| *** siva_krishnan has joined #openstack-swift | 00:00 | |
| *** diogogmt has joined #openstack-swift | 00:00 | |
| *** siva_krishnan has quit IRC | 00:04 | |
| *** zhill has joined #openstack-swift | 00:08 | |
| *** blmartin has joined #openstack-swift | 00:09 | |
| *** jerrygb has joined #openstack-swift | 00:16 | |
| *** ho has joined #openstack-swift | 00:20 | |
| ho | good morning! | 00:22 |
|---|---|---|
| mattoliverau | ho: morning | 00:22 |
| ho | mattoliverau: morning! | 00:23 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 00:39 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 00:39 |
| kota_ | morning | 00:42 |
| notmyname | hello mattoliverau ho kota_. all you people for whom it's already wednesday | 00:42 |
| ho | kota_: notmyname: good morning & hello | 00:45 |
| mattoliverau | notmyname: o/ | 00:54 |
| *** diogogmt has quit IRC | 00:55 | |
| *** zhill has quit IRC | 00:55 | |
| asettle | Ohhhh notmyname I was just thinking I needed to talk to you! | 01:00 |
| notmyname | asettle: hi! | 01:01 |
| asettle | Hello! | 01:01 |
| notmyname | asettle: if it's about that swiftclient docs stuff, that's what I'm doing now ;-) | 01:01 |
| asettle | It is! It's like you know me :p haha. I've been snowed under with RPCO stuff since we got back from Tokyo. But my workload is freeing up... so keen to get started. I'm just about to review Joel's latest patch too :) | 01:02 |
| notmyname | asettle: very good! | 01:02 |
| notmyname | asettle: I've been in the same boat, and this week I'm getting that stuff the attention it needs | 01:02 |
| asettle | Splendid! Well. Keep me updated and let me know where I need to be and what I can do to help :) | 01:03 |
| notmyname | asettle: looking at my notes... | 01:03 |
| notmyname | I've put together a rough outline. I want to clean that up some and then send it to you | 01:04 |
| asettle | Great :) thank you! Looking forward to seeing what you come up with :) | 01:04 |
| notmyname | but that leads to the real question: what's the actual real thing we need to make progress? | 01:04 |
| notmyname | sometimes I feel like I keep talking about getting started without actually getting started. outlines are good, I guess, but then what? :-) | 01:04 |
| asettle | Well it depends how you want to do it. After we have an agreed upon outline - do we want to put it to the swift community? If so, blueprints ahoy. If we just want to get started, then what I'll do is implement the whole change in one patch (i'm talking place holders, new TOC, all that good stuff) and from there we can then have it approved, merged, and we can work on that. Although we'd have to ensure it's not in maste | 01:07 |
| asettle | r so that we don't override the existing documentation. | 01:07 |
| notmyname | the "good thing" (only in the context of this exact statement) is that based on what we have now (http://docs.openstack.org/developer/python-swiftclient/) even if we overwrite everything we don't really lose anything ;-) | 01:08 |
| notmyname | asettle: I think we need to land the ToC in the repo and then work from that as a TODO list | 01:09 |
| notmyname | some parts should be docstrings. some should be new .rst files | 01:09 |
| asettle | Perfect. Well, if you're not afraid of losing anything really... we can start working on it pretty much straight away. I am happy to implement the placeholders for info dumps and ... yep, perfect. | 01:09 |
| asettle | Exactly that. | 01:09 |
| notmyname | ok, I'll clean up my outline and send it over ASAP (if I haven't in 24 hours, you should call me on it) | 01:10 |
| asettle | Hahahaha, okay! Well, my release isn't due for another 8 days. So my head will be in that sand bucket, but I should be able to get back to you soon! :) | 01:10 |
| *** aerwin has quit IRC | 01:25 | |
| *** jerrygb has quit IRC | 01:50 | |
| *** diogogmt has joined #openstack-swift | 01:54 | |
| *** jerrygb has joined #openstack-swift | 01:55 | |
| *** blmartin has quit IRC | 02:02 | |
| *** omkarjoshi has left #openstack-swift | 02:13 | |
| *** wuhg has joined #openstack-swift | 02:15 | |
| openstackgerrit | Merged openstack/swift: Merge branch 'master' into feature/crypto https://review.openstack.org/252024 | 02:28 |
| *** david-lyle has joined #openstack-swift | 02:31 | |
| *** haomaiwa_ has joined #openstack-swift | 02:42 | |
| *** sanchitmalhotra has joined #openstack-swift | 02:47 | |
| openstackgerrit | Merged openstack/swift: Fix crash when a SLO repeats a segment. https://review.openstack.org/250099 | 02:49 |
| *** haomaiwa_ has quit IRC | 02:52 | |
| *** haomaiwang has joined #openstack-swift | 02:53 | |
| openstackgerrit | Janie Richling proposed openstack/swift: WIP - Support for http footers - Replication and EC https://review.openstack.org/237393 | 02:57 |
| *** haomaiwang has quit IRC | 02:57 | |
| *** sgundur has joined #openstack-swift | 03:00 | |
| *** david-lyle has quit IRC | 03:00 | |
| *** sgundur has left #openstack-swift | 03:00 | |
| *** miurahr has joined #openstack-swift | 03:03 | |
| *** venkat_p has joined #openstack-swift | 03:09 | |
| *** changbl has joined #openstack-swift | 03:13 | |
| *** gyee has quit IRC | 03:16 | |
| *** SkyRocknRoll has quit IRC | 03:16 | |
| *** miurahr has quit IRC | 03:20 | |
| *** jerrygb has quit IRC | 03:34 | |
| *** jerrygb has joined #openstack-swift | 03:34 | |
| *** tsg has quit IRC | 03:37 | |
| *** jerrygb has quit IRC | 03:39 | |
| *** ppai has joined #openstack-swift | 03:40 | |
| *** links has joined #openstack-swift | 03:43 | |
| *** mahatic has joined #openstack-swift | 03:54 | |
| *** sanchitmalhotra has quit IRC | 04:01 | |
| *** sanchitmalhotra has joined #openstack-swift | 04:04 | |
| *** km has joined #openstack-swift | 04:06 | |
| *** km is now known as Guest90246 | 04:06 | |
| *** Guest75007 has quit IRC | 04:06 | |
| *** kei_yama_ has joined #openstack-swift | 04:07 | |
| *** kei_yama has quit IRC | 04:08 | |
| *** badari has quit IRC | 04:10 | |
| *** badari has joined #openstack-swift | 04:19 | |
| *** jome has joined #openstack-swift | 04:24 | |
| jome | Hi. Can anyone help me with this scenario. I'd like to have multiple proxy servers configured behind a load balancer that talk to a single object and account servers. Is this even possible? | 04:26 |
| jome | From documentation that I have gathered, this seems to be possible but I've not seen anything docs that explain hos to configure this. | 04:27 |
| *** badari has quit IRC | 04:27 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Modify unit tests to include real crypto https://review.openstack.org/211451 | 04:33 |
| *** venkat_p has quit IRC | 04:33 | |
| *** venkat_p has joined #openstack-swift | 04:33 | |
| *** SkyRocknRoll has joined #openstack-swift | 04:34 | |
| mattoliverau | jome: yeah that's possible, but a single storage server isn't great for object durability. But yeah, the rings just have to container the storage server(s) and all the proxies have the same rings. | 04:43 |
| mattoliverau | also the load balancer would allow you to have many connections, but they will all go to the one storage node, so you might have a bottle neck there too. | 04:44 |
| *** david-lyle has joined #openstack-swift | 04:44 | |
| jome | Okay. maybe I did not make it clear. I do have three storage nodes currently | 04:45 |
| jome | How do I configure multiple proxy servers in that scenario. | 04:46 |
| mattoliverau | jome: ok, well yeah you can have as many proxies as you want, just set them up as normal, and make sure they have a copy of the same rings. | 04:47 |
| mattoliverau | jome: you can then use the healthcheck middleware in the proxies so the load balancer can check proxy health | 04:47 |
| mattoliverau | jome: the rings are what tell a proxy where to connect. So long as your proxies have access to the storage nodes and have the rings, that's it. | 04:49 |
| jome | That simple??? :) | 04:50 |
| mattoliverau | jome: should be, so long as you know how to setup 1 proxy, you can set up n proxies :) | 04:51 |
| mattoliverau | jome: if you had multiple regions then the only difference might be proxy read/write affinity settings, but really there just proxy services with rings that tell them where they can go. | 04:52 |
| jome | So if i set up a proxy on a different machine in the same network for example, all i do is copy the rings from the original server and I'm good to go | 04:53 |
| mattoliverau | jome: yup :) | 04:54 |
| jome | Fantastic. I'll give it a spin now | 04:55 |
| jome | Thanks. | 04:56 |
| mattoliverau | jome: that may you can tune your swift cluster to what ever you need, more connections, scale up the proxies, need more redundently/durabiliy or storage, scale add more storage nodes. | 04:56 |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Modify unit tests to include real crypto https://review.openstack.org/211451 | 04:56 |
| mattoliverau | s/may/way/ | 04:56 |
| mattoliverau | jome: have fun ;) | 04:57 |
| jome | What about auth though? That would have to be one machine, right? | 04:57 |
| mattoliverau | depends on what auth your using | 04:57 |
| mattoliverau | if your using keystone, then it's just middles ware configuration which can be the same. | 04:58 |
| mattoliverau | if it's tempauth.. then you need to have it configured on both, then use something like puppet or ansible to manage (not that tempauth should be used for a production cluster). | 04:59 |
| *** kei_yama_ has quit IRC | 05:00 | |
| *** Guest90246 has quit IRC | 05:01 | |
| *** kei_yama has joined #openstack-swift | 05:01 | |
| *** km has joined #openstack-swift | 05:02 | |
| *** km is now known as Guest64313 | 05:03 | |
| *** darrenc is now known as darrenc_afk | 05:10 | |
| *** thumpba has joined #openstack-swift | 05:14 | |
| *** trifon has joined #openstack-swift | 05:22 | |
| *** klrmn1 has quit IRC | 05:22 | |
| *** jome has quit IRC | 05:27 | |
| *** tsg has joined #openstack-swift | 05:28 | |
| *** jome has joined #openstack-swift | 05:28 | |
| *** darrenc_afk is now known as darrenc | 05:32 | |
| *** thumpba has quit IRC | 05:37 | |
| *** jome has quit IRC | 05:43 | |
| *** links has quit IRC | 05:43 | |
| *** jome has joined #openstack-swift | 05:44 | |
| *** rcernin has joined #openstack-swift | 05:51 | |
| *** thumpba has joined #openstack-swift | 05:52 | |
| *** links has joined #openstack-swift | 05:54 | |
| *** thumpba has quit IRC | 05:59 | |
| *** SkyRocknRoll has quit IRC | 06:03 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Modify unit tests to include real crypto https://review.openstack.org/211451 | 06:15 |
| *** trifon has quit IRC | 06:17 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:20 | |
| *** rcernin has quit IRC | 06:25 | |
| openstackgerrit | Shu Muto proposed openstack/python-swiftclient: Delete python bytecode before every test run https://review.openstack.org/252210 | 06:32 |
| *** ChubYann has quit IRC | 06:38 | |
| *** tsg has quit IRC | 06:40 | |
| *** silor has joined #openstack-swift | 06:57 | |
| *** mahatic has quit IRC | 07:05 | |
| *** silor1 has joined #openstack-swift | 07:06 | |
| *** silor has quit IRC | 07:08 | |
| *** silor1 is now known as silor | 07:08 | |
| *** rcernin has joined #openstack-swift | 07:09 | |
| *** jerrygb has joined #openstack-swift | 07:21 | |
| *** jerrygb has quit IRC | 07:25 | |
| openstackgerrit | Christian Schwede proposed openstack/swift: Add functional test for repeated SLO segments https://review.openstack.org/251909 | 07:58 |
| *** haomaiwang has joined #openstack-swift | 08:04 | |
| *** arnox has joined #openstack-swift | 08:10 | |
| *** kei_yama has quit IRC | 08:11 | |
| *** kei_yama has joined #openstack-swift | 08:17 | |
| *** shakamunyi has quit IRC | 08:27 | |
| *** jmccarthy has quit IRC | 08:27 | |
| *** jmccarthy has joined #openstack-swift | 08:27 | |
| *** jome_ has joined #openstack-swift | 08:29 | |
| *** jome has quit IRC | 08:32 | |
| *** eranrom has joined #openstack-swift | 08:33 | |
| *** ho has quit IRC | 08:41 | |
| *** trifon has joined #openstack-swift | 08:44 | |
| openstackgerrit | renminmin proposed openstack/python-swiftclient: Fix the http request headers logging overwrited https://review.openstack.org/252261 | 08:46 |
| *** jmccarthy has quit IRC | 08:48 | |
| *** jome_ has quit IRC | 08:49 | |
| *** haomaiwang has quit IRC | 08:54 | |
| *** jome has joined #openstack-swift | 08:57 | |
| *** haomaiwa_ has joined #openstack-swift | 08:57 | |
| *** mahatic has joined #openstack-swift | 09:06 | |
| *** jmccarthy has joined #openstack-swift | 09:08 | |
| *** raginbajin has joined #openstack-swift | 09:17 | |
| *** aix has joined #openstack-swift | 09:21 | |
| *** jodah has quit IRC | 09:23 | |
| *** hogepodge has quit IRC | 09:38 | |
| *** hogepodge has joined #openstack-swift | 09:41 | |
| *** jordanP has joined #openstack-swift | 09:43 | |
| *** hseipp has joined #openstack-swift | 09:47 | |
| *** hseipp has quit IRC | 09:47 | |
| *** hseipp has joined #openstack-swift | 09:48 | |
| *** jistr has joined #openstack-swift | 09:50 | |
| *** haomaiwa_ has quit IRC | 09:51 | |
| *** kei_yama has quit IRC | 10:00 | |
| *** jome has quit IRC | 10:06 | |
| *** Guest64313 has quit IRC | 10:07 | |
| *** jmccarthy has quit IRC | 10:08 | |
| *** jmccarthy has joined #openstack-swift | 10:20 | |
| *** weihan has joined #openstack-swift | 10:39 | |
| *** venkat_p has quit IRC | 10:45 | |
| *** jmccarthy has quit IRC | 10:49 | |
| openstackgerrit | Christian Schwede proposed openstack/python-swiftclient: Fix debug and info option parsing https://review.openstack.org/252328 | 10:56 |
| *** acoles_ is now known as acoles | 10:58 | |
| *** venkat_p has joined #openstack-swift | 10:58 | |
| openstackgerrit | Christian Schwede proposed openstack/python-swiftclient: Fix debug and info option parsing https://review.openstack.org/252328 | 10:59 |
| *** jmccarthy has joined #openstack-swift | 11:08 | |
| openstackgerrit | Christian Schwede proposed openstack/python-swiftclient: Fix debug and info option parsing https://review.openstack.org/252328 | 11:15 |
| *** aix has quit IRC | 11:34 | |
| *** venkat_p has quit IRC | 11:36 | |
| *** asettle has quit IRC | 11:43 | |
| *** asettle has joined #openstack-swift | 11:45 | |
| *** venkat_p has joined #openstack-swift | 11:48 | |
| *** lpabon has joined #openstack-swift | 11:59 | |
| *** hseipp has quit IRC | 12:01 | |
| *** trifon has quit IRC | 12:02 | |
| *** trifon has joined #openstack-swift | 12:02 | |
| *** aix has joined #openstack-swift | 12:04 | |
| *** ppai has quit IRC | 12:12 | |
| *** LiuNanke has joined #openstack-swift | 12:16 | |
| *** LiuNanke has quit IRC | 12:22 | |
| *** hseipp has joined #openstack-swift | 12:23 | |
| *** hseipp has quit IRC | 12:23 | |
| *** hseipp has joined #openstack-swift | 12:23 | |
| *** ppai has joined #openstack-swift | 12:26 | |
| *** LiuNanke has joined #openstack-swift | 12:28 | |
| *** LiuNanke has quit IRC | 12:29 | |
| *** jome has joined #openstack-swift | 12:32 | |
| *** silor has quit IRC | 12:34 | |
| *** silor has joined #openstack-swift | 12:37 | |
| *** proteusguy has quit IRC | 12:56 | |
| *** haomaiwa_ has joined #openstack-swift | 12:57 | |
| *** mahatic_ has joined #openstack-swift | 12:57 | |
| *** mahatic has quit IRC | 12:59 | |
| *** hseipp has quit IRC | 13:01 | |
| *** hseipp has joined #openstack-swift | 13:05 | |
| *** mahatic_ has quit IRC | 13:05 | |
| *** jmccarthy has quit IRC | 13:08 | |
| *** proteusguy has joined #openstack-swift | 13:09 | |
| acoles | eranrom: hi, i have a question about your container sync patch 205803 | 13:09 |
| patchbot | acoles: https://review.openstack.org/#/c/205803/ - Container-Sync to iterate only over synced containers | 13:09 |
| *** hseipp has quit IRC | 13:12 | |
| *** tsg has joined #openstack-swift | 13:16 | |
| *** venkat_p has quit IRC | 13:16 | |
| *** lpabon has quit IRC | 13:17 | |
| *** jmccarthy has joined #openstack-swift | 13:25 | |
| *** mahatic_ has joined #openstack-swift | 13:26 | |
| *** mahatic_ has quit IRC | 13:37 | |
| *** thumpba has joined #openstack-swift | 13:39 | |
| *** tsg has quit IRC | 13:52 | |
| *** links has quit IRC | 13:53 | |
| *** sanchitmalhotra has quit IRC | 13:54 | |
| *** jmccarthy has quit IRC | 13:56 | |
| *** venkat_p has joined #openstack-swift | 13:57 | |
| *** venkat_p has quit IRC | 13:58 | |
| *** breitz has quit IRC | 14:05 | |
| *** breitz has joined #openstack-swift | 14:05 | |
| *** thumpba has quit IRC | 14:07 | |
| *** jerrygb has joined #openstack-swift | 14:10 | |
| *** haomaiwa_ has quit IRC | 14:13 | |
| *** petertr7_away is now known as petertr7 | 14:23 | |
| *** haomaiwang has joined #openstack-swift | 14:25 | |
| *** SkyRocknRoll has quit IRC | 14:30 | |
| *** ppai has quit IRC | 14:40 | |
| *** badari has joined #openstack-swift | 14:41 | |
| openstackgerrit | Rico Lin proposed openstack/python-swiftclient: Remove py26 support https://review.openstack.org/252435 | 14:54 |
| *** blmartin has joined #openstack-swift | 14:55 | |
| *** zaitcev has joined #openstack-swift | 14:55 | |
| *** ChanServ sets mode: +v zaitcev | 14:55 | |
| openstackgerrit | Rico Lin proposed openstack/python-swiftclient: Remove py26 support https://review.openstack.org/252435 | 14:56 |
| *** jmccarthy has joined #openstack-swift | 14:57 | |
| petertr7 | Hello! I was wondering if anyone was working on integrating the creation of static large objects into python-swiftclient. Was this something the community would be interested in? | 14:59 |
| petertr7 | I've created a tool that split of a large file into segments, uploads it to swift and creates a manifest file. I was just wondering if potentially contributing this to python-swiftclient would be welcome. The one caveat is that I chose to create pseudo directories on swift to hold the segments. Not sure how you guys feel about that | 15:01 |
| petertr7 | Of course I could always add additional flags to let the user decide where segments are stored. (like in a separate container if they wanted or a specific pseudo folder) | 15:01 |
| *** jordanP has quit IRC | 15:03 | |
| openstackgerrit | Rico Lin proposed openstack/python-swiftclient: Remove py26 support https://review.openstack.org/252435 | 15:03 |
| *** dustins has joined #openstack-swift | 15:04 | |
| petertr7 | Hmm just realized perhaps people in the west coast are no awake yet. | 15:08 |
| openstackgerrit | Rico Lin proposed openstack/python-swiftclient: Remove py26 support https://review.openstack.org/252435 | 15:09 |
| sjmc7 | petertr7, i’m not a swift dev but maybe take a look at http://docs.openstack.org/developer/swift/overview_large_objects.html ? | 15:12 |
| petertr7 | Thanks sjmc7 | 15:12 |
| petertr7 | This is what I'm referring to | 15:12 |
| petertr7 | The tool I've been working on does the steps outlined in the above doc to upload a SLO | 15:13 |
| sjmc7 | ok. the first section mentions support in the swift CLI | 15:13 |
| petertr7 | Ohh, I didn't catch that | 15:13 |
| petertr7 | Thanks, I never realized this | 15:14 |
| sjmc7 | sure | 15:14 |
| petertr7 | Sorry, I just realized, I did know this already | 15:15 |
| petertr7 | The reason I created my own tool | 15:15 |
| petertr7 | was because if the upload of a SLO failed half way through, you'd have to restart | 15:15 |
| petertr7 | My tool does a quick check to pick up where the previous failed upload stopped | 15:15 |
| *** haomaiwang has quit IRC | 15:15 | |
| *** raginbajin has quit IRC | 15:16 | |
| petertr7 | The swift cli I believe does checksum on the segments before attempting to upload it | 15:16 |
| sjmc7 | ah, ok. that sounds like a good addition to have, but you’d be best off talking to a real swift dev :) | 15:16 |
| *** delattec has joined #openstack-swift | 15:16 | |
| petertr7 | haha okay. but thanks for pointing it out | 15:16 |
| sjmc7 | maybe submit a spec and ask for input? | 15:16 |
| *** raginbajin has joined #openstack-swift | 15:17 | |
| petertr7 | I think I'll do that. But I also need to rethink - there is probably a reason why the swift cli does these checksums. Anyways I'll have to think about it more | 15:17 |
| *** delatte has joined #openstack-swift | 15:19 | |
| *** delattec has quit IRC | 15:19 | |
| *** cdelatte has quit IRC | 15:19 | |
| *** jordanP has joined #openstack-swift | 15:21 | |
| *** thumpba has joined #openstack-swift | 15:25 | |
| *** siva_krishnan has joined #openstack-swift | 15:25 | |
| jrichli | petertr7: if you already have the code that does this, its probably best to simply upload the patch as opposed to creating a spec. Specs in swift are typically for large new feature developments. | 15:26 |
| *** petertr7 is now known as petertr7_away | 15:28 | |
| *** thumpba has quit IRC | 15:28 | |
| *** petertr7_away is now known as petertr7 | 15:29 | |
| *** weihan has quit IRC | 15:30 | |
| *** ChanServ has quit IRC | 15:34 | |
| *** raginbajin has quit IRC | 15:34 | |
| *** elligottmc has quit IRC | 15:34 | |
| *** janonymous has quit IRC | 15:34 | |
| *** takashi has quit IRC | 15:34 | |
| *** hogepodge has quit IRC | 15:34 | |
| *** trifon has quit IRC | 15:34 | |
| *** changbl has quit IRC | 15:34 | |
| *** diogogmt has quit IRC | 15:34 | |
| *** tamizh_geek has quit IRC | 15:34 | |
| *** sjmc7 has quit IRC | 15:34 | |
| *** patchbot has quit IRC | 15:34 | |
| *** stevemar has quit IRC | 15:34 | |
| *** sileht has quit IRC | 15:34 | |
| *** chlong has quit IRC | 15:34 | |
| *** amit213 has quit IRC | 15:34 | |
| *** notmyname has quit IRC | 15:34 | |
| *** kota_ has quit IRC | 15:34 | |
| *** okdas has quit IRC | 15:34 | |
| *** peluse has quit IRC | 15:34 | |
| *** jerrygb has quit IRC | 15:34 | |
| *** breitz has quit IRC | 15:34 | |
| *** aix has quit IRC | 15:34 | |
| *** david-lyle has quit IRC | 15:34 | |
| *** chmouel has quit IRC | 15:34 | |
| *** anteaya has quit IRC | 15:35 | |
| *** _fortis has quit IRC | 15:35 | |
| *** dosaboy has quit IRC | 15:35 | |
| *** amandap has quit IRC | 15:35 | |
| *** mattoliverau has quit IRC | 15:35 | |
| *** timur has quit IRC | 15:35 | |
| *** dfg has quit IRC | 15:35 | |
| *** charz has quit IRC | 15:35 | |
| *** hrou has quit IRC | 15:35 | |
| *** lifeless has quit IRC | 15:35 | |
| *** bsdkurt has quit IRC | 15:35 | |
| *** swifterdarrell has quit IRC | 15:35 | |
| *** ndk has quit IRC | 15:35 | |
| *** cschwede has quit IRC | 15:35 | |
| *** glange has quit IRC | 15:35 | |
| *** joearnold has quit IRC | 15:35 | |
| *** alpha_ori has quit IRC | 15:35 | |
| *** a1|away has quit IRC | 15:35 | |
| *** hugokuo has quit IRC | 15:35 | |
| *** portante has quit IRC | 15:35 | |
| *** dabukalam has quit IRC | 15:35 | |
| *** mandarine has quit IRC | 15:35 | |
| *** eikke has quit IRC | 15:35 | |
| *** torgomatic has quit IRC | 15:35 | |
| *** StevenK has quit IRC | 15:35 | |
| *** dmsimard has quit IRC | 15:35 | |
| *** clayg has quit IRC | 15:35 | |
| *** ir2ivps8_ has quit IRC | 15:35 | |
| *** mtreinish has quit IRC | 15:35 | |
| *** bapalm has quit IRC | 15:35 | |
| *** jistr has quit IRC | 15:35 | |
| *** pdardeau has quit IRC | 15:35 | |
| *** MooingLemur has quit IRC | 15:35 | |
| *** bwall has quit IRC | 15:35 | |
| *** zhiyan has quit IRC | 15:35 | |
| *** Guest49572 has quit IRC | 15:35 | |
| *** bkmz has quit IRC | 15:35 | |
| *** onovy has quit IRC | 15:35 | |
| *** ekarlso has quit IRC | 15:35 | |
| *** klrmn has quit IRC | 15:35 | |
| *** fbo_ has quit IRC | 15:35 | |
| *** noqa_v_g1ovnie has quit IRC | 15:35 | |
| *** cppforlife_ has quit IRC | 15:35 | |
| *** zul has quit IRC | 15:35 | |
| *** onder has quit IRC | 15:35 | |
| *** blair has quit IRC | 15:35 | |
| *** briancline has quit IRC | 15:35 | |
| *** acoles has quit IRC | 15:35 | |
| *** tdasilva has quit IRC | 15:35 | |
| *** nottrobin has quit IRC | 15:35 | |
| *** briancurtin has quit IRC | 15:35 | |
| *** jordanP has quit IRC | 15:35 | |
| *** delatte has quit IRC | 15:35 | |
| *** proteusguy has quit IRC | 15:35 | |
| *** asettle has quit IRC | 15:35 | |
| *** arnox has quit IRC | 15:35 | |
| *** gmmaha has quit IRC | 15:35 | |
| *** flaper87 has quit IRC | 15:35 | |
| *** saltsa has quit IRC | 15:35 | |
| *** petertr7 has quit IRC | 15:35 | |
| *** darrenc has quit IRC | 15:35 | |
| *** urth has quit IRC | 15:35 | |
| *** Lickitysplitted_ has quit IRC | 15:35 | |
| *** ejat has quit IRC | 15:35 | |
| *** McMurlock has quit IRC | 15:35 | |
| *** clyps___ has quit IRC | 15:35 | |
| *** j_king_ has quit IRC | 15:35 | |
| *** csmart has quit IRC | 15:35 | |
| *** remix_tj has quit IRC | 15:35 | |
| *** kragniz has quit IRC | 15:35 | |
| *** baffle has quit IRC | 15:35 | |
| *** cebruns has quit IRC | 15:35 | |
| *** balajir has quit IRC | 15:35 | |
| *** swat30 has quit IRC | 15:35 | |
| *** ahale_ has quit IRC | 15:35 | |
| *** jrichli has quit IRC | 15:35 | |
| *** sw3 has quit IRC | 15:35 | |
| *** HenryG has quit IRC | 15:35 | |
| *** zigo has quit IRC | 15:35 | |
| *** zacksh has quit IRC | 15:35 | |
| *** logan- has quit IRC | 15:35 | |
| *** AbyssOne has quit IRC | 15:35 | |
| *** chrisnelson has quit IRC | 15:35 | |
| *** mlanner has quit IRC | 15:35 | |
| *** wbhuber has quit IRC | 15:35 | |
| *** anderstj has quit IRC | 15:35 | |
| *** ctennis has quit IRC | 15:35 | |
| *** timburke has quit IRC | 15:35 | |
| *** EmilienM has quit IRC | 15:35 | |
| *** redbo_ has quit IRC | 15:35 | |
| *** treyd has quit IRC | 15:35 | |
| *** jlhinson has quit IRC | 15:35 | |
| *** wolsen has quit IRC | 15:35 | |
| *** acorwin_ has quit IRC | 15:35 | |
| *** Anticimex has quit IRC | 15:35 | |
| *** jeblair has quit IRC | 15:35 | |
| *** bobby2 has quit IRC | 15:35 | |
| *** hugespoon has quit IRC | 15:35 | |
| *** serverascode has quit IRC | 15:35 | |
| *** sudorandom has quit IRC | 15:35 | |
| *** philipw has quit IRC | 15:35 | |
| *** jroll has quit IRC | 15:35 | |
| *** jamielennox has quit IRC | 15:35 | |
| *** wasmum has quit IRC | 15:35 | |
| *** donagh has quit IRC | 15:35 | |
| *** early has quit IRC | 15:35 | |
| *** CrackerJackMack has quit IRC | 15:35 | |
| *** torgomatic has joined #openstack-swift | 15:37 | |
| *** StevenK has joined #openstack-swift | 15:37 | |
| *** dmsimard has joined #openstack-swift | 15:37 | |
| *** clayg has joined #openstack-swift | 15:37 | |
| *** ir2ivps8_ has joined #openstack-swift | 15:37 | |
| *** wolfe.freenode.net sets mode: +vv torgomatic clayg | 15:37 | |
| *** trifon has joined #openstack-swift | 15:39 | |
| *** changbl has joined #openstack-swift | 15:39 | |
| *** diogogmt has joined #openstack-swift | 15:39 | |
| *** sjmc7 has joined #openstack-swift | 15:39 | |
| *** tamizh_geek has joined #openstack-swift | 15:39 | |
| *** patchbot has joined #openstack-swift | 15:39 | |
| *** stevemar has joined #openstack-swift | 15:39 | |
| *** sileht has joined #openstack-swift | 15:39 | |
| *** chlong has joined #openstack-swift | 15:39 | |
| *** amit213 has joined #openstack-swift | 15:39 | |
| *** notmyname has joined #openstack-swift | 15:39 | |
| *** kota_ has joined #openstack-swift | 15:39 | |
| *** okdas has joined #openstack-swift | 15:39 | |
| *** peluse has joined #openstack-swift | 15:39 | |
| *** wolfe.freenode.net sets mode: +v notmyname | 15:39 | |
| *** jistr has joined #openstack-swift | 15:39 | |
| *** pdardeau has joined #openstack-swift | 15:39 | |
| *** MooingLemur has joined #openstack-swift | 15:39 | |
| *** bwall has joined #openstack-swift | 15:39 | |
| *** zhiyan has joined #openstack-swift | 15:39 | |
| *** Guest49572 has joined #openstack-swift | 15:39 | |
| *** bkmz has joined #openstack-swift | 15:39 | |
| *** onovy has joined #openstack-swift | 15:39 | |
| *** ekarlso has joined #openstack-swift | 15:39 | |
| *** klrmn has joined #openstack-swift | 15:39 | |
| *** fbo_ has joined #openstack-swift | 15:39 | |
| *** noqa_v_g1ovnie has joined #openstack-swift | 15:39 | |
| *** cppforlife_ has joined #openstack-swift | 15:39 | |
| *** zul has joined #openstack-swift | 15:39 | |
| *** onder has joined #openstack-swift | 15:39 | |
| *** blair has joined #openstack-swift | 15:39 | |
| *** raginbajin has joined #openstack-swift | 15:40 | |
| *** zigo has joined #openstack-swift | 15:40 | |
| *** zacksh has joined #openstack-swift | 15:40 | |
| *** logan- has joined #openstack-swift | 15:40 | |
| *** AbyssOne has joined #openstack-swift | 15:40 | |
| *** chrisnelson has joined #openstack-swift | 15:40 | |
| *** mlanner has joined #openstack-swift | 15:40 | |
| *** wbhuber has joined #openstack-swift | 15:40 | |
| *** anderstj has joined #openstack-swift | 15:40 | |
| *** ctennis has joined #openstack-swift | 15:40 | |
| *** timburke has joined #openstack-swift | 15:40 | |
| *** EmilienM has joined #openstack-swift | 15:40 | |
| *** redbo_ has joined #openstack-swift | 15:40 | |
| *** treyd has joined #openstack-swift | 15:40 | |
| *** jlhinson has joined #openstack-swift | 15:40 | |
| *** wolsen has joined #openstack-swift | 15:40 | |
| *** acorwin_ has joined #openstack-swift | 15:40 | |
| *** Anticimex has joined #openstack-swift | 15:40 | |
| *** jeblair has joined #openstack-swift | 15:40 | |
| *** bobby2 has joined #openstack-swift | 15:40 | |
| *** wolfe.freenode.net sets mode: +v timburke | 15:40 | |
| *** mtreinish has joined #openstack-swift | 15:42 | |
| *** bapalm has joined #openstack-swift | 15:42 | |
| *** elligottmc has joined #openstack-swift | 15:42 | |
| *** janonymous has joined #openstack-swift | 15:42 | |
| *** takashi has joined #openstack-swift | 15:42 | |
| *** briancline has joined #openstack-swift | 15:42 | |
| *** acoles has joined #openstack-swift | 15:42 | |
| *** tdasilva has joined #openstack-swift | 15:42 | |
| *** nottrobin has joined #openstack-swift | 15:42 | |
| *** briancurtin has joined #openstack-swift | 15:42 | |
| *** wolfe.freenode.net sets mode: +vv acoles tdasilva | 15:42 | |
| *** hogepodge has joined #openstack-swift | 15:42 | |
| *** hugespoon has joined #openstack-swift | 15:43 | |
| *** serverascode has joined #openstack-swift | 15:43 | |
| *** sudorandom has joined #openstack-swift | 15:43 | |
| *** philipw has joined #openstack-swift | 15:43 | |
| *** jroll has joined #openstack-swift | 15:43 | |
| *** jamielennox has joined #openstack-swift | 15:43 | |
| *** wasmum has joined #openstack-swift | 15:43 | |
| *** donagh has joined #openstack-swift | 15:43 | |
| *** early has joined #openstack-swift | 15:43 | |
| *** CrackerJackMack has joined #openstack-swift | 15:43 | |
| *** trifon has quit IRC | 15:49 | |
| *** aix has joined #openstack-swift | 15:51 | |
| *** _fortis has joined #openstack-swift | 15:52 | |
| *** wshao has joined #openstack-swift | 15:57 | |
| *** proteusguy has joined #openstack-swift | 15:59 | |
| *** chmouel has joined #openstack-swift | 15:59 | |
| *** jerrygb has joined #openstack-swift | 15:59 | |
| *** breitz has joined #openstack-swift | 15:59 | |
| *** david-lyle has joined #openstack-swift | 15:59 | |
| *** eikke has joined #openstack-swift | 15:59 | |
| *** anteaya has joined #openstack-swift | 15:59 | |
| *** dosaboy has joined #openstack-swift | 15:59 | |
| *** amandap has joined #openstack-swift | 15:59 | |
| *** mattoliverau has joined #openstack-swift | 15:59 | |
| *** timur has joined #openstack-swift | 15:59 | |
| *** dfg has joined #openstack-swift | 15:59 | |
| *** charz has joined #openstack-swift | 15:59 | |
| *** lifeless has joined #openstack-swift | 15:59 | |
| *** bsdkurt has joined #openstack-swift | 15:59 | |
| *** swifterdarrell has joined #openstack-swift | 15:59 | |
| *** ndk has joined #openstack-swift | 15:59 | |
| *** glange has joined #openstack-swift | 15:59 | |
| *** cschwede has joined #openstack-swift | 15:59 | |
| *** wolfe.freenode.net sets mode: +vvvv mattoliverau swifterdarrell glange cschwede | 15:59 | |
| *** joearnold has joined #openstack-swift | 15:59 | |
| *** alpha_ori has joined #openstack-swift | 15:59 | |
| *** a1|away has joined #openstack-swift | 15:59 | |
| *** hugokuo has joined #openstack-swift | 15:59 | |
| *** portante has joined #openstack-swift | 15:59 | |
| *** dabukalam has joined #openstack-swift | 15:59 | |
| *** mandarine has joined #openstack-swift | 15:59 | |
| *** rcernin has quit IRC | 15:59 | |
| *** zul has quit IRC | 16:00 | |
| *** dustins has quit IRC | 16:01 | |
| *** kragniz has joined #openstack-swift | 16:03 | |
| *** jmccarthy has quit IRC | 16:07 | |
| openstackgerrit | Christian Schwede proposed openstack/python-swiftclient: Fix debug and info option parsing https://review.openstack.org/252328 | 16:07 |
| *** proteusguy has quit IRC | 16:09 | |
| *** chmouel has quit IRC | 16:09 | |
| *** jerrygb has quit IRC | 16:09 | |
| *** breitz has quit IRC | 16:09 | |
| *** david-lyle has quit IRC | 16:09 | |
| *** anteaya has quit IRC | 16:09 | |
| *** dosaboy has quit IRC | 16:09 | |
| *** amandap has quit IRC | 16:09 | |
| *** mattoliverau has quit IRC | 16:09 | |
| *** timur has quit IRC | 16:09 | |
| *** dfg has quit IRC | 16:09 | |
| *** charz has quit IRC | 16:09 | |
| *** lifeless has quit IRC | 16:09 | |
| *** bsdkurt has quit IRC | 16:09 | |
| *** swifterdarrell has quit IRC | 16:09 | |
| *** ndk has quit IRC | 16:09 | |
| *** cschwede has quit IRC | 16:09 | |
| *** glange has quit IRC | 16:09 | |
| *** joearnold has quit IRC | 16:09 | |
| *** a1|away has quit IRC | 16:09 | |
| *** alpha_ori has quit IRC | 16:09 | |
| *** hugokuo has quit IRC | 16:09 | |
| *** portante has quit IRC | 16:09 | |
| *** dabukalam has quit IRC | 16:09 | |
| *** mandarine has quit IRC | 16:09 | |
| *** eikke has quit IRC | 16:09 | |
| *** diogogmt has quit IRC | 16:17 | |
| *** wshao has quit IRC | 16:31 | |
| *** jome has quit IRC | 16:33 | |
| *** wshao has joined #openstack-swift | 16:34 | |
| *** HenryG has joined #openstack-swift | 16:34 | |
| *** sw3 has joined #openstack-swift | 16:34 | |
| *** jrichli has joined #openstack-swift | 16:34 | |
| *** ahale_ has joined #openstack-swift | 16:34 | |
| *** swat30 has joined #openstack-swift | 16:34 | |
| *** balajir has joined #openstack-swift | 16:34 | |
| *** cebruns has joined #openstack-swift | 16:34 | |
| *** baffle has joined #openstack-swift | 16:34 | |
| *** remix_tj has joined #openstack-swift | 16:34 | |
| *** csmart has joined #openstack-swift | 16:34 | |
| *** j_king_ has joined #openstack-swift | 16:34 | |
| *** clyps___ has joined #openstack-swift | 16:34 | |
| *** McMurlock has joined #openstack-swift | 16:34 | |
| *** ejat has joined #openstack-swift | 16:34 | |
| *** Lickitysplitted_ has joined #openstack-swift | 16:34 | |
| *** urth has joined #openstack-swift | 16:34 | |
| *** darrenc has joined #openstack-swift | 16:34 | |
| *** petertr7 has joined #openstack-swift | 16:34 | |
| *** saltsa has joined #openstack-swift | 16:34 | |
| *** flaper87 has joined #openstack-swift | 16:34 | |
| *** gmmaha has joined #openstack-swift | 16:34 | |
| *** arnox has joined #openstack-swift | 16:34 | |
| *** asettle has joined #openstack-swift | 16:34 | |
| *** delatte has joined #openstack-swift | 16:34 | |
| *** jordanP has joined #openstack-swift | 16:34 | |
| *** hrou has joined #openstack-swift | 16:34 | |
| *** tsg has joined #openstack-swift | 16:34 | |
| *** hseipp has joined #openstack-swift | 16:34 | |
| *** sgundur has joined #openstack-swift | 16:34 | |
| *** diazjf has joined #openstack-swift | 16:34 | |
| *** zul has joined #openstack-swift | 16:34 | |
| *** jmccarthy1 has joined #openstack-swift | 16:34 | |
| *** proteusguy has joined #openstack-swift | 16:34 | |
| *** chmouel has joined #openstack-swift | 16:34 | |
| *** jerrygb has joined #openstack-swift | 16:34 | |
| *** breitz has joined #openstack-swift | 16:34 | |
| *** david-lyle has joined #openstack-swift | 16:34 | |
| *** eikke has joined #openstack-swift | 16:34 | |
| *** anteaya has joined #openstack-swift | 16:34 | |
| *** dosaboy has joined #openstack-swift | 16:34 | |
| *** amandap has joined #openstack-swift | 16:34 | |
| *** mattoliverau has joined #openstack-swift | 16:34 | |
| *** timur has joined #openstack-swift | 16:34 | |
| *** dfg has joined #openstack-swift | 16:34 | |
| *** charz has joined #openstack-swift | 16:34 | |
| *** lifeless has joined #openstack-swift | 16:34 | |
| *** wolfe.freenode.net sets mode: +v mattoliverau | 16:34 | |
| *** bsdkurt has joined #openstack-swift | 16:34 | |
| *** swifterdarrell has joined #openstack-swift | 16:34 | |
| *** ndk has joined #openstack-swift | 16:34 | |
| *** glange has joined #openstack-swift | 16:34 | |
| *** cschwede has joined #openstack-swift | 16:34 | |
| *** joearnold has joined #openstack-swift | 16:34 | |
| *** alpha_ori has joined #openstack-swift | 16:34 | |
| *** a1|away has joined #openstack-swift | 16:34 | |
| *** hugokuo has joined #openstack-swift | 16:34 | |
| *** portante has joined #openstack-swift | 16:34 | |
| *** dabukalam has joined #openstack-swift | 16:34 | |
| *** mandarine has joined #openstack-swift | 16:34 | |
| *** wolfe.freenode.net sets mode: +vvv swifterdarrell glange cschwede | 16:34 | |
| *** jome has joined #openstack-swift | 16:35 | |
| notmyname | good morning | 16:36 |
| *** klrmn1 has joined #openstack-swift | 16:44 | |
| *** jome has quit IRC | 16:44 | |
| *** zul has quit IRC | 16:45 | |
| *** jmccarthy1 has quit IRC | 16:45 | |
| *** proteusguy has quit IRC | 16:45 | |
| *** chmouel has quit IRC | 16:45 | |
| *** jerrygb has quit IRC | 16:45 | |
| *** breitz has quit IRC | 16:45 | |
| *** david-lyle has quit IRC | 16:45 | |
| *** anteaya has quit IRC | 16:45 | |
| *** dosaboy has quit IRC | 16:45 | |
| *** amandap has quit IRC | 16:45 | |
| *** mattoliverau has quit IRC | 16:45 | |
| *** timur has quit IRC | 16:45 | |
| *** dfg has quit IRC | 16:45 | |
| *** charz has quit IRC | 16:45 | |
| *** lifeless has quit IRC | 16:45 | |
| *** bsdkurt has quit IRC | 16:45 | |
| *** swifterdarrell has quit IRC | 16:45 | |
| *** ndk has quit IRC | 16:45 | |
| *** cschwede has quit IRC | 16:45 | |
| *** glange has quit IRC | 16:45 | |
| *** joearnold has quit IRC | 16:45 | |
| *** a1|away has quit IRC | 16:45 | |
| *** alpha_ori has quit IRC | 16:45 | |
| *** hugokuo has quit IRC | 16:45 | |
| *** portante has quit IRC | 16:45 | |
| *** dabukalam has quit IRC | 16:45 | |
| *** mandarine has quit IRC | 16:45 | |
| *** eikke has quit IRC | 16:45 | |
| *** diazjf has quit IRC | 16:45 | |
| *** sgundur has quit IRC | 16:45 | |
| *** hseipp has quit IRC | 16:45 | |
| *** tsg has quit IRC | 16:45 | |
| *** hrou has quit IRC | 16:45 | |
| *** jordanP has quit IRC | 16:45 | |
| *** delatte has quit IRC | 16:45 | |
| *** asettle has quit IRC | 16:45 | |
| *** arnox has quit IRC | 16:45 | |
| *** gmmaha has quit IRC | 16:45 | |
| *** flaper87 has quit IRC | 16:45 | |
| *** saltsa has quit IRC | 16:45 | |
| *** petertr7 has quit IRC | 16:45 | |
| *** darrenc has quit IRC | 16:45 | |
| *** urth has quit IRC | 16:45 | |
| *** Lickitysplitted_ has quit IRC | 16:45 | |
| *** ejat has quit IRC | 16:45 | |
| *** McMurlock has quit IRC | 16:45 | |
| *** clyps___ has quit IRC | 16:45 | |
| *** j_king_ has quit IRC | 16:45 | |
| *** csmart has quit IRC | 16:45 | |
| *** remix_tj has quit IRC | 16:45 | |
| *** baffle has quit IRC | 16:45 | |
| *** cebruns has quit IRC | 16:45 | |
| *** balajir has quit IRC | 16:45 | |
| *** swat30 has quit IRC | 16:45 | |
| *** ahale_ has quit IRC | 16:45 | |
| *** jrichli has quit IRC | 16:45 | |
| *** sw3 has quit IRC | 16:45 | |
| *** HenryG has quit IRC | 16:45 | |
| *** klrmn1 has quit IRC | 16:45 | |
| *** mandarine has joined #openstack-swift | 16:46 | |
| *** dabukalam has joined #openstack-swift | 16:46 | |
| *** portante has joined #openstack-swift | 16:46 | |
| *** hugokuo has joined #openstack-swift | 16:46 | |
| *** alpha_ori has joined #openstack-swift | 16:46 | |
| *** joearnold has joined #openstack-swift | 16:46 | |
| *** cschwede has joined #openstack-swift | 16:46 | |
| *** glange has joined #openstack-swift | 16:46 | |
| *** ndk has joined #openstack-swift | 16:46 | |
| *** swifterdarrell has joined #openstack-swift | 16:46 | |
| *** bsdkurt has joined #openstack-swift | 16:46 | |
| *** lifeless has joined #openstack-swift | 16:46 | |
| *** charz has joined #openstack-swift | 16:46 | |
| *** dfg has joined #openstack-swift | 16:46 | |
| *** timur has joined #openstack-swift | 16:46 | |
| *** mattoliverau has joined #openstack-swift | 16:46 | |
| *** wolfe.freenode.net sets mode: +vvvv cschwede glange swifterdarrell mattoliverau | 16:46 | |
| *** amandap has joined #openstack-swift | 16:46 | |
| *** dosaboy has joined #openstack-swift | 16:46 | |
| *** anteaya has joined #openstack-swift | 16:46 | |
| *** eikke has joined #openstack-swift | 16:46 | |
| *** david-lyle has joined #openstack-swift | 16:46 | |
| *** breitz has joined #openstack-swift | 16:46 | |
| *** jerrygb has joined #openstack-swift | 16:46 | |
| *** chmouel has joined #openstack-swift | 16:46 | |
| *** proteusguy has joined #openstack-swift | 16:46 | |
| *** jmccarthy1 has joined #openstack-swift | 16:46 | |
| *** zul has joined #openstack-swift | 16:46 | |
| *** diazjf has joined #openstack-swift | 16:46 | |
| *** sgundur has joined #openstack-swift | 16:46 | |
| *** hseipp has joined #openstack-swift | 16:46 | |
| *** tsg has joined #openstack-swift | 16:46 | |
| *** hrou has joined #openstack-swift | 16:46 | |
| *** jordanP has joined #openstack-swift | 16:46 | |
| *** delatte has joined #openstack-swift | 16:46 | |
| *** asettle has joined #openstack-swift | 16:46 | |
| *** arnox has joined #openstack-swift | 16:46 | |
| *** gmmaha has joined #openstack-swift | 16:46 | |
| *** flaper87 has joined #openstack-swift | 16:46 | |
| *** saltsa has joined #openstack-swift | 16:46 | |
| *** petertr7 has joined #openstack-swift | 16:46 | |
| *** darrenc has joined #openstack-swift | 16:46 | |
| *** urth has joined #openstack-swift | 16:46 | |
| *** Lickitysplitted_ has joined #openstack-swift | 16:46 | |
| *** ejat has joined #openstack-swift | 16:46 | |
| *** McMurlock has joined #openstack-swift | 16:46 | |
| *** clyps___ has joined #openstack-swift | 16:46 | |
| *** j_king_ has joined #openstack-swift | 16:46 | |
| *** csmart has joined #openstack-swift | 16:46 | |
| *** remix_tj has joined #openstack-swift | 16:46 | |
| *** baffle has joined #openstack-swift | 16:46 | |
| *** cebruns has joined #openstack-swift | 16:46 | |
| *** balajir has joined #openstack-swift | 16:46 | |
| *** swat30 has joined #openstack-swift | 16:46 | |
| *** ahale_ has joined #openstack-swift | 16:46 | |
| *** jrichli has joined #openstack-swift | 16:46 | |
| *** sw3 has joined #openstack-swift | 16:46 | |
| *** HenryG has joined #openstack-swift | 16:46 | |
| notmyname | seems freenode is under some sort of ddos, thus the netsplits | 16:47 |
| notmyname | for those who are here, patch 252308 drops py26 testing from swiftclient. at the last meeting we talked about how that was going to happen | 16:48 |
| patchbot | notmyname: https://review.openstack.org/#/c/252308/ - Remove py26 support from swiftclient | 16:48 |
| *** aix has quit IRC | 16:48 | |
| *** aix has joined #openstack-swift | 16:48 | |
| notmyname | so I propose doing a python-swiftclient release (2.6.1 or 2.7.0) with what's currently HEAD of master | 16:49 |
| openstackgerrit | Alistair Coles proposed openstack/python-swiftclient: Fix debug and info option parsing https://review.openstack.org/252328 | 16:51 |
| jrichli | notmyname: hrou was talkin smack while you cores were not connected | 16:51 |
| *** jordanP has quit IRC | 16:52 | |
| *** chmouel has quit IRC | 16:52 | |
| jrichli | and its not in the online log either! | 16:52 |
| hrou | And apparently there was no history so I guess no one will ever know ;) | 16:52 |
| jrichli | that is, if the 30 people connected at the time can keep the secret ;-) | 16:53 |
| notmyname | heh | 16:53 |
| gmmaha | hehe | 16:53 |
| jrichli | gmmaha was on | 16:54 |
| hrou | I trust that group of folks ;) They'll stay quiet | 16:54 |
| gmmaha | heheh | 16:54 |
| notmyname | hrou: that's ok. while you weren't connected I laid down some new rules for contributing. if you don't follow them to the letter, your code won't be reviewed ;-) | 16:54 |
| gmmaha | i was def on!! | 16:54 |
| *** jordanP has joined #openstack-swift | 16:54 | |
| * gmmaha goes to save the hexchat log to file for the future | 16:54 | |
| jrichli | gmmaha lol | 16:55 |
| jrichli | notmyname: hrou and I were connected most of the time (still not sure why that is). | 16:55 |
| hrou | notmyname, lol (noo I guess there goes the symlink work ;) jk | 16:56 |
| notmyname | probably just depends on which freenode server you're actually connected to | 16:56 |
| hrou | yea I think that was the key | 16:56 |
| jrichli | right, makes sense | 16:56 |
| notmyname | whew. full meeting agenda today | 16:57 |
| *** jistr has quit IRC | 17:01 | |
| *** pgbridge has joined #openstack-swift | 17:02 | |
| *** ChanServ has joined #openstack-swift | 17:04 | |
| *** wolfe.freenode.net sets mode: +o ChanServ | 17:04 | |
| *** nadeem has joined #openstack-swift | 17:05 | |
| *** chmouel has joined #openstack-swift | 17:06 | |
| *** rcernin has joined #openstack-swift | 17:11 | |
| *** wshao has quit IRC | 17:11 | |
| *** gyee has joined #openstack-swift | 17:12 | |
| *** diazjf has quit IRC | 17:19 | |
| timburke | petertr7: we always appreciate better fault tolerance in swiftclient, so i'd be interested in what you had in mind. i also wonder if patch 226897 would address the issue for you | 17:24 |
| patchbot | timburke: https://review.openstack.org/#/c/226897/ - Make LengthWrappers resettable if their _readable ... | 17:24 |
| timburke | (oh, and notmyname: that^^ might be a good candidate to get in before dropping py26 support. and maybe patch 246040? though i have no idea what version of keystoneclient you would need to have py26 support; maybe it would be old enough that it wouldn't emit the warning) | 17:32 |
| patchbot | timburke: https://review.openstack.org/#/c/246040/ - Stop passing attr to keystoneclient when there's n... | 17:32 |
| notmyname | timburke: pretty much for anything that's not landed on master yet, we'll need to have explicit reviewer testing and affirmation that it works on py26 | 17:33 |
| notmyname | technically, that would be after patch 252308 lands, but i'm working under the assumption that it's a done deal | 17:33 |
| patchbot | notmyname: https://review.openstack.org/#/c/252308/ - Remove py26 support from swiftclient | 17:33 |
| notmyname | timburke: so yeah, we could get a few more things in, but it will require more work to do so | 17:33 |
| *** a1|away has joined #openstack-swift | 17:34 | |
| timburke | that's fair. just thought i'd point them out as potentially useful enough to warrant having them in that next release | 17:35 |
| notmyname | ack :-) | 17:35 |
| *** NM has joined #openstack-swift | 17:47 | |
| *** changbl has quit IRC | 17:54 | |
| *** arnox has quit IRC | 17:57 | |
| *** blmartin has quit IRC | 17:58 | |
| zaitcev | I don't understand why encryption wants updateable metadata. Do we have an explanation written down anywhere? | 18:01 |
| *** wuhg has quit IRC | 18:01 | |
| *** klrmn1 has joined #openstack-swift | 18:04 | |
| hrou | zaitcev, what do you mean be updatable metadata ? Are you referring to patch 245885 by any chance ? | 18:05 |
| *** jmccarthy1 has quit IRC | 18:05 | |
| patchbot | hrou: https://review.openstack.org/#/c/245885/ - Append crypto-meta to user-meta | 18:05 |
| *** jordanP has quit IRC | 18:06 | |
| *** changbl has joined #openstack-swift | 18:07 | |
| zaitcev | If I look at the spec (specs/in_progress/at_rest_encryption.rst), it lists a bunch of stuff that it chucks into sysmeta, where it's at risk of being lost if any middleware screws up the update. So 1) I see why we need something stored, 2) I suspect a reason why we want it persist, but we never claim the reason-outcome connection. | 18:15 |
| zaitcev | It looks like it would be possible to live without Massmeta | 18:16 |
| *** dustins has joined #openstack-swift | 18:19 | |
| hrou | zaitcev, In terms what is being stored each piece of user meta has associated crypto meta, namely an IV, this is what needs to be persisted. In terms of the problem with sysmeta, (and the patch above actually does touch on this for fast-post) - the semantics of POST don't allow sysmeta to be updated, and we'd need to set that for the new headers (err new is a bad choice of words, for the headers period given the granularity of user meta). | 18:24 |
| *** sjmc7 has left #openstack-swift | 18:26 | |
| *** jome has joined #openstack-swift | 18:28 | |
| *** hseipp has quit IRC | 18:35 | |
| *** jome has quit IRC | 18:39 | |
| *** jome has joined #openstack-swift | 18:42 | |
| clayg | heyoh! | 18:46 |
| *** jome has quit IRC | 18:54 | |
| *** sgundur has quit IRC | 18:57 | |
| clayg | torgomatic: nice work on the slo segment fix - thanks everyone who reviewe! | 18:57 |
| clayg | anyone have a handle for "Lorcan Browne" ??? | 18:58 |
| clayg | he said he was running ring tests on patch 241571 but gave no indication of how they're going :'( | 18:58 |
| patchbot | clayg: https://review.openstack.org/#/c/241571/ - Put part-replicas where they go | 18:58 |
| acoles | clayg: don't think he hangs out here, he's in my team though | 18:58 |
| acoles | clayg: and they're going ok i believe | 18:59 |
| *** siva_krishnan has quit IRC | 18:59 | |
| *** siva_krishnan has joined #openstack-swift | 18:59 | |
| *** sgundur has joined #openstack-swift | 19:00 | |
| clayg | acoles: phew | 19:00 |
| clayg | acoles: is ec multi frag gee-double-oh-dee-to-oh-gee-oh? | 19:00 |
| *** sgundur has quit IRC | 19:02 | |
| *** klrmn1 has quit IRC | 19:03 | |
| acoles | clayg: if you mean would i like you to review patch 231121 again, then abso-el-triple-o-tee-lee | 19:04 |
| patchbot | acoles: https://review.openstack.org/#/c/231121/ - Make ECDiskFile report all fragments found on disk | 19:04 |
| clayg | acoles: oh - and pick a follow up - there's three patches that depend on that one - i'm not going to get to all of them today | 19:04 |
| acoles | clayg: sure, i am just rebasing them actually. | 19:05 |
| clayg | i'm leaning towards patch 232684, but patch 181407 also seems pretty smart | 19:05 |
| patchbot | clayg: https://review.openstack.org/#/c/232684/ - Don't ssync data when only a durable is missing | 19:05 |
| patchbot | clayg: https://review.openstack.org/#/c/181407/ - EC: Avoid conflicts when ssync'ing fragments | 19:05 |
| *** klrmn1 has joined #openstack-swift | 19:07 | |
| acoles | clayg: 232684 then 181407 is probably a good order to choose | 19:08 |
| * clayg got something right for a change! | 19:08 | |
| clayg | acoles: is really is annoying that you can't use bisect on a reverse sorted list with a custom key :'( | 19:09 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Enable object server to return non-durable data https://review.openstack.org/215276 | 19:09 |
| clayg | acoles: OMG I love _split_gt[e]_timestamp!!! | 19:09 |
| acoles | clayg: i have a philosophy that is a review comment has gone past 3 or 4 lines then the code probably needs changing | 19:10 |
| clayg | it's like a 10 second grok - gah - so freaking good | 19:10 |
| acoles | clayg: so i am looking at Lorcan's spreadsheet, which needs some cleansing before we can share. there's a couple of case where he sees balance 999.99 with the new builder | 19:12 |
| clayg | acoles: I think your philosophy services you well - I also have this nose for kwargs to methods that made it do two drastically different things - most of the time IME it means it'd look better as two methods | 19:12 |
| clayg | acoles: that might happen when there's only a single device in a server/zone for sure | 19:13 |
| acoles | clayg specifically when setting weight to zero on a bunch of devices | 19:13 |
| clayg | oic | 19:13 |
| clayg | like not all parts get shed - that's not great - unless min_part_hours is just keeping them from moving? which would be fine/expected | 19:13 |
| acoles | i'm pretty sure he used the trick to force min_part_hours to zero (there's some kind of flag for that i think?) cos he's run a gazillion scenarios in last day or so | 19:16 |
| clayg | acoles: awesome! ok so maybe there's an issue - that's great! I wonder how we can tease out the minimum viable test case.. | 19:17 |
| clayg | torgomatic: in your work with fuzz or the ring analyzier did you go over anything with removing devices and checking balance? | 19:18 |
| acoles | clayg: so he is stepping weight down to zero before then removing the devices, and when they are removed balance is 0, its just weight zero where he sees 999 | 19:20 |
| *** aix has quit IRC | 19:20 | |
| clayg | acoles: sure does *sound* like min_part_hours? | 19:20 |
| acoles | i can ask him to repeat to double check | 19:20 |
| clayg | acoles: I think if weight is 0 even one part on the devices is going to say the balance is *terrible* - and if min_part_hours is locking it up there's not much to be done | 19:21 |
| acoles | makes sense | 19:21 |
| clayg | OTOH it's quite possible those last few parts locked on the zero wieght device are just in the "swapped part" scenario - and the new gathers aren't picking them up because they think they won't have anywhere else to go | 19:21 |
| acoles | ok so will ask him to check min_part_hours for those. and if it repeats than we could look at sharing the ring data. | 19:22 |
| clayg | but the removed device case is different - removed devices just have their parts gobbled up no matter what - we could do something similar for zero weight devices - leaving the only condition that would leave a part on a dev with zero weight as min_part_hours | 19:22 |
| clayg | it depends why you're reducing the weight to zero operationally I guess - maybe - it's probably better to reassign them - I think we could write a test for this | 19:23 |
| clayg | s/anywhere else/anywhere better/g | 19:23 |
| clayg | but maybe any other device is better than a zero weight device | 19:24 |
| clayg | acoles: if these rings are artificial the easiest thing might be to share one with me via google drive (or a tempurl) | 19:24 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Don't ssync data when only a durable is missing https://review.openstack.org/232684 | 19:25 |
| acoles | clayg: ^^ rebased | 19:25 |
| clayg | if it's the swapped-part-problem (there's a couple of tests in the change that talk about it) I won't be able to say for sure unless I can see the replica2part2dev | 19:25 |
| clayg | acoles: and tell this cat to get him some IRC!? | 19:26 |
| openstackgerrit | Christian Schwede proposed openstack/python-swiftclient: Fix debug and info option parsing https://review.openstack.org/252328 | 19:26 |
| acoles | clayg: i'll speak with lorcan tomorrow. in some case the rings are from production. | 19:26 |
| clayg | acoles: good for him! | 19:26 |
| acoles | clayg: gotta go get son from geetaar lesson, back for meeting time later | 19:27 |
| clayg | acoles: k, awesome | 19:27 |
| *** acoles is now known as acoles_ | 19:27 | |
| *** briancurtin has quit IRC | 19:33 | |
| *** serverascode has quit IRC | 19:33 | |
| *** nottrobin has quit IRC | 19:33 | |
| *** zhiyan has quit IRC | 19:33 | |
| jrichli | zaitcev: you bring up a good point, though. the spec is a bit out of date. Crypto is no longer adding 'massmeta'. It is making use of sysmeta, though, to store data required for the decryption. | 19:34 |
| clayg | yay sysmeta! | 19:37 |
| clayg | RIP massmeta | 19:38 |
| zaitcev | That means... Alasdair is going to make it so sysmeta persists across POST somehow? Right? | 19:38 |
| zaitcev | Or we're just asking all middleware to be careful about not deleting encryption sysmeta? | 19:39 |
| clayg | zaitcev: it sorta already does - it's stored with the .datafile and bleeds through .meta files - on POST-AS-COPY it'll slurp up any sysmeta it finds and keep it with the new datafile | 19:39 |
| zaitcev | clayg: I figured that out, but I wanted to review fast-post | 19:39 |
| clayg | zaitcev: middleware should namespace it's sysmeta, there's no protection against middleware an operator puts in the pipeline being incompatible (stoping on each others keys) | 19:40 |
| clayg | the *encryption* middleware has to be internally consistent about storing encyrption data that goes wiht the data and metadata in deifferent keys | 19:40 |
| clayg | IIRC | 19:40 |
| clayg | I'm should just let jrichli explain it ;) | 19:40 |
| jrichli | clayg: you are doing fine :-) zaitcev: there was one small change made to sysmeta behavior. one sec | 19:44 |
| jrichli | zaitcev: in proxy/controllers/obj, there is a place that says "# post-as-copy: ignore new sysmeta, copy existing sysmeta" | 19:46 |
| notmyname | I've got another meeting about to start that will last right up to the swift meeting. might be one or two minutes late starting | 19:46 |
| jrichli | it used to do "remove_items(sink_req.headers, condition)", crypto branch has removed that action | 19:46 |
| jrichli | notmyname: sounds like a fun afternoon :-) | 19:47 |
| jrichli | oh, of course, the meeting is at 3 now. i was thinking 4 | 19:48 |
| zaitcev | jrichli: thanks | 19:48 |
| jrichli | zaitcev: np | 19:48 |
| zaitcev | Since we're at it, does anyone actually use the encryption already? | 19:48 |
| openstackgerrit | Tim Burke proposed openstack/python-swiftclient: Stop passing attr to keystoneclient when there's no filter_value https://review.openstack.org/246040 | 19:49 |
| zaitcev | Not afraid of setting the middleware into pipelines and then one day upgrade the cluster to Mitaka and not being able to decrypt anything? | 19:49 |
| jrichli | zaitcev: no, nobody's using it since not all the functests pass yet (big things like COPY wouldnt work). | 19:50 |
| jrichli | zaitcev: as far as future-proofing, we have had talks about what would happen if crypto is taken in and out of the pipeline. We are trying to make the best decisions for these types of scenarios | 19:51 |
| jrichli | cluster upgrades would only be an issue if we make a change at some point that is not backwards incompatible. | 19:52 |
| zaitcev | jrichli: obviously taking it out is going to ruin GETs (although perhaps not DELETEs). I'm more concerned about better sysmeta being added/removed and such. | 19:52 |
| jrichli | zaitcev: ah, so the effects of changing the sysmeta behavior? I would think acoles is the best person to speak with about those concerns | 19:54 |
| zaitcev | okay | 19:55 |
| zaitcev | I'll just go along with fast-post as is, then. If you need changes in the future, we'll take them in when crypto branch gets in, I suppose. | 19:55 |
| jrichli | zaitcev: there are issues with crypto + fast-post right now that pchng is working on. So, again, the spec is out of date, and does not capture some of the latest details of how all the crypto-meta will actually work. | 19:57 |
| jrichli | zaitcev: some details are here https://trello.com/c/018BLfj1/58-fix-errors-when-using-fast-post | 19:58 |
| *** cppforlife_ has quit IRC | 19:59 | |
| *** diazjf has joined #openstack-swift | 20:01 | |
| hrou | jrichli, for fast post the patch 245885 - some interesting discussions there. | 20:06 |
| patchbot | hrou: https://review.openstack.org/#/c/245885/ - Append crypto-meta to user-meta | 20:06 |
| jrichli | hrou: thanks. I had seen all the comments there except for your recent ones. The first couple comments there are what sparked the updates to the trello card. | 20:08 |
| jrichli | hrou: i will read your comments now | 20:08 |
| *** sgundur has joined #openstack-swift | 20:09 | |
| hrou | jrichli, and don't read too much to them, just some random thoughts :) | 20:09 |
| *** nottrobin has joined #openstack-swift | 20:11 | |
| *** silor has quit IRC | 20:11 | |
| *** blmartin has joined #openstack-swift | 20:15 | |
| jrichli | hrou: for the case you describe here "you can POST an object that is plain txt after enabling crypto and we'll want to be just encrypting that header. ", the body crypto-meta being absent should cause override flag to be set, so no encrypting of hdrs. | 20:16 |
| *** serverascode has joined #openstack-swift | 20:17 | |
| jrichli | hrou: that is something i mentioned in the trello card - that it is sort of a requirement that we have unintentionally put onto a keymaster, i think | 20:17 |
| hrou | jrichli, oh so in that scenario a plain txt object, that you POST (with fast-post or not) will never be encrypted ? | 20:18 |
| jrichli | hrou: IIRC. I need to make sure we have a test that asserts this behavior. | 20:18 |
| *** zhiyan has joined #openstack-swift | 20:19 | |
| hrou | jrichli, interesting as I think that was acoles_ concern " ... But we cannot assume that e.g. where there are existing objects that pre-date crypto being enabled, or when crypto has been selectively applied. ..." | 20:19 |
| jrichli | hrou: I believe acoles concerns for this are all involving a scenario where crypto middleware is taken in and out of the pipeline | 20:20 |
| *** tsg has quit IRC | 20:22 | |
| jrichli | hrou: so, the problem is more about if you have an existing encrypted thing, then POST when crypto is not present, then put crypto back in and GET. | 20:22 |
| *** tsg has joined #openstack-swift | 20:22 | |
| *** klrmn1 has quit IRC | 20:22 | |
| jrichli | hrou: but maybe i need to read all this again later tonight and think deeply :-) | 20:23 |
| hrou | jrichli, its funny I was about to type the very same thing as I didn't think that was the original concern, but mind you I see why that's an issue too : ) | 20:23 |
| jrichli | hrou: I think that its never a problem if an existing object body is unencrypted, because we have always planned for that | 20:24 |
| hrou | jrichli, less fast post ? | 20:24 |
| jrichli | hrou: I guess what I mean is that by design, if its an unencrypted body, there is no way anything for that object will ever be encrypted | 20:25 |
| jrichli | with any feature | 20:25 |
| jrichli | only because the keymaster looks to the body crypto-meta to make the decision | 20:25 |
| jrichli | so ... again, it seems like a requirement of the keymaster to check that | 20:26 |
| *** cppforlife_ has joined #openstack-swift | 20:26 | |
| *** briancurtin has joined #openstack-swift | 20:27 | |
| eranrom | acoles: ok, gotcha. I got convinced that the update can take place from PUT/POST - just call the update after doing broker.update_metadata | 20:30 |
| eranrom | acoles: please ignore my last comment over the patch, I will post a new one with your suggestion soon. | 20:31 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 20:32 |
| clayg | eranrom: don't forget to take the extra replicate step out of the probe test for container sync? | 20:32 |
| clayg | or at least - it *should* work without it ;) | 20:32 |
| eranrom | clayg: Yeah I was just thinking about it. | 20:35 |
| jrichli | for those following the crypto talk, hrou pointed out that the post as copy is a GET + PUT. So, we have both remembered the reason crypto will still leave an unencrypted thing alone is that keymaster needs to look for PUT to be from a copy. | 20:37 |
| *** ho has joined #openstack-swift | 20:55 | |
| * notmyname is hoping this meeting finishes in the next 3 minutes | 20:57 | |
| clayg | notmyname: drop the mic | 20:57 |
| kota_ | mornin | 20:58 |
| openstackgerrit | Peter Chng proposed openstack/swift: Add round-trip encrypter/decrypter unit tests https://review.openstack.org/251606 | 20:58 |
| mattoliverau | morning | 20:58 |
| notmyname | analyst briefing about swift and swiftstack | 20:58 |
| notmyname | clayg: "swift is awesome. /me drops mic" | 20:58 |
| jrichli | the meeting before us is still going on anyway | 20:59 |
| *** acoles_ is now known as acoles | 20:59 | |
| jrichli | they ended now | 21:00 |
| notmyname | right on time, my other meeting ended too | 21:00 |
| notmyname | let me move back to my desk and we'll do our swift meeting | 21:00 |
| *** ober37 has joined #openstack-swift | 21:01 | |
| *** ober37 has left #openstack-swift | 21:01 | |
| *** ober37 has joined #openstack-swift | 21:01 | |
| *** cutforth has joined #openstack-swift | 21:02 | |
| notmyname | all right | 21:02 |
| notmyname | meeting time in #openstack-meeting | 21:02 |
| openstackgerrit | Peter Chng proposed openstack/swift: Add round-trip encrypter/decrypter unit tests https://review.openstack.org/251606 | 21:04 |
| acoles | eranrom: ack, sounds good, look forward to next patchset | 21:05 |
| acoles | zaitcev: you are asking great questions! fully updateable sysmeta on object POSTs is tricky. they have something on feature/crypto that meets their needs, but IIRC I was nervous about it. | 21:08 |
| *** klrmn1 has joined #openstack-swift | 21:12 | |
| *** diazjf1 has joined #openstack-swift | 21:15 | |
| zaitcev | acoles: I was just wondering... Once we decree that parts of metadata (user or system) persist, then previously unnecessary explicit delete becomes necessary. The usual approach is to overload with e.g. empty key: post with A= causes A to be deleted. Then a compatibility issue possibly arises. I'm probably naive considering this | 21:17 |
| *** diazjf has quit IRC | 21:17 | |
| zaitcev | acoles: we may say that IV and cipher are never to be deleted anyway, so the fact that persistent meta cannot be deleted (only updated) is of no concern | 21:17 |
| acoles | zaitcev: problem with posting A= to delete A=old to an object is that we can't go back and delete from the object file, so we either remember A= for ever, or at some point A=old comes back to haunt | 21:22 |
| acoles | which is ok if a piece of middleware can be reliedon to always set some value for A, even if it is empth (deleted) | 21:24 |
| acoles | empty* | 21:25 |
| *** diazjf1 has quit IRC | 21:31 | |
| notmyname | tsg: #openstack-meeting for pyeclib update | 21:31 |
| acoles | zaitcev: btw thanks for looking at the fast post stuff, i hope to update the patch soon, when i get time | 21:31 |
| notmyname | tsg: thanks | 21:32 |
| clayg | acoles: how quick do you gotta run off after the meeting? I'm looking at _verify_on_disk_files using the '.ts' keys and am totally confused? | 21:37 |
| clayg | acoles: this doesn't bode well => https://gist.github.com/clayg/82411b8076303a2cd1a0 | 21:38 |
| acoles | clayg: i can hang around for a bit | 21:39 |
| acoles | i hate those verify methods | 21:40 |
| clayg | acoles: heheheh | 21:40 |
| clayg | acoles: i just wanted to confirm that a) .ts .meta .data files in the results dict are gone (diff sorta verifies that) and b) see if you have any pointers to tests that *should* be raising asserts for invalid files on disk that apparenlty aren't hitting the code right | 21:41 |
| *** takashi_k has joined #openstack-swift | 21:42 | |
| acoles | clayg: oh thats bad :/ just shows why its such a bad idea to pass untyped dict things around in a language rather than use proper objects with like real interface :0 | 21:44 |
| * acoles goes off to rant | 21:44 | |
| clayg | acoles: tdd says even with strong types your code is broken unless you test it | 21:45 |
| clayg | acoles: but... i feel ya | 21:45 |
| clayg | makes refactors harder :\ - but bags of dicts are just so damn handy! | 21:46 |
| clayg | acoles: why didn't you make the file_info results a class if you were so sure strong types are better! :P | 21:46 |
| clayg | acoles: py3 has some loosey goosey type support you might be into | 21:46 |
| acoles | clayg: you know i NEARLY did !! | 21:48 |
| acoles | clayg: but yeah, there's a hole in tests for sure | 21:48 |
| torgomatic | I love strongly typed languages with interfaces and contracts and compiler-enforced correctness that *still* let every reference be null and blow up your code at runtime | 21:48 |
| clayg | acoles: well if you have an idea where to look i'm all ears - otherwise I'll keep digging and leave you a crumb with my comments | 21:49 |
| clayg | go go code review! | 21:49 |
| clayg | torgomatic: I think you said that because you *don't* mean it? | 21:49 |
| clayg | torgomatic: sometimes in this line of work we forget what love is | 21:50 |
| acoles | clayg: i'll look for the test that should cover it after meeting | 21:50 |
| clayg | acoles: *hugs* | 21:52 |
| acoles | clayg: i said i'd look, i didn't say i'd find ;) | 21:53 |
| clayg | lol | 21:54 |
| ho | acoles: thanks for reviewing of patch 202411. i have a question about your expectation value for the test. now it takes 70s in your env. | 22:00 |
| patchbot | ho: https://review.openstack.org/#/c/202411/ - Add functional test for access control (RBAC) with... | 22:00 |
| notmyname | next meeting time. be back in 30-45 | 22:01 |
| acoles | ho: hi! | 22:02 |
| *** NM has quit IRC | 22:02 | |
| *** ober37 has quit IRC | 22:03 | |
| timburke | speaking of swiftclient (from the meeting), i'm interested in people's opinions on patch 184957 | 22:03 |
| patchbot | timburke: https://review.openstack.org/#/c/184957/ - Add --decode-content option | 22:03 |
| timburke | bit of background: it used to be that if you uploaded for example a .tar.gz file and included a Content-Encoding: gzip header, swiftclient would (1) decode the content, leaving you with just a tar file, and (2) claim that the download failed, because the MD5 (of the *decoded* content) no longer matched the ETag (from the *encoded* content) | 22:04 |
| acoles | clayg: so re. (a) yes .ts .meta and .data keys should be gone from results dict | 22:04 |
| timburke | this was fixed in patch 184659, but it occurred to me that some users may have been expecting the decode-it-for-me behavior and just ignored the error (since the file would still get written). so i wrote 184957 to add an option to restore the previous behavior sans error | 22:04 |
| patchbot | timburke: https://review.openstack.org/#/c/184659/ - Stop decoding object content (MERGED) | 22:04 |
| timburke | so i have two questions for people: first, does this sound like something you would actually want and use? or do you feel more like "*shrug* seems cool; couldn't hurt"? | 22:04 |
| timburke | second, what are your opinions on swiftclient's role here? should it act primarily as a transport layer and just move bytes between server and client, or should it try to interpret those bytes? if the latter, under what circumstances? | 22:04 |
| ho | acoles: hi, i would like to know targeted time. 35s? | 22:05 |
| *** petertr7 is now known as petertr7_away | 22:06 | |
| *** NM has joined #openstack-swift | 22:06 | |
| pdardeau | peluse: thx for reviewing 239534 | 22:07 |
| peluse | pdardeau: welcome! | 22:08 |
| acoles | ho: i don't have a target time in mind. i think we just have to try our best to make the tests as time-efficient as possible. for some of the scenarios i thought it would be ok to re-use the same resource and that did seem to save significant time. | 22:08 |
| jrichli | acoles zaitcev: I know acoles is gonna turn into a pumpkin soon, so I dont want to stir too much discussion. but I was wondering if we could designate a sub-set of sysmeta ... like designate "persisted" sysmeta | 22:10 |
| ho | acoles: i see. i prefer to not having dependency b/w cases so my idea is using random.samples() to pick up small num of cases based on normal distribution matrix. | 22:11 |
| ho | acoles: what do you think? | 22:11 |
| zaitcev | jrichli: we could, the problem is that I can't imagine what the long-term consequences are going to be. | 22:12 |
| jrichli | zaitcev: the list would be carefully controlled. only things that must persist like crypto-meta necessary for decrypting | 22:12 |
| pdardeau | timburke: re swiftclient's role, i think by default it should just be transport only. only interpret (do value-add processing) if requested (by whichever mechanism) | 22:13 |
| acoles | ho: i see two problems with randomizing test scenarios - 1) can you be confident that a patch should merge if not all scenarios were tested? and 2) its hard to reproduce a test failure if the next run chooses different scenarios. | 22:13 |
| *** dustins has quit IRC | 22:15 | |
| pdardeau | timburke: i wouldn't expect any value-add processing just by setting Content-Encoding header | 22:15 |
| *** takashi_k has quit IRC | 22:16 | |
| *** jome has joined #openstack-swift | 22:16 | |
| ho | acoles: i see. if i have pooling resources it lose a chance to execute the test parallel with other way (testr) | 22:20 |
| timburke | pdardeau: that was basically where i arrived while fixing the original bug; we *definitely* shouldn't decode by default. but should the client even have options to do decoding? or should we intentionally limit our scope, and leave such interpretation to the user/application? | 22:20 |
| pdardeau | timburke: if i'm writing a swift application, i'd expect to handle it myself. not to say it shouldn't be there, but i think it opens pandora's box | 22:22 |
| acoles | jrichli: zaitcev its ok to "persist" object sysmeta is it is set in every POST. The problem comes when you try to keep an old sysmeta item and 'merge' it with a new set of metadata from a new POST. you risk violating the principle that every piece of state MUST be maintained with the timestamp at which it was created. account/container servers do just that. It is possible that object servers could but its imho a lot of wo | 22:22 |
| acoles | rk. See https://review.openstack.org/#/c/109314/1/specs/swift/updateable-obj-sysmeta.rst | 22:22 |
| hrou | acoles, and that's only problematic with fast post (with none fast post we're just getting a new set of everything) ? | 22:24 |
| acoles | ho: I don't think the func tests suite can run in parallel without a lot of changes since the tests shares account and rely on setting up state in those accounts (maybe even shared containers). | 22:24 |
| acoles | hrou: correct | 22:24 |
| zaitcev | acoles: That's what I don't get about what you said earlier, about A=. If we do _not_ have updateable metadata, then every metadata has its own timestamp (except Content-Type), and therefore lasting A= is unnecessary. May simply remove it. But you said "A=old comes back to haunt". How? | 22:25 |
| zaitcev | acoles: If we do, then perfect, the stuck A= means "deleted" in both API and implementation. | 22:25 |
| timburke | pdardeau: that's my fear. i wrote 184957 because i saw that 184659 potentially broke existing behavior for people, but i'm happy to abandon it if we don't want to set a precedent of pushing application logic into the client | 22:26 |
| acoles | zaitcev: if I have A=old in to.data file and A=new in t1.meta, GET will show me A=new. If I now replace t1.meta with t2.meta that has no A, GET will show me A=old :( Unless I always copy the complete set of metadata from to.data to t1.meta, then from t1.meta to t2.meta and never read metadata from t0.data. But to do that I would need to tag every piece of metadata with its timestamp when copying into tx.meta. Which is do- | 22:29 |
| acoles | able. | 22:29 |
| acoles | BUT... t2.meta on server x could have different content from t2.meta on server y if there have been failed requests to one or other server. | 22:29 |
| acoles | so they are inconsistent, but their names are the same. So that spec linked above tried to deal with that. | 22:30 |
| zaitcev | *scribble scribble* | 22:30 |
| acoles | zaitcev: now, if middleware *always* tells object server the value of A in every POST then object server never needs to copy it from an old .meta to a new .meta and that works just fine. | 22:31 |
| acoles | ...until one day the middleware goes missing and a post creeps in with no A...how does the object server know NOT to use the old A value in the .data file? | 22:32 |
| acoles | ...if as jrichli suggests we declare a sysmeta namespace that will only be returned in GETs if it is in the latest .meta file then that would work...which is pretty much what the massmeta proposal was ! | 22:33 |
| *** cutforth has quit IRC | 22:33 | |
| ho | acoles: in the test cases, it is necessary to configure acl on containers for each test. if i share the container, it's neccesary to remove acl and put acl ... the framework became complex and is it really fast than creating new container with acl??? | 22:35 |
| jrichli | acoles: I do realize this suggestion is like massmeta, but I thought maybe it was more palatable to clayg since it wasn't a new namespace? | 22:35 |
| jrichli | or, i mean ... designate a subset somehow - i a way that clayg is more comfortable wiht | 22:35 |
| *** ianbrown has joined #openstack-swift | 22:36 | |
| hrou | acoles, jrichli - so instead of a new namespace, there's just a list of meta prefixed (e.g. crypto) you treat differently ? In the sense of their lack of presence in the newest meta implies they're not set? | 22:36 |
| acoles | clayg: ok, back to testing the verify methods - on reflection, I don't there are tests that cause the _verify_ondisk_files asserts to blow up, because I *think* they only blow up if there is a bug in the code, where bug could be an ondisk file scenario that we haven't thought of and therefore haven't dealt with in get_ondisk_files. | 22:36 |
| hrou | That was more a jrichli question. | 22:36 |
| acoles | clayg: make sense? I never liked it but I guess it does mean we might get a log if there is a scenario we haven't coded for. | 22:36 |
| jrichli | hrou: something like that. i haven't thought it out fully yet | 22:37 |
| acoles | clayg: of course, the current patchset is bad. but i'm not sure how to test other than to call the method directly and pass in a dict with both ts_file and data_file set etc | 22:38 |
| acoles | actually thats quite reasonable ^^ | 22:38 |
| jrichli | me thinks clayg went to the meeting with notmyname | 22:39 |
| acoles | jrichli: well you better hope clayg isn't following along or he'll spot the flanking manouver :-) | 22:39 |
| ho | acoles: thank for this conversation. I will think it more. thanks! | 22:39 |
| hrou | acoles, lol (its renamed thats all that matters ;) | 22:40 |
| acoles | ho: i agree that with ACLs the resource needs to be setup | 22:40 |
| * jrichli looks up flanking maneuver | 22:40 | |
| jrichli | ah, nice | 22:40 |
| notmyname | back from meetings | 22:41 |
| acoles | ho: no problem, have a good day! | 22:41 |
| acoles | jrichli: maybe its out-flanking, idk | 22:42 |
| jrichli | notmyname: you missed out on acoles having 3 discussions or more at once | 22:42 |
| acoles | jrichli: and thats just in irc! | 22:42 |
| jrichli | wow! | 22:42 |
| jrichli | you should feel very needed :-) | 22:43 |
| acoles | jrichli: actually there is nothing else happening | 22:44 |
| jrichli | acoles: lol :-) | 22:44 |
| acoles | clayg: ok i was going to push rebased patch 181407 but there's no point since 231121 needs fixing | 22:47 |
| patchbot | acoles: https://review.openstack.org/#/c/181407/ - EC: Avoid conflicts when ssync'ing fragments | 22:47 |
| *** eranrom has quit IRC | 22:47 | |
| acoles | clayg: lemme know what you think about adding a test to directly call the _verify_ondisk_files method with crazy dicts. | 22:49 |
| *** lyrrad has joined #openstack-swift | 22:51 | |
| *** ianbrown has quit IRC | 22:54 | |
| *** sgundur has quit IRC | 23:00 | |
| clayg | acoles: not pushing the rebase yet is ok by me | 23:02 |
| clayg | acoles: I have some comments on patch 231121 too - it's getting late for you yeah? | 23:02 |
| patchbot | clayg: https://review.openstack.org/#/c/231121/ - Make ECDiskFile report all fragments found on disk | 23:02 |
| clayg | acoles: if you think calling _verify directly is the best test maybe that's ok - but I was thinking it be better to have a [(<<list_of_files_on_disk>, <{expected_attributes}|Exception>), ...] test | 23:04 |
| ho | tdasilva: around? | 23:07 |
| acoles | clayg: understand, but there *ought* to be no list of files on disk that could generate the assertion error | 23:10 |
| * acoles now waits to be proven badly wrong | 23:11 | |
| *** km has joined #openstack-swift | 23:11 | |
| *** km is now known as Guest36914 | 23:12 | |
| *** dmorita has joined #openstack-swift | 23:16 | |
| *** acoles has quit IRC | 23:20 | |
| *** dmorita has quit IRC | 23:22 | |
| *** dmorita has joined #openstack-swift | 23:23 | |
| *** acoles_ has joined #openstack-swift | 23:27 | |
| *** ChanServ sets mode: +v acoles_ | 23:27 | |
| *** acoles_ is now known as acoles | 23:28 | |
| *** kei_yama has joined #openstack-swift | 23:33 | |
| *** siva_krishnan has quit IRC | 23:42 | |
| asettle | notmyname just got your email! Thank you :) i'll review today :) | 23:52 |
| notmyname | thanks | 23:53 |
| asettle | No, thank you :) | 23:53 |
| *** cryptsting has joined #openstack-swift | 23:56 | |
| *** nadeem has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!