| clayg | bah, I think the hope of regression is out the window - it's been like that since it was added in 156f27c9 | 00:04 |
|---|---|---|
| clayg | i'm going to go back and see if I can get a pipeline that works just to be sure | 00:05 |
| *** thumpba_ has quit IRC | 00:06 | |
| clayg | bah - no conf.d support :\ | 00:07 |
| *** ho has joined #openstack-swift | 00:09 | |
| ho | morning! | 00:09 |
| clayg | ImportError: No module named webob -> wow this is getting silly quickly :\ | 00:12 |
| *** kota_ has joined #openstack-swift | 00:16 | |
| tdasilva | clayg: started this etherpad just to help out get all scenarios listed out: https://etherpad.openstack.org/p/object_version_and_ACL_use_cases | 00:17 |
| *** bgmccollum has left #openstack-swift | 00:23 | |
| clayg | good idea | 00:24 |
| tdasilva | might be helpful to others to understand whatever solution we come up with :) | 00:25 |
| *** david-lyle_afk has quit IRC | 00:34 | |
| *** zhill has quit IRC | 00:34 | |
| mattoliverau | ho: morning | 00:35 |
| ho | mattoliverau: morning | 00:40 |
| clayg | tdasilva: that was helpful I think | 00:41 |
| tdasilva | clayg: +1 | 00:41 |
| tdasilva | clayg: mind if I clear authorship colors? | 00:42 |
| clayg | you can do that! | 00:44 |
| tdasilva | just find them annoying to read | 00:44 |
| clayg | tdasilva: in general, if there is ever any oppertunity to make something better and the only downside is that "it might make clayg mad" - DO IT - without giving it a second thought; I'll get over it | 00:45 |
| *** Nadeem has quit IRC | 00:45 | |
| tdasilva | haha | 00:45 |
| clayg | so much better - thanks | 00:45 |
| clayg | so i like that we have four states - and we're only using three of them | 00:46 |
| clayg | i also like that the only state that makes sense is if you have full access | 00:46 |
| clayg | I also like that either the write-only or read-only states both have some nice property and some bad property | 00:46 |
| tdasilva | i guess that would be the fifth state | 00:47 |
| clayg | i see no-access, read-only, write-only, read-write | 00:47 |
| clayg | no I think it's fair to ignore read access to the container because versioned writes just doesn't even come into play with those requests | 00:49 |
| tdasilva | oh true true | 00:49 |
| clayg | it only modifies behavior on put or delete | 00:49 |
| yuanz | torgomatic, clayg, so swift actuall does not fully support the if-match/if-none-match, only if-none-match * supported right now? | 00:52 |
| torgomatic | yuanz: on PUT, yes; on GET, it's got full support | 00:52 |
| *** zul has joined #openstack-swift | 00:54 | |
| yuanz | torgomatic, thx, got it. I didn't know this. In obj-server PUT looks like it's supported | 00:55 |
| torgomatic | yeah, it kind of does, but I think it's only * because of eventual-consistency issues making it really hard to do otherwise | 00:55 |
| clayg | torgomatic: I think eventual-consistency means it's best effort regardless | 00:58 |
| clayg | torgomatic: which is sorta weird because the use-case it was intended for and the only one it really works with is when you have multiple actors that can potentially generate the same content - but you want to save the bandwidth if someone beat you to it - but re-uploading (and over-writing any previous contents) will never be *wrong* it's just inefficent | 01:00 |
| yuanz | right, thx for the education. Question on the x-newest for EC: do we still need to query each FA in EC, or just return the timestamp of ts.durable? | 01:05 |
| *** david-lyle_afk has joined #openstack-swift | 01:21 | |
| dmorita | peluse: are you around? | 01:21 |
| *** lcurtis has joined #openstack-swift | 01:26 | |
| clayg | yuanz: x-newest currently goes 2 * replica count into handoffs, hitting every primary node on EC X-Newest is the least of our worries | 01:29 |
| *** devlaps has joined #openstack-swift | 01:31 | |
| *** cdelatte has quit IRC | 01:49 | |
| *** devlaps has quit IRC | 01:51 | |
| clayg | how did tdasilva make the colors go away on https://etherpad.openstack.org/p/object_version_and_ACL_use_cases | 01:55 |
| clayg | does anyone in here know how to etherpad? | 01:55 |
| kota_ | seems to press the bottun at the center of the page header. | 02:00 |
| kota_ | clayg: the icon consists of circle and colors | 02:01 |
| peluse | dmorita, I'm around for just a few minutes but will be available almost all of tomorrow - what's up? | 02:03 |
| dmorita | peluse: I updated a patch https://review.openstack.org/#/c/141252/ . I just would like you to remove -2 of this patch because I separated out codes relating to replicator. | 02:04 |
| clayg | kota_: you the man! | 02:05 |
| clayg | oh if peluse -2'd something I *definately* want to see it ;) | 02:05 |
| kota_ | clayg: :) | 02:05 |
| peluse | ha! | 02:06 |
| peluse | OK, will looks right now | 02:06 |
| dmorita | peluse: thanks :) | 02:06 |
| clayg | dmorita: peluse: so i was acctually going the other way with the auditor to be more like the updater/replicator | 02:09 |
| clayg | it turns out that if an object-replicator ever ran against an EC policies datadir we won't be happy - and the EC auditor will probably have to do something to reap old .data files that don't have a .durable | 02:10 |
| clayg | so really if you see an objects dir for a policy that's not listed in your swift.conf it's best to raise a warning and stay the fuck out before you make things worse | 02:11 |
| peluse | clayg, sure, we could have the auditor do some cleanup however we had sorta planned on reconstructor doing that (but still up in the air) | 02:11 |
| peluse | clayg, but agree on your fuck comment :) | 02:11 |
| peluse | dmorita, I'm still having some heartburn with the whole conf_check thing and I'm not sure that pepcis AC will fix it | 02:12 |
| clayg | peluse: oh maybe it does reclaiming - either way acoles and I are pretty keen on going with a EcPolicy diskfile (for all the same reasons the EcObjectController is a good idea) and if you have a per-policy diskfile no one gets to pretend they know what to do with some policy that's not in their swift.conf | 02:13 |
| clayg | object-server has to 400, consistency daemons need to skip that datadir - and everyone complains about not having swift.conf up-to-date | 02:13 |
| peluse | clayg, yeah I'm very interested in hearing more about the ECPolicy diskfile on Thu afternoon | 02:14 |
| clayg | encryption is another example - torgomatic keeps pointing out that if you let unencrypted objects sit in the same datadir as encrypted objects an attacker can replace your encrypted data with unencrypted data and you'd never know - so an encrypted policy specific diskfile might have some auditing behaviors that are different | 02:15 |
| peluse | dmorita, are you cool if we just keep your patch "on hold" until we sort out some of the EC things we're working through now? Tey do have material impact on your patch (may solve your issues without needing an additional change)? | 02:15 |
| peluse | clayg, good point - or torgomatic I guess | 02:15 |
| peluse | dmorita, and the timeframe I'm talking about is just another week or sop | 02:16 |
| peluse | sop=so | 02:16 |
| clayg | well it's all a pipe dream atm - we *need* to get the EcObjectController squared - and do something about internal routing requests (COPY and object versions) getting their controllers right (yuan has an idea on this) - THEN I *want* to see if we can make diskfile better with per policy diskfile routing | 02:16 |
| dmorita | peluse: No problem. I do not rush for it and we should fix EC related patches first :) | 02:17 |
| peluse | clayg, did you say something about a pipe? | 02:17 |
| peluse | dmorita, great. thanks. Feel free to bring it up again end of next week if we don't - would appreciate the reminder | 02:18 |
| peluse | OK, the chef needs to go make dinner for the family.... later folks | 02:19 |
| *** jrichli has joined #openstack-swift | 02:24 | |
| *** haomaiwang has joined #openstack-swift | 02:28 | |
| *** haomaiwang has quit IRC | 02:29 | |
| *** haomaiwang has joined #openstack-swift | 02:29 | |
| *** haomaiwang has quit IRC | 02:29 | |
| *** haomaiwa_ has joined #openstack-swift | 02:30 | |
| *** haomaiwa_ has quit IRC | 02:30 | |
| *** haomaiwang has joined #openstack-swift | 02:31 | |
| *** david-lyle_afk has quit IRC | 02:36 | |
| *** anticw has quit IRC | 03:09 | |
| *** anticw has joined #openstack-swift | 03:09 | |
| *** NM has joined #openstack-swift | 03:12 | |
| *** thumpba has joined #openstack-swift | 03:17 | |
| *** david-lyle_afk has joined #openstack-swift | 03:17 | |
| *** NM has quit IRC | 03:20 | |
| *** donagh has quit IRC | 03:26 | |
| *** thumpba_ has joined #openstack-swift | 03:30 | |
| *** thumpba has quit IRC | 03:34 | |
| *** donagh has joined #openstack-swift | 03:34 | |
| *** ppai has joined #openstack-swift | 03:47 | |
| *** lcurtis has quit IRC | 04:01 | |
| *** thumpba_ has quit IRC | 04:02 | |
| *** jrichli has quit IRC | 04:05 | |
| *** thumpba has joined #openstack-swift | 04:43 | |
| *** thumpba has quit IRC | 04:47 | |
| *** thumpba has joined #openstack-swift | 04:57 | |
| *** thumpba has quit IRC | 05:03 | |
| *** thumpba has joined #openstack-swift | 05:08 | |
| *** thumpba has quit IRC | 05:18 | |
| clayg | tdasilva: what is Marula Kruger? | 05:35 |
| *** panbalag has quit IRC | 05:46 | |
| *** bkopilov has quit IRC | 06:08 | |
| *** bkopilov has joined #openstack-swift | 06:09 | |
| *** panbalag has joined #openstack-swift | 06:20 | |
| *** silor has joined #openstack-swift | 06:27 | |
| *** silor has quit IRC | 06:37 | |
| *** bkopilov has quit IRC | 07:38 | |
| *** bkopilov has joined #openstack-swift | 07:45 | |
| *** bkopilov has quit IRC | 07:50 | |
| *** chlong has quit IRC | 08:01 | |
| *** bkopilov has joined #openstack-swift | 08:02 | |
| openstackgerrit | Yuan Zhou proposed openstack/swift: Update contianer sync to use internal client https://review.openstack.org/143791 | 08:04 |
| *** rledisez has joined #openstack-swift | 08:09 | |
| *** thumpba has joined #openstack-swift | 08:19 | |
| *** thumpba has quit IRC | 08:23 | |
| *** geaaru has joined #openstack-swift | 08:27 | |
| *** bkopilov has quit IRC | 08:49 | |
| *** bkopilov has joined #openstack-swift | 08:51 | |
| *** acoles_away is now known as acoles | 08:58 | |
| clayg | yawn - i gotta turn in | 09:01 |
| clayg | g'night | 09:01 |
| acoles | clayg: whoa, you still around? good night! | 09:02 |
| *** ho has quit IRC | 09:15 | |
| openstackgerrit | Yuan Zhou proposed openstack/swift: List valid EC types in error https://review.openstack.org/142311 | 09:21 |
| *** jistr has joined #openstack-swift | 09:25 | |
| *** jistr has quit IRC | 09:50 | |
| *** jistr has joined #openstack-swift | 09:52 | |
| *** nshaikh has joined #openstack-swift | 09:56 | |
| *** SkyRocknRoll has joined #openstack-swift | 10:11 | |
| *** chlong has joined #openstack-swift | 10:23 | |
| *** jistr has quit IRC | 10:38 | |
| *** knl has joined #openstack-swift | 10:38 | |
| *** jistr has joined #openstack-swift | 10:39 | |
| *** knl has quit IRC | 10:39 | |
| *** haomaiwang has quit IRC | 10:49 | |
| straycat | Does the proxy-server auto create accounts, or the account server? | 10:53 |
| straycat | My instance is creating accounts but failing to create containers and I'm trying to figure out why. | 10:54 |
| kota_ | straycat: that behavior is configurabe | 10:59 |
| kota_ | straycat: when you set account_autocreate to be true, proxy-sever will create the account automatically. | 11:00 |
| kota_ | straycat: but it is false in default. | 11:00 |
| straycat | Right, yes I have that set. | 11:00 |
| kota_ | straycat: what response code did you get as the failure? | 11:01 |
| kota_ | straycat: the response code would be helpful to figure out where the failure comes from. | 11:03 |
| straycat | kota_, swift upload gives a 404 | 11:04 |
| kota_ | straycat: ok, I'm thinking... | 11:05 |
| kota_ | straycat: Basically, when 404 appeared at PUT container, account deosn't exist in the Swift. | 11:07 |
| acoles | straycat: did you do 'swift post <containername>' before the upload | 11:09 |
| straycat | kota_, ok | 11:10 |
| straycat | acoles, no | 11:10 |
| acoles | straycat: swift upload will attempt to create the container but just curious if you see 404 from straightforward 'swift post <container>' | 11:10 |
| *** silor has joined #openstack-swift | 11:11 | |
| kota_ | straycat: did you do 'swift list'? | 11:12 |
| kota_ | straycat: i.e. get a list of containers. | 11:13 |
| acoles | kota_: hi! | 11:14 |
| kota_ | acoles: Hi! Thanks to help us :) | 11:14 |
| kota_ | acoles: How's going? Maybe, good morning! | 11:14 |
| acoles | kota_: i don't think i have helped yet :) | 11:14 |
| acoles | kota_: yes its morning! thanks | 11:15 |
| straycat | sorry, i'm going to have to try this later i've screwed the config somehow | 11:16 |
| straycat | running swift-init all start now gives me ValueError: invalid literal for int() with base 10: '' | 11:16 |
| straycat | thanks though :] | 11:16 |
| kota_ | starycat: ok, no worried | 11:17 |
| *** dencaval has joined #openstack-swift | 11:17 | |
| straycat | okay, it's back, some pid files got left behind in /var/run, so i can swift list, and swift put swift-container | 11:25 |
| straycat | should have mentioned there's a 400 as well as a 404 | 11:25 |
| straycat | s/put/post/ | 11:25 |
| *** o5k has joined #openstack-swift | 11:25 | |
| *** knl has joined #openstack-swift | 11:25 | |
| straycat | well, swift post returns with 0, but looking at the log the POST got a 400 | 11:27 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Make functional tests tolerate pre-existing versioned container https://review.openstack.org/163385 | 11:38 |
| acoles | straycat: swiftclient had a bug with return status not indicating failures. Its fixed upstream but don't think it is yet in last release | 11:44 |
| acoles | straycat: 'swift --debug post c1' will show you http response status | 11:45 |
| openstackgerrit | Daisuke Morita proposed openstack/swift: object-replicator warns when object dirs mismatch with policy settings https://review.openstack.org/163389 | 11:51 |
| straycat | acoles, that's cool | 11:52 |
| straycat | so i see that running curl -i http://127.0.0.1:8080/v1/AUTH_3eda71c0765f43478ba9e44256bfa701/c1 -X POST -H "Content-Lengt | 11:53 |
| straycat | h: 0" -H "X-Auth-Token: 40dea1c964bd4e2ead1f60a52b8705d1" | 11:53 |
| straycat | gives me a 400 bad request | 11:53 |
| straycat | i guess the more important part is RESP BODY: Invalid path: /sdb/719/AUTH_3eda71c0765f43478ba9e44256bfa701/c1 | 11:55 |
| *** dmorita has quit IRC | 11:56 | |
| kota_ | straycat: wow, curious. Looks like as from container-sever... | 11:58 |
| openstackgerrit | Prashanth Pai proposed openstack/swift: Make object creation more atomic in Linux https://review.openstack.org/162243 | 12:02 |
| *** erlon has joined #openstack-swift | 12:04 | |
| *** cdelatte has joined #openstack-swift | 12:14 | |
| *** delattec has joined #openstack-swift | 12:14 | |
| kota_ | straycat: Sorry I am running out of time to look at. It looks curious but tentaitively I have no idea. | 12:16 |
| straycat | kota_, I think it might be related to https://answers.launchpad.net/swift/+question/241916 i'm checking | 12:17 |
| kota_ | straycat: great. I hope it would be what you want. | 12:19 |
| straycat | thanks :) | 12:19 |
| acoles | straycat: could be. can you look in your logs, /var/log/swift/storage<n>.log and see if you can locate the 400 there, and for which server? | 12:19 |
| straycat | acoles, redploying currently, but i recall seeing a 400 from the account-server, there were no messages from container-server or object-server | 12:20 |
| *** o5k_ has joined #openstack-swift | 12:20 | |
| acoles | straycat: well that would be consistent with that bug, if a path with 4 parts e.g. /sdb/719/AUTH_3eda71c0765f43478ba9e44256bfa701/c1 is sent to account server it will 400. | 12:21 |
| *** o5k has quit IRC | 12:22 | |
| straycat | yep looks like i borked the ring configs, swift-ring-builder tells me they're all set to listen on 6002 | 12:29 |
| * straycat should have checked that ages ago | 12:29 | |
| acoles | straycat: :( well lets hope that fixes things for you | 12:29 |
| *** ppai has quit IRC | 12:31 | |
| *** mahatic has joined #openstack-swift | 12:36 | |
| *** kota_ has quit IRC | 12:38 | |
| *** SkyRocknRoll has quit IRC | 12:48 | |
| *** SkyRocknRoll has joined #openstack-swift | 12:51 | |
| *** SkyRocknRoll has joined #openstack-swift | 12:51 | |
| *** NM has joined #openstack-swift | 12:51 | |
| *** mariusv has joined #openstack-swift | 12:58 | |
| *** mariusv has quit IRC | 12:58 | |
| *** mariusv has joined #openstack-swift | 12:58 | |
| *** mariusv has quit IRC | 13:01 | |
| *** mariusv has joined #openstack-swift | 13:02 | |
| *** mariusv has quit IRC | 13:02 | |
| *** mariusv has joined #openstack-swift | 13:02 | |
| *** mariusv has quit IRC | 13:02 | |
| *** mariusv has joined #openstack-swift | 13:02 | |
| *** aix has joined #openstack-swift | 13:13 | |
| *** nellysmitt has joined #openstack-swift | 13:24 | |
| m_han | hi, i'm deploing SAIO and i get some errors in the the functional tests, can anybody help? | 13:29 |
| m_han | ping 0.0.0.0 :3 | 13:31 |
| *** jrichli has joined #openstack-swift | 13:36 | |
| *** nellysmitt has quit IRC | 13:39 | |
| *** nellysmi_ has joined #openstack-swift | 13:42 | |
| *** nellysmi_ has quit IRC | 13:44 | |
| *** annegentle has joined #openstack-swift | 13:49 | |
| straycat | Okay this works now | 13:50 |
| straycat | acoles, thanks for your help :) | 13:50 |
| acoles | straycat: welcome | 13:50 |
| m_han | acoles : i'm getting some errors in the the functional tests of SAIO, can you help me ? | 13:53 |
| acoles | m_han : paste your errors somewhere e.g. http://paste.openstack.org/ and put a link here (don't paste the errors here!) | 13:54 |
| *** zaitcev has joined #openstack-swift | 13:55 | |
| *** ChanServ sets mode: +v zaitcev | 13:55 | |
| *** tellesnobrega_ has joined #openstack-swift | 14:00 | |
| *** tsg_ has joined #openstack-swift | 14:05 | |
| *** tgohad has joined #openstack-swift | 14:05 | |
| *** tgohad has quit IRC | 14:05 | |
| *** SkyRocknRoll has quit IRC | 14:05 | |
| m_han | acoles : http://paste.openstack.org/show/191597/ tell me there is a need for the log files, thank's in advance man | 14:08 |
| *** mahatic has quit IRC | 14:09 | |
| *** mahatic has joined #openstack-swift | 14:09 | |
| mahatic | acoles, hi! if you are around and free, could you tell me how are the policies accessed here:https://review.openstack.org/#/c/138697/3/swift/cli/recon.py | 14:14 |
| zaitcev | m_han: looks like proxy-server.conf has some of TempAuth entries wrong | 14:14 |
| mahatic | acoles, I mean, as I see it, it is accessed by the url['policies'] but I'm not sure how that is valid. Would you be able to help? | 14:15 |
| acoles | m_han: like zaitcev says, check the tempauth account config in proxy-server.conf matches those in test.conf | 14:20 |
| acoles | m_han: the sample configs should work (with tempauth) | 14:20 |
| acoles | mahatic: looking | 14:20 |
| mahatic | acoles, great. thanks! | 14:20 |
| acoles | mahatic : looks to me like url is a key into all_scan dict. recon.scout requests recon info from host which i presume is a dict with a key 'policies'. Looks like object auditor also has a change to add a 'policies' key in recon_cache_dict. | 14:31 |
| *** lcurtis has joined #openstack-swift | 14:32 | |
| m_han | this is my /etc/swift/proxy-server.conf http://paste.openstack.org/show/191610/ | 14:33 |
| m_han | it shoud work like that or i have to change the entries in the [filter:tempauth] directive? | 14:34 |
| zaitcev | no, that part looks okay | 14:41 |
| zaitcev | so, maybe the test.conf is wrong or whatever it is that you use | 14:41 |
| mahatic | acoles, hosts are retrieved in main() by calling get_devices. But there is no [policies] key in Ring class which is called in that method. | 14:42 |
| zaitcev | I have SWIFT_TEST_CONFIG_FILE=/etc/swift/func_test.conf | 14:42 |
| *** o5k_ has quit IRC | 14:42 | |
| zaitcev | If that one checks out, I don't have any good ideas. The previous paste looked like the tenant "test" worked, but "test2" failed. | 14:43 |
| mahatic | acoles, sorry if that was confusing. In cli/recon.py the hosts info is collected in main() by calling get_devices method, which in turn calls Ring class | 14:43 |
| m_han | i've followed the official doc step by step docs.openstack.org/developer/swift/development_saio.html i don't see where is the prb :// | 14:47 |
| acoles | mahatic: i'm confused, is your query with line 616 in https://review.openstack.org/#/c/138697/3/swift/cli/recon.py | 14:52 |
| acoles | ? | 14:52 |
| mahatic | yeah that's right | 14:52 |
| mahatic | my question is, I do no really see a key as 'policies' anywhere. Am I missing something? | 14:53 |
| mahatic | do not* | 14:53 |
| *** jkugel has joined #openstack-swift | 14:54 | |
| acoles | m_han: i just ran functional tests ok with your proxy-server.conf. I did have to change the user to 'swift' in DEFAULT section. maybe check you have correct user permissions in your setup? If you post your test.conf i will verify that too, but the sample.conf should work. | 14:55 |
| tsg_ | yuanz, still around? | 14:58 |
| acoles | mahatic: all_scan[url] is set equal to the response content from a call to Scout.scout_host() which loads json, so the 'policies' key will be in that response body json (I presume - I am not hugely familiar with this) - so as long as the object server recon middleware returns json with a 'policies' key then the key will be in the dict stored under all_scans[url]. | 14:59 |
| m_han | acoles : this is my test.conf http://paste.openstack.org/show/191615/ | 14:59 |
| m_han | i'm cheking my permissions in the setup | 15:02 |
| mahatic | acoles, hmm okay. I am actually aware of the flow (that it calls scout_host() and that it loads json). But I just don't see anywhere policies as a key. As in a way to know which policy the URL belongs to. | 15:04 |
| *** tellesnobrega_ has quit IRC | 15:07 | |
| mahatic | acoles, line 115 in https://review.openstack.org/#/c/138697/3/swift/obj/auditor.py -> is creating I guess. | 15:07 |
| acoles | mahatic: obj/auditor.py line 122 in that patch adds key ['policies'] to recon_cache_dict - is that not the dict that gets returned in response to a recon request? | 15:07 |
| acoles | mahatic: right :) | 15:07 |
| mahatic | acoles, yeah :) | 15:09 |
| *** bkopilov has quit IRC | 15:09 | |
| acoles | mahatic: so if that dict has a key ['policies'], then cli recon effectively all_scan[url] = json.loads(dict with a 'policies' key in it) so all_scan[url]['policies'] is cool? | 15:11 |
| mahatic | acoles, yup! :) I was all the time looking at the middleware/recon.py So was confused in the way things are called. But that is not relevant here. | 15:13 |
| mahatic | acoles, thanks! :) | 15:13 |
| acoles | m_han: your test.conf works for me too. like i said check you have followed setup substituting with user 'med' | 15:14 |
| acoles | mahatic: ok! | 15:15 |
| *** bkopilov has joined #openstack-swift | 15:23 | |
| *** david-lyle_afk is now known as david-lyle | 15:28 | |
| *** Nadeem has joined #openstack-swift | 15:41 | |
| *** nellysmitt has joined #openstack-swift | 15:45 | |
| *** Nadeem_ has joined #openstack-swift | 15:49 | |
| *** nellysmitt has quit IRC | 15:50 | |
| *** Nadeem_ has quit IRC | 15:50 | |
| *** Nadeem_ has joined #openstack-swift | 15:50 | |
| *** Nadeem has quit IRC | 15:53 | |
| *** tellesnobrega has quit IRC | 15:58 | |
| *** tellesnobrega has joined #openstack-swift | 15:59 | |
| *** doxavore has joined #openstack-swift | 16:03 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 16:09 |
| *** jrichli has quit IRC | 16:10 | |
| *** jrichli has joined #openstack-swift | 16:11 | |
| *** donagh has quit IRC | 16:13 | |
| *** briancurtin has quit IRC | 16:13 | |
| *** matt__ has joined #openstack-swift | 16:14 | |
| *** zacksh has quit IRC | 16:14 | |
| *** jroll has quit IRC | 16:14 | |
| *** classicsnail has quit IRC | 16:14 | |
| *** CrackerJackMack has quit IRC | 16:14 | |
| *** mattoliverau has quit IRC | 16:14 | |
| *** zigo_ has quit IRC | 16:14 | |
| *** goodes has quit IRC | 16:14 | |
| *** zigo_ has joined #openstack-swift | 16:14 | |
| *** zacksh has joined #openstack-swift | 16:15 | |
| *** jroll has joined #openstack-swift | 16:15 | |
| *** CrackerJackMack has joined #openstack-swift | 16:16 | |
| *** classicsnail has joined #openstack-swift | 16:16 | |
| *** jroll has quit IRC | 16:16 | |
| *** jroll has joined #openstack-swift | 16:16 | |
| *** annegentle has quit IRC | 16:18 | |
| *** goodes has joined #openstack-swift | 16:19 | |
| *** briancurtin has joined #openstack-swift | 16:19 | |
| *** mahatic has quit IRC | 16:19 | |
| *** donagh has joined #openstack-swift | 16:23 | |
| *** nshaikh has left #openstack-swift | 16:26 | |
| *** SkyRocknRoll has joined #openstack-swift | 16:28 | |
| *** mahatic has joined #openstack-swift | 16:31 | |
| *** SkyRocknRoll has quit IRC | 16:37 | |
| *** SkyRocknRoll has joined #openstack-swift | 16:37 | |
| *** thumpba has joined #openstack-swift | 16:41 | |
| *** thumpba has quit IRC | 16:46 | |
| *** thumpba has joined #openstack-swift | 16:48 | |
| *** openstack has joined #openstack-swift | 16:55 | |
| *** thumpba has quit IRC | 16:57 | |
| *** gvernik has joined #openstack-swift | 17:05 | |
| *** rledisez has quit IRC | 17:15 | |
| *** jordanP has joined #openstack-swift | 17:16 | |
| *** SkyRocknRoll has quit IRC | 17:17 | |
| *** chlong has quit IRC | 17:21 | |
| openstackgerrit | Tushar Gohad proposed openstack/swift: Bring EC API calls in sync with PyECLib 1.0 https://review.openstack.org/163533 | 17:22 |
| tsg_ | torgomatic, clayg, peluse: not sure yet if Jenkins has picked up the pyeclib update, but we'll need this ^^^ change once it does | 17:24 |
| peluse | OK | 17:31 |
| *** tsg_ has quit IRC | 17:31 | |
| *** zhill has joined #openstack-swift | 17:34 | |
| *** bkopilov has quit IRC | 17:35 | |
| clayg | tsg_ does https://review.openstack.org/#/c/163533/1 also work with 0.98? | 17:36 |
| *** bkopilov has joined #openstack-swift | 17:39 | |
| *** annegentle has joined #openstack-swift | 17:41 | |
| *** devlaps has joined #openstack-swift | 17:42 | |
| *** nellysmitt has joined #openstack-swift | 17:46 | |
| *** knl has quit IRC | 17:47 | |
| *** nellysmitt has quit IRC | 17:50 | |
| *** Trixboxer has quit IRC | 17:53 | |
| *** bkopilov has quit IRC | 17:55 | |
| *** bkopilov has joined #openstack-swift | 17:57 | |
| clayg | is this thing on? | 17:57 |
| klrmn | *tap* *tap*tap* | 17:58 |
| clayg | hehe | 17:59 |
| acoles | peluse: if you haven't already applied it i am about to pusj another version of write_durable patch 162717 | 18:00 |
| patchbot | acoles: https://review.openstack.org/#/c/162717/ | 18:00 |
| acoles | thanks patchbot | 18:00 |
| acoles | peluse: assuming it passes local test ;) | 18:01 |
| *** reed has joined #openstack-swift | 18:03 | |
| *** reed has quit IRC | 18:03 | |
| *** reed has joined #openstack-swift | 18:03 | |
| torgomatic | No meeting this week? | 18:04 |
| *** remix_tj has quit IRC | 18:04 | |
| acoles | torgomatic: nope notmyname cancelled | 18:07 |
| acoles | torgomatic: you'd be 53 mins early if there were :) | 18:07 |
| torgomatic | To heck with everything about time changes. :p | 18:07 |
| peluse | acoles, heh, yeah I just applied it. funny, this is the one that I was trying to add as a dependency the other day that was giving me fits. | 18:08 |
| peluse | acoles, and I was doing it for the reasons you state in the comments before you stated them :) | 18:08 |
| peluse | (git was giving me fits, not the code) | 18:08 |
| acoles | peluse: lol | 18:08 |
| peluse | acoles, about to test and see if .durable is now correctly restored on a reconstruction.... | 18:09 |
| acoles | peluse: i just added an fsync in there, but noticed something else i'm not sure is right...hash_cleanup_listdir is called after data file is written but before durable, so nothing will be cleaned out :/ think we need to delay calling hash_cleanup_listdir til durable is in place? | 18:10 |
| *** remix_tj has joined #openstack-swift | 18:10 | |
| peluse | acoles, well, its called like all over the place though so isn't it the case that things will be cleaned up just not til next time someone calls it (because of a put or a reconstructor run or whatever)? | 18:11 |
| acoles | peluse: i guess so but it seems like a waste of a listdir if we know it won't actually clear out the older data files | 18:12 |
| peluse | acoles, yeah, true. BTW, the test just now worked as expected | 18:13 |
| acoles | peluse: cool (re test). re HCL, I'll mull some more. | 18:13 |
| peluse | acoles, cool. I've got a race condition I'm dealing with that I need to get back to that is killing me. Will likely leave the .durable code as is for now until I figure out WTF is happening there | 18:16 |
| acoles | peluse: bummer | 18:16 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Diskfile decides if durable is written based on policy https://review.openstack.org/162717 | 18:17 |
| *** gyee has joined #openstack-swift | 18:17 | |
| acoles | peluse: ^^ latest. so torgomatic had some concern about this, which i am not disregarding, but wanted to add the fsync and error handling (which i hope i got right - clayg??) | 18:18 |
| *** mahatic has quit IRC | 18:19 | |
| *** aix has quit IRC | 18:21 | |
| *** mmcardle has quit IRC | 18:21 | |
| *** remix_tj has quit IRC | 18:22 | |
| peluse | acoles, OK thanks | 18:25 |
| *** remix_tj has joined #openstack-swift | 18:26 | |
| acoles | clayg: around? | 18:26 |
| *** silor has quit IRC | 18:26 | |
| *** nellysmitt has joined #openstack-swift | 18:27 | |
| *** remix_tj has quit IRC | 18:36 | |
| *** remix_tj has joined #openstack-swift | 18:37 | |
| *** tsg_ has joined #openstack-swift | 18:37 | |
| *** jistr has quit IRC | 18:46 | |
| *** geaaru has quit IRC | 18:50 | |
| *** ho has joined #openstack-swift | 18:52 | |
| *** nellysmitt has quit IRC | 18:59 | |
| *** nellysmitt has joined #openstack-swift | 18:59 | |
| acoles | clayg: nm, guess i missed you, maybe tomorrow. | 19:00 |
| *** acoles is now known as acoles_away | 19:01 | |
| *** doxavore has quit IRC | 19:03 | |
| *** jordanP has quit IRC | 19:08 | |
| gvernik | is there Swift meeting today? | 19:15 |
| ho | gvernik: I checked the IRC log and today no meeting. | 19:17 |
| ho | gvernik: 2015-03-10T15:31:53 <notmyname> reminder that there is no meeting tomorrow | 19:17 |
| gvernik | thanks | 19:17 |
| *** gvernik has quit IRC | 19:22 | |
| *** dencaval has quit IRC | 19:31 | |
| *** thumpba_ has quit IRC | 19:31 | |
| tdasilva | clayg: around? | 19:34 |
| torgomatic | blah, ranged GETs are annoying | 19:38 |
| torgomatic | every time I write a new test case I find a new bug :/ | 19:38 |
| *** Fin1te has joined #openstack-swift | 19:39 | |
| *** nellysmitt has quit IRC | 19:46 | |
| *** annegentle has quit IRC | 19:46 | |
| *** annegentle has joined #openstack-swift | 19:47 | |
| *** annegentle has quit IRC | 19:52 | |
| peluse | torgomatic, stop writing new tests then :) | 20:03 |
| *** annegentle has joined #openstack-swift | 20:08 | |
| *** thumpba has joined #openstack-swift | 20:13 | |
| openstackgerrit | Tushar Gohad proposed openstack/swift: Bring EC API calls in sync with PyECLib 1.0 https://review.openstack.org/163533 | 20:24 |
| *** silor has joined #openstack-swift | 20:25 | |
| *** silor has quit IRC | 20:30 | |
| *** delattec has quit IRC | 20:31 | |
| *** cdelatte has quit IRC | 20:31 | |
| clayg | peluse: lol! | 20:31 |
| clayg | acoles_away: tdasilva: nope, but am now | 20:31 |
| openstackgerrit | paul luse proposed openstack/swift: Erasure Code Reconstructor https://review.openstack.org/131872 | 20:35 |
| *** jistr has joined #openstack-swift | 20:35 | |
| peluse | ^ more rebase and bug fix. should be some good stuff to talk about in addition to the multi FI patch over the next few days | 20:36 |
| peluse | .. and another half dozen good patches too! | 20:36 |
| *** cdelatte has joined #openstack-swift | 20:42 | |
| notmyname | hello | 20:57 |
| peluse | hola | 20:58 |
| peluse | back in CA yet? | 20:58 |
| notmyname | nope. waiting in the Philly airport. | 20:59 |
| notmyname | will be a 6.5 hour flight home :-( | 20:59 |
| peluse | oh boy | 20:59 |
| notmyname | peluse: since we didn't have the team meeting today, can you send me a quick email about EC status this week? | 21:00 |
| peluse | sure, or we can talk tomorrow afternoon in SanFran as there are a few things up in the air that we need to talk about wrt what/when/how | 21:01 |
| peluse | your call | 21:01 |
| notmyname | peluse: I can glean from trello and from what clayg and torgomatic have told me. just looking to send out my weekly summary in the morning | 21:05 |
| peluse | ahh, OK. I'll send a note to you and the two of them in case they want to add some detail/correct something | 21:07 |
| *** tellesnobrega_ has joined #openstack-swift | 21:09 | |
| notmyname | peluse: I just forwarded you the email I was discussing it with them in. just add to it | 21:10 |
| notmyname | well, whenever this really slow network gets around to sending the email | 21:10 |
| *** jistr has quit IRC | 21:10 | |
| *** tellesnobrega_ has quit IRC | 21:11 | |
| matt__ | Morning | 21:24 |
| ho | matto_: morning. your irc name is differenct from usual :) | 21:25 |
| *** matt__ is now known as mattoliverau | 21:25 | |
| *** ChanServ sets mode: +v mattoliverau | 21:25 | |
| mattoliverau | yeah, rackspace must have rebooted my quassel server again during the night | 21:26 |
| ho | mattoliverau: now you recovered. | 21:28 |
| mattoliverau | ho: yup :) | 21:28 |
| mattoliverau | notmyname: how's the ops summit going? | 21:28 |
| openstackgerrit | Samuel Merritt proposed openstack/swift: EC: support single ranges for GET requests https://review.openstack.org/163620 | 21:29 |
| *** Fin1te has quit IRC | 21:34 | |
| torgomatic | now to figure out how to demimulate object-server responses | 21:34 |
| *** movielovers has joined #openstack-swift | 21:41 | |
| *** NM has quit IRC | 21:53 | |
| clayg | peluse: I think i'm going to look at squashing policy-everywhere down with the multi fi patch - I think the fi patch does so much plubming of policy_index, only for policy-everywhere to go and replumb as policy it may well be just as easy to review in one change - anyway I'll look at the diff and put it up directly on feature/ec | 21:55 |
| *** jkugel has quit IRC | 22:02 | |
| peluse | clayg, I just got done going over your policy-everywhere patch in more detail and I agree | 22:03 |
| peluse | clayg, good call | 22:03 |
| peluse | mattoliverau, I believe notmyname is wheels-up right about now (or close to it anyways) | 22:04 |
| clayg | hrmm... it is a little bigger | 22:04 |
| peluse | yeah, that's what she said | 22:04 |
| *** jrichli has quit IRC | 22:05 | |
| *** annegentle has quit IRC | 22:05 | |
| *** ho__ has joined #openstack-swift | 22:12 | |
| *** Nadeem_ has quit IRC | 22:20 | |
| *** annegentle has joined #openstack-swift | 22:25 | |
| torgomatic | did something change in pyeclib-land recently? all my stuff is failing with "ImportError: No module named pyeclib.ec_iface" | 22:27 |
| torgomatic | peluse: tsg_: ^^ | 22:27 |
| tsg_ | torgomatic: something jenkins is not happy about .. I have been looking at it | 22:28 |
| tsg_ | about to push hopefully the final update | 22:28 |
| torgomatic | tsg_: that's good; I'm glad someone with a clue is looking at it, because I'm staring at it like a dog watching a magic trick | 22:28 |
| *** srsakhamuri has joined #openstack-swift | 22:29 | |
| tsg_ | torgomatic: yeah the slave logs are kind of pointless | 22:29 |
| tsg_ | but I do think I know what's going on | 22:29 |
| tsg_ | this library integration into debian/ubuntu/jenkins has been a PITA | 22:30 |
| tsg_ | torgomatic: I will do rechecks on the stuff that failed | 22:30 |
| torgomatic | tsg_: sounds good to me | 22:32 |
| *** annegentle has quit IRC | 22:34 | |
| *** tellesnobrega_ has joined #openstack-swift | 22:40 | |
| *** tellesnobrega_ has quit IRC | 22:41 | |
| *** lcurtis has quit IRC | 22:42 | |
| openstackgerrit | Samuel Merritt proposed openstack/swift: EC: support single ranges for GET requests https://review.openstack.org/163620 | 22:48 |
| *** panbalag has quit IRC | 22:53 | |
| *** echevemaster has joined #openstack-swift | 22:56 | |
| *** movielovers has left #openstack-swift | 22:57 | |
| notmyname | back online (in the air) | 22:57 |
| tsg_ | torgomatic: things looking better now https://jenkins02.openstack.org/job/gate-swift-python27/4311/console | 23:08 |
| tsg_ | this is for https://review.openstack.org/#/c/163533/ | 23:08 |
| torgomatic | tsg_: ok, cool... so assuming that lands and then I rebase on it, I should be good, yes? | 23:08 |
| tsg_ | yes | 23:09 |
| torgomatic | sweet! | 23:09 |
| tsg_ | torgomatic: waiting for your review :) | 23:09 |
| tsg_ | needless to say, in addition to the rebase, you'll need an update to pyeclib 1.0.3 off of pypi | 23:10 |
| torgomatic | tsg_: I'll +2/+A it as soon as Jenkins is done with it | 23:10 |
| tsg_ | thank you! | 23:10 |
| torgomatic | that way I can get all these other patches back in action :) | 23:11 |
| tsg_ | yep .. I am going to trigger rechecks as soon as 163533 is green | 23:11 |
| *** thumpba_ has joined #openstack-swift | 23:13 | |
| *** thumpba has quit IRC | 23:15 | |
| *** annegentle has joined #openstack-swift | 23:32 | |
| *** chlong has joined #openstack-swift | 23:41 | |
| *** annegentle has quit IRC | 23:43 | |
| *** thumpba_ has quit IRC | 23:45 | |
| *** km has joined #openstack-swift | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!