| *** haomaiwa_ has quit IRC | 00:01 | |
| *** haomaiwang has joined #openstack-swift | 00:01 | |
| *** thurloat_isgone is now known as thurloat | 00:08 | |
| *** thurloat is now known as thurloat_isgone | 00:09 | |
| *** thumpba has joined #openstack-swift | 00:13 | |
| notmyname | here's a comparison https://gist.github.com/notmyname/a65f61aaa6fb1375a748 | 00:14 |
|---|---|---|
| *** thurloat_isgone is now known as thurloat | 00:15 | |
| *** thumpba has quit IRC | 00:17 | |
| *** _JZ_ has quit IRC | 00:20 | |
| *** number80 has quit IRC | 00:25 | |
| notmyname | timburke: got it. now I have an unreviewed patches count. right now there are 23 unreviewed patches in swift | 00:26 |
| timburke | mean/median age? stddev? | 00:26 |
| notmyname | no, just counting. so in your case, that counts as one patch. I'm not tracking the time you've spent on it. it counts against the reviewer team because it's effectively infinitely long | 00:27 |
| mattoliverau | asettle: add me as a reviewer, I'd love to review it.. even if it is long. | 00:28 |
| asettle | mattoliverau: no problemo | 00:28 |
| asettle | doing now | 00:28 |
| mattoliverau | woot! | 00:28 |
| timburke | notmyname: also, it might be worth considering the time from merge conflict to new patch as part of the owner-responsiveness (assuming it actually *was* the owner that submitted the new patch, and assuming the owner was the last person to leave a comment) | 00:30 |
| *** mingyu_ has joined #openstack-swift | 00:34 | |
| *** mingyu_ has quit IRC | 00:34 | |
| notmyname | timburke: instead, I just added the is:mergable filter. so stuff that is in merge conflict won't be counted either way | 00:34 |
| *** arthus has quit IRC | 00:43 | |
| *** klrmn1 has quit IRC | 00:49 | |
| *** number80 has joined #openstack-swift | 00:55 | |
| openstackgerrit | Merged openstack/swift: Removed default value for log_statsd_host https://review.openstack.org/278502 | 00:55 |
| *** Jeffrey4l has joined #openstack-swift | 00:56 | |
| *** patchbot has quit IRC | 00:59 | |
| *** haomaiwang has quit IRC | 01:01 | |
| *** haomaiwa_ has joined #openstack-swift | 01:02 | |
| *** patchbot has joined #openstack-swift | 01:03 | |
| *** Jeffrey4l has quit IRC | 01:07 | |
| *** esker has joined #openstack-swift | 01:08 | |
| *** garthb has quit IRC | 01:12 | |
| *** klrmn has joined #openstack-swift | 01:12 | |
| *** chlong has joined #openstack-swift | 01:18 | |
| *** Jeffrey4l has joined #openstack-swift | 01:20 | |
| *** esker has quit IRC | 01:27 | |
| clayg | have I just not ever *noticed* the review riders before? I feel like I've heard people talk about this mysterious people that +1 things because stackalytics measures things - but I rarely noticed them hanging around swift? | 01:30 |
| clayg | Can i track down Sivasathurappan Radhakrishnan and demand he explain his review process? | 01:30 |
| clayg | I feel like it's "my bot noticed it only changes tests or docstrings" - which... could be helpful if there was a comment to the effect? | 01:31 |
| clayg | but what if it's just like "i randomly +1 things" - that could *also* go in a comment so it's more obviousl that's it's less useful? | 01:33 |
| clayg | i'm just so confused! | 01:33 |
| *** raginbajin has quit IRC | 01:35 | |
| *** raginbajin has joined #openstack-swift | 01:37 | |
| *** lyrrad has quit IRC | 01:37 | |
| timburke | clayg: i think siva_krishnan is still getting oriented; looks like he started in on openstack stuff in november? he's certainly been helpful in getting some essentially-abandoned swift3 patches back into shape | 01:38 |
| redbo | Oh god... I accidentally breathed near my SAIO, and now nothing works. | 01:41 |
| *** mrmoje has joined #openstack-swift | 01:42 | |
| *** gyee has quit IRC | 01:48 | |
| *** stantonnet has joined #openstack-swift | 01:52 | |
| *** asettle has quit IRC | 01:53 | |
| *** asettle has joined #openstack-swift | 01:57 | |
| *** haomaiwa_ has quit IRC | 02:01 | |
| *** haomaiwang has joined #openstack-swift | 02:03 | |
| *** mragupat has joined #openstack-swift | 02:08 | |
| *** mragupat has quit IRC | 02:08 | |
| *** Jeffrey4l has quit IRC | 02:08 | |
| *** mragupat has joined #openstack-swift | 02:08 | |
| pdardeau | clayg: timburke comment on siva_krishnan is accurate. he is new to project as are the rest of us in osic | 02:11 |
| *** daemontool has quit IRC | 02:11 | |
| *** Jeffrey4l has joined #openstack-swift | 02:11 | |
| *** vinsh has quit IRC | 02:12 | |
| *** vinsh has joined #openstack-swift | 02:13 | |
| *** fbo_ has joined #openstack-swift | 02:15 | |
| *** fbo has quit IRC | 02:15 | |
| *** openstackgerrit has quit IRC | 02:15 | |
| *** chlong has quit IRC | 02:15 | |
| Mingyu | Clayg: I reviewed patch 207790 again and did some tests. It works. I suggest to merge this soon as people raise questions about saio from time to time. | 02:18 |
| patchbot | Mingyu: https://review.openstack.org/#/c/207790/ - swift - Update SAIO setup for Ubuntu 14.04 | 02:18 |
| *** vinsh has quit IRC | 02:18 | |
| *** vinsh has joined #openstack-swift | 02:18 | |
| *** asettle has quit IRC | 02:19 | |
| *** daemontool has joined #openstack-swift | 02:23 | |
| *** klrmn has quit IRC | 02:23 | |
| *** mragupat has quit IRC | 02:24 | |
| *** openstackgerrit has joined #openstack-swift | 02:24 | |
| *** Jeffrey4l has quit IRC | 02:25 | |
| *** chlong has joined #openstack-swift | 02:28 | |
| *** asettle has joined #openstack-swift | 02:29 | |
| *** zaitcev has quit IRC | 02:31 | |
| *** asettle has quit IRC | 02:34 | |
| *** arch-nemesis has joined #openstack-swift | 02:39 | |
| *** km has joined #openstack-swift | 02:49 | |
| *** Jeffrey4l has joined #openstack-swift | 02:49 | |
| *** forcebanana has quit IRC | 02:52 | |
| clayg | Mingyu: neat! | 02:57 |
| clayg | timburke: pdardeau: GOOD TO KNOW! if you see siva_krishnan ask him to leave comments when he does reviews so other reviewers can get a feel for what he's doing | 02:59 |
| *** haomaiwang has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-swift | 03:01 | |
| *** zaitcev has joined #openstack-swift | 03:02 | |
| *** ChanServ sets mode: +v zaitcev | 03:02 | |
| *** arch-nemesis has quit IRC | 03:07 | |
| *** daemontool has quit IRC | 03:08 | |
| *** links has joined #openstack-swift | 03:10 | |
| clayg | redbo: lol | 03:10 |
| redbo | I figured it out, for some reason there was a wrong version pbr.egg-info in ~/python-swiftclient. I don't know how it got there or why python decided it'd be the version to import. | 03:13 |
| *** bill_az_ has quit IRC | 03:15 | |
| siva_krishnan | clayg sorry for that when I reviewed patch 278615 gate check results of first patch was showing up. new gate check results wasn't updated. I would make sure that this wouldn't happen again | 03:16 |
| patchbot | siva_krishnan: https://review.openstack.org/#/c/278615/ - swift - Docstring omission in class BaseDiskFileManager. | 03:16 |
| clayg | siva_krishnan: hey, no problem - sorry i was just confused! | 03:17 |
| clayg | siva_krishnan: but... i'm easily confused :'( | 03:17 |
| siva_krishnan | :) | 03:18 |
| *** sanchitmalhotra has joined #openstack-swift | 03:19 | |
| clayg | redbo: that sounds awesome! | 03:22 |
| openstackgerrit | Ankur proposed openstack/swift: Docstring omission in class BaseDiskFileManager. https://review.openstack.org/278615 | 03:23 |
| clayg | redbo: so the joke in slack this morning was "OpenStack recently released Bandit, a static analyzer that traverses abstract syntax trees (ASTs) of Python code.” Because of course it did. | 03:24 |
| clayg | redbo: but then the follow up joke was "I wonder how the managed to get it installed!?" | 03:24 |
| clayg | #pbrlife | 03:24 |
| *** arch-nemesis has joined #openstack-swift | 03:27 | |
| redbo | We haven't upgraded in a while, it's going to be a huge project getting the dependencies sorted out on our various operating systems. | 03:27 |
| redbo | I mean aside from cherry picking in security updates, we do that. | 03:27 |
| *** klrmn has joined #openstack-swift | 03:28 | |
| clayg | redbo: *totally* | 03:29 |
| *** esker has joined #openstack-swift | 03:32 | |
| Mingyu | redbo: I also encountered that problem yesterday (GMT+8 time). I guess that if you install swiftclient first and then you try to install pbr, that issue appears. | 03:34 |
| clayg | if you install swift with `pip install -e .` instead of `python setup.py develop` you shouldn't *need* to install pbr | 03:36 |
| clayg | i mean god only knows what pip will do under the hood - but i'm pretty sure it has something to do with working | 03:37 |
| clayg | maybe `pip install --upgrade -e .` upgrade is always a good idea new bugs are better than the bugs that already got fixed | 03:38 |
| *** esker has quit IRC | 03:53 | |
| *** venkat has joined #openstack-swift | 03:58 | |
| *** esker has joined #openstack-swift | 03:58 | |
| *** asettle has joined #openstack-swift | 04:00 | |
| *** haomaiwang has quit IRC | 04:01 | |
| *** haomaiwang has joined #openstack-swift | 04:01 | |
| *** esker has quit IRC | 04:03 | |
| Mingyu | clayg: LOL. God knows. Yes, i used `pip install -e .` but still pip refused to install pbr, until I (re)moved the "pbr-1.8.1-py2.7.egg" directory from swiftclient. | 04:04 |
| *** asettle has quit IRC | 04:05 | |
| *** arthus has joined #openstack-swift | 04:08 | |
| notmyname | good evening | 04:08 |
| *** arch-nemesis has quit IRC | 04:08 | |
| mattoliverau | notmyname: evening | 04:12 |
| Mingyu | good afternoon from Shenzhen :-p | 04:15 |
| notmyname | :-) | 04:15 |
| *** mrmoje has quit IRC | 04:18 | |
| *** klrmn has quit IRC | 04:19 | |
| *** klrmn has joined #openstack-swift | 04:19 | |
| *** mrmoje has joined #openstack-swift | 04:19 | |
| *** mragupat has joined #openstack-swift | 04:21 | |
| *** janonymous has joined #openstack-swift | 04:22 | |
| *** trifon has joined #openstack-swift | 04:24 | |
| *** diazjf has joined #openstack-swift | 04:28 | |
| clayg | Mingyu: my job is so weird - i had no idea you were in China - it's just like "helpful person on the internet" - has absolutely no relation so any physical space or nationality - you could have been on Mars or South Bay | 04:30 |
| clayg | notmyname: do you remmeber how far inland alpha_ori went when were out there? Shenzhen looks delightful | 04:31 |
| notmyname | I think he stayed in HK | 04:33 |
| *** forcebanana has joined #openstack-swift | 04:35 | |
| clayg | Mingyu: how long does it take you to get to HK (by train? drive?) - do you go often (yearly? ever?) | 04:35 |
| Mingyu | by subway, about 1 hour, but I've never been there before, lol | 04:36 |
| *** ppai has joined #openstack-swift | 04:37 | |
| Mingyu | actually, I've just moved to to Shenzhen in Dec. I based in beijing before then. | 04:38 |
| *** asettle has joined #openstack-swift | 04:40 | |
| *** asettle has quit IRC | 04:41 | |
| *** asettle has joined #openstack-swift | 04:41 | |
| notmyname | Mingyu: it's someplace I'd like to visit, especially because of all the hardware hacking stuff i've seen out of there | 04:42 |
| clayg | Mingyu: wow that's such a big move! and so recently! cool! | 04:43 |
| *** arthus has quit IRC | 04:43 | |
| Mingyu | notmyname: yeah, many hardware experts here. Welcome! | 04:46 |
| clayg | onovy: tomorrow I'm going to have to pick your brain how to do an inline with upstream install of newer setuptools for old distros like trusty (and even precise god help me) | 04:48 |
| clayg | onovy: even typing that sounds so INSANE - it *must* be easier to just hack the setup.py to remove those silly python version environ marker things? | 04:48 |
| Mingyu | clayg: thanks to the internet! I can choose to leave in the city I like continuing my work:-) | 04:49 |
| Mingyu | Job | 04:49 |
| clayg | setuptools .... 20.0 !? | 04:53 |
| notmyname | wat? | 04:54 |
| notmyname | wasn't it at version 0. something like 6 months ago? | 04:54 |
| *** arthus has joined #openstack-swift | 04:54 | |
| clayg | notmyname: if every change is a new major version then none of them are | 04:54 |
| notmyname | export VERSION=`git shortlog | wc -l` # bigger is better! | 04:56 |
| *** zhill has joined #openstack-swift | 04:58 | |
| StevenK | notmyname: -1 for use of `` :-P | 04:58 |
| notmyname | oh is that not allowed? ;-) | 04:59 |
| StevenK | (And setuptools uses hg, not git) | 04:59 |
| *** haomaiwang has quit IRC | 05:01 | |
| *** haomaiwa_ has joined #openstack-swift | 05:01 | |
| *** mrmoje has quit IRC | 05:02 | |
| *** jamielennox is now known as jamielennox|away | 05:06 | |
| Mingyu | I | 05:15 |
| Mingyu | I'm also surprised by the version number of Setup tools | 05:15 |
| Mingyu | was | 05:15 |
| Mingyu | I remember the latest version of setuptools you can get via apt-get is 3.3 ? | 05:22 |
| *** asettle has quit IRC | 05:28 | |
| openstackgerrit | Merged openstack/python-swiftclient: Support --os-identity-api-version option https://review.openstack.org/275719 | 05:28 |
| *** garthb has joined #openstack-swift | 05:28 | |
| zaitcev | In the race for the most ridiculous command line option, this loses to --yes-i-really-mean-it of ceph. | 05:29 |
| zaitcev | But it's quite close. | 05:30 |
| *** jamielennox|away is now known as jamielennox | 05:31 | |
| *** zaitcev has quit IRC | 05:35 | |
| *** zaitcev has joined #openstack-swift | 05:35 | |
| *** ChanServ sets mode: +v zaitcev | 05:35 | |
| *** asettle has joined #openstack-swift | 05:37 | |
| *** asettle has quit IRC | 05:42 | |
| *** klrmn has quit IRC | 05:45 | |
| *** haomaiwa_ has quit IRC | 06:01 | |
| *** 18VAACNL5 has joined #openstack-swift | 06:01 | |
| *** ChubYann has quit IRC | 06:03 | |
| *** wasmum has quit IRC | 06:05 | |
| openstackgerrit | Merged openstack/swift: Script for checking sanity of manpages https://review.openstack.org/276280 | 06:14 |
| *** venkat has quit IRC | 06:32 | |
| *** ppai has quit IRC | 06:33 | |
| *** bkumar has joined #openstack-swift | 06:34 | |
| *** venkat has joined #openstack-swift | 06:36 | |
| *** asettle has joined #openstack-swift | 06:37 | |
| openstackgerrit | Merged openstack/swift: go: GREP middleware https://review.openstack.org/217381 | 06:37 |
| *** ppai has joined #openstack-swift | 06:37 | |
| *** asettle has quit IRC | 06:41 | |
| *** asettle has joined #openstack-swift | 06:42 | |
| *** jith_ has joined #openstack-swift | 06:53 | |
| jith_ | hi all, Is there any possibility to reduce the number of replicas, so that one disk device will be free | 06:54 |
| mattoliverau | jith_: you could reduce the weight of that disk device. Cause just changing the number of replicas wont mean that disk wont have any objects on it. | 06:56 |
| jith_ | mattoliverau: thanks.. so i need to reduce the weight of the device to 0.. so that while rebalancing,,, no data will be there?? | 06:58 |
| jith_ | my set up is 3 disk with three replicas... i want to remove one disk.. two replica is enough for me... | 06:59 |
| mattoliverau | jith_: yeah, but if its a prod system you might want to lower it slowly, then rebalance.. then wait for the replicators to move data off... then lower the weight some more, otherwise you'll have alot of data movement at once. | 06:59 |
| mattoliverau | well, you'll have less disks then replicas, so that isn't nice. | 07:00 |
| jith_ | oh ok thanks.. | 07:00 |
| mattoliverau | you could reduce the replica count, then replicate, then lower weight, etc. | 07:00 |
| jith_ | thanks a lot | 07:00 |
| *** 18VAACNL5 has quit IRC | 07:01 | |
| redbo | Are there any tools to reduce the replica count? | 07:01 |
| *** haomaiwang has joined #openstack-swift | 07:01 | |
| redbo | oh swift-ring-builder <file> set_replicas <replicas> | 07:02 |
| mattoliverau | I thought we allowed the change or replica counts now | 07:02 |
| mattoliverau | redbo: maybe you guys dont use it :) | 07:02 |
| redbo | True, I've never had a reason to use it. | 07:03 |
| *** zaitcev has quit IRC | 07:04 | |
| jith_ | if i want to reduce the replica, first i need to swift-ring-builder <file> set_replicas 2.. then the data in three disk will be equally moved and three disk will be having part of data.. is it so.. | 07:05 |
| *** cbartz has joined #openstack-swift | 07:06 | |
| redbo | You have to do set_replicas, then rebalance, then distribute the ring out. | 07:06 |
| jith_ | after data movement i can remove the replica using swift-ring-builder remove ??? | 07:07 |
| jith_ | thanks | 07:07 |
| redbo | Wait. You might want to set_replicas and remove the node before you rebalance. I don't know, I'm not a ringxpert. | 07:08 |
| redbo | that might let you do it all at once. | 07:09 |
| *** baojg has joined #openstack-swift | 07:13 | |
| *** asettle has quit IRC | 07:15 | |
| *** asettle has joined #openstack-swift | 07:18 | |
| jith_ | redbo: thanks | 07:19 |
| jith_ | If I am reducing the replica..swift-ring-builder <file> set-replicas <value>.. this should be done for all three builder files (object,container and account) right?? | 07:19 |
| redbo | If you're running all of them on the same disks and you want to get rid of one, yes. | 07:21 |
| jith_ | redbo: yjamls | 07:24 |
| jith_ | thanks* | 07:24 |
| *** asettle has quit IRC | 07:34 | |
| *** baojg has quit IRC | 07:35 | |
| *** baojg has joined #openstack-swift | 07:35 | |
| *** km has quit IRC | 07:39 | |
| *** silor has joined #openstack-swift | 07:40 | |
| *** baojg has quit IRC | 07:42 | |
| *** diazjf has quit IRC | 07:42 | |
| *** asettle has joined #openstack-swift | 07:44 | |
| *** silor has quit IRC | 07:44 | |
| *** silor has joined #openstack-swift | 07:45 | |
| *** asettle has quit IRC | 07:49 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Decrypt multipart mime doc https://review.openstack.org/248377 | 07:50 |
| *** silor has quit IRC | 07:52 | |
| *** silor has joined #openstack-swift | 07:53 | |
| *** dmorita has quit IRC | 07:54 | |
| *** baojg has joined #openstack-swift | 07:55 | |
| *** dmorita has joined #openstack-swift | 07:55 | |
| *** dmorita has quit IRC | 07:56 | |
| *** dmorita has joined #openstack-swift | 07:57 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** dmorita has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-swift | 08:01 | |
| *** silor1 has joined #openstack-swift | 08:01 | |
| *** silor has quit IRC | 08:02 | |
| *** silor1 is now known as silor | 08:02 | |
| *** arthus has quit IRC | 08:05 | |
| *** jaakkos has quit IRC | 08:08 | |
| *** rledisez has joined #openstack-swift | 08:12 | |
| *** arthus has joined #openstack-swift | 08:12 | |
| *** dmorita has joined #openstack-swift | 08:15 | |
| *** garthb has quit IRC | 08:15 | |
| *** dmorita has quit IRC | 08:19 | |
| *** geaaru has joined #openstack-swift | 08:27 | |
| openstackgerrit | Merged openstack/swift: Enable in-process func tests to optionally use fast-post https://review.openstack.org/274086 | 08:29 |
| *** asettle has joined #openstack-swift | 08:30 | |
| *** rcernin has joined #openstack-swift | 08:35 | |
| *** mragupat has quit IRC | 08:40 | |
| *** patchbot has quit IRC | 08:45 | |
| *** patchbot` has joined #openstack-swift | 08:45 | |
| *** bkumar4 has joined #openstack-swift | 08:46 | |
| *** patchbot` is now known as patchbot | 08:46 | |
| *** openstackgerrit has quit IRC | 08:47 | |
| *** openstackgerrit_ has joined #openstack-swift | 08:47 | |
| *** stantonnet_ has joined #openstack-swift | 08:47 | |
| *** openstackgerrit_ is now known as openstackgerrit | 08:48 | |
| *** arthus has quit IRC | 08:51 | |
| *** geaaru has quit IRC | 08:52 | |
| *** bkumar has quit IRC | 08:52 | |
| *** stantonnet has quit IRC | 08:52 | |
| *** zhill has quit IRC | 08:56 | |
| *** geaaru has joined #openstack-swift | 08:57 | |
| *** jordanP has joined #openstack-swift | 09:00 | |
| jordanP | /join #openstack-meeting | 09:00 |
| jordanP | oups | 09:00 |
| *** haomaiwang has quit IRC | 09:01 | |
| *** haomaiwa_ has joined #openstack-swift | 09:01 | |
| *** arthus has joined #openstack-swift | 09:05 | |
| *** jordanP has quit IRC | 09:12 | |
| openstackgerrit | Michael Barton proposed openstack/swift: go: object engine abstraction layer https://review.openstack.org/278932 | 09:15 |
| *** mvk has joined #openstack-swift | 09:27 | |
| openstackgerrit | Ondřej Nový proposed openstack/swift: Bump PyECLib requirement to >= 1.2.0 https://review.openstack.org/278944 | 09:27 |
| *** jaakkos has joined #openstack-swift | 09:46 | |
| *** acoles_ is now known as acoles | 09:52 | |
| *** baojg_ has joined #openstack-swift | 09:58 | |
| *** baojg has quit IRC | 09:58 | |
| *** haomaiwa_ has quit IRC | 10:01 | |
| *** haomaiwang has joined #openstack-swift | 10:01 | |
| *** mingyu_ has joined #openstack-swift | 10:03 | |
| *** mingyu_ has quit IRC | 10:03 | |
| *** stantonnet_ has quit IRC | 10:03 | |
| *** stantonnet has joined #openstack-swift | 10:09 | |
| jith_ | hi all.. i have reduced the replica count... but still i couldnt find the free space in disk | 10:10 |
| jith_ | mattoliverau: . i have reduced the replica count... but still i couldnt find the space changes in disk | 10:21 |
| jith_ | altered the device weight too | 10:21 |
| ahale | jith_: it'll take some time for the replication to work through the change after rebalance and ring deploy to storage. maybe you could speed it up changing handoffs_first and handoff_delete options. kinda scary doing that on 2 replicas tho | 10:36 |
| *** venkat has quit IRC | 10:46 | |
| acoles | onovy: do you mind if I push a small change to add a comment to explain the unit test in patch 258158? then I can +2 ... | 10:48 |
| patchbot | acoles: https://review.openstack.org/#/c/258158/ - swift - Keystone middleware deprecated option is_admin rem... | 10:48 |
| onovy | acoles: of course | 10:50 |
| onovy | go for it | 10:50 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Keystone middleware deprecated option is_admin removed https://review.openstack.org/258158 | 10:53 |
| acoles | onovy: tell me if you don't agree ^^ | 10:53 |
| *** mvk has quit IRC | 10:54 | |
| onovy | https://review.openstack.org/#/c/258158/5..6/test/unit/common/middleware/test_keystoneauth.py // so this change, right? | 10:55 |
| patchbot | onovy: https://review.openstack.org/#/c/258158/ - swift - Keystone middleware deprecated option is_admin rem... | 10:55 |
| onovy | acoles: +1, thanks | 10:57 |
| *** arthus has quit IRC | 10:58 | |
| *** haomaiwang has quit IRC | 11:01 | |
| acoles | onovy: great, thanks for working on that | 11:01 |
| *** haomaiwa_ has joined #openstack-swift | 11:01 | |
| acoles | onovy: now I have to update my SAIO conf :) | 11:01 |
| *** arthus has joined #openstack-swift | 11:02 | |
| *** venkat has joined #openstack-swift | 11:07 | |
| pchng | acoles: When you have time, could you have a look at patch 266545? I made some changes but it's still a WIP and wanted some feedback before doing more. | 11:12 |
| patchbot | pchng: https://review.openstack.org/#/c/266545/ - swift (feature/crypto) - Ensure sysmeta is written/updated on (fast) POST | 11:12 |
| *** baojg_ has quit IRC | 11:12 | |
| *** baojg has joined #openstack-swift | 11:16 | |
| acoles | pchng: ack | 11:18 |
| *** daemontool has joined #openstack-swift | 11:21 | |
| *** mvk has joined #openstack-swift | 11:26 | |
| *** jmccarthy has quit IRC | 11:27 | |
| *** jmccarthy has joined #openstack-swift | 11:28 | |
| jith_ | ahale: thanks | 11:29 |
| jith_ | ahale: is there any way to find that swift cluster is under the process of data movement or rebalancing?? | 11:39 |
| *** haomaiwa_ has quit IRC | 12:01 | |
| *** haomaiwang has joined #openstack-swift | 12:01 | |
| *** asettle has quit IRC | 12:05 | |
| *** McMurlock has quit IRC | 12:08 | |
| *** McMurlock2 has joined #openstack-swift | 12:08 | |
| *** ig0r_ has joined #openstack-swift | 12:12 | |
| venkat | How many concurrent requests are handled by the proxy? | 12:12 |
| venkat | if workers= 1 | 12:12 |
| venkat | hi, someone can answer | 12:19 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Add comments to role-based access control functional tests https://review.openstack.org/279015 | 12:28 |
| *** bkumar4 has quit IRC | 12:30 | |
| cschwede | notmyname: clayg: i reviewed the first rbac test this morning, and approved it. i added a follow up, based on the discussion yesterday to improve readability ^^ | 12:31 |
| acoles | cschwede: good work | 12:32 |
| cschwede | acoles: thx :) | 12:33 |
| cschwede | i starred the follow up patches, will review them in the next few days too | 12:33 |
| cschwede | and probably add some more doc lines | 12:33 |
| acoles | :) | 12:33 |
| acoles | cschwede: you must have that 11-tuple mapping burnt into your memory now :) | 12:35 |
| venkat | How many concurrent requests are handled by the proxy? | 12:35 |
| venkat | if workers= 1 | 12:35 |
| venkat | hi, someone can answer | 12:35 |
| cschwede | acoles: definetely, and i see „None, None, None“ everywhere ;) | 12:36 |
| Mingyu | venkat: 1024 by default | 12:38 |
| *** daemontool has quit IRC | 12:39 | |
| *** daemontool has joined #openstack-swift | 12:39 | |
| Mingyu | <venkat> you can change this value via the option "max_clients" at etc/proxy-server.conf | 12:39 |
| Mingyu | venkat: you can change this value via the option "max_clients" at etc/proxy-server.conf | 12:41 |
| *** arthus has quit IRC | 12:42 | |
| *** arthus has joined #openstack-swift | 12:47 | |
| *** daemontool_ has joined #openstack-swift | 12:49 | |
| *** daemontool__ has joined #openstack-swift | 12:49 | |
| *** daemontool has quit IRC | 12:51 | |
| *** venkat has quit IRC | 12:51 | |
| *** daemontool_ has quit IRC | 12:54 | |
| *** haomaiwang has quit IRC | 13:01 | |
| *** bkumar4 has joined #openstack-swift | 13:01 | |
| *** haomaiwang has joined #openstack-swift | 13:01 | |
| *** bkumar4 has quit IRC | 13:01 | |
| *** bkumar4 has joined #openstack-swift | 13:02 | |
| *** venkat has joined #openstack-swift | 13:03 | |
| daemontool__ | hi anyone around? | 13:06 |
| daemontool__ | I wanted to get some feedback on a bug we had https://launchpad.net/bugs/1537364 | 13:06 |
| openstack | Launchpad bug 1537364 in Freezer "freezer-swift-restore-error" [Critical,Fix released] - Assigned to Fausto Marzi (fausto-marzi) | 13:06 |
| daemontool__ | it looks like the get_object method now accept only str as object name | 13:06 |
| daemontool__ | in python-swiftclient==2.7.0 | 13:06 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Add comments to role-based access control functional tests https://review.openstack.org/279015 | 13:07 |
| daemontool__ | before that version the object name was converted as string | 13:07 |
| daemontool__ | by the module | 13:07 |
| daemontool__ | just wanted to ask if that is an intended behaviour | 13:07 |
| *** delattec has joined #openstack-swift | 13:10 | |
| *** arthus has quit IRC | 13:10 | |
| *** bkumar4 has quit IRC | 13:10 | |
| *** delatte has quit IRC | 13:13 | |
| *** nadeem has joined #openstack-swift | 13:19 | |
| *** nadeem has quit IRC | 13:19 | |
| *** venkat has quit IRC | 13:19 | |
| *** esker has joined #openstack-swift | 13:20 | |
| *** nadeem has joined #openstack-swift | 13:20 | |
| *** arthus has joined #openstack-swift | 13:22 | |
| *** daemontool__ has quit IRC | 13:23 | |
| *** delatte has joined #openstack-swift | 13:32 | |
| *** delattec has quit IRC | 13:35 | |
| *** daemontool__ has joined #openstack-swift | 13:35 | |
| *** daemontool_ has joined #openstack-swift | 13:42 | |
| *** daemontool__ has quit IRC | 13:46 | |
| *** links has quit IRC | 13:48 | |
| *** daemontool_ is now known as daemontool | 13:56 | |
| *** arthus has quit IRC | 13:57 | |
| *** haomaiwang has quit IRC | 14:01 | |
| *** haomaiwang has joined #openstack-swift | 14:01 | |
| *** ppai has quit IRC | 14:06 | |
| *** trifon has quit IRC | 14:06 | |
| *** esker has quit IRC | 14:08 | |
| *** sanchitmalhotra has quit IRC | 14:08 | |
| *** arthus has joined #openstack-swift | 14:10 | |
| tdasilva | good morning | 14:17 |
| *** esker has joined #openstack-swift | 14:18 | |
| *** esker has quit IRC | 14:23 | |
| *** links has joined #openstack-swift | 14:27 | |
| *** baojg has quit IRC | 14:37 | |
| *** proteusguy_ has quit IRC | 14:49 | |
| *** dmorita has joined #openstack-swift | 14:55 | |
| *** dmorita has quit IRC | 14:55 | |
| *** haomaiwang has quit IRC | 15:01 | |
| *** haomaiwang has joined #openstack-swift | 15:01 | |
| onovy | notmyname: hi. did you received my email about debian CVE pls? | 15:01 |
| *** proteusguy_ has joined #openstack-swift | 15:07 | |
| openstackgerrit | Merged openstack/swift: Add functional test for access control (RBAC) with Keystone https://review.openstack.org/202411 | 15:07 |
| *** _JZ_ has joined #openstack-swift | 15:07 | |
| acoles | ho ho_away ^^ congrats! | 15:09 |
| acoles | oh ho has no bouncer :( | 15:09 |
| acoles | tdasilva: hi, i have a question about versioned_writes config, specifically this statement https://github.com/openstack/swift/blob/master/swift/common/middleware/versioned_writes.py#L60 | 15:10 |
| tdasilva | acoles: yeah, that was kind of messy | 15:10 |
| tdasilva | but what's your question? | 15:11 |
| acoles | tdasilva: our product guys are concerned that at some point the allow_versions option may be retired and some mass reset of metadata would be required - was there any plan to ever deprecate it? | 15:11 |
| acoles | tdasilva: wondering if we should force users to reset containers now, or just leave allow_versions in place forever? | 15:11 |
| tdasilva | acoles: yeah, the idea is that in the future it would be deprecated. the problem is that we typically don't like to force users "to do" things when they upgrade versions | 15:12 |
| tdasilva | upgrades should just work without much need for configuration, right? | 15:12 |
| acoles | tdasilva: ah :/ so there is a moment in time in the future when users have to reset x-versions-location?? is there a problem with leaving allow_versions in place (apart from it looking like an anomaly I guess) | 15:13 |
| *** haypo has joined #openstack-swift | 15:14 | |
| acoles | wait, deprecated is ok, our concern is if deprecation is followed by removal of the feature so that only containers with sysmeta versions location would work | 15:14 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 15:15 |
| openstackgerrit | Alistair Coles proposed openstack/python-swiftclient: Add test for --debug taking precedence over --info https://review.openstack.org/279132 | 15:16 |
| haypo | cschwede, acoles, torgomatic, clayg, notmyname, timburke : hi! last call for python 3 :-) would you mind reviewing https://review.openstack.org/#/c/236998/ & https://review.openstack.org/#/c/237027/ ? | 15:16 |
| patchbot | haypo: https://review.openstack.org/#/c/236998/ - swift - Port get_hmac() and hash_path() to Python 3 | 15:16 |
| patchbot | haypo: https://review.openstack.org/#/c/237027/ - swift - Port parse_mime_headers() to Python 3 | 15:16 |
| haypo | otherwise, i will abandon them next, as announced on openstack-dev, and stop bugging you with my py3 patches ;-) | 15:16 |
| tdasilva | acoles: hmmm..I guess I don't see the difference | 15:17 |
| *** mragupat has joined #openstack-swift | 15:18 | |
| *** arthus has quit IRC | 15:19 | |
| acoles | haypo: i'll try! | 15:20 |
| *** klrmn has joined #openstack-swift | 15:20 | |
| *** ig0r_ has quit IRC | 15:21 | |
| haypo | acoles: cool :) | 15:21 |
| acoles | tdasilva: 'deprecated' suggests 'you should no longer use this but it still works' which doesn't really apply here because you must use allow_versions for versioning of "older" container to still work | 15:22 |
| acoles | tdasilva: 'removed' to me implies it no longer works, which in this case would mean that allow_versions is ignored and container server never returns x-versions-location headers from an older versions container | 15:23 |
| acoles | tdasilva: ^^ that is what I fear and would prefer to never happen | 15:23 |
| *** klrmn has quit IRC | 15:25 | |
| tdasilva | acoles: sounds to me like the "best" solution would be to remove the option, while still allowing the code to check if anything is set in "x-versions-location" | 15:27 |
| tdasilva | what do you think? | 15:27 |
| *** garthb has joined #openstack-swift | 15:31 | |
| acoles | tdasilva: heh - I was wondering why that was not done before? is there an issue with an existing container that has x-versions-location but its server is config'd with allow_versions=False, so the container is not versioned, but then if we stop checking allow_versions and return x-versions-location for that container, it suddenly starts to do versioning? | 15:31 |
| *** petertr7 is now known as petertr7_away | 15:31 | |
| *** daemontool_ has joined #openstack-swift | 15:32 | |
| acoles | tdasilva: i wish I had paid more attention when you guys were figuring out the backwards compatibilty :) | 15:32 |
| *** daemontool has quit IRC | 15:34 | |
| *** garthb has quit IRC | 15:35 | |
| *** garthb has joined #openstack-swift | 15:35 | |
| *** zul_ has joined #openstack-swift | 15:36 | |
| *** zul has quit IRC | 15:36 | |
| *** lcurtis has joined #openstack-swift | 15:49 | |
| *** clyps_ has quit IRC | 15:53 | |
| *** Zyric has quit IRC | 15:53 | |
| *** etienneme has quit IRC | 15:53 | |
| *** EmilienM has quit IRC | 15:53 | |
| *** HenryG has quit IRC | 15:55 | |
| *** haypo has quit IRC | 15:55 | |
| *** Zyric has joined #openstack-swift | 15:57 | |
| *** bjkeller has joined #openstack-swift | 15:58 | |
| *** fthiagogv has joined #openstack-swift | 15:59 | |
| *** jordanP has joined #openstack-swift | 16:00 | |
| *** haomaiwang has quit IRC | 16:01 | |
| *** haomaiwa_ has joined #openstack-swift | 16:01 | |
| *** HenryG has joined #openstack-swift | 16:02 | |
| *** vstinner has joined #openstack-swift | 16:02 | |
| *** links has quit IRC | 16:04 | |
| *** clyps__ has joined #openstack-swift | 16:06 | |
| *** EmilienM has joined #openstack-swift | 16:06 | |
| *** etienneme has joined #openstack-swift | 16:08 | |
| *** trifon has joined #openstack-swift | 16:09 | |
| *** diazjf has joined #openstack-swift | 16:18 | |
| *** garthb has quit IRC | 16:20 | |
| tdasilva | acoles: the initial goal was to not require users to have to touch the config files when they did an upgrade | 16:20 |
| *** diazjf1 has joined #openstack-swift | 16:20 | |
| *** geaaru has quit IRC | 16:22 | |
| tdasilva | acoles: I think as part of "deprecating" allow_versions, would be to have users set 'allow_versioned_writes' to either true or false | 16:22 |
| *** diazjf has quit IRC | 16:22 | |
| *** mvk has quit IRC | 16:24 | |
| acoles | tdasilva: understood, so the preferred way to enable/disable versioning is 'allow_versioned_writes' | 16:26 |
| *** thumpba has joined #openstack-swift | 16:26 | |
| *** jordanP has quit IRC | 16:27 | |
| tdasilva | acoles: yes, ideally that would be the only method of enable and disable, but we have 'allow_versions' for older configurations, and the problem currently is that to check the container metadata, allow_versions must be enabled | 16:28 |
| tdasilva | so we could change that part | 16:28 |
| acoles | tdasilva: yep, was just typing the same | 16:28 |
| notmyname | good morning | 16:29 |
| tdasilva | need to think more about it, but my guess is that we could change the container code to "always" return x-versions-location | 16:29 |
| tdasilva | or maybe make that dependent on 'allow_versioned_writes' | 16:29 |
| tdasilva | notmyname: hello! :) | 16:29 |
| notmyname | cschwede: yay! (rbac) | 16:30 |
| notmyname | onovy: no, I haven't checked emails yet. IRC first ;-) | 16:30 |
| acoles | but if we change it so that container servers *always* return x-versions-location then there might be a container with x-versions-location set that currently is 'disabled' by allow_versions=False, then suddenly becomes enabled. hmmm... | 16:31 |
| tdasilva | acoles: so yeah, that what I was thinking that we would have to force users to set 'allow_versioned_writes' | 16:32 |
| acoles | tdasilva: well, they may already have allow_versioned_writes=True and new containers with sysmeta versions location are enabled, but allow_versions=False is disabling older containers, so we cannot get rid of allow_versions without possibly enabling previously disabled containers. But...arguably that's ok, versioning is enabled for all containers or its not, in some cases it could just lead to surprises. | 16:37 |
| *** cppforlife_ has quit IRC | 16:38 | |
| *** arch-nemesis has joined #openstack-swift | 16:38 | |
| acoles | tdasilva: anyway, my initial concern was whether there was any plan to one day make container servers NEVER return x-versions-location and force clients to reset x-versions-location. That is what concerns our product guys. That would be a bad thing. But sounds like that was not the "plan"? | 16:39 |
| *** Mingyu has quit IRC | 16:39 | |
| *** nottrobin has quit IRC | 16:39 | |
| acoles | notmyname: good morning | 16:39 |
| *** wasmum has joined #openstack-swift | 16:39 | |
| *** chsc has joined #openstack-swift | 16:41 | |
| acoles | notmyname: and other hackathon attendees - FYI we will be providing a bus between Marriott and HPE site at start and end of each day | 16:41 |
| *** esker has joined #openstack-swift | 16:42 | |
| *** nadeem has quit IRC | 16:42 | |
| *** nottrobin has joined #openstack-swift | 16:43 | |
| *** cppforlife_ has joined #openstack-swift | 16:43 | |
| *** Mingyu has joined #openstack-swift | 16:44 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Decrypt multipart mime doc https://review.openstack.org/248377 | 16:46 |
| *** dmorita has joined #openstack-swift | 16:47 | |
| openstackgerrit | Ankur proposed openstack/swift: Docstring omission in class BaseDiskFileManager. https://review.openstack.org/278615 | 16:48 |
| tdasilva | acoles: wow, thanks! | 16:48 |
| tdasilva | acoles: planning to look at patch 260179 today, I also need to go back a couple of patchsets and answer some of the questions there | 16:49 |
| patchbot | tdasilva: https://review.openstack.org/#/c/260179/ - swift - decouple versioned writes from COPY | 16:49 |
| acoles | tdasilva: i heard some stories of nova people getting lost during their midcycle here... | 16:49 |
| tdasilva | acoles: they couldn't process the direction? | 16:50 |
| acoles | tdasilva: cool. you may want to refer to patch 272724 to understand some of the changes I made in 260179 | 16:50 |
| patchbot | acoles: https://review.openstack.org/#/c/272724/ - swift - Copy headers correctly when copying object | 16:50 |
| tdasilva | acoles: sorry, that was my lame attempt at a pun | 16:50 |
| acoles | tdasilva: heh. apparently google directions gets it wrong | 16:51 |
| tdasilva | what's the distance? | 16:52 |
| acoles | approx 5 miles | 16:52 |
| acoles | but nova used a hotel about 1.5 miles way | 16:52 |
| tdasilva | mmm | 16:53 |
| tdasilva | ok | 16:53 |
| acoles | tdasilva: there is approximately nothing to do in evening within 1.5 miles of our site ;) | 16:54 |
| acoles | so we chose to use a hotel in centre | 16:55 |
| *** ankurgupta1 has joined #openstack-swift | 16:55 | |
| vinsh | So.. cinder volume backup- deletes.. they don't work well when using write_affinity/read_affnity in a multi region swift cluster. | 16:55 |
| vinsh | Wonder if this is known or I should file a bug. | 16:55 |
| * vinsh searches | 16:55 | |
| tdasilva | acoles: makes sense, how's the weather been? | 16:55 |
| *** rcernin has quit IRC | 16:56 | |
| tdasilva | acoles: early in the week there was a powerful wind storm there, right? | 16:56 |
| *** daemontool_ has quit IRC | 16:57 | |
| *** jmccarthy1 has joined #openstack-swift | 16:57 | |
| *** esker has quit IRC | 16:58 | |
| acoles | tdasilva: yes. so far we have had a relatively warm and very wet winter. right now its clear and cooler. | 16:58 |
| Mingyu | hi vinsh: what did you mean by "don't work well" could you plz give more information about it? | 17:00 |
| *** haomaiwa_ has quit IRC | 17:01 | |
| vinsh | Well, it seems that the cinder volume backup "delete" doesn't complete. Some objects could reamin.. and the backup container remains. If you remove the write-affinity from the proxy node.. the backup is deleted almost immediately without issue. | 17:01 |
| vinsh | This would be 4 zones. 2 regions. (2 zones/region) | 17:01 |
| *** haomaiwang has joined #openstack-swift | 17:01 | |
| *** esker has joined #openstack-swift | 17:02 | |
| Mingyu | There seems a known issue about that, but I'm not sure whether it is a bug or not. | 17:03 |
| vinsh | Yeah, Its been bothering my QA team for a long time. Figured I should get to the bottom of it :) | 17:03 |
| vinsh | I don't understand exactly why it occurs. | 17:03 |
| *** cbartz has quit IRC | 17:04 | |
| vinsh | I finally deployed 2.5.0 and it still exits, so its on my list to look at. | 17:04 |
| Mingyu | If you "delete" something in a multi-region cluster with write affinity, object still can be read for a short while. | 17:04 |
| Mingyu | for sometimes. | 17:05 |
| vinsh | Right... inconsistent. Sometimes object segments can remain. | 17:05 |
| *** gyee has joined #openstack-swift | 17:05 | |
| *** dmorita has quit IRC | 17:06 | |
| vinsh | I wonder if write_affinity is really benefiting our deployment that much anyway. Our cross region link is pretty fast, might not be needed. | 17:06 |
| Mingyu | Does it occurs only when you create the object and then delete it very soon? | 17:06 |
| *** esker has quit IRC | 17:07 | |
| vinsh | Yes, that is the common test case QA would hit it with. Vs a typical backup would not be deleted so quickly after creation. | 17:07 |
| *** petertr7_away is now known as petertr7 | 17:07 | |
| vinsh | where very soon == <45 seconds | 17:08 |
| vinsh | My worry is write_affinity is penalizing us in other places I don't know about. :) | 17:08 |
| Mingyu | So whether to use write affinity depends on your requirements. IMO, it is good for typical backup scenario, but if in the tests, the backup objects are deleted soon, it maybe problematic. | 17:10 |
| vinsh | a typical backup scenario would be where its not read immediately right? | 17:11 |
| vinsh | Vs some user requiring read right after write - cross region. | 17:11 |
| Mingyu | not *deleted* immediately. | 17:11 |
| vinsh | Ah. ya | 17:12 |
| vinsh | I wonder, would more frequent container-updater runs or something help this? | 17:12 |
| *** patchbot has quit IRC | 17:14 | |
| *** esker has joined #openstack-swift | 17:14 | |
| *** esker has joined #openstack-swift | 17:14 | |
| *** patchbot has joined #openstack-swift | 17:14 | |
| *** rledisez has quit IRC | 17:15 | |
| Mingyu | vinsh: I though there was a talk at the Tokyo Summit may help you. | 17:15 |
| *** lyrrad has joined #openstack-swift | 17:15 | |
| Mingyu | thought | 17:15 |
| vinsh | I tried to attend as many as I could :) | 17:15 |
| vinsh | I'll go look at the youtubes.. I think I know what you are talking about.. some "lessons learned " one. | 17:15 |
| notmyname | hello, again | 17:16 |
| vinsh | uh-oh. | 17:16 |
| vinsh | ;) | 17:16 |
| notmyname | acoles: aren't you getting us a bus or something so we don't get lost? | 17:16 |
| notmyname | vinsh: IMO, write affinity is something you almost never want | 17:16 |
| Mingyu | vinsh: yeah, this one: https://www.youtube.com/watch?v=zJmaLCxBLyA | 17:17 |
| vinsh | Ah-hah | 17:17 |
| acoles | notmyname: yes. 16:41:57 ;-) | 17:17 |
| vinsh | notmyname: It sounded like a good thing.. but in the end.. I lean towards your read of it. | 17:17 |
| acoles | 08:41:57 for you I guess | 17:17 |
| vinsh | Mingyu: Thank you :) | 17:17 |
| Mingyu | you're welcome | 17:18 |
| *** jmccarthy1 has joined #openstack-swift | 17:19 | |
| *** dmorita has joined #openstack-swift | 17:19 | |
| *** klrmn has joined #openstack-swift | 17:20 | |
| Mingyu | notmyname: Is write affinity so bad? | 17:21 |
| notmyname | well..."bad" is such a strong word | 17:21 |
| notmyname | ;-) | 17:21 |
| cschwede | Mingyu: depends on your incoming traffic and how fast your replication is | 17:22 |
| vinsh | Maybe more like "only desirable in very specific cases" | 17:22 |
| notmyname | write affinity explicitly forces swift to place your data on handoff nodes | 17:22 |
| notmyname | and it's really only useful if you've got bursty workloads where you can have enough network to move the data later (in the downtime between gaps of client activity) | 17:23 |
| *** dmorita has quit IRC | 17:23 | |
| vinsh | Man. that tokyo talk is spot on. +1 | 17:23 |
| vinsh | -1 for me not doing enough summit homework. | 17:23 |
| *** petertr7 is now known as petertr7_away | 17:23 | |
| Mingyu | haha. too many presentations there. | 17:24 |
| *** klrmn has quit IRC | 17:25 | |
| Mingyu | notmyname: So sometimes it is helpful, but most of the time ... | 17:26 |
| notmyname | Mingyu: IMO yes | 17:27 |
| Mingyu | I enabled write affinity in some deployments and also learned some lessons. I proposed an presentation for Austin Summit on multi-region clusters: https://www.openstack.org/summit/austin-2016/vote-for-speakers/presentation/8012 | 17:28 |
| vinsh | I'll vote for that! | 17:30 |
| Mingyu | Thank you :-) | 17:30 |
| *** esker has quit IRC | 17:31 | |
| Mingyu | Maybe we can have a full discussion on that topic in Austin then, if this proposal is selected. | 17:32 |
| *** esker has joined #openstack-swift | 17:33 | |
| *** arthus has joined #openstack-swift | 17:36 | |
| Mingyu | I agree that sometimes the affinity really helps but not in every multi-region swift cluster. | 17:36 |
| *** nadeem has joined #openstack-swift | 17:36 | |
| notmyname | right | 17:36 |
| *** nadeem has quit IRC | 17:36 | |
| *** nadeem has joined #openstack-swift | 17:37 | |
| *** esker has quit IRC | 17:39 | |
| *** esker has joined #openstack-swift | 17:39 | |
| *** chirag has joined #openstack-swift | 17:41 | |
| *** garthb has joined #openstack-swift | 17:42 | |
| *** nadeem has quit IRC | 17:42 | |
| vinsh | now. "read_affinity" is probably less controversial. | 17:42 |
| notmyname | yeah. read affinity is really great | 17:43 |
| Mingyu | especially for some internet applications where read >> write | 17:43 |
| chirag | I was uploading approx 800 parallel PUT req to mu swift cluster when one of my request failed. | 17:44 |
| chirag | File "/usr/lib64/python2.6/httplib.py", line 391, in begin#012 version, status, reason = self._read_status()#012 File "/usr/lib64/python2.6/httplib.py", line 355, in _read_status#012 raise BadStatusLine(line)#012BadStatusLine | 17:44 |
| *** esker has quit IRC | 17:45 | |
| chirag | can someone help me with this ? | 17:45 |
| Mingyu | vinsh: I guess your scenario is to backup in one datacenter and recover the data in a offsite disaster recovery center? | 17:45 |
| *** garthb has quit IRC | 17:45 | |
| *** garthb has joined #openstack-swift | 17:46 | |
| vinsh | Mingyu: My scenario is truly a wide range of users/use cases. DR is important.. but this same cluster backs web servers/content. SO cinder-voulume-backups from the rest of the openstack cluster.. as well as many other swift uses. | 17:46 |
| *** ankurgupta1 has quit IRC | 17:47 | |
| notmyname | onovy: ah. there's your email. sorry, I had missed it in my travel | 17:47 |
| *** vstinner has quit IRC | 17:48 | |
| notmyname | onovy: what version of swift is in jessie/stable? | 17:49 |
| *** dmorita has joined #openstack-swift | 17:50 | |
| *** nadeem has joined #openstack-swift | 17:50 | |
| Mingyu | vinsh: understood. | 17:51 |
| *** dmorita has quit IRC | 17:52 | |
| *** dmorita_ has joined #openstack-swift | 17:52 | |
| *** esker has joined #openstack-swift | 17:53 | |
| *** klrmn has joined #openstack-swift | 17:54 | |
| *** haomaiwang has quit IRC | 18:01 | |
| *** ChubYann has joined #openstack-swift | 18:01 | |
| *** haomaiwang has joined #openstack-swift | 18:01 | |
| openstackgerrit | Ben Keller proposed openstack/swift: Headers now use a case incensitive dictionary https://review.openstack.org/274906 | 18:04 |
| notmyname | nice to see this thought written down http://www.mikeperham.com/2016/02/09/kill-your-dependencies/ | 18:04 |
| *** esker has quit IRC | 18:10 | |
| *** rcernin has joined #openstack-swift | 18:16 | |
| *** ankurgupta1 has joined #openstack-swift | 18:21 | |
| *** esker has joined #openstack-swift | 18:21 | |
| *** bjkeller has quit IRC | 18:24 | |
| *** Jeffrey4l has quit IRC | 18:30 | |
| *** petertr7_away is now known as petertr7 | 18:33 | |
| *** nottrobin has quit IRC | 18:34 | |
| *** garthb has quit IRC | 18:37 | |
| *** garthb has joined #openstack-swift | 18:37 | |
| *** jlvillal has quit IRC | 18:38 | |
| *** jlvillal has joined #openstack-swift | 18:39 | |
| *** garthb has quit IRC | 18:48 | |
| *** garthb has joined #openstack-swift | 18:49 | |
| chirag | File "/usr/lib64/python2.6/httplib.py", line 391, in begin#012 version, status, reason = self._read_status()#012 File "/usr/lib64/python2.6/httplib.py", line 355, in _read_status#012 raise BadStatusLine(line)#012BadStatusLine | 18:49 |
| chirag | can someone help me with this ? | 18:49 |
| *** garthb has quit IRC | 18:50 | |
| *** esker has quit IRC | 18:51 | |
| *** garthb has joined #openstack-swift | 18:51 | |
| *** nottrobin has joined #openstack-swift | 18:52 | |
| *** gyee has quit IRC | 18:53 | |
| *** daemontool has joined #openstack-swift | 18:54 | |
| *** ig0r_ has joined #openstack-swift | 18:58 | |
| acoles | notmyname: clayg I'll try to finish review of patch 215867 tomorrow | 18:58 |
| patchbot | acoles: https://review.openstack.org/#/c/215867/ - swift - Make handoffs_first a more useful "mode" | 18:58 |
| *** ankurgupta1 has quit IRC | 18:59 | |
| openstackgerrit | Samuel Merritt proposed openstack/swift: Fix typos in comments https://review.openstack.org/279233 | 19:00 |
| *** zhill has joined #openstack-swift | 19:00 | |
| *** haomaiwang has quit IRC | 19:01 | |
| *** haomaiwang has joined #openstack-swift | 19:01 | |
| acoles | torgomatic: looks like maybe you will beat me to it | 19:02 |
| torgomatic | acoles: yep :) | 19:02 |
| acoles | torgomatic: looks like you did :) | 19:02 |
| * acoles strikes first item from tomorrow's todo list | 19:03 | |
| *** ig0r_ has quit IRC | 19:03 | |
| *** arthus has quit IRC | 19:05 | |
| *** bjkeller has joined #openstack-swift | 19:05 | |
| *** acoles is now known as acoles_ | 19:07 | |
| *** ig0r_ has joined #openstack-swift | 19:08 | |
| *** bgmccollum has joined #openstack-swift | 19:14 | |
| *** klrmn1 has joined #openstack-swift | 19:18 | |
| *** klrmn has quit IRC | 19:20 | |
| *** peterlisak has quit IRC | 19:32 | |
| *** peterlisak has joined #openstack-swift | 19:33 | |
| *** zaitcev has joined #openstack-swift | 19:35 | |
| *** ChanServ sets mode: +v zaitcev | 19:35 | |
| *** zhill has quit IRC | 19:49 | |
| *** kprabhuv has joined #openstack-swift | 19:50 | |
| *** trifon has quit IRC | 19:51 | |
| *** nadeem has quit IRC | 19:54 | |
| *** vinsh has quit IRC | 20:00 | |
| *** vinsh has joined #openstack-swift | 20:00 | |
| *** haomaiwang has quit IRC | 20:01 | |
| *** haomaiwa_ has joined #openstack-swift | 20:01 | |
| *** nadeem has joined #openstack-swift | 20:01 | |
| *** zhill has joined #openstack-swift | 20:02 | |
| *** vinsh_ has joined #openstack-swift | 20:07 | |
| *** vinsh has quit IRC | 20:07 | |
| *** esker has joined #openstack-swift | 20:09 | |
| *** dmorita_ has quit IRC | 20:18 | |
| *** dmorita has joined #openstack-swift | 20:20 | |
| openstackgerrit | Merged openstack/python-swiftclient: Fix debug and info option parsing https://review.openstack.org/252328 | 20:22 |
| openstackgerrit | Merged openstack/python-swiftclient: Add test for --debug taking precedence over --info https://review.openstack.org/279132 | 20:22 |
| *** rcernin has quit IRC | 20:24 | |
| openstackgerrit | Tim Burke proposed openstack/swift: By default, disallow inbound X-Timestamp headers https://review.openstack.org/279256 | 20:33 |
| timburke | notmyname: torgomatic: ^^ | 20:34 |
| openstackgerrit | Ankur proposed openstack/swift: Docstring omission in class BaseDiskFileManager. https://review.openstack.org/278615 | 20:35 |
| *** dmorita has quit IRC | 20:39 | |
| onovy | notmyname: 2.2.0 | 20:47 |
| onovy | 1.4.8 in old-stable :) | 20:47 |
| onovy | notmyname: https://security-tracker.debian.org/tracker/source-package/swift | 20:47 |
| *** daemontool_ has joined #openstack-swift | 20:52 | |
| *** daemontool has quit IRC | 20:55 | |
| openstackgerrit | Sarafraj Singh proposed openstack/swift: Fixed params https://review.openstack.org/279264 | 20:56 |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Fixed params on autocreate_account docstring https://review.openstack.org/279264 | 21:00 |
| *** haomaiwa_ has quit IRC | 21:01 | |
| *** haomaiwang has joined #openstack-swift | 21:01 | |
| clayg | redbo: *DiskFile* !? I'm in the twilight zone | 21:06 |
| tdasilva | lol | 21:07 |
| *** dmorita has joined #openstack-swift | 21:07 | |
| *** mfalatic has joined #openstack-swift | 21:08 | |
| klrmn1 | would trying to delete a non-empty container cause a 409 error? could trying to delete said container too-soon after deleting the last object run afoul of eventual consistency and return the 409? | 21:09 |
| onovy | tdasilva: cschwede: hi, can you look to https://review.openstack.org/#/c/256873/ and https://review.openstack.org/#/c/256875/ pls? | 21:11 |
| patchbot | onovy: https://review.openstack.org/#/c/256873/ - swift-ceph-backend - Deprecated tox -downloadcache option removed | 21:11 |
| patchbot | onovy: https://review.openstack.org/#/c/256875/ - swift-ceph-backend - .gitreview file | 21:11 |
| timburke | klrmn1: definitely yes on the first Q; probably yes on the second | 21:15 |
| mattoliverau | morning | 21:18 |
| notmyname | onovy: so I think it's cool you want to fix those bugs, but on the other hand, those aren't upstream supported versions. just curious, but how long is debian support for 1.4.8 and 2.2.0? | 21:20 |
| notmyname | mattoliverau: looks like your abandoner checker got some coffee this morning and woke up :-) | 21:21 |
| onovy | Debian 7 “Wheezy” from February 2016 to May 2018 | 21:21 |
| onovy | Debian 8 “Jessie“ from May 2018 to April/May 2020 | 21:21 |
| onovy | clayg: you can do +2+A before glob, and then write "recheck" here, after globs merge | 21:22 |
| *** pdardeau| has quit IRC | 21:22 | |
| notmyname | so that means that for the next 4 years we could be backporting stuff to swift 2.2.0 (which was released in 2014)? | 21:22 |
| clayg | onovy: "can" or "should"? | 21:23 |
| *** esker has quit IRC | 21:23 | |
| onovy | clayg: it's working, don't know what is "correct" :) | 21:24 |
| onovy | notmyname: from debian point, yes | 21:24 |
| onovy | stuff=CVE only | 21:24 |
| notmyname | onovy: what normally happens for other projects (non-openstack)? how do you deal with backports there? | 21:26 |
| onovy | notmyname: it's not backports. it's just security fixies | 21:26 |
| *** esker has joined #openstack-swift | 21:26 | |
| onovy | every maintainer tries to contact upstream if they can/want provide patch | 21:27 |
| clayg | klrmn1: yes 409 - yes if container is not updated it might "mistakenly" return 409 | 21:27 |
| onovy | if this is not possible, maitainer can try to do it, but it's not simple | 21:27 |
| notmyname | well, ok. but security fixes are backported patches from stuff that lands on master | 21:27 |
| onovy | yep :] | 21:27 |
| notmyname | so you ask the project to help you, if they say no you try it, and if that doesn't work, it stays unpatches? | 21:27 |
| notmyname | *unpatched | 21:28 |
| clayg | klrmn1: "good news" is that when the async updates get processed and it turns out there is no more objects in the container - there's a non-zero change the container will "poof" (despite the proxy returning 409 based on quorum) | 21:28 |
| *** silor has quit IRC | 21:28 | |
| *** trifon has joined #openstack-swift | 21:31 | |
| *** esker has quit IRC | 21:31 | |
| mattoliverau | notmyname: ? If something strange happened, rackspace probably decided to reboot the cloud server the abandoner sits on. | 21:31 |
| onovy | notmyname: yep, it stays unpatched then | 21:33 |
| notmyname | mattoliverau: ah. I just saw about 2 dozen emails just now | 21:33 |
| notmyname | onovy: hmmm | 21:33 |
| onovy | it's community distro | 21:33 |
| notmyname | onovy: so right now, with swift and this issue, you've done a basic backport attempt, but without some upstream help to get it cleanly merged, it will stay unpatched. right? | 21:34 |
| onovy | notmyname: right | 21:34 |
| mattoliverau | notmyname: ahhh haven't checked the abandoner in a while, my bad. | 21:35 |
| notmyname | mattoliverau: no worries. me neither | 21:35 |
| onovy | notmyname: btw: zigo told me you are open to do this backports :) | 21:37 |
| notmyname | onovy: oh, well then I better get started ;-) | 21:37 |
| onovy | something about Atlanta at SwiftStack party :) | 21:37 |
| onovy | notmyname: it's up to you, i know it's EOL | 21:38 |
| notmyname | onovy: I'm interested in seeing that bug fixes in a distro release are patched | 21:39 |
| onovy | perfect :) | 21:39 |
| notmyname | that's not exactly the same thing as "I'll backport any fixes to whatever release you've got in your distro" ;-) | 21:39 |
| onovy | :)) | 21:40 |
| *** esker has joined #openstack-swift | 21:46 | |
| *** thumpba has quit IRC | 21:54 | |
| *** daemontool_ has quit IRC | 21:56 | |
| *** garthb has quit IRC | 21:57 | |
| *** garthb has joined #openstack-swift | 21:58 | |
| *** pgbridge has quit IRC | 21:58 | |
| *** delatte has quit IRC | 22:00 | |
| *** haomaiwang has quit IRC | 22:01 | |
| *** haomaiwang has joined #openstack-swift | 22:01 | |
| *** esker has quit IRC | 22:02 | |
| *** chlong has quit IRC | 22:03 | |
| *** esker has joined #openstack-swift | 22:03 | |
| *** pgbridge has joined #openstack-swift | 22:04 | |
| tdasilva | onovy: re the swift-cephbackend patches, there's an issue with that project regarding the use of librados | 22:08 |
| onovy | which is not related to my change, and gating is working :) | 22:08 |
| tdasilva | onovy: when storing "swift" objects on rados, those objects need to be striped. they should not be stored "whole" as we typically do in swift | 22:09 |
| tdasilva | rados likes objects that are about 4mb in size | 22:10 |
| *** trifon has quit IRC | 22:10 | |
| tdasilva | so the current implementation there is incorrect | 22:10 |
| openstackgerrit | Ondřej Nový proposed openstack/swift: Additional error log message for drive-audit https://review.openstack.org/263691 | 22:11 |
| tdasilva | onovy: it's not related to your patch, but i'm just trying to give you a heads up | 22:11 |
| tdasilva | onovy: i've been playing with a different implementation here: https://github.com/thiagodasilva/swift/tree/bird | 22:12 |
| onovy | ah :) | 22:12 |
| tdasilva | that uses hummingbird and libradosstriper library | 22:12 |
| onovy | interesting | 22:12 |
| tdasilva | onovy: you can also checkout this: https://github.com/thiagodasilva/gorados | 22:13 |
| tdasilva | but it's all prototype code | 22:13 |
| onovy | hmmm, go everywhere :) | 22:14 |
| tdasilva | onovy: hehe, in this case it was nice to be able to just link libradosstriper: https://github.com/thiagodasilva/gorados/blob/master/stripedobject.go#L22 | 22:15 |
| onovy | ah. you are just wrapping C library? | 22:15 |
| onovy | "just" :) | 22:15 |
| tdasilva | right | 22:15 |
| tdasilva | and with go interfaces, it makes it all very easy to plug into the hummingbird code | 22:16 |
| onovy | looks cool | 22:16 |
| tdasilva | patches are welcome :) | 22:17 |
| onovy | btw: what is "usecase" for this kind of projects? (have other then file storage for swift) | 22:17 |
| onovy | *tahn | 22:17 |
| *** esker has quit IRC | 22:17 | |
| onovy | *than | 22:17 |
| *** petertr7 is now known as petertr7_away | 22:18 | |
| tdasilva | add the ability to store data on other backend storage systems using swift api | 22:18 |
| *** kprabhuv has quit IRC | 22:19 | |
| onovy | yes, that's description, but real world example? | 22:19 |
| tdasilva | onovy: here's a similar project: https://github.com/openstack/swiftonfile | 22:19 |
| tdasilva | in this case to store data on gluster, also used by ibm to store data on gpfs | 22:20 |
| onovy | so we have Swift REST API with replication, redundancy, etc. | 22:21 |
| onovy | so why store it in ceph? :) | 22:21 |
| tdasilva | one use case is for folks that already have a small ceph cluster and would like to store data in it using swift api | 22:22 |
| *** trifon has joined #openstack-swift | 22:22 | |
| onovy | ah, that's the point! :) | 22:22 |
| *** mragupat has quit IRC | 22:23 | |
| tdasilva | onovy: what's your use case? | 22:24 |
| onovy | not using it at all :) | 22:24 |
| onovy | i was just doing global changes in all OS projects | 22:25 |
| *** mragupat has joined #openstack-swift | 22:29 | |
| openstackgerrit | Tim Burke proposed openstack/swift: By default, disallow inbound X-Timestamp headers https://review.openstack.org/279256 | 22:29 |
| *** fthiagogv has quit IRC | 22:31 | |
| *** daemontool has joined #openstack-swift | 22:34 | |
| *** mfalatic has quit IRC | 22:34 | |
| *** cdelatte has joined #openstack-swift | 22:42 | |
| *** rickyrem has joined #openstack-swift | 22:44 | |
| *** arch-nemesis has quit IRC | 22:53 | |
| *** km has joined #openstack-swift | 22:58 | |
| *** zhill has quit IRC | 23:00 | |
| *** haomaiwang has quit IRC | 23:01 | |
| *** haomaiwang has joined #openstack-swift | 23:01 | |
| *** asettle has joined #openstack-swift | 23:02 | |
| *** rickyrem has quit IRC | 23:05 | |
| openstackgerrit | John Dickinson proposed openstack/swift-bench: authors/changelog update for 1.1. release https://review.openstack.org/279302 | 23:07 |
| *** mragupat has quit IRC | 23:11 | |
| *** rickyrem has joined #openstack-swift | 23:13 | |
| *** bjkeller has left #openstack-swift | 23:21 | |
| openstackgerrit | John Dickinson proposed openstack/swift-bench: version bump for 1.1. release https://review.openstack.org/279310 | 23:23 |
| openstackgerrit | John Dickinson proposed openstack/swift-bench: version bump to open 1.2 dev work https://review.openstack.org/279311 | 23:23 |
| *** _JZ_ has quit IRC | 23:24 | |
| notmyname | I wonder if this will work | 23:28 |
| *** chlong has joined #openstack-swift | 23:30 | |
| *** nadeem has quit IRC | 23:33 | |
| *** trifon has quit IRC | 23:35 | |
| *** kei_yama has joined #openstack-swift | 23:38 | |
| *** lcurtis has quit IRC | 23:41 | |
| *** ho_ has joined #openstack-swift | 23:42 | |
| *** chsc has quit IRC | 23:42 | |
| ho_ | good morning! | 23:44 |
| *** marcusvrn_ has quit IRC | 23:47 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!