| *** gyee has joined #openstack-swift | 00:02 | |
| *** hoonetorg has quit IRC | 00:22 | |
| *** dja has joined #openstack-swift | 00:29 | |
| *** catintheroof has joined #openstack-swift | 00:32 | |
| *** hoonetorg has joined #openstack-swift | 00:35 | |
| *** Renich_ has joined #openstack-swift | 00:37 | |
| Renich_ | timburke: btw, thanks for the help | 00:39 |
|---|---|---|
| *** Renich has quit IRC | 00:39 | |
| Renich_ | timburke: it didn't work... no idea what's up. Might try, later on, at #rackspace or something | 00:39 |
| Renich_ | bye | 00:39 |
| *** Renich_ has quit IRC | 00:39 | |
| *** Renich has joined #openstack-swift | 00:40 | |
| *** zhurong has joined #openstack-swift | 00:53 | |
| *** m_kazuhiro has joined #openstack-swift | 00:59 | |
| m_kazuhiro | good morning | 01:05 |
| *** Renich has quit IRC | 01:05 | |
| *** Renich has joined #openstack-swift | 01:06 | |
| notmyname | mattoliverau: nice! | 01:08 |
| mattoliverau | m_kazuhiro: morning | 01:08 |
| m_kazuhiro | mattoliverau: morning! | 01:14 |
| *** hoonetorg has quit IRC | 01:16 | |
| *** Renich has quit IRC | 01:20 | |
| *** catintheroof has quit IRC | 01:24 | |
| *** vint_bra has joined #openstack-swift | 01:27 | |
| tone_zrt | morning! | 01:28 |
| mattoliverau | tone_zrt: mornin | 01:30 |
| *** hoonetorg has joined #openstack-swift | 01:30 | |
| *** catintheroof has joined #openstack-swift | 01:30 | |
| notmyname | I'd like to welcome mahatic to swift core :-) | 01:31 |
| mattoliverau | \o/ | 01:31 |
| mattoliverau | mahatic: congrats \o/ | 01:31 |
| tone_zrt | mattoliverau: good afternoon :) | 01:31 |
| mattoliverau | ^ when you wake up :) | 01:32 |
| mahatic | mattoliverau: thanks, I'm up for a (very) early morning meeting internally :) | 01:32 |
| notmyname | yeah, it's pretty early for mahatic, but I wanted to say something for her morning and before tomorrow's team meeting when she'll be asleep | 01:32 |
| *** catintheroof has quit IRC | 01:32 | |
| *** JimCheung has quit IRC | 01:33 | |
| m_kazuhiro | tone_zrt: morning! | 01:35 |
| tone_zrt | m_kazuhiro: morning! | 01:35 |
| m_kazuhiro | mahatic: congrats! | 01:36 |
| mahatic | m_kazuhiro: thanks! | 01:38 |
| *** vint_bra has quit IRC | 01:51 | |
| *** vint_bra has joined #openstack-swift | 01:54 | |
| *** aluria has quit IRC | 01:56 | |
| *** sams-gleb has joined #openstack-swift | 01:57 | |
| *** vint_bra has quit IRC | 02:01 | |
| *** sams-gleb has quit IRC | 02:01 | |
| jrichli | mahatic: Congrats :-) | 02:04 |
| openstackgerrit | Thiago da Silva proposed openstack/swift master: Symlink implementation. https://review.openstack.org/232162 | 02:06 |
| clayg | mahatic: welcome to teh suck | 02:07 |
| mahatic | clayg: lol, thanks? | 02:07 |
| tdasilva | mahatic: congratulations! | 02:11 |
| *** tanee is now known as tanee_away | 02:14 | |
| *** dja has quit IRC | 02:14 | |
| *** tanee_away is now known as tanee | 02:14 | |
| *** dja has joined #openstack-swift | 02:16 | |
| *** sweeper has quit IRC | 02:40 | |
| *** dja has quit IRC | 02:40 | |
| *** dja has joined #openstack-swift | 02:42 | |
| *** SkyRocknRoll has quit IRC | 02:43 | |
| *** sweeper has joined #openstack-swift | 02:53 | |
| *** bkopilov has quit IRC | 02:58 | |
| *** gyee has quit IRC | 03:01 | |
| pdardeau_ | mahatic: congratulations!! :) | 03:07 |
| *** JimCheung has joined #openstack-swift | 03:10 | |
| *** JimCheung has quit IRC | 03:15 | |
| *** gyee has joined #openstack-swift | 03:17 | |
| *** redbo has quit IRC | 03:19 | |
| openstackgerrit | Janie Richling proposed openstack/swift master: Log the correct request type of a subrequest downstream of copy mw https://review.openstack.org/444604 | 03:20 |
| *** dmorita has quit IRC | 03:25 | |
| *** dmorita has joined #openstack-swift | 03:27 | |
| *** redbo has joined #openstack-swift | 03:27 | |
| *** ChanServ sets mode: +v redbo | 03:27 | |
| *** Jeffrey4l has quit IRC | 03:28 | |
| *** dmorita has quit IRC | 03:31 | |
| *** Jeffrey4l has joined #openstack-swift | 03:41 | |
| jrichli | how do you edit a commit message directly from gerrit? I can never find the button for that ... | 03:41 |
| clayg | jrichli: I always have to click on the commit message in the list of files | 03:42 |
| clayg | then next to the list of patch sets - there's a notepad/edit button | 03:42 |
| jrichli | clayg: thanks, i had done that, but wasn't on the latest patch. i see it now | 03:42 |
| *** m_kazuhiro has quit IRC | 03:43 | |
| openstackgerrit | Janie Richling proposed openstack/swift master: Log the correct request type of a subrequest downstream of copy https://review.openstack.org/444604 | 03:44 |
| *** dmorita has joined #openstack-swift | 03:50 | |
| *** dmorita has quit IRC | 03:54 | |
| *** links has joined #openstack-swift | 03:56 | |
| *** klrmn has quit IRC | 03:56 | |
| *** sams-gleb has joined #openstack-swift | 03:59 | |
| *** sams-gleb has quit IRC | 04:03 | |
| *** m_kazuhiro has joined #openstack-swift | 04:05 | |
| *** dmorita has joined #openstack-swift | 04:16 | |
| *** dmorita_ has joined #openstack-swift | 04:18 | |
| *** dmorita has quit IRC | 04:18 | |
| *** psachin has joined #openstack-swift | 04:20 | |
| *** dmorita_ has quit IRC | 04:22 | |
| *** dmorita has joined #openstack-swift | 04:23 | |
| *** dmorita_ has joined #openstack-swift | 04:25 | |
| *** dmorita has quit IRC | 04:27 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift master: Updated from global requirements https://review.openstack.org/88736 | 04:29 |
| *** dmorita_ has quit IRC | 04:29 | |
| *** gkadam has joined #openstack-swift | 04:45 | |
| *** psachin has quit IRC | 05:00 | |
| *** zhurong has quit IRC | 05:00 | |
| *** bkopilov has joined #openstack-swift | 05:05 | |
| kota_ | good afternoon | 05:06 |
| mattoliverau | kota_: good arvo :) | 05:06 |
| kota_ | mahatic: welcome to the swift core team :-D | 05:07 |
| kota_ | mattoliverau: o/ | 05:10 |
| *** dja has quit IRC | 05:16 | |
| *** psachin has joined #openstack-swift | 05:17 | |
| *** JimCheung has joined #openstack-swift | 05:27 | |
| mahatic | jrichli: tdasilva pdardeau_ kota_ thank you! | 05:28 |
| *** psachin has quit IRC | 05:49 | |
| *** gyee has quit IRC | 05:51 | |
| *** JimCheung has quit IRC | 05:55 | |
| *** JimCheung has joined #openstack-swift | 05:55 | |
| *** psachin has joined #openstack-swift | 05:57 | |
| *** adriant has quit IRC | 05:59 | |
| *** JimCheung has quit IRC | 06:00 | |
| *** sams-gleb has joined #openstack-swift | 06:01 | |
| *** cshastri has joined #openstack-swift | 06:05 | |
| *** sams-gleb has quit IRC | 06:05 | |
| clayg | notmyname: look I typed things! Moar Better Faster Rebalance -- https://etherpad.openstack.org/p/swift-rebalance | 06:24 |
| *** ChubYann has quit IRC | 06:28 | |
| *** dmorita has joined #openstack-swift | 06:43 | |
| mattoliverau | clayg: nice.. but also your realise there is a high chance that _any_ temp name may end up being the actual name.. long live tsync ;P | 06:47 |
| *** dmorita has quit IRC | 06:48 | |
| *** zhurong has joined #openstack-swift | 06:58 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: TestObjectController refactoring https://review.openstack.org/440466 | 07:07 |
| *** m_kazuhiro has quit IRC | 07:07 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift master: Fix (un)patch_policies https://review.openstack.org/440936 | 07:10 |
| kota_ | just rebased | 07:11 |
| *** tonanhngo_ has joined #openstack-swift | 07:12 | |
| *** tonanhngo has quit IRC | 07:14 | |
| *** gkadam is now known as gkadam-lunch | 07:29 | |
| *** hseipp has joined #openstack-swift | 07:39 | |
| *** silor has joined #openstack-swift | 07:40 | |
| *** sams-gleb has joined #openstack-swift | 07:41 | |
| *** silor1 has joined #openstack-swift | 07:43 | |
| *** tesseract has joined #openstack-swift | 07:44 | |
| *** silor has quit IRC | 07:46 | |
| *** silor1 is now known as silor | 07:46 | |
| *** tonanhngo_ has quit IRC | 07:47 | |
| *** SkyRocknRoll has joined #openstack-swift | 07:50 | |
| *** silor has quit IRC | 07:50 | |
| *** tonanhngo has joined #openstack-swift | 07:53 | |
| *** tonanhngo has quit IRC | 07:54 | |
| *** SkyRocknRoll has quit IRC | 07:56 | |
| *** pcaruana has joined #openstack-swift | 08:05 | |
| *** jcaron has joined #openstack-swift | 08:15 | |
| *** zhurong has quit IRC | 08:28 | |
| *** d0ugal has joined #openstack-swift | 08:32 | |
| *** geaaru has joined #openstack-swift | 08:35 | |
| *** amoralej|off is now known as amoralej | 08:37 | |
| acoles | good morning | 08:43 |
| acoles | mattoliverau: great write up! | 08:45 |
| acoles | mahatic: congrats and welcome to the team :D | 08:45 |
| acoles | mattoliverau: I have a script that installs the swift test users into keystone. It may be a little dated but I used it recently. https://gist.github.com/alistairncoles/b5b905f7f68d82ae242e59982ed902bd | 08:47 |
| acoles | tdasilva: did I promise you that link ^^ a long time ago ? | 08:48 |
| *** dja has joined #openstack-swift | 08:55 | |
| *** tonanhngo has joined #openstack-swift | 08:56 | |
| *** tonanhngo has quit IRC | 08:57 | |
| *** kei_yama has quit IRC | 08:58 | |
| *** cbartz has joined #openstack-swift | 09:01 | |
| acoles | kota_: thanks for your comments on the composite ring etherpad | 09:09 |
| kota_ | acoles: yeah, i'm still now writing though | 09:09 |
| mahatic | acoles: thanks :D | 09:13 |
| *** furlongm has joined #openstack-swift | 09:20 | |
| *** furlongm has quit IRC | 09:20 | |
| mattoliverau | acoles: cool, mind if I steal that? ;) | 09:27 |
| acoles | mattoliverau: nope, go ahead | 09:27 |
| *** zhurong has joined #openstack-swift | 09:28 | |
| acoles | mattoliverau: there is also this which includes setting up the endpoints, services etc https://gist.github.com/alistairncoles/eace39b1e19c6ce708edd8cd9e951ff2 | 09:30 |
| acoles | but not the keystone install/uwsgi goodness you have! | 09:30 |
| acoles | kota_: (brainstorming...) do we need a different ring builder file suffix/type for composite ring? could we add keys to the builder dict that is pickled and use those to differentiate composite from normal ring when loading the builder file? | 09:33 |
| *** gkadam-lunch is now known as gkadam | 09:34 | |
| kota_ | acoles: making a brand-new builder file? ah maybe like as `swift-ring-builder composite-ring.builder zip composite-ring.json` | 09:35 |
| kota_ | or, `swift-ring-builder composite-ring.builder add_ring <some ring>`? | 09:36 |
| *** zhurong has quit IRC | 09:38 | |
| acoles | kota_: yeah, something like that. TBH I'm not sure what clayg had in mind (he'll tell me I am talking nonsense!) | 09:39 |
| acoles | but my thinking is that composite-ring.builder could maintain list of component rings | 09:40 |
| kota_ | ah, so we will maintaine 2 types of builder file with same suffix | 09:40 |
| *** tonanhngo has joined #openstack-swift | 09:41 | |
| acoles | well, we add keys to the dicts that are serialized to builder file. but yes, effectively two kinds of builder files. | 09:41 |
| *** tonanhngo has quit IRC | 09:41 | |
| *** zhurong has joined #openstack-swift | 09:43 | |
| * kota_ is thinking | 09:43 | |
| acoles | then maybe decorate each builder command with something like @composite_unsupported which raises error if the new composite-related keys are found in the dict (hmmm, or something more elegant) | 09:44 |
| kota_ | sounds nice | 09:46 |
| kota_ | and some command should be also @normal_unsupported | 09:46 |
| kota_ | currently, i'm thinking the entry point instaed of *create* command | 09:47 |
| kota_ | ah... | 09:47 |
| kota_ | it may be ok, we set create as the entry point even if it's composite | 09:47 |
| kota_ | no | 09:48 |
| kota_ | no? | 09:48 |
| kota_ | part_power should be ok to specify because the rings in the composite ring should have same part power | 09:48 |
| kota_ | min_part_hours too | 09:49 |
| kota_ | ah, it may be ok if difference on min_part_hours exit | 09:49 |
| kota_ | probably the pain point is *replicas* for the create command | 09:50 |
| acoles | or we use add_ring to bootstrap if builder file does not exist? | 09:50 |
| acoles | hmmm, not great because we might like add_ring to error if builder file does not exist :) | 09:51 |
| kota_ | acoles: true | 09:52 |
| *** zhurong has quit IRC | 09:53 | |
| kota_ | or, if we can take an option like --composite and then, skip <replicas> and <min_part_hours> for args? | 09:53 |
| kota_ | `swift-ring-builder c.builder create --composite <part_power>` | 09:53 |
| *** dmellado has quit IRC | 09:54 | |
| *** amoralej has quit IRC | 09:54 | |
| kota_ | or even, we don't need to take any args on --composite option | 09:55 |
| kota_ | might be | 09:55 |
| acoles | kota_: we'll find a way | 09:57 |
| acoles | replicas == 0 -> composite | 09:57 |
| acoles | maybe not obvious enough | 09:57 |
| acoles | just add a new command | 09:58 |
| acoles | or like you say, --composite | 09:58 |
| kota_ | ok, let's write down it to the etherpad | 10:00 |
| *** amoralej has joined #openstack-swift | 10:00 | |
| *** dmellado has joined #openstack-swift | 10:02 | |
| acoles | kota_: thanks! | 10:08 |
| kota_ | acoles: done to write as "possible solutions -> idea 1". | 10:18 |
| *** openstackgerrit has quit IRC | 10:18 | |
| kota_ | acoles: if you find something different, please correct that directly | 10:18 |
| kota_ | and I could start to modify my patch since tommorow morning if you and clayg confirmed that idea (and support it) | 10:19 |
| acoles | ok kota_ have a good evening | 10:19 |
| kota_ | acoles: yeah, thanks a lot for the discussion | 10:20 |
| *** tonanhngo has joined #openstack-swift | 10:23 | |
| *** tonanhngo has quit IRC | 10:24 | |
| *** geaaru has quit IRC | 10:32 | |
| *** janonymous has joined #openstack-swift | 10:42 | |
| *** geaaru has joined #openstack-swift | 10:46 | |
| *** tonanhngo has joined #openstack-swift | 10:52 | |
| *** furlongm has joined #openstack-swift | 10:52 | |
| *** tonanhngo has quit IRC | 10:53 | |
| *** dmorita has joined #openstack-swift | 11:01 | |
| *** dmorita has quit IRC | 11:05 | |
| *** geaaru has quit IRC | 11:10 | |
| *** dmorita has joined #openstack-swift | 11:14 | |
| *** sams-gleb has quit IRC | 11:14 | |
| *** dmorita has quit IRC | 11:15 | |
| *** sams-gleb has joined #openstack-swift | 11:15 | |
| *** dmorita has joined #openstack-swift | 11:15 | |
| *** dmorita_ has joined #openstack-swift | 11:18 | |
| *** sams-gleb has quit IRC | 11:19 | |
| *** dmorita has quit IRC | 11:19 | |
| *** dmorita_ has quit IRC | 11:19 | |
| *** dmorita has joined #openstack-swift | 11:19 | |
| *** dmorita has quit IRC | 11:20 | |
| *** dmorita has joined #openstack-swift | 11:20 | |
| *** dmorita has quit IRC | 11:25 | |
| *** psachin has quit IRC | 11:30 | |
| *** psachin has joined #openstack-swift | 11:30 | |
| *** joeljwright has joined #openstack-swift | 11:34 | |
| *** ChanServ sets mode: +v joeljwright | 11:34 | |
| tdasilva | acoles: maybe??? ;) | 11:35 |
| tdasilva | mattoliverau, acoles now i'm really looking to put this stuff into an ansible script together with mathiasb scripts for barbican | 11:36 |
| tdasilva | it's coming all together :) | 11:36 |
| acoles | tdasilva: :) | 11:41 |
| *** gkadam has quit IRC | 11:42 | |
| *** dmorita has joined #openstack-swift | 11:48 | |
| *** dmorita has quit IRC | 11:52 | |
| *** jordanP has joined #openstack-swift | 11:52 | |
| *** tonanhngo has joined #openstack-swift | 11:55 | |
| *** tonanhngo has quit IRC | 11:56 | |
| *** dmorita has joined #openstack-swift | 11:57 | |
| *** sams-gleb has joined #openstack-swift | 11:58 | |
| *** dmorita has quit IRC | 12:02 | |
| *** bkopilov has quit IRC | 12:27 | |
| *** tonanhngo has joined #openstack-swift | 12:37 | |
| *** tonanhngo has quit IRC | 12:38 | |
| *** janonymous has quit IRC | 12:44 | |
| *** klamath has joined #openstack-swift | 12:45 | |
| *** klamath has joined #openstack-swift | 12:46 | |
| *** jaosorior has joined #openstack-swift | 12:47 | |
| *** catintheroof has joined #openstack-swift | 12:59 | |
| *** tonanhngo has joined #openstack-swift | 13:00 | |
| *** tonanhngo has quit IRC | 13:01 | |
| *** links has quit IRC | 13:02 | |
| *** geaaru has joined #openstack-swift | 13:06 | |
| *** cshastri has quit IRC | 13:09 | |
| *** tongli has joined #openstack-swift | 13:23 | |
| *** psachin has quit IRC | 13:24 | |
| *** winggundamth has joined #openstack-swift | 13:29 | |
| *** amoralej is now known as amoralej|lunch | 13:34 | |
| rledisez | for an object upload, the proxy is providing the account/container/container-servers+devices where to register the object | 13:49 |
| rledisez | for the delete-at feature, the proxy is providing the account/container/container-servers+devices where to register the "task" | 13:49 |
| rledisez | is there like a rule that the object-server must be able to work without the container ring ? | 13:49 |
| *** vint_bra has joined #openstack-swift | 13:50 | |
| acoles | rledisez: I think the goal is to avoid every object server process needing to load the ring data | 13:54 |
| *** winggundamth has quit IRC | 13:55 | |
| rledisez | acoles: ok, it makes sense. thx for the info | 13:56 |
| *** zhurong_ has joined #openstack-swift | 14:03 | |
| *** links has joined #openstack-swift | 14:05 | |
| *** zhurong_ has quit IRC | 14:06 | |
| *** bkopilov has joined #openstack-swift | 14:12 | |
| *** cbartz has left #openstack-swift | 14:18 | |
| *** tonanhngo has joined #openstack-swift | 14:23 | |
| *** tonanhngo has quit IRC | 14:25 | |
| *** cbartz has joined #openstack-swift | 14:29 | |
| *** jordanP has quit IRC | 14:39 | |
| *** jordanP has joined #openstack-swift | 14:39 | |
| *** d0ugal has quit IRC | 14:42 | |
| *** links has quit IRC | 14:42 | |
| *** d0ugal has joined #openstack-swift | 14:43 | |
| *** sams-gleb has quit IRC | 14:50 | |
| *** sams-gleb has joined #openstack-swift | 14:51 | |
| *** sams-gleb has quit IRC | 14:55 | |
| *** JimCheung has joined #openstack-swift | 14:57 | |
| notmyname | good morning | 15:01 |
| *** JimCheung has quit IRC | 15:02 | |
| notmyname | clayg: thanks for the writeup! | 15:02 |
| *** amoralej|lunch is now known as amoralej | 15:02 | |
| *** sams-gleb has joined #openstack-swift | 15:04 | |
| timss | clayg: read your rebalance etherpad and wondering if I should be scared now :o Have yet to do a big rebalance on my cluster | 15:07 |
| *** cbartz has left #openstack-swift | 15:07 | |
| *** dja has quit IRC | 15:08 | |
| *** tongli has quit IRC | 15:13 | |
| thurloat | morning ! | 15:14 |
| notmyname | timss: don't be too worried (probably ;-) ) | 15:15 |
| *** tonanhngo has joined #openstack-swift | 15:17 | |
| notmyname | timss: but that being said, it's generally easier to add a little bit of capacity to swift frequently instead of "saving it up" and doubling your capacity all at once | 15:18 |
| *** tonanhngo has quit IRC | 15:18 | |
| timss | notmyname: I'll put those worrying thoughts somewhere hidden until the day of doom then! :) | 15:22 |
| *** klrmn has joined #openstack-swift | 15:22 | |
| notmyname | has anyone been following the "use etcd for storing configs" ML thread? (and has a short summary?) | 15:23 |
| timss | notmyname: yup thanks for the tip, might not match with whatever hardware I'll get my hands on, but disks could be added in chunks using rings/weights I suppose | 15:23 |
| *** Jeffrey4l has quit IRC | 15:38 | |
| *** Jeffrey4l has joined #openstack-swift | 15:39 | |
| timburke | good morning | 15:53 |
| timburke | welcome and congratulations, mahatic! glad to have someone new to pester for reviews ;-) | 15:54 |
| *** jaosorior has quit IRC | 16:02 | |
| *** silor has joined #openstack-swift | 16:05 | |
| *** silor1 has joined #openstack-swift | 16:10 | |
| timburke | tdasilva: i had a thought on symlinks this morning, although it doesn't need to be a part of the current patchset | 16:10 |
| *** silor has quit IRC | 16:10 | |
| tdasilva | timburke: sure, no problem, i'm sure we will still require change | 16:10 |
| *** silor1 is now known as silor | 16:10 | |
| timburke | i wonder if slo should dereference symlinks during a manifest PUT | 16:10 |
| tdasilva | changes | 16:10 |
| timburke | this would let us avoid (at least) one extra backend request during GETs, and would play well with a symlink-based versioned_writes (the SLO would reference the current parts in the archive container rather than the symlink which could change) | 16:12 |
| *** klrmn has quit IRC | 16:18 | |
| timburke | at the same time, i feel like it maintains the spirit of an slo being static | 16:19 |
| tdasilva | timburke: just to make sure I understand. by " dereference symlinks during a manifest PUT" you mean actually write down the target path on the manifest we write to disk | 16:19 |
| *** openstackgerrit has joined #openstack-swift | 16:20 | |
| openstackgerrit | Alistair Coles proposed openstack/swift master: WIP - Make EC GET defer nodes likely to have duplicate fragments https://review.openstack.org/446050 | 16:20 |
| tdasilva | timburke: if that's the case, then my concern would be for users that are using symlinks to point changing targets and they used symlinks in the slo manifest for that reason | 16:22 |
| tdasilva | timburke: I think versioning is a good example. One could argue that the user wants the most current version and not the archived version??? idk | 16:22 |
| *** JimCheung has joined #openstack-swift | 16:22 | |
| timburke | yeah, that was my thinking. and this debate is why i'm partial toward it being a follow-up | 16:23 |
| tdasilva | timburke: but then does that mean an api / behavior change that we never want to do it because of backwards compatibility? | 16:23 |
| timburke | but having the manifest point to "whatever the current version is" will cause the slo to break whenever anything's updated | 16:23 |
| timburke | i would be content to describe it as a bug fix | 16:24 |
| timburke | (because of the aforementioned breakage) | 16:24 |
| tdasilva | hehehehe | 16:24 |
| jrichli | It's been awhile since I thought about symlinks, but I thought we were restricting the interface so that the target could not change. I realize that doesn't mesh with wanting to maybe use them for auto-tiering, however | 16:24 |
| tdasilva | jrichli: target doesn't change on POSTs | 16:25 |
| tdasilva | but user can always do a new PUT | 16:25 |
| jrichli | tdasilva: ah, right. thx | 16:25 |
| tdasilva | that's how I envision we would do versioning for example | 16:25 |
| timburke | bingo | 16:26 |
| timburke | i should really try to find some time to take a stab at that (symlink-backed versioning)... | 16:26 |
| tdasilva | sometimes i really dislike the fact that we got the API so restricted and no new versions of the API. It forces us to get new APIs right the first time or we end up with bad apis and some weird hacks to fix things | 16:27 |
| timburke | yup :-( | 16:28 |
| tdasilva | that's why i'd prefer that we at least be more comfortable doing "beta" apis | 16:28 |
| tdasilva | or alpha, whatever | 16:28 |
| timburke | the only trouble is that the "alpha/beta" api inevitably ends up in production | 16:28 |
| tdasilva | probabably alpha is better cause it really tells operators/users that things are expected to change | 16:29 |
| tdasilva | yeah, but if somebody put something alpha in prod, then i'm sorry, but don't complain if there was a change | 16:29 |
| notmyname | turns out that people don't use it when you tell them it could break :-) | 16:30 |
| tdasilva | I think, I'd prefer that situation then forcing us to get the API right the very first time it lands on master | 16:30 |
| timburke | it's even worse when you're talking about new on-disk formats. an api could conceivably change; there'd be the period of pain as clients get fixed, but it eventually passes. we have no story for migrating on-disk formats, though | 16:30 |
| timburke | and symlinks is both | 16:31 |
| notmyname | but we *have* migrated on-disk formats. those are "easy" to change, compared to APIs. we've reverted several changes because some client somewhere got broken because we started following the RFC or something | 16:31 |
| notmyname | (easy in the sense that carrying away everest is easy compared to terraforming mars) | 16:32 |
| tdasilva | heh | 16:32 |
| *** joeljwright has quit IRC | 16:35 | |
| timburke | notmyname: when? and could we ever feel comfortable ripping out the support for the old format? how can you ever feel reasonably certain that you converted everything? | 16:35 |
| notmyname | timburke: ring format changes. .durable changes. storage policies. adding columns to DBs | 16:36 |
| timburke | i've been thinking about this a bit in the context of key rotation, or making the etag of a slo in container listings match the etag in a GET/HEAD response, or... | 16:36 |
| timburke | storage policies don't have a migration story! we've been talking about how to handle that for ages! | 16:37 |
| tdasilva | timburke: i think he means that we actually added storage policies | 16:38 |
| notmyname | oh I just mean that we used to have everything in objects/ and now we have objects-N/ | 16:38 |
| tdasilva | objects objects-1 etc | 16:38 |
| notmyname | yeah, what tdasilva | 16:38 |
| notmyname | said | 16:38 |
| *** SkyRocknRoll has joined #openstack-swift | 16:39 | |
| notmyname | timburke: but I'll grant you that everything is terrible, and we can't change anything. we weren't smart enough back then to do the right thing, and we aren't smart enough now to fix it. | 16:39 |
| notmyname | ;-) | 16:40 |
| tdasilva | notmyname: but we keep not being smart thinking that we are smart to get new apis right this time | 16:40 |
| tdasilva | read that really fast | 16:40 |
| notmyname | lol | 16:41 |
| notmyname | THIS TIME IT'S DIFFERENT! | 16:41 |
| tdasilva | more than fixing the broken apis, i wish we would change our process with the knowledge that we will probably mess up again on new apis | 16:41 |
| tdasilva | hehehehe | 16:42 |
| tdasilva | sorry, i'm done ranting, back to fixing symlinks | 16:42 |
| notmyname | yeah, I agree. the only thing I take issue with is that we haven't or can't change internal formats. we're pretty good on that front, considering that changing has to take into account enormous amounts of data that can't go away | 16:43 |
| notmyname | and I'm scared that if we finally arrive at the "one day we'll version the api...", then we'll go too far and have version 38 in six months | 16:44 |
| tdasilva | lol, that's terrible | 16:44 |
| notmyname | we've had v1 of the API for 7 years (and it predates swift itself a couple of years). so if we have a v2, I'd love for it to last at least another 10 more years :-) | 16:44 |
| notmyname | but maybe that's the wrong attitude | 16:44 |
| notmyname | idk | 16:44 |
| tdasilva | but yeah, hence, my current naive thinking is that 'alpha apis' is a middle ground | 16:45 |
| tdasilva | but i'm saying this more on the fact that others seem to be going that way rather than knowing from experience that it is a good thing | 16:47 |
| timburke | notmyname: yeah, we'd totally hit some "large" version number pretty quickly. i'd wager if we ever wanted to go down the "v2" path, we'd need to throw it all on a (*very* long-lived) feature branch, do all of our breaking, let people run it in prod anyway, find all the *other places* we needed or wanted to break, and wait for it to get vaguely stable before brining it back in and calling it "v2" | 16:47 |
| notmyname | or we could do microversions! ;-) | 16:47 |
| notmyname | swift api v1, now with symlinks v2.13, tempurls v1.1, POST v3, crypto v1.0, ec v3, and staticweb v2.17 | 16:48 |
| notmyname | swift api compatibility matrix v2.14! | 16:49 |
| tdasilva | lol, that sounds awful too, where did i see that recently? | 16:49 |
| tdasilva | oh yeah, qnap nas | 16:49 |
| jrichli | the problem then is regression testing with different combinations | 16:50 |
| *** gyee has joined #openstack-swift | 16:50 | |
| jrichli | i have worked in a world where we had hundreds of different services versioned separately. it was craziness. | 16:51 |
| jrichli | (and yes, i did know you were joking) | 16:52 |
| tdasilva | going to do some more snow cleanup...brb... | 16:52 |
| *** tesseract has quit IRC | 16:53 | |
| *** cbartz has joined #openstack-swift | 17:10 | |
| *** klrmn has joined #openstack-swift | 17:10 | |
| *** cbartz has left #openstack-swift | 17:19 | |
| *** pcaruana has quit IRC | 17:28 | |
| *** jordanP has quit IRC | 17:31 | |
| *** JimCheung has quit IRC | 17:36 | |
| acoles | clayg: kota_ left some comments on the etherpad re builder files for composite rings, https://etherpad.openstack.org/p/composite_rings | 17:37 |
| acoles | clayg: we were chatting about it earlier today | 17:38 |
| *** szaher has joined #openstack-swift | 17:39 | |
| *** JimCheung has joined #openstack-swift | 17:43 | |
| timburke | does anyone remember why https://github.com/openstack/swift/blob/2.13.0/swift/common/middleware/versioned_writes.py#L473-L475 isn't a bug? tdasilva, maybe? i'm sure it must've come up in the middleware refactor... | 17:43 |
| *** SkyRocknRoll has quit IRC | 17:44 | |
| tdasilva | timburke: if I remember correctly this was maintained for backwards compatibility "it's always been this way so let's not change it" | 17:44 |
| timburke | clayg! i wanna change it! | 17:44 |
| tdasilva | especially during the refactor, we tried to maintain the behavior very much the same as pre-refactor | 17:45 |
| timburke | tdasilva: yup... as torgomatic would say, it was for "hysterical raisins" | 17:46 |
| tdasilva | hhahahahaha | 17:46 |
| tdasilva | timburke: as clayg would say, do you actually have a use case for wanting to change it | 17:47 |
| tdasilva | ? | 17:47 |
| notmyname | to be discussed at today's meeting https://wiki.openstack.org/wiki/Swift/Fixing-rebalance-and-golang | 17:49 |
| timburke | tdasilva: yup, related to history mode. i want to give a less-privileged user read/write access to a container with history-mode versioning enabled, and i don't want to ever have to worry about losing data as a result | 17:49 |
| *** lespaul_ has quit IRC | 17:50 | |
| timburke | that non-account-owner can do whatever they want inside that container (including creating large objects) and currently if they create a DLO, they can overwrite an object without having an archive copy created | 17:51 |
| timburke | SLO behavior is great (or, as great as can be expected). DLO behavior starts to feel like a security hole | 17:52 |
| *** hseipp has quit IRC | 17:54 | |
| tdasilva | timburke: just trying to follow up: "if they create a DLO, they can overwrite..." but isn't that code preventing them from creating a DLO in a versioned container? | 17:57 |
| *** SkyRocknRoll has joined #openstack-swift | 17:57 | |
| tdasilva | oh, it doesn't prevent, it just doesn't version what's there | 17:57 |
| timburke | yep! just let it through! | 17:58 |
| tdasilva | so they can overwrite without an archive copy, got it | 17:58 |
| timburke | then there's also the awkward delete behavior acoles documented at https://bugs.launchpad.net/swift/+bug/1626989 ... | 17:58 |
| openstack | Launchpad bug 1626989 in OpenStack Object Storage (swift) "versioning history mode treats DLO manifests inconsistently" [Undecided,New] | 17:58 |
| tdasilva | I think we discussed that and were generally ok, like "naahh..it's just a header" | 17:58 |
| tdasilva | lol at that bug | 17:59 |
| *** ChubYann has joined #openstack-swift | 18:00 | |
| tdasilva | we create delete markers but no versions | 18:00 |
| timburke | i wanna be able to roll back state. if we don't version dlos, it's impossible; there's no way of knowing where that dlo pointed, if we even know that it ever existed! | 18:01 |
| acoles | timburke: as I remember it, DLO manifest weren't versioned before the refactor to middleware, so we kept it that way, for better or worse | 18:01 |
| tdasilva | timburke: yep, makes sense | 18:01 |
| tdasilva | acoles: yes, that's how i remember too | 18:01 |
| timburke | if we *do* version dlos (and whatever container(s) have the segments) there's a chance at rolling back. it'll be a bit painful, but doable | 18:01 |
| timburke | so, any opposition to versioning dlos? notmyname, clayg? | 18:02 |
| *** tonanhngo has joined #openstack-swift | 18:03 | |
| *** tonanhngo has quit IRC | 18:07 | |
| *** tonanhngo has joined #openstack-swift | 18:11 | |
| *** gyee has quit IRC | 18:18 | |
| notmyname | I ... don't knwo | 18:21 |
| *** shivanim has joined #openstack-swift | 18:24 | |
| shivanim | Hello, I am Shivani Mehendarge. I am interested in working on swift for the outreachy program. Can anyone please tell me how to start with it? | 18:29 |
| notmyname | shivanim: hello! | 18:29 |
| *** ganders has joined #openstack-swift | 18:29 | |
| shivanim | notmyname, :) | 18:29 |
| notmyname | shivanim: I've been an outreachy mentor a couple of times in the past. what are you interested in for swift? | 18:29 |
| shivanim | I have cloned the project just now. I have very litte idea about it. | 18:30 |
| notmyname | shivanim: do you have any ideas of what you'd like to work on? what made you consider swift in the first place? | 18:31 |
| shivanim | I wanted to contribute to it. Maybe by fixing a bug or documentation perhaps. But I want to learn how swift works. | 18:31 |
| notmyname | shivanim: sure (and that's great). but why swift as opposed to some other project? | 18:32 |
| shivanim | Actually I am interested in cloud computing and seeing that swift can help retrieve a lot of data through simple API, makes me interested. :) | 18:34 |
| notmyname | cool | 18:34 |
| notmyname | shivanim: have you looked through the outreachy site and talked to any of the coordinators yet? | 18:34 |
| notmyname | shivanim: what time zone are you in? | 18:36 |
| shivanim | notmyname, I have looked through the site but haven't talked to any coordinators. | 18:36 |
| notmyname | https://wiki.openstack.org/wiki/Outreachy seems to be the openstack landing page for outreachy | 18:37 |
| shivanim | notmyname, Iam in IST. | 18:37 |
| notmyname | shivanim: ok. mahatic would be a great person to talk to. she's in your time zone and a swift core and also one of the main outreachy coordinators for openstack | 18:38 |
| shivanim | notmyname, I have visited the wiki page for Openstack :) | 18:38 |
| shivanim | notmyname, ok :) | 18:38 |
| shivanim | notmyname, After cloning the swift project, how can I go about it? What should I preferably do next? | 18:39 |
| notmyname | you should set up a swift all in one (SAIO) https://docs.openstack.org/developer/swift/development_saio.html | 18:40 |
| notmyname | shivanim: we don't have any outreachy projects listed for swift right now. that's not to say you can't be helpful (quite the opposite) | 18:40 |
| shivanim | notmyname, Oh, I didn't notice that. | 18:41 |
| notmyname | shivanim: we do have several significant things going on that will take a lot of work from a lot of people to do. these are some of the places where you could help out | 18:41 |
| notmyname | shivanim: what's your dev background and experience? are you just getting started or have you been doing dev work for a while? | 18:42 |
| shivanim | notmyname, I am a beginner in open source contributions. | 18:42 |
| notmyname | that's totally ok :-) | 18:43 |
| shivanim | I have made 2 contributions to pagure recently during pycon pune :) | 18:43 |
| notmyname | have you done stuff on your own or for an employer that's not open source? | 18:43 |
| notmyname | pagure? /me goes to google | 18:43 |
| shivanim | notmyname, yes I have :) | 18:43 |
| notmyname | shivanim: the first steps for you will be to set up the SAIO, start using the API, and understand how the code is put together. swift.openstack.org is where the dev docs are hosted, and of course you can always ask questions in here | 18:46 |
| shivanim | notmyname, Thanks for your help :) I will surely go through the docs . | 18:46 |
| shivanim | notmyname, If I want to participate in outreachy, I will have to contribute to one of the projects openstack that is participating in it? | 18:47 |
| notmyname | shivanim: great! you can also check out https://wiki.openstack.org/wiki/Meetings/Swift for a list of some stuff that's being worked on and https://wiki.openstack.org/wiki/Swift/ideas for some things that people have written related to some of their work | 18:47 |
| notmyname | shivanim: as I understand it, to participate in outreachy in openstack, you need a project and a mentor. and the mentor is the most important (ie the mentor helps define the project with you and says it's good or not) | 18:48 |
| *** catinthe_ has joined #openstack-swift | 18:48 | |
| notmyname | shivanim: but there's also an application process. I'm not as familiar with that step. mahatic would be the person to talk to | 18:49 |
| shivanim | notmyname, okay , I will surely talk to her :) | 18:49 |
| timburke | as i recall, it does involve having at least one proposed change, just to make sure you've gotten as far as checking out the repo, reading the contributor docs, etc. | 18:50 |
| *** catintheroof has quit IRC | 18:51 | |
| shivanim | timburke, yes, you have to make atleast one contribution to the project you are opting for. | 18:52 |
| *** SkyRocknRoll has quit IRC | 18:53 | |
| shivanim | notmyname, where can I find the list of projects in openstack that are there in outreachy? | 18:55 |
| notmyname | shivanim: all I've got is the wiki page I linked earlier | 18:55 |
| shivanim | I just checked this link https://wiki.gnome.org/Outreachy/2017/MayAugust#Participating_Organizations . Here there is a project for swift. | 18:55 |
| timburke | https://wiki.openstack.org/wiki/Internship_ideas maybe? | 18:56 |
| notmyname | yeah, there's one listed for swift3 (which is closely related to swift--swift3 is an external extension to swift, so you gotta know both) | 18:57 |
| notmyname | and to be clear, despite there not being a "dp project XYZ in swift" listed, there's a lot of work to do in swift. but in the short term there's probably not a lot of projects in swift that need to be done solo by an outreachy intern. but an outreachy intern could help with the other big stuff going on | 18:58 |
| shivanim | notmyname, okay :) | 18:59 |
| shivanim | timburke, notmyname Thank you :) | 18:59 |
| *** lespaul has quit IRC | 19:16 | |
| *** amoralej is now known as amoralej|off | 19:20 | |
| *** geaaru has quit IRC | 19:32 | |
| *** shivanim has quit IRC | 19:32 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Version DLOs, just like every other type of object https://review.openstack.org/446142 | 19:38 |
| *** ganders has quit IRC | 19:44 | |
| *** sgundur has quit IRC | 19:45 | |
| *** pumaranikar has quit IRC | 19:45 | |
| *** mmotiani has quit IRC | 19:45 | |
| *** ntata has quit IRC | 19:45 | |
| clayg | timss: If you're going to be rebalancing a large replicated cluster don't worry too much - you're in good company - there's tons of stuff you can do to make that go well for you and you'll get to learn cool stuff along the way! | 19:48 |
| clayg | timss: if you're running a large EC cluster that needs to rebalance... well... me too! | 19:49 |
| *** chsc has joined #openstack-swift | 19:57 | |
| clayg | Did we point shivanim at low-hanging-fruit? https://bugs.launchpad.net/swift/+bugs?field.assignee=&field.tag=low-hanging-fruit&field.status%3Alist=NEW&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS | 20:05 |
| *** silor has quit IRC | 20:07 | |
| timburke | clayg: did we ever come to a conclusion on https://bugs.launchpad.net/swift/+bug/1537811 ? should we just close that as Won't Fix / Invalid? | 20:07 |
| openstack | Launchpad bug 1537811 in OpenStack Object Storage (swift) "204 No Content responses have Content-Length specified" [Low,In progress] - Assigned to Trevor McCasland (twm2016) | 20:07 |
| *** sgundur has joined #openstack-swift | 20:10 | |
| *** sweeper has quit IRC | 20:11 | |
| clayg | timburke: no decision was made to my knowledge - i belive it it is still low priority - but probably should be "not assigned" | 20:26 |
| clayg | timburke: re-reading the varnish ticket it's like reading my future - "we did the what spec said - pretty sure this will be fine" "yeah you broke me ... me too ... me as well" "gd, ^&*# you web - you are a cesspool of ill behaved clients - it's unfathomable anything anywhere works at all!" "everything was working fine till you read the damn spec and wanted to look smart" "touche; revert" | 20:31 |
| timburke | right? so it seems like we probably want to keep Content-Length | 20:32 |
| clayg | i'm ok with pulling it out if there's a good reason to? So far all we've got is "it's not reliable" | 20:33 |
| clayg | i.e. some proxies strip it out and some proxies add it in even when not there | 20:33 |
| clayg | so there's a strong case that clients SHOULD not expect a content-length header on requests which don't have a body (esp. 204) | 20:33 |
| clayg | our current respones don't make that especially apparent - so there's some risk we reenforce the cesspool of ill behaved clients | 20:35 |
| clayg | in varnishes case they went with liberal what you accept and conservative of what you send | 20:36 |
| clayg | they don't get mad (anymore) if a backend *sends them* a 204 w/ content-length - but they also don't go spewing that back out into the world | 20:36 |
| notmyname | IMO we should strive for the same | 20:37 |
| clayg | we fixed tempest - which was the only client we *know of* that expected the header there | 20:37 |
| clayg | I really do feel like we *could* pull it out | 20:37 |
| clayg | i just wish we had *any other* demonstrable harm - but it's such a small innocuous thing | 20:38 |
| clayg | so... I'm torn... good thing it's also a really low priority - helps significantly to minimzie the require soul searching | 20:39 |
| clayg | maybe at somepoint someone will bump into something annoying find that bug and ding us | 20:39 |
| clayg | oh oh oh! how bout this! since 7230 *obsoletes* 2616 we should expect new proxies to implemnt 7230 - which *is* (surprisingly?) prescriptive about servers sending content-length with 204 responses | 20:41 |
| mattoliverau | Morning | 20:41 |
| clayg | ... so if the assumption is that new proxies getting built will follow 7230 - we can also expect the mistake varnish made to be common going forward | 20:41 |
| timburke | fwiw, it looks like 204s aren't the only times proxies will mess with content-length; see https://bugs.launchpad.net/python-swiftclient/+bug/1586690 and https://bugs.launchpad.net/python-swiftclient/+bug/1621581 | 20:42 |
| openstack | Launchpad bug 1586690 in python-swiftclient "Uploading empty(0 B) file fails" [Undecided,Incomplete] - Assigned to Uday Swami (swamius) | 20:42 |
| timburke | doesn't really indicate (to me) which way we ought to go here, just putting it out there | 20:42 |
| openstack | Launchpad bug 1621581 in python-swiftclient "swiftclient returns response headers without 'Content-Length' param, thus causing upload object to fail" [Undecided,In progress] - Assigned to Arun Mani (arun-mani) | 20:42 |
| clayg | esentially swift (and other old school webservers like apache) will be the cesspool of clients that make it impossible to implement the spec | 20:42 |
| clayg | ... until we 1) get someone to update 7230 or 2) stop sending content-length | 20:42 |
| clayg | one of those two things seems significantly easier than the other ;) | 20:42 |
| notmyname | 1) someone else 2) us | 20:43 |
| notmyname | ;-) | 20:43 |
| timburke | always be waiting on someone else ;-) | 20:43 |
| *** sgundur has quit IRC | 20:43 | |
| notmyname | or we could just stop sending it and think about reverting it someone complains | 20:44 |
| * notmyname stepping out for 10 minutes. be back for the meeting | 20:44 | |
| clayg | timburke: i'm not sure what either lp bug #1586690 or lp bug #1621581 where getting at exactly | 20:45 |
| openstack | Launchpad bug 1586690 in python-swiftclient "Uploading empty(0 B) file fails" [Undecided,Incomplete] https://launchpad.net/bugs/1586690 - Assigned to Uday Swami (swamius) | 20:45 |
| openstack | Launchpad bug 1621581 in python-swiftclient "swiftclient returns response headers without 'Content-Length' param, thus causing upload object to fail" [Undecided,In progress] https://launchpad.net/bugs/1621581 - Assigned to Arun Mani (arun-mani) | 20:45 |
| clayg | I think there was also a bug about 304 returning content-length and some if-match response returning accept-ranges | 20:46 |
| timburke | mostly, i'm just sad that proxies keep mangling our headers :'( | 20:47 |
| *** sgundur has joined #openstack-swift | 20:56 | |
| notmyname | team meeting in 2 minutes in #openstack-meeting | 20:58 |
| *** dmorita has joined #openstack-swift | 20:59 | |
| *** m_kazuhiro has joined #openstack-swift | 20:59 | |
| kota_ | good morning | 20:59 |
| m_kazuhiro | kota_: good morning | 21:00 |
| kota_ | m_kazuhiro: o/ | 21:00 |
| *** sgundur has quit IRC | 21:04 | |
| *** ntata has joined #openstack-swift | 21:06 | |
| *** dja has joined #openstack-swift | 21:08 | |
| notmyname | thurloat: talking about your patch in -meeting | 21:08 |
| *** lespaul has joined #openstack-swift | 21:12 | |
| *** dja has quit IRC | 21:23 | |
| *** sgundur has joined #openstack-swift | 21:33 | |
| *** Jeffrey4l_ has joined #openstack-swift | 21:35 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Version DLOs, just like every other type of object https://review.openstack.org/446142 | 21:36 |
| *** sgundur has quit IRC | 21:41 | |
| *** sgundur has joined #openstack-swift | 21:41 | |
| openstackgerrit | Tim Burke proposed openstack/swift master: Factor out a bunch of common testing setup https://review.openstack.org/446185 | 21:43 |
| *** m_kazuhiro has quit IRC | 21:52 | |
| *** ntata has quit IRC | 21:53 | |
| *** jamielennox has quit IRC | 21:55 | |
| *** sgundur has quit IRC | 21:59 | |
| acoles | kota_: before you head for breakfast... | 21:59 |
| kota_ | acoles: ? | 21:59 |
| notmyname | https://www.openstack.org/blog/2017/03/helping-ptg-attendees-and-other-developers-get-to-the-openstack-summit/ | 21:59 |
| acoles | kota_: it might be helpful to split the composite rings patch into two : 1. the functions to perform prevalidation and perform the zipping, along with necessary tests; followed by 2. some kind of CLI support. | 22:01 |
| * tdasilva is sorry about missing swift meeting, catching up now | 22:01 | |
| acoles | kota_: that way we can make progress on 1 while debating 2 :) and decouple testing of the functionality from any particular CLI interface | 22:02 |
| notmyname | tdasilva: no worries | 22:02 |
| notmyname | tdasilva: I think we only volunteered you for a few things | 22:02 |
| tdasilva | rofl | 22:02 |
| kota_ | acoles: that's what I've been thinking since the last now > spliting patch as chain ;-) | 22:02 |
| acoles | kota_: two agree -> decision ! :D | 22:02 |
| acoles | kota_: let me know where I can help | 22:03 |
| kota_ | acoles: let me make sure where you think as "functions"? | 22:04 |
| acoles | kota_: also FYI (still very much WIP) https://review.openstack.org/446050 | 22:04 |
| patchbot | patch 446050 - swift - WIP - Make EC GET defer nodes likely to have dupli... | 22:04 |
| kota_ | acoles: the boader could not be concrete to me for now | 22:05 |
| kota_ | thanks the GET thing! | 22:05 |
| acoles | kota_: so for example as clayg commented here https://review.openstack.org/#/c/441921/2/swift/cli/ringzipper.py@167 | 22:06 |
| patchbot | patch 441921 - swift - Composite Ring Zipper Implementation | 22:06 |
| kota_ | oic | 22:06 |
| kota_ | the *MAGIC* is the first functionality to implement | 22:07 |
| acoles | so split out the validation/zipping to separate module e.g. common/ring/composite_ring.py | 22:08 |
| kota_ | from... builder file? ok, that's a difference | 22:08 |
| kota_ | acoles: ok, and no new bin and cli tool | 22:08 |
| acoles | kota_: well, that would be patch 2, dependent. | 22:09 |
| acoles | and we can continue to debate new cli tool vs integrate into swift-ring-builder as we gather more opinions | 22:09 |
| acoles | but make progress on composite_ring.py (or whatever its called) | 22:10 |
| kota_ | sounds great plan | 22:10 |
| kota_ | that works in parallel, make sense, it will make it speed up | 22:10 |
| kota_ | let's do that | 22:11 |
| acoles | kota_: ok :) | 22:12 |
| kota_ | thanks acoles to bring it up ;-D | 22:12 |
| acoles | kota_: NP have a good day, I'm away now. | 22:13 |
| kota_ | acoles: good night, have a good sleep! | 22:13 |
| * kota_ is going to find breakfirst | 22:13 | |
| *** adriant has joined #openstack-swift | 22:14 | |
| *** jamielennox has joined #openstack-swift | 22:15 | |
| *** m_kazuhiro has joined #openstack-swift | 22:16 | |
| *** sams-gleb has quit IRC | 22:24 | |
| *** dja has joined #openstack-swift | 22:26 | |
| timss | clayg: Running replication, and that's reassuring to hear, thanks! Trying to pick up tips and tricks here and there, but I'll probably be around when the day arrives :) | 22:31 |
| clayg | lol | 22:45 |
| *** m_kazuhiro has quit IRC | 22:47 | |
| *** dja has quit IRC | 22:49 | |
| *** gyee has joined #openstack-swift | 22:52 | |
| *** klamath has quit IRC | 22:59 | |
| *** tonanhngo has quit IRC | 23:02 | |
| *** catinthe_ has quit IRC | 23:02 | |
| *** tonanhngo has joined #openstack-swift | 23:02 | |
| *** tonanhngo has quit IRC | 23:07 | |
| *** sams-gleb has joined #openstack-swift | 23:24 | |
| *** sams-gleb has quit IRC | 23:29 | |
| *** lespaul has quit IRC | 23:34 | |
| *** chsc has quit IRC | 23:36 | |
| *** kei_yama has joined #openstack-swift | 23:40 | |
| openstackgerrit | Daisuke Morita proposed openstack/swift master: WIP: Changing Policies https://review.openstack.org/209329 | 23:43 |
| *** mmotiani has joined #openstack-swift | 23:49 | |
| *** ntata has joined #openstack-swift | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!