| *** jamielennox is now known as jamielennox|away | 00:05 | |
| *** annegentle has joined #openstack-swift | 00:05 | |
| *** jamielennox|away is now known as jamielennox | 00:07 | |
| *** jamielennox is now known as jamielennox|away | 00:10 | |
| *** zhill has quit IRC | 00:10 | |
| *** kevinc_ has quit IRC | 00:12 | |
| *** sanchitmalhotra has joined #openstack-swift | 00:18 | |
| *** sanchitmalhotra1 has quit IRC | 00:21 | |
| *** haigang has joined #openstack-swift | 00:26 | |
| *** charz has quit IRC | 00:28 | |
| *** charz has joined #openstack-swift | 00:32 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 00:42 | |
| *** zhill has joined #openstack-swift | 00:43 | |
| *** sanchitmalhotra has quit IRC | 00:45 | |
| *** nakagawamsa has quit IRC | 00:45 | |
| *** zhill has quit IRC | 00:47 | |
| *** haigang has quit IRC | 00:50 | |
| *** lcurtis has joined #openstack-swift | 01:17 | |
| *** zhill has joined #openstack-swift | 01:17 | |
| lcurtis | @all...is it safe to change weights again before ring is done balancing? | 01:24 |
|---|---|---|
| *** sanchitmalhotra has joined #openstack-swift | 01:25 | |
| *** zhill has quit IRC | 01:25 | |
| lcurtis | i know there are min_part hours etc, but is there logic to prevent something reallly bad from happening if you change weights mid balance AS LONG AS MIN-PART hours has passed? (ie, things are pretty close to finishing) | 01:26 |
| *** sanchitmalhotra1 has quit IRC | 01:26 | |
| *** lcurtis has quit IRC | 01:32 | |
| *** lcurtis has joined #openstack-swift | 01:33 | |
| openstackgerrit | Merged openstack/python-swiftclient: Update mock to get away from env markers https://review.openstack.org/205150 | 01:36 |
| *** haigang has joined #openstack-swift | 01:39 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 01:52 | |
| *** sanchitmalhotra has quit IRC | 01:54 | |
| *** haigang has quit IRC | 01:56 | |
| *** haigang has joined #openstack-swift | 01:59 | |
| *** haomaiwang has joined #openstack-swift | 02:08 | |
| *** jamielennox|away is now known as jamielennox | 02:08 | |
| *** sanchitmalhotra has joined #openstack-swift | 02:11 | |
| *** sanchitmalhotra1 has quit IRC | 02:13 | |
| *** haomaiwang has quit IRC | 02:15 | |
| *** haomaiwang has joined #openstack-swift | 02:16 | |
| *** zhill has joined #openstack-swift | 02:19 | |
| *** zhill has quit IRC | 02:23 | |
| *** tsg has quit IRC | 02:28 | |
| *** gyee_500 has quit IRC | 02:29 | |
| *** annegentle has quit IRC | 02:38 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 02:45 | |
| *** annegentle has joined #openstack-swift | 02:45 | |
| *** annegentle has quit IRC | 02:46 | |
| *** sanchitmalhotra has quit IRC | 02:47 | |
| *** haomaiwang has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-swift | 03:01 | |
| *** sanchitmalhotra has joined #openstack-swift | 03:09 | |
| *** sanchitmalhotra1 has quit IRC | 03:11 | |
| *** tsg has joined #openstack-swift | 03:24 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 03:33 | |
| *** sanchitmalhotra has quit IRC | 03:35 | |
| *** annegentle has joined #openstack-swift | 03:38 | |
| *** SkyRocknRoll has joined #openstack-swift | 03:51 | |
| *** sanchitmalhotra has joined #openstack-swift | 03:56 | |
| *** annegentle has quit IRC | 03:58 | |
| *** sanchitmalhotra1 has quit IRC | 03:58 | |
| *** haomaiwang has quit IRC | 04:01 | |
| *** haomaiwang has joined #openstack-swift | 04:02 | |
| *** jamielennox is now known as jamielennox|away | 04:02 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 04:08 | |
| *** sanchitmalhotra has quit IRC | 04:10 | |
| *** sanchitmalhotra has joined #openstack-swift | 04:10 | |
| *** sanchitmalhotra1 has quit IRC | 04:12 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 04:25 | |
| *** sanchitmalhotra has quit IRC | 04:27 | |
| *** sanchitmalhotra has joined #openstack-swift | 04:32 | |
| *** sanchitmalhotra1 has quit IRC | 04:34 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 04:34 | |
| *** sanchitmalhotra has quit IRC | 04:36 | |
| *** sanchitmalhotra has joined #openstack-swift | 04:41 | |
| *** early has quit IRC | 04:42 | |
| *** morgan_404 is now known as morgan_204 | 04:42 | |
| *** morgan_204 is now known as morgan_206 | 04:44 | |
| *** sanchitmalhotra1 has quit IRC | 04:44 | |
| *** morgan_206 is now known as morgan_503 | 04:45 | |
| *** jamielennox|away is now known as jamielennox | 04:46 | |
| *** SkyRocknRoll has quit IRC | 04:50 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 04:58 | |
| *** early has joined #openstack-swift | 05:00 | |
| *** sanchitmalhotra has quit IRC | 05:01 | |
| *** haomaiwang has quit IRC | 05:01 | |
| *** haomaiwa_ has joined #openstack-swift | 05:04 | |
| *** hrou has joined #openstack-swift | 05:05 | |
| *** lcurtis has quit IRC | 05:09 | |
| *** sanchitmalhotra has joined #openstack-swift | 05:09 | |
| *** tsg has quit IRC | 05:10 | |
| *** sanchitmalhotra1 has quit IRC | 05:11 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:32 | |
| *** chlong has quit IRC | 05:41 | |
| *** chlong has joined #openstack-swift | 05:44 | |
| *** haigang has quit IRC | 05:45 | |
| *** haigang has joined #openstack-swift | 05:46 | |
| *** haigang has quit IRC | 05:47 | |
| *** haigang has joined #openstack-swift | 05:47 | |
| *** proteusguy has quit IRC | 05:53 | |
| *** proteusguy has joined #openstack-swift | 05:54 | |
| *** fifieldt_ has quit IRC | 06:00 | |
| *** haomaiwa_ has quit IRC | 06:01 | |
| *** haomaiwang has joined #openstack-swift | 06:02 | |
| *** akle has quit IRC | 06:11 | |
| *** hrou has quit IRC | 06:24 | |
| *** akle has joined #openstack-swift | 06:26 | |
| *** hrou has joined #openstack-swift | 06:28 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 06:47 | |
| *** sanchitmalhotra has quit IRC | 06:49 | |
| *** hrou has quit IRC | 06:51 | |
| *** haomaiwang has quit IRC | 07:01 | |
| *** sanchitmalhotra has joined #openstack-swift | 07:01 | |
| *** haomaiwang has joined #openstack-swift | 07:02 | |
| *** sanchitmalhotra1 has quit IRC | 07:03 | |
| *** rledisez has joined #openstack-swift | 07:15 | |
| *** jamielennox is now known as jamielennox|away | 07:17 | |
| *** haigang has quit IRC | 07:21 | |
| *** chlong has quit IRC | 07:36 | |
| *** akle has quit IRC | 07:36 | |
| *** chlong has joined #openstack-swift | 07:39 | |
| *** sanchitmalhotra1 has joined #openstack-swift | 07:59 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** sanchitmalhotra has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-swift | 08:01 | |
| *** jistr has joined #openstack-swift | 08:13 | |
| *** sanchitmalhotra has joined #openstack-swift | 08:16 | |
| *** sanchitmalhotra1 has quit IRC | 08:19 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements https://review.openstack.org/89250 | 08:21 |
| *** lyrrad has quit IRC | 08:32 | |
| *** SkyRocknRoll has quit IRC | 08:32 | |
| *** SkyRocknRoll has joined #openstack-swift | 08:48 | |
| *** sakaYK has joined #openstack-swift | 08:53 | |
| *** haomaiwang has quit IRC | 09:01 | |
| *** haomaiwang has joined #openstack-swift | 09:01 | |
| *** janonymous has quit IRC | 09:53 | |
| *** haigang has joined #openstack-swift | 09:56 | |
| *** SkyRocknRoll has quit IRC | 09:59 | |
| *** sakaYK has quit IRC | 09:59 | |
| *** haomaiwang has quit IRC | 10:01 | |
| *** haomaiwang has joined #openstack-swift | 10:02 | |
| *** haigang has quit IRC | 10:03 | |
| *** baojg has joined #openstack-swift | 10:07 | |
| *** sakaYK has joined #openstack-swift | 10:18 | |
| *** sakaYK has quit IRC | 10:23 | |
| *** sakaYK has joined #openstack-swift | 10:24 | |
| *** sakaYK has quit IRC | 10:26 | |
| *** baojg has quit IRC | 10:27 | |
| *** baojg has joined #openstack-swift | 10:27 | |
| *** sakaYK has joined #openstack-swift | 10:29 | |
| *** ho has quit IRC | 10:29 | |
| *** ho has joined #openstack-swift | 10:30 | |
| *** baojg has quit IRC | 10:32 | |
| *** sanchitmalhotra has quit IRC | 10:33 | |
| *** sakaYK has quit IRC | 10:35 | |
| *** sakaYK has joined #openstack-swift | 10:36 | |
| *** sakaYK has quit IRC | 10:38 | |
| *** sanchitmalhotra has joined #openstack-swift | 10:39 | |
| *** sanchitmalhotra has quit IRC | 10:40 | |
| *** sakaYK has joined #openstack-swift | 10:43 | |
| *** sakaYK has quit IRC | 10:45 | |
| *** ho_ has joined #openstack-swift | 10:51 | |
| *** ho has quit IRC | 10:51 | |
| *** haomaiwang has quit IRC | 10:53 | |
| *** haigang has joined #openstack-swift | 10:55 | |
| *** ho_ has quit IRC | 10:58 | |
| *** ho has joined #openstack-swift | 10:58 | |
| *** robefran has joined #openstack-swift | 11:25 | |
| *** Kennan2 has joined #openstack-swift | 11:36 | |
| *** Kennan has quit IRC | 11:37 | |
| *** bkopilov has quit IRC | 11:39 | |
| *** ho has quit IRC | 11:39 | |
| *** ho has joined #openstack-swift | 11:40 | |
| *** delattec has joined #openstack-swift | 11:51 | |
| *** cdelatte has quit IRC | 11:53 | |
| *** bkopilov has joined #openstack-swift | 11:56 | |
| *** delattec has quit IRC | 12:02 | |
| *** delattec has joined #openstack-swift | 12:02 | |
| *** haigang has quit IRC | 12:05 | |
| *** lcurtis has joined #openstack-swift | 12:23 | |
| *** petertr7_away is now known as petertr7 | 12:29 | |
| *** marcusvrn has joined #openstack-swift | 12:43 | |
| *** bkopilov has quit IRC | 12:44 | |
| *** jkugel has joined #openstack-swift | 12:46 | |
| *** nexusz99 has quit IRC | 12:46 | |
| *** lcurtis has quit IRC | 12:55 | |
| *** lcurtis has joined #openstack-swift | 13:01 | |
| *** bill_az has joined #openstack-swift | 13:05 | |
| *** dustins has joined #openstack-swift | 13:12 | |
| *** rodrigods has quit IRC | 13:13 | |
| *** lcurtis has quit IRC | 13:14 | |
| mattoliverau | Morning | 13:26 |
| *** fifieldt has joined #openstack-swift | 13:30 | |
| *** dustins has quit IRC | 13:34 | |
| *** dustins has joined #openstack-swift | 13:34 | |
| openstackgerrit | paul luse proposed openstack/swift: EC GET path: support use of alternate fragments on object servers https://review.openstack.org/213147 | 13:38 |
| peluse_ | yo | 13:38 |
| peluse_ | how was Thursday? | 13:38 |
| *** bill_az has quit IRC | 13:41 | |
| *** pbourke has joined #openstack-swift | 13:44 | |
| pbourke | hi, when using the authtoken config items listed at http://docs.openstack.org/developer/swift/overview_auth.html#configuring-swift-to-use-keystone, I get "Use of the auth_admin_prefix, [...] and admin_tenant_name configuration options is deprecated in favor of auth_plugin and related options and may be removed in a future release." | 13:45 |
| pbourke | is there an example somewhere of the latest config options I should be using? | 13:45 |
| *** jrichli has joined #openstack-swift | 13:45 | |
| *** breitz has quit IRC | 13:53 | |
| *** breitz has joined #openstack-swift | 13:53 | |
| *** joeljwright has joined #openstack-swift | 13:54 | |
| *** ChanServ sets mode: +v joeljwright | 13:54 | |
| *** petertr7 is now known as petertr7_away | 13:58 | |
| *** jrichli has quit IRC | 14:00 | |
| *** petertr7_away is now known as petertr7 | 14:00 | |
| *** mahatic has joined #openstack-swift | 14:08 | |
| *** hrou has joined #openstack-swift | 14:13 | |
| *** jrichli has joined #openstack-swift | 14:13 | |
| *** chlong has quit IRC | 14:19 | |
| *** haomaiwang has joined #openstack-swift | 14:25 | |
| *** jlhinson has joined #openstack-swift | 14:29 | |
| *** SkyRocknRoll has joined #openstack-swift | 14:30 | |
| *** haomaiwang has quit IRC | 14:35 | |
| *** annegentle has joined #openstack-swift | 14:37 | |
| *** wbhuber has joined #openstack-swift | 14:41 | |
| *** alejandrito has joined #openstack-swift | 14:44 | |
| *** kevinc_ has joined #openstack-swift | 14:54 | |
| *** petertr7 is now known as petertr7_away | 15:06 | |
| *** annegentle has quit IRC | 15:11 | |
| *** annegentle has joined #openstack-swift | 15:12 | |
| *** baojg has joined #openstack-swift | 15:15 | |
| mattoliverau | peluse: hey man! Thursday was great, shame you had to leave. | 15:16 |
| *** minwoob has joined #openstack-swift | 15:21 | |
| *** jistr is now known as jistr|mtg | 15:29 | |
| *** openstackgerrit has quit IRC | 15:31 | |
| *** openstackgerrit has joined #openstack-swift | 15:31 | |
| *** lcurtis has joined #openstack-swift | 15:33 | |
| *** nadeem has joined #openstack-swift | 15:52 | |
| *** nadeem has quit IRC | 15:57 | |
| *** nadeem has joined #openstack-swift | 15:58 | |
| *** wbhuber_ has joined #openstack-swift | 16:02 | |
| *** jistr|mtg is now known as jistr | 16:03 | |
| *** nadeem has quit IRC | 16:04 | |
| *** nadeem has joined #openstack-swift | 16:05 | |
| *** wbhuber has quit IRC | 16:05 | |
| *** pbourke has quit IRC | 16:09 | |
| *** pbourke has joined #openstack-swift | 16:10 | |
| peluse_ | mattoliverau, yeah, but sometimes life gets in the way of work :) | 16:17 |
| *** bill_az has joined #openstack-swift | 16:22 | |
| *** rledisez has quit IRC | 16:23 | |
| *** baojg has quit IRC | 16:23 | |
| *** baojg has joined #openstack-swift | 16:24 | |
| openstackgerrit | paul luse proposed openstack/swift: EC GET path: support use of alternate fragments on object servers https://review.openstack.org/213147 | 16:28 |
| *** bkopilov has joined #openstack-swift | 16:28 | |
| *** baojg has quit IRC | 16:28 | |
| *** baojg has joined #openstack-swift | 16:29 | |
| *** wbhuber_ has quit IRC | 16:30 | |
| *** jistr has quit IRC | 16:32 | |
| *** ho has quit IRC | 16:33 | |
| *** ho has joined #openstack-swift | 16:33 | |
| *** baojg has quit IRC | 16:35 | |
| *** petertr7_away is now known as petertr7 | 16:36 | |
| *** gyee has joined #openstack-swift | 16:37 | |
| MooingLemur | peluse_: I think the problems I saw with EC are definitively gone with all the fixes that have been committed in the meantime. I can't get it to break anymore :) | 16:40 |
| *** mahatic_ has joined #openstack-swift | 16:42 | |
| *** mahatic has quit IRC | 16:42 | |
| *** joeljwright has left #openstack-swift | 16:43 | |
| *** dustins has quit IRC | 16:47 | |
| *** ho has quit IRC | 16:55 | |
| *** ho has joined #openstack-swift | 16:55 | |
| *** bill_az has quit IRC | 16:55 | |
| *** dustins has joined #openstack-swift | 16:55 | |
| peluse_ | MooingLemur, that's really good news! We had a great hackathon this past week and half 3-4 patches coming that we think will really help solidy things as well | 16:56 |
| *** jkugel has quit IRC | 16:57 | |
| peluse_ | s/half/have | 16:57 |
| *** wbhuber has joined #openstack-swift | 16:58 | |
| robefran | does anyone know how admin can set account-sysmeta from the command line (curl)? I'm investigating ratelimit in an SAIO env, and would like to set X-Account-Sysmeta-Global-Write-Ratelimit | 17:01 |
| robefran | I've tried POST with and adming Auth_token, but the account-sysmeta does not show up in a subsequent HEAD | 17:02 |
| *** lyrrad has joined #openstack-swift | 17:04 | |
| *** Kennan2 has quit IRC | 17:05 | |
| MooingLemur | peluse_: though I do think there still might be a memory leak that is apparent if you have an object that can never be fully reconstructed. I'm going to intentionally break one of my objects to confirm. | 17:10 |
| peluse_ | robefran, sysmeta is for use by Swift so you can't set it from outside of the proxy (well, its SW so there are ways to work around that but then you're doing things outside of the design) | 17:15 |
| * peluse_ offline for a while | 17:15 | |
| *** david-lyle is now known as cbrown | 17:17 | |
| *** hrou has quit IRC | 17:17 | |
| robefran | peluse_: from a closer reading of the ratelimit doc I found , This header will be hidden from the user, because of the gatekeeper middleware, and can only be set using a direct client to the account nodes | 17:17 |
| *** chsc has joined #openstack-swift | 17:17 | |
| robefran | peluse_: so, how do I use a direct client to the account nodes ? | 17:18 |
| *** jrichli has quit IRC | 17:18 | |
| *** jlhinson has quit IRC | 17:18 | |
| *** Kennan has joined #openstack-swift | 17:21 | |
| *** jkugel has joined #openstack-swift | 17:21 | |
| *** hrou has joined #openstack-swift | 17:23 | |
| *** hezhiqiang has joined #openstack-swift | 17:39 | |
| *** hrou has quit IRC | 17:41 | |
| tamizh_geek | Hello people., does anybody have a idea why this piece of code, updates all headers except a few? | 17:45 |
| tamizh_geek | https://github.com/openstack/swift/blob/master/swift/proxy/controllers/base.py#L70 | 17:45 |
| openstackgerrit | Clément Contini proposed openstack/swift: Keep user id and project id in subrequests env https://review.openstack.org/212149 | 17:56 |
| *** annegentle has quit IRC | 18:00 | |
| openstackgerrit | Merged openstack/swift: FakeFooters middleware https://review.openstack.org/165517 | 18:06 |
| *** zhill has joined #openstack-swift | 18:07 | |
| clayg | tamizh_geek: I'd have to track down where it's called to understand "why" | 18:38 |
| clayg | tamizh_geek: beyond - "apparently we had to update all of the headers from somewhere accept for the same few in a couple of places so it got put into a function" | 18:39 |
| tamizh_geek | clayg: There is more detailed question here, with the trace - https://review.openstack.org/#/c/204196/ | 18:39 |
| tamizh_geek | clayg: Where did you that qouted text from? | 18:40 |
| clayg | I was quoting it as the "obvious" answer to the question of why a function would work the way it does - because the code that it was extracted from needed to work that way :P | 18:41 |
| tamizh_geek | clayg: and where would that be? :) | 18:46 |
| clayg | the code it was extracted from? - git archeology i suppose | 18:46 |
| clayg | I was looking at it ... it's not super obvious | 18:47 |
| clayg | obviously part of the problem was the inconsistent behavior in the storage servers wrt accept headers and error responses | 18:47 |
| tamizh_geek | oh the commit message! damn.. | 18:48 |
| clayg | but I'm drawing a blank on when best response would ever be dealing with some 2XX's and 4/5XX's | 18:48 |
| clayg | i mean I guess for a DELETE would have some 202's and 404's maybe - but those would have used to have all been text/html - account & contaienr's GET's are really the only place I can think of where Accept header was going to be honored - and those don't use best_response I don't think | 18:51 |
| tamizh_geek | that patch isn't a simple on as I expected. There is a lot of inconsistencies around. One for example is a non-existent container DELETE. Now, that returns a HTML response and text/html header, even when the accept is application/json | 18:51 |
| *** mahatic_ has quit IRC | 18:53 | |
| clayg | I have a hunch that it might be best if we the proxy were to try and extract the "reason" from the backend responses and then do it's own content-type negotiation based on accept headers | 18:54 |
| tamizh_geek | clayg: yes, at least for the error responses. | 18:55 |
| clayg | e.g. in the 503 case most of the time none of the actual "bodies" from the backend storage servers end up as part of the constructed client response. | 18:55 |
| clayg | tamizh_geek: yup | 18:55 |
| clayg | tamizh_geek: well... DON'T GIVE UP! | 18:56 |
| tamizh_geek | so in the specific case we discuss, if there are bunch of 2XX and just a single 4XX then the best_response would actually select a 4XX only right? | 18:56 |
| clayg | tamizh_geek: people getting board with it is the only way it's lasted this long :\ | 18:56 |
| clayg | timburke: no - if there was a bunch of 2XX's it end up building some kind of 2XX looking response | 18:56 |
| clayg | tamizh_geek: ^^ | 18:56 |
| clayg | timburke: ignore me | 18:56 |
| tamizh_geek | clayg: heh | 18:57 |
| tamizh_geek | clayg: I am not giving up, at least for now :-) | 18:57 |
| clayg | that's the spirit! you can do it! | 18:57 |
| wbhuber | clayg: glad to know that you returned back home safe! regarding the quorum on durable response being too low, doesn't it come down from the GET path (i.e. _get_or_head_response)? | 18:59 |
| tamizh_geek | see, if its cooking up a 2XX response from the mixed response - then it would be slightly tricky to do it at the proxy layer. If its always like the fail response wins, like a 503, it would be much simple! | 19:00 |
| tamizh_geek | clayg: ^^ | 19:00 |
| clayg | wbhuber: yeah I think the current value of 2 for quorum durable is too low - I think acoles_ pointed out that parity+1 is best | 19:00 |
| wbhuber | right - i am looking into it | 19:01 |
| clayg | wbhuber: where *is* that calculated? somehwere in ECObjectController - but it must be in PUT somewhere... maybe the PUTTER? | 19:01 |
| *** SkyRocknRoll has quit IRC | 19:01 | |
| clayg | tamizh_geek: heh, yes perhaps easier - but not correct :P | 19:01 |
| clayg | tamizh_geek: I'm trying to think if maybe it'd be better if the proxy always requested json from the backend - and then traslated at the last minute for the client response (that might be totally crazy) | 19:02 |
| tamizh_geek | clayg: how is it done for the success response? I mean the listings? Who serializes the final output format? proxy or the respective servers? | 19:03 |
| clayg | tamizh_geek: servers | 19:04 |
| tamizh_geek | clayg: Its actually a good idea, we can push all the SerDer logic to a module in proxy. Have it extendable - so that people can write their own format serializers :-) | 19:04 |
| clayg | tamizh_geek: so... it's crazy... it's just that on most of GET paths it's not a "combine multiple responses to generate the appropriate client response" sort of situation - we just ask the backend for a response and return it to the client | 19:05 |
| *** hezhiqiang has quit IRC | 19:08 | |
| tamizh_geek | clayg: I should probably first document all these inconsistencies in a matrix. endpoints vs status codes vs accept : how they behave/how they should have behaved. | 19:10 |
| tamizh_geek | What do you think? | 19:10 |
| *** zhill has quit IRC | 19:12 | |
| *** esker has quit IRC | 19:13 | |
| *** moo has joined #openstack-swift | 19:18 | |
| *** zhill has joined #openstack-swift | 19:23 | |
| moo | Hi. Anybody here who can tell me how to rollback the store within the swift middleware pipeline? I want to delete the before saved object when an error occurs within the pipeline. Currently I try this with a subrequest but get a HTTP 409 Error | 19:24 |
| *** NM has joined #openstack-swift | 19:27 | |
| clayg | tamizh_geek: it would definately be helpful to understand the scope of the problem | 19:44 |
| clayg | tamizh_geek: I might suggest that you try to write unittests for it tho - just because once you get it written down you'll want to start fixing it - and then once it's fixed you'll want to merge it - and maintainers will want tests to prevent regression | 19:44 |
| clayg | moo: sounds dangerous - the proxy isn't trasnactional - if the process is brutally killed there would be no one to do this clean up | 19:46 |
| clayg | OTOH, if you do end up needing to write someone to do the cleanup - you might consider just letting them do it and no bother doing it in the proxy (although sometimes it's helpful to be opptomistic and have both places do it) | 19:47 |
| clayg | moo: either way - you need to write a tombstone "at the same time" as a the .data file you could use internal offset timestamps to create a tombstone that's "newer" utils.Timestamp(x-timestamp-from-request-you-want-to-kill, offset=1) might work | 19:48 |
| *** delattec has quit IRC | 19:53 | |
| peluse_ | clayg, ping | 19:53 |
| clayg | peluse_: so I was playing with the "re-insert alt-frags to running node iter" idea on https://gist.github.com/clayg/bf6edd1d79026dc3e92f | 19:53 |
| peluse_ | yeah, I saw that - thanks. That's nearly identical to what I tried at first if you can believe it! | 19:53 |
| peluse_ | The sleep() that you have in there I think is what I see missing in the real implementaion and why all the connects seem to happen before the for loop that enums the parts_iter | 19:54 |
| clayg | it's sort of curious how the variance in connects/timeouts is going to make it so sometimes you get multiple 404s before the nodes with multiples get re-isnerted - sometimes you wont! | 19:54 |
| clayg | peluse_: maybe - you don't need the sleep in the for-realyz case when there is *actual* network yielding the reactor | 19:55 |
| clayg | and I think the FakeConn in __init__ has the sleep in there for you with a comment about "allowing the node_iter to yeild" | 19:55 |
| peluse_ | clayg, yeah I thought maybe a real network connection (or exception) might cause a yield but in the test code that doesn't happen | 19:55 |
| clayg | peluse_: also, I totally believe you | 19:55 |
| peluse_ | hmmm... | 19:55 |
| peluse_ | hmmm... on the fakeConn sleep thin that is | 19:55 |
| peluse_ | I'l dork with it some more - I'm prety hapy that the node wrapper I came up with is so close to yours though :) | 19:56 |
| clayg | # be nice to trixy bits with node_iter's | 19:56 |
| clayg | ^ so a clayg comment - why do you people let me merge this crap :'( | 19:56 |
| clayg | peluse_: good work! | 19:57 |
| clayg | lunch | 19:58 |
| peluse_ | clayg, anyway the patch that's there now doesn't modify the node_iter it just spawns more threads on the same pile and might be better since there's no chance of missing the ability to modify the node_iter beacuse the threads are done. Plus its pretty small and pretty simple | 19:59 |
| peluse_ | https://review.openstack.org/#/c/213147/ | 19:59 |
| *** alejandrito has quit IRC | 20:04 | |
| *** petertr7 is now known as petertr7_away | 20:05 | |
| *** peluse_ is now known as peluse | 20:07 | |
| *** ChanServ sets mode: +v peluse | 20:07 | |
| *** hezhiqiang has joined #openstack-swift | 20:09 | |
| *** hezhiqiang has quit IRC | 20:13 | |
| *** geaaru has joined #openstack-swift | 20:19 | |
| *** zhill has quit IRC | 20:26 | |
| *** zhill has joined #openstack-swift | 20:32 | |
| *** moo has quit IRC | 20:35 | |
| *** jkugel has left #openstack-swift | 20:46 | |
| *** dustins has quit IRC | 20:58 | |
| MooingLemur | X-Container-Read: .r:*,account:nonadmin without .rlistings will make the ACL code ignore that the account:nonadmin ACL exists, because .r:* will pass the test. | 21:02 |
| MooingLemur | Bug, or design decision? :) | 21:02 |
| MooingLemur | I mean, account:nonadmin should be able to list the container, but now they can't because .r:* takes precendence. | 21:04 |
| *** gyee has quit IRC | 21:09 | |
| *** cbrown is now known as david-lyle | 21:20 | |
| *** jerrygb has joined #openstack-swift | 21:23 | |
| *** jerrygb has quit IRC | 21:30 | |
| *** geaaru has quit IRC | 21:32 | |
| peluse | clayg, did you and acoles_ talk about a plan/timeline for opt. GET? I think I was going to work on the proxy side and acoles_ on the obj server side so unless I hear different will start preparing that patch here shortly... | 21:34 |
| wbhuber | clayg: peluse: acoles_: i've started digging on upgrading the quorum on durable response that is currently too low to parity + 1. i don't think that's related to opt. GET though. | 21:40 |
| peluse | wbhuber, cool, thanks. yeah they're different (well, independent patches at least) | 21:41 |
| *** zhill has quit IRC | 21:43 | |
| *** chsc has quit IRC | 21:50 | |
| *** zhill has joined #openstack-swift | 21:54 | |
| *** nadeem has quit IRC | 22:02 | |
| *** minwoob has quit IRC | 22:09 | |
| *** wbhuber has quit IRC | 22:14 | |
| *** lcurtis has quit IRC | 22:29 | |
| *** marcusvrn has quit IRC | 22:43 | |
| *** NM has quit IRC | 22:44 | |
| *** pgbridge has quit IRC | 22:59 | |
| *** zhill has quit IRC | 23:08 | |
| *** pgbridge has joined #openstack-swift | 23:08 | |
| *** zhill has joined #openstack-swift | 23:17 | |
| *** kutija has quit IRC | 23:32 | |
| *** breitz has quit IRC | 23:39 | |
| *** breitz has joined #openstack-swift | 23:39 | |
| *** zhill has quit IRC | 23:45 | |
| openstackgerrit | Kai Zhang proposed openstack/swift: Fix some minor typos https://review.openstack.org/213378 | 23:49 |
| *** zhill has joined #openstack-swift | 23:54 | |
| *** zhill has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!