clayg | timburke: oh no you broke it!? | 00:03 |
---|---|---|
*** catintheroof has quit IRC | 00:03 | |
timburke | clayg: i guess? looks like there are some issues during teardown if initialization fails for some backends. jerasure_rs_cauchy in particular? | 00:06 |
*** NM has joined #openstack-swift | 00:11 | |
*** caiobrentano_ has quit IRC | 00:13 | |
*** caiobrentano has joined #openstack-swift | 00:14 | |
openstackgerrit | Clay Gerrard proposed openstack/swift: Shuffle disks and parts in reconstructor https://review.openstack.org/425468 | 00:16 |
*** vinsh has quit IRC | 00:18 | |
openstackgerrit | Clay Gerrard proposed openstack/swift: Shuffle disks and parts in reconstructor https://review.openstack.org/425468 | 00:18 |
*** dmorita has quit IRC | 00:19 | |
*** dmorita has joined #openstack-swift | 00:22 | |
openstackgerrit | Tim Burke proposed openstack/liberasurecode: Make free_rs_cauchy_desc tolerate partially-initialized descriptors https://review.openstack.org/425471 | 00:26 |
timburke | interesting that jerasure_rs_vand and liberasurecode_rs_vand seem to work fine with num_parity=5... | 00:27 |
timburke | and strange that you can't initialize jerasure_rs_cauchy 12 + 5 | 00:28 |
timburke | or 11 + 6 -- does it have a hard cap at 16 total fragments, maybe? | 00:30 |
*** sams-gleb has joined #openstack-swift | 00:31 | |
*** sams-gleb has quit IRC | 00:35 | |
*** NM has quit IRC | 00:46 | |
clayg | timburke: yeah there's a bug up on isa-l that i guess explains the matrix being used by liberasure isn't correct for all transformations or something ... it seemed like maybe isa-l could do better; but liberasurecode missed some error handling; regardless *everyone* agreed there wasn't nothing to de done about the data in-place encoded with the the bad implementation | 00:53 |
clayg | https://github.com/01org/isa-l/issues/10 (and then there was some emails too) | 00:55 |
*** _JZ_ has quit IRC | 00:57 | |
zaitcev | I have a basic question about the way Swift works. Very basic. | 00:57 |
zaitcev | Object auditors check etags, correct? If so, etags have to be saved alongside objects. | 00:58 |
zaitcev | But, it's completely valid to do a PUT with no etag (e.g. with curl). | 00:58 |
*** NM has joined #openstack-swift | 00:59 | |
zaitcev | I always thought that in such case, proxies calculated etags and sent them along. But we never had footers before EC. | 00:59 |
zaitcev | So, how did it actually work? | 00:59 |
timburke | zaitcev: as i recall, object-server does the md5 on upload. if there was an etag, it might 422, but otherwise it just writes down the md5 (that was calculated anyway) as the etag | 01:02 |
zaitcev | timburke: thanks.... I'm blind like a mole as usual. It says etag=md5() right there in PUT() | 01:03 |
zaitcev | but when you mentioned that I saw HTTPUnprocessableEntity(request=request) | 01:04 |
*** m_kazuhiro has joined #openstack-swift | 01:12 | |
m_kazuhiro | good morning | 01:12 |
*** chsc has joined #openstack-swift | 01:14 | |
*** chsc has quit IRC | 01:19 | |
*** tqtran has quit IRC | 01:26 | |
*** clu_ has quit IRC | 01:28 | |
*** mingyu has joined #openstack-swift | 01:40 | |
*** catintheroof has joined #openstack-swift | 01:44 | |
*** mingyu has quit IRC | 01:44 | |
*** Jeffrey4l_ has quit IRC | 01:49 | |
*** catintheroof has quit IRC | 01:49 | |
*** Jeffrey4l has joined #openstack-swift | 01:49 | |
*** ejat has quit IRC | 01:51 | |
*** ejat has joined #openstack-swift | 01:51 | |
*** ejat has quit IRC | 01:51 | |
*** ejat has joined #openstack-swift | 01:51 | |
openstackgerrit | Ha Dung proposed openstack/pyeclib: Fixes some current typo errors on pyeclib. https://review.openstack.org/425492 | 02:05 |
openstackgerrit | Clay Gerrard proposed openstack/swift: Make the reconstructor handoffs_first work (and useful) https://review.openstack.org/425493 | 02:05 |
openstackgerrit | Clay Gerrard proposed openstack/swift: Give spawn a cycle https://review.openstack.org/425495 | 02:09 |
openstackgerrit | Tim Burke proposed openstack/swift: Warn about using EC with isa_l_rs_vand and nparity >= 5 https://review.openstack.org/425496 | 02:10 |
*** caiobrentano has quit IRC | 02:11 | |
*** cdelatte has quit IRC | 02:11 | |
*** klrmn has quit IRC | 02:14 | |
portante | clayg: you are welcome | 02:16 |
timburke | wait, we vendor enum34 in pyeclib? huh... | 02:18 |
*** NM has quit IRC | 02:19 | |
*** winggundamth_ has joined #openstack-swift | 02:19 | |
*** two_tired has joined #openstack-swift | 02:29 | |
openstackgerrit | Clay Gerrard proposed openstack/swift: Shuffle disks and parts in reconstructor https://review.openstack.org/425468 | 02:30 |
openstackgerrit | Clay Gerrard proposed openstack/swift: Make the reconstructor handoffs_first work (and useful) https://review.openstack.org/425493 | 02:31 |
*** sams-gleb has joined #openstack-swift | 02:33 | |
*** lespaul has quit IRC | 02:34 | |
*** sams-gleb has quit IRC | 02:38 | |
*** Jeffrey4l has quit IRC | 02:45 | |
*** dmorita has quit IRC | 02:47 | |
*** dmorita has joined #openstack-swift | 02:47 | |
openstackgerrit | Clay Gerrard proposed openstack/swift: Do not revert fragments to handoffs https://review.openstack.org/425441 | 02:49 |
clayg | timburke: the header sounds like we made some changes/fixes? | 02:50 |
clayg | # Original credit: enum34 project https://pypi.python.org/pypi/enum34 | 02:51 |
clayg | # Changes for PyECLib (c) Tushar Gohad (tusharsg@gmail.com) | 02:51 |
clayg | ... but I never noticed - TIL | 02:51 |
*** calebb has quit IRC | 02:55 | |
*** bkopilov has quit IRC | 03:00 | |
*** d0ugal has quit IRC | 03:00 | |
*** d0ugal has joined #openstack-swift | 03:17 | |
*** calebb has joined #openstack-swift | 03:22 | |
clayg | quiet in here... | 03:24 |
jidar | what do people do for embedding credentials in scripts to do stuff like get keystone tokens to download/put files into/out of swift? Is there nothing like a machine credentials or a way to give a instance/user access that's only related to swift? | 03:25 |
jidar | is there a best practice I'm missing out on? | 03:26 |
clayg | public or private cloud? | 03:36 |
clayg | private cloud - best option is probably create a user/account for that application/task - then share the data with other users via acl's | 03:37 |
clayg | if it's public cloud and you can't get another account/user just for that task - probably just give the tool/app it's own container and tempurl-key and make the file not world readable and rotate it every so often with temp-url-key2 or /we | 03:38 |
openstackgerrit | Merged openstack/swift: remove func-fast-post from tox.ini https://review.openstack.org/425264 | 03:41 |
*** mingyu has joined #openstack-swift | 03:41 | |
openstackgerrit | Merged openstack/swift: Bring docs inline with changes to tox envs https://review.openstack.org/425311 | 03:42 |
jidar | clayg: thanks for that, I'm looking over the ACLs, this is for a private cloud, I guess the idea of creating a regular _member_ user to access swift felt off | 03:42 |
clayg | cschwede: i just got the "warning openstack CFP closes in 2 weeks" | 03:42 |
clayg | we gunna do another talk this year!? | 03:42 |
clayg | i'm not sure I have anything interesting left to say | 03:43 |
clayg | jidar: yeah i like it! you're using keystone then? | 03:43 |
jidar | I am | 03:43 |
clayg | I don't think it has account ACL's unfortunately - but container ACLs are pretty good - everyone has temp-url-keys! | 03:44 |
jidar | yea, right now I'm just authenticating keystone, pulling the token out and using it as a param to curl | 03:44 |
jidar | not fond of putting those username/passwords into git obviously | 03:44 |
clayg | yeah, hard to get them out of source control | 03:45 |
jidar | fair enough! | 03:45 |
clayg | having the script `source .keys` and doing a little something to get that file on the node isn't *so* terrible | 03:45 |
clayg | most puppet/chef soft of situations have encrypted databags or something | 03:46 |
jidar | this is all being generated out of heat | 03:46 |
*** mingyu has quit IRC | 03:46 | |
clayg | so how does heat normally deal with secrets? | 03:46 |
jidar | heh, afaik it doesn't | 03:46 |
clayg | so whatever it does for your mysql password is probably good enough | 03:47 |
jidar | you can pass the password in as a param to the template | 03:47 |
jidar | so it only gets passed at run time | 03:47 |
clayg | ah... so maybe that then? | 03:47 |
jidar | I guess that can work | 03:48 |
*** JimCheung has quit IRC | 03:52 | |
jidar | again, clayg thanks for all of the help | 03:52 |
*** JimCheung has joined #openstack-swift | 03:52 | |
*** JimCheung has quit IRC | 03:57 | |
*** klrmn has joined #openstack-swift | 03:59 | |
*** psachin has joined #openstack-swift | 04:03 | |
*** m_kazuhiro has quit IRC | 04:05 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 04:06 |
*** varunsomani has joined #openstack-swift | 04:07 | |
*** SkyRocknRoll has joined #openstack-swift | 04:19 | |
*** varunsomani has quit IRC | 04:23 | |
*** bkopilov has joined #openstack-swift | 04:28 | |
*** sams-gleb has joined #openstack-swift | 04:35 | |
*** sams-gleb has quit IRC | 04:40 | |
*** SkyRocknRoll has quit IRC | 04:41 | |
*** mingyu has joined #openstack-swift | 04:43 | |
*** klrmn has quit IRC | 04:44 | |
*** klrmn has joined #openstack-swift | 04:46 | |
*** mingyu has quit IRC | 04:47 | |
*** klrmn has quit IRC | 05:02 | |
*** Guest58531 has quit IRC | 05:04 | |
*** zacksh has quit IRC | 05:05 | |
*** sgundur has quit IRC | 05:06 | |
*** zacksh has joined #openstack-swift | 05:07 | |
*** mgagne has joined #openstack-swift | 05:07 | |
*** mgagne is now known as Guest33539 | 05:07 | |
*** sgundur has joined #openstack-swift | 05:09 | |
*** SkyRocknRoll has joined #openstack-swift | 05:35 | |
*** mmmucky has quit IRC | 05:39 | |
*** mmmucky has joined #openstack-swift | 05:40 | |
*** winggundamth_ has quit IRC | 05:42 | |
*** two_tired has quit IRC | 05:44 | |
*** klrmn has joined #openstack-swift | 05:52 | |
*** klrmn has quit IRC | 05:53 | |
*** varunsomani has joined #openstack-swift | 06:24 | |
*** JimCheung has joined #openstack-swift | 06:25 | |
*** varunsomani has quit IRC | 06:28 | |
*** JimCheung has quit IRC | 06:30 | |
*** sams-gleb has joined #openstack-swift | 06:37 | |
*** dmorita has quit IRC | 06:38 | |
*** sams-gleb has quit IRC | 06:39 | |
*** sams-gleb has joined #openstack-swift | 06:39 | |
*** sams-gleb has quit IRC | 06:40 | |
*** sams-gleb has joined #openstack-swift | 06:41 | |
*** sams-gleb has quit IRC | 06:41 | |
*** mingyu has joined #openstack-swift | 06:44 | |
*** mingyu has quit IRC | 06:49 | |
*** ChubYann has quit IRC | 06:55 | |
*** sams-gleb has joined #openstack-swift | 07:04 | |
*** sams-gleb has quit IRC | 07:04 | |
*** sams-gleb has joined #openstack-swift | 07:04 | |
*** dmorita has joined #openstack-swift | 07:04 | |
*** dmorita has quit IRC | 07:06 | |
*** dmorita has joined #openstack-swift | 07:06 | |
*** tesseract has joined #openstack-swift | 07:11 | |
*** tdasilva has quit IRC | 07:15 | |
*** PavelK has joined #openstack-swift | 07:21 | |
*** dmorita has quit IRC | 07:36 | |
*** dmorita has joined #openstack-swift | 07:37 | |
*** dmorita_ has joined #openstack-swift | 07:38 | |
*** dmorita_ has quit IRC | 07:39 | |
*** dmorita has quit IRC | 07:41 | |
*** tdasilva has joined #openstack-swift | 07:55 | |
*** rledisez has joined #openstack-swift | 08:11 | |
*** tdasilva has quit IRC | 08:22 | |
*** kei_yama has quit IRC | 08:22 | |
*** geaaru has joined #openstack-swift | 08:26 | |
*** tdasilva has joined #openstack-swift | 08:29 | |
*** oshritf has joined #openstack-swift | 08:31 | |
*** geaaru has joined #openstack-swift | 08:35 | |
*** mingyu has joined #openstack-swift | 08:45 | |
*** glb1 has joined #openstack-swift | 08:46 | |
*** mingyu has quit IRC | 08:50 | |
*** glb1 has quit IRC | 08:54 | |
*** hseipp has joined #openstack-swift | 09:08 | |
*** cbartz has joined #openstack-swift | 09:10 | |
*** hseipp has quit IRC | 09:10 | |
*** hseipp has joined #openstack-swift | 09:16 | |
*** oshritf has quit IRC | 09:17 | |
*** jcaron has joined #openstack-swift | 09:18 | |
*** oshritf has joined #openstack-swift | 09:24 | |
*** oshritf has quit IRC | 09:27 | |
*** oshritf has joined #openstack-swift | 09:30 | |
*** SkyRocknRoll has quit IRC | 09:33 | |
*** oshritf has quit IRC | 09:36 | |
*** dmorita has joined #openstack-swift | 09:39 | |
*** oshritf has joined #openstack-swift | 09:40 | |
*** oshritf has quit IRC | 09:42 | |
*** dmorita has quit IRC | 09:44 | |
*** mvk has quit IRC | 09:47 | |
cschwede | clayg: sure, i'd be more than happy to do a talk with you! i have 1,2 ideas, will ping you direct | 09:54 |
*** SkyRocknRoll has joined #openstack-swift | 09:56 | |
*** oshritf has joined #openstack-swift | 10:00 | |
*** jordanP has joined #openstack-swift | 10:05 | |
*** SkyRocknRoll has quit IRC | 10:09 | |
*** oshritf has quit IRC | 10:12 | |
*** acoles_ is now known as acoles | 10:28 | |
acoles | clayg: well spotted on bug 1659147 | 10:33 |
openstack | bug 1659147 in OpenStack Object Storage (swift) "swift-get-nodes not working in "full" mitaka " [Undecided,New] https://launchpad.net/bugs/1659147 | 10:33 |
patchbot | Error: No closing quotation | 10:33 |
*** oshritf has joined #openstack-swift | 10:35 | |
*** oshritf has quit IRC | 10:37 | |
*** sams-gleb has quit IRC | 10:44 | |
*** oshritf has joined #openstack-swift | 10:46 | |
*** mingyu has joined #openstack-swift | 10:47 | |
*** mingyu has quit IRC | 10:51 | |
*** sams-gleb has joined #openstack-swift | 11:02 | |
*** oshritf_ has joined #openstack-swift | 11:10 | |
*** oshritf has quit IRC | 11:13 | |
*** openstackgerrit has quit IRC | 11:17 | |
*** NM has joined #openstack-swift | 11:32 | |
*** silor has joined #openstack-swift | 11:43 | |
*** silor has quit IRC | 11:54 | |
*** silor has joined #openstack-swift | 11:54 | |
*** silor has quit IRC | 11:57 | |
*** silor has joined #openstack-swift | 11:59 | |
*** bkopilov has quit IRC | 12:05 | |
*** mvk has joined #openstack-swift | 12:16 | |
*** vint_bra has joined #openstack-swift | 12:18 | |
*** caiobrentano has joined #openstack-swift | 12:24 | |
*** catintheroof has joined #openstack-swift | 12:27 | |
*** catintheroof has quit IRC | 12:38 | |
*** catintheroof has joined #openstack-swift | 12:38 | |
*** mingyu has joined #openstack-swift | 12:48 | |
*** mingyu has quit IRC | 12:53 | |
*** JimCheung has joined #openstack-swift | 12:54 | |
*** oshritf__ has joined #openstack-swift | 12:57 | |
*** catinthe_ has joined #openstack-swift | 12:57 | |
*** JimCheung has quit IRC | 12:58 | |
*** oshritf__ has quit IRC | 12:58 | |
*** catintheroof has quit IRC | 12:59 | |
*** oshritf_ has quit IRC | 13:00 | |
*** oshritf__ has joined #openstack-swift | 13:02 | |
*** oshritf__ has quit IRC | 13:08 | |
*** SkyRocknRoll has joined #openstack-swift | 13:08 | |
*** oshritf__ has joined #openstack-swift | 13:09 | |
*** NM has quit IRC | 13:22 | |
*** varunsomani has joined #openstack-swift | 13:25 | |
*** varunsomani has quit IRC | 13:29 | |
*** acoles is now known as acoles_ | 13:33 | |
*** NM has joined #openstack-swift | 13:34 | |
*** Jeffrey4l has joined #openstack-swift | 13:35 | |
*** acoles_ is now known as acoles | 13:36 | |
*** gabor_antal has quit IRC | 13:45 | |
*** bkopilov has joined #openstack-swift | 13:45 | |
*** PavelK has quit IRC | 14:15 | |
*** dmorita has joined #openstack-swift | 14:18 | |
*** dmorita has quit IRC | 14:20 | |
*** dmorita has joined #openstack-swift | 14:20 | |
*** oshritf__ has quit IRC | 14:41 | |
*** SkyRocknRoll has quit IRC | 14:44 | |
*** Jeffrey4l has quit IRC | 14:44 | |
*** Jeffrey4l has joined #openstack-swift | 14:44 | |
*** SkyRocknRoll has joined #openstack-swift | 14:56 | |
*** dmorita has quit IRC | 15:03 | |
*** sams-gleb has quit IRC | 15:11 | |
*** sams-gleb has joined #openstack-swift | 15:12 | |
*** sams-gleb has quit IRC | 15:16 | |
*** jarbod has quit IRC | 15:28 | |
*** jarbod has joined #openstack-swift | 15:28 | |
*** sams-gleb has joined #openstack-swift | 15:28 | |
*** psachin has quit IRC | 15:29 | |
*** sams-gle_ has joined #openstack-swift | 15:30 | |
*** sams-gleb has quit IRC | 15:30 | |
*** stevemar has quit IRC | 15:36 | |
*** stevemar has joined #openstack-swift | 15:37 | |
*** catintheroof has joined #openstack-swift | 15:39 | |
*** catinthe_ has quit IRC | 15:42 | |
*** chsc has joined #openstack-swift | 15:48 | |
*** chsc has joined #openstack-swift | 15:48 | |
*** chsc has quit IRC | 15:54 | |
*** mvk has quit IRC | 15:55 | |
*** lxsli has quit IRC | 15:57 | |
*** lxsli has joined #openstack-swift | 16:01 | |
*** _JZ_ has joined #openstack-swift | 16:04 | |
*** chsc has joined #openstack-swift | 16:24 | |
*** chsc has joined #openstack-swift | 16:24 | |
*** ukaynar has joined #openstack-swift | 16:28 | |
*** dmorita has joined #openstack-swift | 16:38 | |
*** dmorita has quit IRC | 16:43 | |
*** tesseract has quit IRC | 16:55 | |
*** zaitcev has quit IRC | 17:03 | |
*** tqtran has joined #openstack-swift | 17:04 | |
*** openstackgerrit has joined #openstack-swift | 17:06 | |
openstackgerrit | Christopher Bartz proposed openstack/python-swiftclient: ISO 8601 timestamps for tempurl https://review.openstack.org/423377 | 17:06 |
*** dmorita has joined #openstack-swift | 17:10 | |
*** cbartz has left #openstack-swift | 17:11 | |
notmyname | good morning | 17:14 |
*** ukaynar has quit IRC | 17:15 | |
*** ukaynar has joined #openstack-swift | 17:16 | |
*** zaitcev has joined #openstack-swift | 17:19 | |
*** ChanServ sets mode: +v zaitcev | 17:19 | |
*** SkyRocknRoll has quit IRC | 17:22 | |
*** klrmn has joined #openstack-swift | 17:27 | |
*** dmorita has quit IRC | 17:29 | |
*** dmorita has joined #openstack-swift | 17:29 | |
*** ukaynar has quit IRC | 17:30 | |
*** ukaynar has joined #openstack-swift | 17:30 | |
*** silor has quit IRC | 17:32 | |
*** JimCheung has joined #openstack-swift | 17:36 | |
*** hseipp has quit IRC | 17:36 | |
*** rledisez has quit IRC | 17:44 | |
notmyname | if I don't see any comments on https://review.openstack.org/#/c/425454/ in the next few hours, I'll land it so we have time to do the release repo dance before the end of the week | 17:48 |
patchbot | patch 425454 - python-swiftclient - 3.3.0 authors/changelog update | 17:48 |
notmyname | clayg: nice reconstructor patches :-) | 17:51 |
*** dmorita has quit IRC | 17:53 | |
*** david-lyle has quit IRC | 17:59 | |
*** zackmdavis has quit IRC | 18:03 | |
*** zackmdavis has joined #openstack-swift | 18:03 | |
*** arch-nemesis has joined #openstack-swift | 18:05 | |
*** arch-nemesis has quit IRC | 18:12 | |
*** dmorita has joined #openstack-swift | 18:18 | |
*** varunsomani has joined #openstack-swift | 18:19 | |
clayg | notmyname: they're AMAZING | 18:20 |
timburke | go sgundur! thanks for validating my hypothesis! | 18:27 |
clayg | acoles: i'm sure you already expected that when you were doing it | 18:29 |
clayg | acoles: re lp bug #1659147 | 18:29 |
openstack | Launchpad bug 1659147 in OpenStack Object Storage (swift) "swift-get-nodes not working in "full" mitaka " [Undecided,New] https://launchpad.net/bugs/1659147 | 18:29 |
patchbot | Error: No closing quotation | 18:29 |
clayg | acoles: I wasn't sure how to take on how the "bug report" now has a question about keystone? | 18:30 |
clayg | i mean I guess I should set the status to invalid and suggest she use the mailing listing or get on IRC? | 18:31 |
clayg | or maybe use ask? | 18:31 |
clayg | people use ask right? | 18:31 |
*** vinsh has joined #openstack-swift | 18:32 | |
timburke | respond by pointing to cosbench / ssbench issue trackers and #openstack-keystone, then close as invalid? | 18:32 |
*** rcernin has joined #openstack-swift | 18:32 | |
sgundur | timburke, thanks; I am yet to go through his patch completely | 18:32 |
clayg | timburke: would you? :D | 18:33 |
clayg | timburke: well, no except | 18:33 |
clayg | timburke: I don't think this person should report bugs anymore :'( | 18:34 |
timburke | maybe after i've got the EC warning squared away | 18:34 |
timburke | i mean, we *could* totally fix ssbench to support v3 | 18:34 |
clayg | does it not? I didn't look. I acctually don't know the differences between keystone v2 and v3 - isn't that what was requested? | 18:35 |
acoles | clayg: nor me. not too familiar with *bench but if they support v2 that should still work with keystone v3 I think. oh, unless the keystone service is only offering a v3 API. | 18:36 |
acoles | clayg: biggest difference from v2 to v3 IIRC is support for domains | 18:37 |
*** dmorita has quit IRC | 18:37 | |
timburke | clayg: it doesn't -- v1/v2 only. see https://github.com/swiftstack/ssbench/blob/master/ssbench/swift_client.py#L203-L328 -- pretty sure it's just an old (pre-v3) fork of swiftclient | 18:38 |
timburke | clayg: the differences between v2 and v3 are why i'd point them to #openstack-keystone, but i think the real ask is to have these benchmarks support v3 | 18:39 |
clayg | timburke: ok, well let's hope I wasn't too off-putting | 18:40 |
*** dmorita has joined #openstack-swift | 18:42 | |
*** Nicolaas has joined #openstack-swift | 18:42 | |
*** mvk has joined #openstack-swift | 18:42 | |
*** geaaru has quit IRC | 18:43 | |
*** dmorita has quit IRC | 18:43 | |
clayg | acoles: so this was me on the train reading PavelKs comments on patch 419787 "bullshit. gd $%^& srly!? well maybe it's not so bad. I'm never going to get it right." | 18:43 |
patchbot | https://review.openstack.org/#/c/419787/ - swift - Better optimistic lock in get_hashes | 18:43 |
clayg | still haven't made it to acceptance :'( | 18:43 |
*** Nicolaas has quit IRC | 18:44 | |
clayg | I swear this cat has a parrallel processor installed in his brain :'( | 18:45 |
timburke | maybe it's the new baby making him acutely aware of the need to context-switch, and just how long it can take to get back to what you were working on before? :P | 18:49 |
clayg | rofl | 18:49 |
*** varunsomani has quit IRC | 18:49 | |
clayg | spoken like a true father | 18:49 |
acoles | clayg: yeah i think he has a valid concern | 18:50 |
clayg | "as soon as I shut down my laptop I will realise it is crazy!" | 18:50 |
clayg | lol | 18:50 |
acoles | clayg: I just posted my comments on gerrit | 18:50 |
clayg | *been there* | 18:50 |
acoles | clayg: so what I actually thought is "can I even shut down *before* clay tells me this is crazy" :) | 18:51 |
*** jordanP has quit IRC | 18:53 | |
*** mingyu has joined #openstack-swift | 18:53 | |
acoles | timburke: I once heard (no idea if true) that fighter pilots were blasted with recordings of their own kids crying while training, to test their ability to not be distracted | 18:54 |
acoles | g'night | 18:56 |
*** acoles is now known as acoles_ | 18:56 | |
notmyname | that sounds terrifying | 18:57 |
*** mingyu has quit IRC | 18:58 | |
*** varunsomani has joined #openstack-swift | 19:10 | |
*** JimCheung has quit IRC | 19:20 | |
*** JimCheung has joined #openstack-swift | 19:20 | |
*** clu_ has joined #openstack-swift | 19:35 | |
openstackgerrit | Tim Burke proposed openstack/swift: Warn about using EC with isa_l_rs_vand and nparity >= 5 https://review.openstack.org/425496 | 19:36 |
*** david-lyle has joined #openstack-swift | 19:45 | |
*** varunsomani has quit IRC | 19:54 | |
notmyname | SDSC gave a presentation last night on Swift https://github.com/OpenStackSanDiego/Presentations/tree/master/2017-01%20Swift <-- pdf of slides in there | 19:54 |
*** Renich has joined #openstack-swift | 20:00 | |
clayg | mattoliverau: cschwede: you guys need to get the openstack swift ansible stuff promoted better - SDSC thinks Swift deployment automation tools are maybe not the best in the world!? (i sorta have my own angle for making swift deployment/mangement better ;) ) | 20:01 |
*** Guest33539 is now known as mgagne | 20:01 | |
*** mgagne has quit IRC | 20:01 | |
*** mgagne has joined #openstack-swift | 20:01 | |
tdasilva | heh | 20:02 |
notmyname | tdasilva: has some cools ansible swift stuff too that I found | 20:02 |
tdasilva | lol, that was all mattoliverau | 20:02 |
tdasilva | there's been a ton of great work with tripleo too | 20:03 |
notmyname | tdasilva: actually, i was going to talk to you about it at some point (not right now though). I'm working on something similar too | 20:03 |
tdasilva | but yeah, reading about the rings issue reminded me of an email on the openstack mailing list recently | 20:03 |
tdasilva | notmyname: cool! | 20:03 |
clayg | tdasilva: their little pseudo code script was so cute tho! I was like see - you *like* it - it's job security for devops! | 20:04 |
tdasilva | hehe, yeah, i saw that, "goto #1" | 20:05 |
* tdasilva wishes ovh could talk more about their ring managing bots, that sounded really cool | 20:05 | |
clayg | but then there like "it's a big con for swift and pro for other systems that you have to do this" | 20:05 |
clayg | tdasilva: I think it's the same as what is in the psuedo script there? | 20:06 |
tdasilva | clayg: maybe?? it sounded super complex with databases and managing the weights on multiples disks and nodes and what not, not sure...maybe it is a simple script... | 20:07 |
*** NM has quit IRC | 20:07 | |
clayg | *I* feel like no matter how you do it - it's hard to *configure* something like a ring manager - the code to "do" the ring stuff is brain dead simple | 20:07 |
clayg | but it's like - how do you get dispersion information - how do you get your nodes to talk to where the rings are | 20:07 |
clayg | so anything we do that's opinionated on that makes it *harder* to configure? | 20:08 |
clayg | but... you know... the longer I feel this way the more I have to admit I must be wrong | 20:08 |
clayg | people would rather be told "do it this way" than "it's easy do it however you want" | 20:08 |
*** JimCheung has quit IRC | 20:09 | |
tdasilva | clayg: similarly, there's the issue of monitoring | 20:09 |
clayg | tdasilva: i mean yeah - it did sorta sound like you need some sort of data store - you could use the builder file - or a file system - or json - or a database - whatever you want! ;) | 20:09 |
*** JimCheung has joined #openstack-swift | 20:10 | |
tdasilva | there are so many out there and everybody wants to use their own thing | 20:10 |
clayg | well it did sorta sound like they were not in love with the openstack telementry integration with swift - but then it also looked like they had a pretty good solution in the end? | 20:10 |
clayg | maybe they've had experiences with other systems that had more community stuff they could pull of the shelf? I... well *I* sort of have my own thing going for monitoring :D | 20:11 |
tdasilva | yeah, I do wonder how what version of the telemetry software they are using, my understanding is that there's been a complete overhaul of ceilometer, but it happened very recently | 20:11 |
clayg | cool! | 20:11 |
notmyname | wasn't there a split of ceilometer into different projects? | 20:12 |
notmyname | one had an unpronouncable name with lots of vowels. aodh? aeiou? eieio? | 20:13 |
tdasilva | notmyname: i think so, now there's gnocchi and aodh | 20:13 |
notmyname | ok | 20:13 |
tdasilva | aodh is for "alarms" ??? | 20:13 |
notmyname | idk | 20:14 |
tdasilva | gnocchi for metrics, i believe?? | 20:14 |
notmyname | I have seen no feedback on the swiftclient authors/changelog patch (https://review.openstack.org/#/c/425454/) so I'm going to go ahead and merge it | 20:14 |
patchbot | patch 425454 - python-swiftclient - 3.3.0 authors/changelog update | 20:14 |
*** JimCheung has quit IRC | 20:14 | |
notmyname | if you don't like it, you've got <gate time> to say something | 20:15 |
clayg | timburke: is that raymond hettinger on your screen over there!? | 20:16 |
clayg | timburke: also - your monitor is huge ;) | 20:16 |
timburke | clayg: why wouldn't it be? | 20:16 |
timburke | as expected, it's a great talk -- https://www.youtube.com/watch?v=voXVTjwnn-U | 20:27 |
tdasilva | lol, "I don't tweet when I get off the plane, I tweet to teach python" | 20:39 |
tdasilva | the cool think i noticed about the LCA videos is how they switch from the presentation screen and back to the speaker | 20:41 |
tdasilva | i like that better than the openstack videos | 20:41 |
tdasilva | s/think/thing | 20:41 |
clayg | this isn't LCA is it? | 20:41 |
tdasilva | lol, it is not | 20:42 |
timburke | nope, PyBay2016 apparently. found it while watching some lca videos though | 20:42 |
tdasilva | i guess two conferences do the same | 20:42 |
tdasilva | ;) | 20:42 |
*** mingyu has joined #openstack-swift | 20:54 | |
*** mingyu has quit IRC | 20:59 | |
*** cebreidian has quit IRC | 20:59 | |
timburke | anybody know enough about the unit test structure to help me add a test for https://review.openstack.org/#/c/425471/? i noticed the bug when trying to init an ECDriver like pyeclib.ec_iface.ECDriver(ec_type='jerasure_rs_cauchy', k=10, m=10) | 21:31 |
patchbot | patch 425471 - liberasurecode - Make free_rs_cauchy_desc tolerate partially-initia... | 21:31 |
*** Jeffrey4l_ has joined #openstack-swift | 21:34 | |
*** Jeffrey4l has quit IRC | 21:35 | |
*** cebreidian has joined #openstack-swift | 21:38 | |
*** cebreidian has quit IRC | 21:43 | |
timburke | i kinda expected https://github.com/openstack/liberasurecode/blob/master/test/liberasurecode_test.c#L136-L142 to make it bomb out? trying those args through pyeclib makes python seg fault or abort | 21:43 |
*** catintheroof has quit IRC | 21:47 | |
*** sams-gle_ has quit IRC | 21:50 | |
*** vint_bra has quit IRC | 22:01 | |
*** chsc has quit IRC | 22:03 | |
*** chsc has joined #openstack-swift | 22:16 | |
*** chsc has joined #openstack-swift | 22:16 | |
*** mingyu has joined #openstack-swift | 22:56 | |
openstackgerrit | Tim Burke proposed openstack/swift: Expose encryption status in GET/HEAD/PUT/POST responses https://review.openstack.org/348603 | 22:58 |
openstackgerrit | Tim Burke proposed openstack/swift: Expose encryption status to client https://review.openstack.org/348604 | 22:58 |
clayg | timburke: where are those args used tho? | 22:59 |
clayg | timburke: maybe pyeclib has it's own bug that causes it to segfault when liberasure is just erroring in some less than helpful way? | 23:00 |
*** mingyu has quit IRC | 23:01 | |
timburke | pretty sure it's a problem with libec. pyeclib maybe *also* needs work, but one step at a time. try something like this: | 23:05 |
timburke | echo -e 'r\nbt' | gdb -x - --args python -c 'import pyeclib.ec_iface; pyeclib.ec_iface.ECDriver(ec_type="jerasure_rs_cauchy", k=10, m=10)' | 23:05 |
patchbot | -e 'r\nbt' | gdb -x - --args python -c 'import pyeclib.ec_iface; pyeclib.ec_iface.ECDriver(ec_type="jerasure_rs_cauchy", k=10, m=10)' | 23:05 |
* timburke sighs | 23:05 | |
timburke | you'll see a stack like jerasure_rs_cauchy_init calling free_rs_cauchy_desc which tries to free something it has no business freeing | 23:06 |
* clayg installs gdb and gets his libec/pyeclib repos up-to-date | 23:08 | |
clayg | oh... no I don't think i *ever* had jerasure installed :\ | 23:08 |
clayg | not sure I'm going to :\ | 23:08 |
*** NM has joined #openstack-swift | 23:08 | |
clayg | I need like "plugins" or something for vsaio - I think I added some bs support to the `reec` script *if* you have isa-l checked out - maybe I could do that for jerasure? | 23:09 |
timburke | clayg: it's fairly easy. go look at https://review.openstack.org/#/c/408286/1/tools/pip-install-with-liberasurecode.sh and ignore all the prefixing bs | 23:09 |
patchbot | patch 408286 - pyeclib - Install jerasure from source when testing liberasu... | 23:09 |
clayg | timburke: thanks! | 23:13 |
clayg | timburke: on the list -> https://github.com/swiftstack/vagrant-swift-all-in-one/issues/56 | 23:16 |
clayg | timburke: I wonder if there's a way to just make vsaio use the liberasurecode/tools scripts? Can they be consumed like that? the seem to be all up in the VIRTUAL_ENV | 23:17 |
timburke | ? that's in pyeclib, and hasn't merged (in no small part 'cause we want to remove the need to test against libec master entirely) | 23:18 |
*** NM has quit IRC | 23:20 | |
clayg | oic | 23:21 |
clayg | ok, well I'll stick trying to add something to vsaio then - sorry I can't be more help! | 23:22 |
timburke | i could probably do that fairly easily if you like | 23:22 |
clayg | not a priority for me atm - let your todo list be your guide :D | 23:23 |
*** kei_yama has joined #openstack-swift | 23:31 | |
*** mingyu has joined #openstack-swift | 23:42 | |
timburke | ah! it's the .w = 8 that makes it work! looks like specifying w=8 to pyeclib still has it come out the default 4 | 23:45 |
timburke | so the particular problem i see is the result of https://github.com/openstack/liberasurecode/blob/cb0daba975dfb3b7bffe402d7aa73fac19180f73/src/backends/jerasure/jerasure_rs_cauchy.c#L263 and us trying to clean up desc before all of its members have been allocated | 23:46 |
clayg | ... file a bug and let someone who cares about jerasure square it? | 23:49 |
clayg | some times it's hard once you have the bit in your teeth tho | 23:49 |
clayg | I'm rooting for you timburke!!! | 23:50 |
clayg | grrrrrrr | 23:50 |
timburke | like, i know exactly where the problem is, i know exactly how to fix it, i *have* fixed it, but i don't know enough about how we run tests to write a *good* patch | 23:51 |
*** david-lyle has quit IRC | 23:51 | |
clayg | sounds like everyone contributing to a project for the first time ever? it's why launhcpad comes with a checkbox "this file is a patch" when you upload bug reports? | 23:52 |
clayg | maybe kota_ or kevin or tsg or someone can add some guidance? | 23:52 |
clayg | ... to the bug report ... | 23:52 |
clayg | I don't even know how to *run* tests (that exercise jerasure) | 23:53 |
clayg | so I'm way down your slider of possible being helpful | 23:53 |
clayg | tdasilva maybe knows/cares about jerasure? or someone from hp? | 23:53 |
*** david-lyle has joined #openstack-swift | 23:54 | |
timburke | mainly i just hate having code that *we own* that can and will *crash your python* | 23:54 |
clayg | suck it python! | 23:55 |
clayg | that's what you get for using jerasure | 23:55 |
timburke | on the plus side, jerasure doesn't suffer from https://bugs.launchpad.net/swift/+bug/1639691 so suck it ISA-L! | 23:55 |
openstack | Launchpad bug 1639691 in PyECLib "EC: Swift can return corrupted Data and be able to go data lost at isa_l_rs_vand policy with >=5 parities" [Undecided,In progress] | 23:55 |
clayg | yeah... poor isa_l_rs_vand :'( isa_l_rs_cauchy is great tho! | 23:56 |
*** rcernin has quit IRC | 23:56 | |
*** oshritf__ has joined #openstack-swift | 23:57 | |
*** rcernin has joined #openstack-swift | 23:57 | |
*** chsc has quit IRC | 23:58 | |
*** rcernin has quit IRC | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!