-swift | 17:00 | |
| *** ujjain has joined #openstack-swift | 17:00 | |
| *** klrmn has joined #openstack-swift | 17:04 | |
| *** ouchkernel has joined #openstack-swift | 17:07 | |
| *** hseipp has quit IRC | 17:10 | |
| *** ouchkern_ has joined #openstack-swift | 17:11 | |
| *** ouchkernel has quit IRC | 17:13 | |
| tdasilva | clayg: hi, not sure about cschwede's bouncer?? he mentioned this morning he was on a new bouncer, so maybe he is still adjusting that... | 17:18 |
| clayg | I don't really know how to reply to lp bug #1700585 anymore | 17:20 |
| openstack | Launchpad bug 1700585 in OpenStack Object Storage (swift) "Slow replication with lack of visibility can cause availability problems (esp. in multi-region)" [Undecided,Confirmed] https://launchpad.net/bugs/1700585 | 17:20 |
| clayg | I want to say: https://gist.github.com/clayg/e87e2b9c69abbe25f45cc09d570687ec (but only for therapy, I don't think it acctually adds anything to the bug report) | 17:22 |
| clayg | jrichli: any ideas? | 17:22 |
| *** ouchkern_ has quit IRC | 17:22 | |
| *** silor has joined #openstack-swift | 17:23 | |
| *** aselius has left #openstack-swift | 17:23 | |
| *** tonanhngo has joined #openstack-swift | 17:23 | |
| *** ouchkernel has joined #openstack-swift | 17:24 | |
| *** tonanhngo has quit IRC | 17:28 | |
| *** ouchkern_ has joined #openstack-swift | 17:39 | |
| tdasilva | clayg: fwiw, I do like the idea of maybe incorporating the partition handoff counts somewhere (e.g., recon) but as you said, then it really becomes up to operators how to tune their environment based on that data | 17:39 |
| *** ouchkernel has quit IRC | 17:42 | |
| *** tonanhngo has joined #openstack-swift | 17:42 | |
| *** links has joined #openstack-swift | 17:46 | |
| *** tonanhngo has quit IRC | 17:47 | |
| *** tonanhngo has joined #openstack-swift | 17:53 | |
| clayg | doing the thing with get_more_nodes to classify handoff/mis-placed is a fairly new thing - i haven't even plummbed that idea in swiftstack telemetry yet | 17:53 |
| clayg | but it was extremely relevant last time I looked at replication issues on a multi-region cluster! | 17:54 |
| *** klamath has joined #openstack-swift | 17:54 | |
| *** tonanhngo has quit IRC | 17:58 | |
| *** tonanhngo has joined #openstack-swift | 17:59 | |
| *** tonanhngo has quit IRC | 18:03 | |
| *** tonanhngo has joined #openstack-swift | 18:05 | |
| *** tonanhngo has quit IRC | 18:10 | |
| *** Sukhdev_ has joined #openstack-swift | 18:20 | |
| *** mvk has quit IRC | 18:22 | |
| *** links has quit IRC | 18:30 | |
| *** Sukhdev_ has quit IRC | 18:38 | |
| *** ChubYann has joined #openstack-swift | 18:42 | |
| *** chsc has joined #openstack-swift | 18:47 | |
| *** klamath has quit IRC | 18:50 | |
| *** klamath has joined #openstack-swift | 18:52 | |
| *** ouchkern_ has quit IRC | 18:56 | |
| openstackgerrit | Timur Alperovich proposed openstack/python-swiftclient master: Turn stdin uploads in SLO above 10MB. https://review.openstack.org/478611 | 19:05 |
| openstackgerrit | Timur Alperovich proposed openstack/python-swiftclient master: Turn stdin uploads into SLO above 10MB. https://review.openstack.org/478611 | 19:05 |
| *** peterlisak has quit IRC | 19:11 | |
| *** onovy has quit IRC | 19:11 | |
| *** tonanhngo has joined #openstack-swift | 19:12 | |
| *** tesseract has quit IRC | 19:13 | |
| *** tonanhngo has quit IRC | 19:16 | |
| *** Guest28796 is now known as mgagne | 19:19 | |
| *** mgagne has quit IRC | 19:19 | |
| *** mgagne has joined #openstack-swift | 19:19 | |
| *** tonanhngo has joined #openstack-swift | 19:24 | |
| *** tonanhngo has quit IRC | 19:29 | |
| *** tonanhngo has joined #openstack-swift | 19:33 | |
| *** hseipp has joined #openstack-swift | 19:46 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: WIP: Ring rebalance respects co-builders' last_part_moves https://review.openstack.org/477000 | 20:04 |
| acoles | clayg: ^^ still work to do but progress | 20:04 |
| *** peterlisak has joined #openstack-swift | 20:08 | |
| *** onovy has joined #openstack-swift | 20:09 | |
| *** silor has quit IRC | 20:14 | |
| *** lucasxu has quit IRC | 20:30 | |
| *** ouchkernel has joined #openstack-swift | 20:36 | |
| *** JimCheung has quit IRC | 20:43 | |
| *** JimCheung has joined #openstack-swift | 20:43 | |
| *** hseipp has quit IRC | 20:46 | |
| *** MVenesio has quit IRC | 20:48 | |
| *** pcaruana has quit IRC | 21:01 | |
| *** ouchkernel has quit IRC | 21:06 | |
| *** caiobrentano has joined #openstack-swift | 21:17 | |
| clayg | torgomatic: https://bugs.launchpad.net/swift/+bug/1697543 | 21:22 |
| openstack | Launchpad bug 1697543 in OpenStack Object Storage (swift) "Ring refuses to save even when 100% parts move" [Medium,Confirmed] | 21:22 |
| *** klamath has quit IRC | 21:25 | |
| *** klamath has joined #openstack-swift | 21:26 | |
| *** mat128 has quit IRC | 21:26 | |
| *** itlinux has quit IRC | 21:35 | |
| jrichli | clayg: sorry i wasn't online earlier. I haven't followed the exchange, but ill ramp up if you think i could be helpful | 21:53 |
| openstackgerrit | Samuel Merritt proposed openstack/swift master: Improve debug logging from ring builder. https://review.openstack.org/478657 | 21:56 |
| torgomatic | clayg ^^ | 21:56 |
| openstackgerrit | Samuel Merritt proposed openstack/swift master: Improve debug logging from ring builder. https://review.openstack.org/478657 | 22:09 |
| *** klamath has quit IRC | 22:10 | |
| *** klamath has joined #openstack-swift | 22:11 | |
| *** tonanhngo has quit IRC | 22:16 | |
| *** mvk has joined #openstack-swift | 22:17 | |
| *** chlong_ has quit IRC | 22:18 | |
| *** tonanhngo has joined #openstack-swift | 22:21 | |
| *** tonanhngo has quit IRC | 22:25 | |
| *** caiobrentano has quit IRC | 22:28 | |
| *** caiobrentano has joined #openstack-swift | 22:29 | |
| *** caiobrentano has quit IRC | 22:29 | |
| *** klamath has quit IRC | 22:35 | |
| *** JimCheung has quit IRC | 22:37 | |
| *** vint_bra has quit IRC | 22:40 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Extend device tier pretty print normalization https://review.openstack.org/478667 | 22:41 |
| clayg | torgomatic: ^ | 22:41 |
| torgomatic | clayg: LGTM. You want to squash those together? | 22:46 |
| clayg | sure atomic commits are good thing right? | 22:47 |
| torgomatic | beats a dependency chain | 22:47 |
| clayg | who else wants to welcome torgomatic back into channel / apologize he has to look at rings!? | 22:48 |
| timur_ | whoo! torgomatic is back | 22:48 |
| *** timur_ is now known as timur | 22:49 | |
| *** JimCheung has joined #openstack-swift | 22:53 | |
| *** JimCheung has quit IRC | 22:58 | |
| *** ukaynar has joined #openstack-swift | 22:58 | |
| *** klrmn1 has joined #openstack-swift | 23:08 | |
| *** klrmn has quit IRC | 23:08 | |
| *** JimCheung has joined #openstack-swift | 23:09 | |
| *** caiobrentano has joined #openstack-swift | 23:10 | |
| caiobrentano | Hi all! Simple question... Is Swift fully ready for IPv6? | 23:12 |
| clayg | FULLY | 23:13 |
| clayg | except maybe keystone | 23:13 |
| *** JimCheung has quit IRC | 23:13 | |
| clarkb | clayg: keystone should be fine with it too | 23:13 |
| clayg | whoo! | 23:13 |
| clarkb | puppet-openstack modules test ipv6'd services | 23:14 |
| openstackgerrit | Clay Gerrard proposed openstack/swift master: Try out a different interface https://review.openstack.org/478676 | 23:15 |
| caiobrentano | clag Nice! Thanks | 23:21 |
| *** caiobrentano has quit IRC | 23:23 | |
| *** kei_yama has joined #openstack-swift | 23:31 | |
| *** chsc has quit IRC | 23:34 | |
| *** JimCheung has joined #openstack-swift | 23:35 | |
| *** JimCheung has quit IRC | 23:46 | |
| *** JimCheung has joined #openstack-swift | 23:47 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!