| *** haomaiwa_ has quit IRC | 00:01 | |
| *** haomaiwang has joined #openstack-swift | 00:01 | |
| *** wbhuber has joined #openstack-swift | 00:02 | |
| clayg | briancline: np | 00:05 |
|---|---|---|
| *** pchng_ghost has joined #openstack-swift | 00:10 | |
| *** pchng_ has quit IRC | 00:13 | |
| jrichli | acoles_ : I now remember having the same issue before with the content-type in container listing becoming "<content-type>;meta=". I got around that by encoding the content-type value | 00:16 |
| jrichli | acoles_ hrou: and I think I now remember hrou asking about that extra encoding, and I just didn't remember why it was there. It was removed. Now I think we have discovered why ;-) | 00:17 |
| clayg | our current part-placement algo sucks: | 00:23 |
| clayg | Tier Parts % Max 0 1 2 3 | 00:23 |
| clayg | -------------------------------------------------------------------------- | 00:23 |
| clayg | r1z3-127.0.0.1/sdb3 668 14.97 1 356 568 100 0 | 00:23 |
| clayg | someone should really fix that | 00:23 |
| clayg | acoles_: I might need you to explain to me why ssync needs to keep two .meta's around | 00:24 |
| wbhuber | clayg: uve a launchpad bug for that? | 00:24 |
| clayg | acoles_: i saw that when I had a server who had the latest metadata but needed the ctype - it got a duplciate meta file with a bigger -<ctime> delta | 00:25 |
| clayg | acoles_: it was happy to leave both files on disk until it needed to move the part | 00:25 |
| clayg | very strange... | 00:25 |
| clayg | well, maybe not - i guess creating a new meta file for the ctype is probably no worse that creating a new meta with the combined info and removing the old - maybe it's even better a little | 00:26 |
| clayg | just weird to have to obj dirs that represent the same information but have different on disk representations (maybe?) | 00:26 |
| clayg | wbhuber: lp bug #1452431 and patch #241571 | 00:27 |
| openstack | Launchpad bug 1452431 in OpenStack Object Storage (swift) "some parts replicas assigned to duplicate devices in the ring" [High,Confirmed] https://launchpad.net/bugs/1452431 - Assigned to Samuel Merritt (torgomatic) | 00:27 |
| *** diogogmt has joined #openstack-swift | 00:29 | |
| serverascode | does anyone have recommendations on what documentation I can look at that will help me understand what happens when I add a region to a swift cluster? like how much data will move around and such? | 00:34 |
| clayg | serverascode: i'm not sure that's well documented | 00:39 |
| serverascode | yeah I'm having a hard time trying to figure out what will happen | 00:40 |
| clayg | serverascode: you take your prod rings into a temp dir and play around with adding the devices and see whats-what with "dispersion -v" | 00:40 |
| clayg | serverascode: well same thing that happens when you add a zone, except if you have more replicas than regions a lot of parts are going to *want* to move - but maybe won't be able to (yet) until you add more capacity (weight) to the remote region | 00:42 |
| serverascode | clayg: thanks, ok, I will look into that :) | 00:42 |
| *** m_kazuhiro has joined #openstack-swift | 00:43 | |
| clayg | ok, i'm pretty happy working through the failure modes on fast-POST functionally | 00:44 |
| clayg | containers and object trees all seem to consistent stuff up - POST is doing it's thing with the asyncs | 00:44 |
| *** jamielennox is now known as jamielennox|away | 00:46 | |
| clayg | oh - i guess I didn't do anything with different data files on disk - let me see what I can cook up there | 00:46 |
| kota_ | good morning | 00:47 |
| *** jamielennox|away is now known as jamielennox | 00:47 | |
| *** daemontool has quit IRC | 00:48 | |
| *** daemontool has joined #openstack-swift | 00:48 | |
| clayg | oh i totally forgot about swift-init *-replicator status | 00:52 |
| *** jamielennox is now known as jamielennox|away | 00:58 | |
| *** haomaiwang has quit IRC | 01:01 | |
| *** haomaiwang has joined #openstack-swift | 01:01 | |
| *** jamielennox|away is now known as jamielennox | 01:06 | |
| *** badari has quit IRC | 01:06 | |
| jrichli | kota_ : good morning | 01:07 |
| kota_ | jrichli: :-) | 01:07 |
| *** mzhou has quit IRC | 01:08 | |
| *** cschwede has quit IRC | 01:17 | |
| *** badari has joined #openstack-swift | 01:18 | |
| *** badari has quit IRC | 01:24 | |
| *** zhill has quit IRC | 01:24 | |
| *** garthb has quit IRC | 01:25 | |
| *** asettle has quit IRC | 01:36 | |
| clayg | acoles_: phew - that was fun! | 01:37 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Add missing docs for ring.builder.rebalance https://review.openstack.org/242295 | 01:41 |
| openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/swift: Added unit test cases for server Added three unit test cases for REPLICATE and __call_ magic methods in account and container https://review.openstack.org/242300 | 01:52 |
| *** badari has joined #openstack-swift | 01:53 | |
| *** janonymous has joined #openstack-swift | 01:57 | |
| *** haomaiwang has quit IRC | 02:01 | |
| *** haomaiwang has joined #openstack-swift | 02:01 | |
| *** janonymous has quit IRC | 02:09 | |
| *** breitz2 has quit IRC | 02:11 | |
| *** breitz has joined #openstack-swift | 02:12 | |
| *** asettle has joined #openstack-swift | 02:27 | |
| *** gyee has quit IRC | 02:31 | |
| *** mahatic has joined #openstack-swift | 02:45 | |
| mahatic | good morning | 02:50 |
| *** hrou has joined #openstack-swift | 02:54 | |
| openstackgerrit | Merged openstack/swift: Device marked to be removed in info about the ring. https://review.openstack.org/238917 | 03:00 |
| *** haomaiwang has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-swift | 03:02 | |
| janonymous_ | morning | 03:08 |
| *** aagrawal has joined #openstack-swift | 03:21 | |
| *** badari has quit IRC | 03:25 | |
| *** wbhuber has quit IRC | 03:38 | |
| *** wbhuber has joined #openstack-swift | 03:39 | |
| *** wbhuber has quit IRC | 03:47 | |
| *** wbhuber has joined #openstack-swift | 03:48 | |
| *** haomaiwang has quit IRC | 03:51 | |
| *** wbhuber has quit IRC | 03:55 | |
| *** links has joined #openstack-swift | 03:59 | |
| kota_ | mahatic, janonymous: good morning | 03:59 |
| openstackgerrit | Merged openstack/swift: Add unit tests for utils.validate_hash_path https://review.openstack.org/231905 | 04:00 |
| janonymous_ | kota_ : o/ | 04:01 |
| *** hrou has quit IRC | 04:11 | |
| *** ppai has joined #openstack-swift | 04:19 | |
| openstackgerrit | Bill Huber proposed openstack/swift: ObjectControllers return application errors as 499 on bad read https://review.openstack.org/236007 | 04:23 |
| *** mwheckmann has joined #openstack-swift | 04:32 | |
| *** wuhg has joined #openstack-swift | 04:43 | |
| *** jrichli_ has joined #openstack-swift | 05:02 | |
| *** m_kazuhiro has quit IRC | 05:03 | |
| *** jrichli has quit IRC | 05:05 | |
| *** zhill has joined #openstack-swift | 05:08 | |
| *** jrichli_ has quit IRC | 05:10 | |
| *** klrmn1 has joined #openstack-swift | 05:11 | |
| *** klrmn2 has quit IRC | 05:11 | |
| *** zhill has quit IRC | 05:15 | |
| *** jerrygb_ has quit IRC | 05:18 | |
| *** klrmn1 has quit IRC | 05:26 | |
| *** asettle has quit IRC | 05:27 | |
| *** jlhinson has joined #openstack-swift | 05:27 | |
| *** venkat_p has joined #openstack-swift | 05:30 | |
| *** mahatic_ has joined #openstack-swift | 05:50 | |
| *** mahatic has quit IRC | 05:54 | |
| *** jlhinson has quit IRC | 06:01 | |
| *** mwheckmann has quit IRC | 06:02 | |
| *** m_kazuhiro has joined #openstack-swift | 06:14 | |
| *** haomaiwang has joined #openstack-swift | 06:20 | |
| *** openstackgerrit has quit IRC | 06:31 | |
| *** openstackgerrit has joined #openstack-swift | 06:31 | |
| *** zhill has joined #openstack-swift | 06:32 | |
| *** david-lyle has joined #openstack-swift | 06:40 | |
| *** peterlisak has joined #openstack-swift | 06:44 | |
| *** david-lyle has quit IRC | 06:45 | |
| *** mahatic_ has quit IRC | 06:47 | |
| *** mahatic has joined #openstack-swift | 06:49 | |
| *** jamielennox is now known as jamielennox|away | 06:59 | |
| *** haomaiwang has quit IRC | 07:01 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Read the response body, if response status is greater than 300. https://review.openstack.org/183899 | 07:01 |
| *** haomaiwang has joined #openstack-swift | 07:01 | |
| *** zhill has quit IRC | 07:20 | |
| clayg | great patch 241571 is in conflict | 07:20 |
| patchbot | clayg: https://review.openstack.org/#/c/241571/ - Put part-replicas where they go | 07:20 |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Put part-replicas where they go https://review.openstack.org/241571 | 07:35 |
| clayg | could have been worse, we'll see what gerrit says | 07:35 |
| *** SkyRocknRoll has joined #openstack-swift | 07:37 | |
| *** cschwede has joined #openstack-swift | 07:51 | |
| *** ChanServ sets mode: +v cschwede | 07:55 | |
| *** arnox has joined #openstack-swift | 08:00 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-swift | 08:01 | |
| hugokuo | not sure if this is a bug or my mistake... How will you use POST to add header on an manifest file ? Let's say add X-Delete-At on a manifest file | 08:07 |
| hugokuo | While issue a POST to manifest file, it pulls all segments and store it as a single object in server side. | 08:07 |
| *** haomaiwang has quit IRC | 08:11 | |
| hugokuo | btw... it's for DLO | 08:17 |
| *** rledisez has joined #openstack-swift | 08:20 | |
| mahatic | hugokuo: https://bugs.launchpad.net/swift/+bug/1487791 - this might be relevant | 08:22 |
| openstack | Launchpad bug 1487791 in OpenStack Object Storage (swift) "POST to DLO squashes data without fast-POST" [Undecided,New] | 08:22 |
| *** geaaru has joined #openstack-swift | 08:24 | |
| hugokuo | @mahatic yes... it is. thx | 08:25 |
| clayg | wooo look at mahatic knows all the bugs | 08:30 |
| mahatic | clayg: :D no, I kinda stumbled onto it (when the other day I was asking about slo dlo) | 08:31 |
| clayg | oh wtf ever cli tests! | 08:34 |
| clayg | who merged that crap?! | 08:34 |
| clayg | kota_'s gotta update the doc strings | 08:34 |
| clayg | i have a bunch of Replica count of 3.0 requires more than 2 devices failures that aren't even in the tests that are supposed to test ring building - they're just checking stupid cli output | 08:35 |
| clayg | grrrr | 08:35 |
| kota_ | what's | 08:35 |
| clayg | kota_: do you have a patch to fix the doc string on rebalance cause the return got changed? | 08:35 |
| *** venkat_p has quit IRC | 08:36 | |
| kota_ | clayg: just docs change i have | 08:36 |
| *** hseipp has joined #openstack-swift | 08:36 | |
| *** zigo has quit IRC | 08:37 | |
| clayg | kota_: it's no big deal "swift-ring-builder can't remove a device with zero weight" has bit me like three times now; but the chance was fine; shit happens; i'm just being whiny | 08:38 |
| clayg | *ignore me* | 08:39 |
| mahatic | how do you debug a functional test again? (not print but trace) | 08:39 |
| kota_ | patch 233096? | 08:39 |
| patchbot | kota_: https://review.openstack.org/#/c/233096/ - swift-ring-builder can't remove a device with zero... | 08:39 |
| clayg | mahatic: pdb? | 08:42 |
| clayg | pdb.set_trace() | 08:42 |
| *** zigo has joined #openstack-swift | 08:42 | |
| mahatic | doesn't work in func | 08:42 |
| mahatic | that works for unit tests | 08:42 |
| clayg | oh... well you can use inprocess | 08:42 |
| clayg | i think.. | 08:42 |
| clayg | it may be that when it gets to that point I write a unittest | 08:42 |
| kota_ | ah, the conflicted one might be patch 238917 | 08:42 |
| patchbot | kota_: https://review.openstack.org/#/c/238917/ - Device marked to be removed in info about the ring. | 08:42 |
| kota_ | clayg: sorry I added +A for that. | 08:42 |
| clayg | kota_: oh yeah that one too | 08:42 |
| clayg | no it's also a good change | 08:42 |
| clayg | oh was that the one that added these cli tests? | 08:42 |
| clayg | no... | 08:42 |
| *** openstackstatus has quit IRC | 08:42 | |
| kota_ | that one added cli/ring_buider test for making sure the output. | 08:42 |
| openstackgerrit | Merged openstack/swift: Read the response body, if response status is greater than 300. https://review.openstack.org/183899 | 08:44 |
| clayg | oh | 08:47 |
| mahatic | clayg: maybe a unittest. not sure of inprocess - an option in set_trace? | 08:48 |
| *** jordanP has joined #openstack-swift | 08:51 | |
| clayg | mahatic: I dont' even remember how to run inprocess func tests - I think you can just set an env flag | 08:54 |
| clayg | or something with tox? | 08:54 |
| mahatic | yeah, i'm looking at it. thanks! | 08:54 |
| clayg | good merge that interal client fix - thanks kota! | 08:58 |
| kota_ | clayg: :) | 08:58 |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Put part-replicas where they go https://review.openstack.org/241571 | 08:59 |
| *** venkat_p has joined #openstack-swift | 08:59 | |
| clayg | mahatic: where was the first place you looked for "SWIFT_TEST_IN_PROCESS=true .functests" | 09:02 |
| clayg | or is that even what you came up with? | 09:02 |
| clayg | mahatic: we should do a doc update | 09:02 |
| clayg | i checked the README - then searched the saio docs - then I looked in test/functional/__init__.py (and eventually backed into the env var) | 09:04 |
| clayg | but i was advanaged knowing that in process functests were even a thing... | 09:04 |
| mahatic | clayg: true. saio can have a doc update maybe. or the first contribution | 09:05 |
| clayg | mahatic: but fwiw pdb.set_trace totally worked | 09:05 |
| mahatic | clayg: and I didn't come up with that :) I was toggling with pycharm capabilities | 09:05 |
| clayg | i'm up in my debug on an object server in the middle of a functional tests | 09:05 |
| clayg | those things are crazy | 09:06 |
| clayg | python is amazingawesomesause | 09:06 |
| mahatic | clayg: really? the console just stops when I put set_trace in func/tests.py | 09:06 |
| clayg | -vsx (nose tries to capture stdout by default) | 09:06 |
| clayg | how do you invoke the tests? | 09:06 |
| clayg | pycharm magic? | 09:06 |
| mahatic | haha, nope. In this case nosetests /test/functional/tests.py | 09:07 |
| clayg | sweet! | 09:07 |
| clayg | yeah just -s (--no-capture) | 09:07 |
| mahatic | clayg: awesome! that works! | 09:08 |
| clayg | werd | 09:10 |
| *** links has quit IRC | 09:11 | |
| clayg | next time some one sees pdardeau we need tell him to get a bouncer setup | 09:11 |
| clayg | pdardeau: patch 1513761 | 09:12 |
| clayg | er... lp bug #1513761 | 09:12 |
| openstack | Launchpad bug 1513761 in OpenStack Object Storage (swift) "in-process functests are not documented" [Undecided,New] https://launchpad.net/bugs/1513761 | 09:12 |
| clayg | ^ cc portante | 09:12 |
| mahatic | cool, you raiseda bug | 09:13 |
| *** jistr has joined #openstack-swift | 09:19 | |
| mahatic | clayg: it's actually documented here: http://docs.openstack.org/developer/swift/development_guidelines.html | 09:20 |
| clayg | well there you go! | 09:22 |
| *** jmccarthy has quit IRC | 09:22 | |
| mahatic | lesson leanrt: keep a tab of three docs - saio, first_contribution, development_guidelines (unless there are more?) | 09:23 |
| *** jmccarthy has joined #openstack-swift | 09:24 | |
| clayg | heh | 09:24 |
| clayg | i can tell which git review was for fast-POST because the number is so damn low | 09:25 |
| *** acoles_ is now known as acoles | 09:26 | |
| acoles | clayg: good morning | 09:27 |
| * acoles comments on launchpad before connecting to irc :) | 09:27 | |
| clayg | acoles: I think we cross posted - i marked it as invalid - problem wasn't the information wasn't written down - just that I coudln't find it - happens all the time | 09:28 |
| acoles | clayg: true | 09:31 |
| acoles | clayg: thanks for the review!! i owe you a whisky or two! | 09:32 |
| clayg | for doing my job? | 09:33 |
| clayg | it's fine - i owe you a whisky or two as well I'm sure | 09:33 |
| clayg | acoles: I'm acctually still on about this ssync thing... | 09:33 |
| acoles | well i'm always up for beers but you never want to stop talking work :P | 09:33 |
| clayg | look at these .meta's -> https://gist.github.com/clayg/a378a6ac4405d81d0aa3 | 09:33 |
| acoles | clayg: yeah i need to go and ponder that | 09:34 |
| acoles | clayg: so do you observe 94116-29010c.meta created *after* 94116.meta in same dir?? (i realise maybe its hard for you to know the order) | 09:36 |
| clayg | no not hard - that's exactly what I see for sure | 09:38 |
| acoles | 94116.meta *should* get removed - that may be a bug if it is not. Its contents would be same as 94116-29010c.meta except for the content-type | 09:38 |
| acoles | so its redundant once 94116-29010c.meta is written | 09:38 |
| acoles | hmmm | 09:38 |
| clayg | confirmed with swift-object-info the metadata from 94116 makes its way into 94116-29010c along with the older content-type | 09:40 |
| *** haomaiwa_ has joined #openstack-swift | 09:40 | |
| clayg | 94116 has no conent-type (but it was the newest user metadata) - 94116-29010c has the new user metadata (but also the older content type update that was missing in 94116) | 09:41 |
| clayg | acoles: do you expect hash_listdir_cleanup to sweep it out? maybe it's just a < vs <= thing on the .meta time portion of the timestamp | 09:41 |
| acoles | right. so is suspect its a glitch in determining which .metas are obsolete, maybe i'm only comparing the primary timestamps, which are equal | 09:41 |
| acoles | clayg: yeah exactly what i am suspecting | 09:42 |
| clayg | well it's definately a problematic bug because it causes ssync partitions to have different hashes :\ | 09:42 |
| clayg | (updated the gist) | 09:42 |
| acoles | need a <= in place of <, or perhaps needs a comparison of ctype time | 09:43 |
| clayg | acoles: I'll try to write you a test | 09:43 |
| acoles | clayg: thanks! | 09:44 |
| *** haomaiwa_ has quit IRC | 09:45 | |
| clayg | acoles: https://gist.github.com/clayg/6a455fd71943256023fb | 09:51 |
| clayg | I think something that exposes how ssync can end up shipping a combined udpate to a node that already has the newest meta would be a little more behavior driven | 09:52 |
| clayg | but if possible, it'd help with abstractions and holding it all in your head if you could count on hash_listdir_cleanup to just "do the right thing" no matter how the .meta files got in there | 09:53 |
| acoles | clayg: this ssync test *ought* to expose this, its probably just not asserting the dir contents - test/unit/obj/test_ssync_sender.py:2453 | 09:57 |
| acoles | clayg: yeah, agree, HCL should do the right thing. I'll look into it today. | 09:57 |
| clayg | /tmp/tmpOmC5V7/tmp_test_ssync_receiver/dev/objects/9/3d1/4f6587504178bff82772d66708d083d1/1446804093.00000-186a0.meta | 10:02 |
| clayg | /tmp/tmpOmC5V7/tmp_test_ssync_receiver/dev/objects/9/3d1/4f6587504178bff82772d66708d083d1/1446804093.00000.meta | 10:02 |
| clayg | ^ rx_testdir - wtg test_content_type_sync | 10:03 |
| clayg | acoles: ^ | 10:03 |
| clayg | ok, so you've got two good tests to work from | 10:03 |
| acoles | clayg: got it | 10:03 |
| acoles | clayg: k, i have the baton, nice catch | 10:03 |
| clayg | acoles: awesome - do you need me to ding you on the review? | 10:03 |
| clayg | acoles: ok, great, yeah np - it was totaly luck | 10:04 |
| clayg | g'night | 10:04 |
| acoles | i'll WIP it | 10:04 |
| acoles | you too! | 10:04 |
| *** joeljwright has joined #openstack-swift | 10:08 | |
| *** ChanServ sets mode: +v joeljwright | 10:08 | |
| *** venkat_p has quit IRC | 10:11 | |
| *** dabukalam has quit IRC | 10:12 | |
| *** mitz has quit IRC | 10:12 | |
| *** dabukalam has joined #openstack-swift | 10:13 | |
| *** mfalatic has joined #openstack-swift | 10:13 | |
| *** mitz has joined #openstack-swift | 10:14 | |
| *** mfalatic_ has quit IRC | 10:15 | |
| *** venkat_p has joined #openstack-swift | 10:24 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Guess content-type on PUT https://review.openstack.org/232572 | 10:40 |
| acoles | mahatic: you around? could you check the change i made to the tests ^^? | 10:40 |
| mahatic | acoles: sure, looking. I thought you'd have a simpler version of it - like a 2 line change or something :) | 10:42 |
| acoles | mahatic: is that a challenge - to compress it to 2 lines? :) | 10:43 |
| mahatic | acoles: nope :D, my version of it looked complicated (more line changes and a possible use of ast) which is why I didn't push | 10:44 |
| mahatic | acoles: happy to see your change also involved quite a few line change though ;) Thanks! Looks good, will run it on my local | 10:45 |
| mahatic | acoles: not sure if json.loads(param[5:]) won't result differently on diff machines | 10:49 |
| *** jamielennox|away is now known as jamielennox | 10:49 | |
| mahatic | will have to wait for jenkins I guess | 10:49 |
| acoles | mahatic: which part? json.loads will return a dict. dict comparison is independent of how the keys might be ordered in the dicts. | 10:51 |
| mahatic | acoles: oh yeah, alright. nevermind | 10:51 |
| acoles | mahatic: the problem we ran into is when comparing serializations of dicts when the items get serialized in unpredictable order. | 10:52 |
| acoles | mahatic: a similar pitfall is to assert dict1.keys() == [a,b,c] - it may be [c,a,b] or whatever. | 10:54 |
| mahatic | acoles: correct. dict1.keys() == similar_dict1.keys() wouldn't be a prob i think | 10:55 |
| acoles | mahatic: i think that's safe, i'd want to check before trusting my life on it though ;) | 10:57 |
| mahatic | acoles: heh, sure | 11:02 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Decrypting Container Listing https://review.openstack.org/214438 | 11:03 |
| acoles | ^^ thats just a rebase | 11:03 |
| mahatic | okay | 11:05 |
| *** aix has quit IRC | 11:07 | |
| *** jamielennox is now known as jamielennox|away | 11:11 | |
| *** lpabon has joined #openstack-swift | 11:43 | |
| *** aagrawal has quit IRC | 11:50 | |
| *** jerrygb has joined #openstack-swift | 11:54 | |
| *** kei_yama has quit IRC | 11:58 | |
| *** joeljwright has quit IRC | 11:59 | |
| *** joeljwright has joined #openstack-swift | 12:00 | |
| *** ChanServ sets mode: +v joeljwright | 12:00 | |
| * portante causes too much trouble! | 12:27 | |
| *** Guest7190 has quit IRC | 12:30 | |
| *** wuhg has quit IRC | 12:31 | |
| acoles | portante: howdy! | 12:31 |
| *** SkyRocknRoll has quit IRC | 12:34 | |
| *** zigo has quit IRC | 12:36 | |
| *** zigo has joined #openstack-swift | 12:37 | |
| *** weihan has joined #openstack-swift | 12:38 | |
| portante | acoles: how are things? | 12:38 |
| kota_ | fyi, I'll be offline until next Tuesday for a vacation. just notification. | 12:44 |
| *** venkat_p has quit IRC | 12:49 | |
| *** SkyRocknRoll has joined #openstack-swift | 12:51 | |
| *** weihan has quit IRC | 12:51 | |
| *** m_kazuhiro has quit IRC | 12:56 | |
| *** SkyRocknRoll_ has joined #openstack-swift | 12:58 | |
| *** NM has quit IRC | 13:28 | |
| *** NM has joined #openstack-swift | 13:28 | |
| *** aix has joined #openstack-swift | 13:29 | |
| acoles | kota_: have fun | 13:36 |
| acoles | portante: good thanks | 13:36 |
| kota_ | acoles: thx | 13:38 |
| *** NM has quit IRC | 13:39 | |
| *** peterlisak has quit IRC | 13:40 | |
| *** peterlisak has joined #openstack-swift | 13:41 | |
| *** peterlisak_ has joined #openstack-swift | 13:42 | |
| *** pchng_ghost is now known as pchng | 13:46 | |
| *** peterlisak has quit IRC | 13:47 | |
| *** peterlisak_ is now known as peterlisak | 13:47 | |
| *** peterlisak13 has joined #openstack-swift | 13:47 | |
| *** mwheckmann has joined #openstack-swift | 13:48 | |
| *** diogogmt has quit IRC | 13:49 | |
| *** peterlisak has quit IRC | 13:51 | |
| *** peterlisak has joined #openstack-swift | 13:51 | |
| *** diogogmt has joined #openstack-swift | 13:53 | |
| *** peterlisak has quit IRC | 13:53 | |
| *** ppai has quit IRC | 13:54 | |
| *** hrou has joined #openstack-swift | 13:59 | |
| *** rohit_ has joined #openstack-swift | 14:01 | |
| *** dustins has joined #openstack-swift | 14:03 | |
| *** badari has joined #openstack-swift | 14:04 | |
| *** diazjf has joined #openstack-swift | 14:07 | |
| *** NM has joined #openstack-swift | 14:10 | |
| *** jrichli has joined #openstack-swift | 14:16 | |
| *** vinsh has joined #openstack-swift | 14:20 | |
| *** jlhinson has joined #openstack-swift | 14:21 | |
| *** diogogmt has quit IRC | 14:21 | |
| *** diogogmt has joined #openstack-swift | 14:26 | |
| *** petertr7_away is now known as petertr7 | 14:27 | |
| *** diazjf has quit IRC | 14:29 | |
| hrou | mahatic, I saw the convo earlier about testing ! I think some info about unit/func test debugging in the first contribution guide is still a great idea, I know often that's one of the first things people ask here ! | 14:32 |
| *** jlhinson has quit IRC | 14:43 | |
| *** diazjf has joined #openstack-swift | 14:49 | |
| *** pdardeau has joined #openstack-swift | 14:56 | |
| *** NM1 has joined #openstack-swift | 14:58 | |
| *** NM has quit IRC | 14:58 | |
| *** okdas has quit IRC | 14:59 | |
| *** okdas has joined #openstack-swift | 14:59 | |
| *** okdas has joined #openstack-swift | 14:59 | |
| *** blmartin has joined #openstack-swift | 15:00 | |
| *** alejandrito has joined #openstack-swift | 15:04 | |
| *** blmartin has quit IRC | 15:07 | |
| *** amoturi has joined #openstack-swift | 15:08 | |
| *** wbhuber has joined #openstack-swift | 15:10 | |
| hurricanerix | y | 15:10 |
| hurricanerix | whoops, wrong window :) | 15:10 |
| mahatic | hrou: yeah, i think so. Or at least improve what is in development_guidelines | 15:12 |
| acoles | mahatic: you're still here! are you +1 on https://review.openstack.org/232572 ? just checking cos you are the author :) | 15:15 |
| mahatic | acoles: I was just notified by jrichli's comment and was wondering if I can +1 it :) | 15:15 |
| jrichli | unit test question: How are the number of nodes determined for ObjectController tests in unit/proxy/controllers/test_obj.py? | 15:16 |
| mahatic | acoles: there you go, I +1'd it | 15:16 |
| mahatic | thanks for checking | 15:17 |
| acoles | mahatic: thanks | 15:17 |
| *** NM1 has quit IRC | 15:20 | |
| *** NM has joined #openstack-swift | 15:20 | |
| *** tongli has joined #openstack-swift | 15:20 | |
| *** diogogmt has quit IRC | 15:32 | |
| *** sileht has quit IRC | 15:32 | |
| peterlisak13 | Hi, I made a patch 241342 to add content-type in swiftclient list of container content. Is there someone who parses that output and can be potentially affected by the change? | 15:34 |
| patchbot | peterlisak13: https://review.openstack.org/#/c/241342/ - Add content-type in list of container content | 15:34 |
| peterlisak13 | I'm deciding between new option like --content-type or adding to existing --long (IMO quite generic and suitable). | 15:34 |
| *** NM has quit IRC | 15:36 | |
| *** hseipp has left #openstack-swift | 15:37 | |
| *** juzuluag has quit IRC | 15:38 | |
| *** diogogmt has joined #openstack-swift | 15:38 | |
| clayg | good morning | 15:40 |
| wbhuber | jrichli: if i understand your question correctly, i'd say the number of nodes is determined by the config file that the unit test picks up | 15:41 |
| mahatic | clayg: good morning (already!) | 15:43 |
| clayg | mahatic: acoles: prettysure dict.keys() == [a, b, c] runs the risk that dict.keys() returns [c, a, b] | 15:43 |
| clayg | mahatic: acoles: use sets | 15:43 |
| clayg | os sorts | 15:43 |
| *** wbhuber has quit IRC | 15:44 | |
| *** wbhuber has joined #openstack-swift | 15:44 | |
| *** jlhinson has joined #openstack-swift | 15:44 | |
| mahatic | clayg: yes, or json? | 15:45 |
| *** venkat_p has joined #openstack-swift | 15:47 | |
| *** SkyRocknRoll_ has quit IRC | 15:49 | |
| clayg | mahatic: I think even if d1 == d2 you can't be sure dumps(d1) == dumps(d2) | 15:49 |
| mahatic | clayg: oh I see | 15:52 |
| *** mzhou has joined #openstack-swift | 15:53 | |
| openstackgerrit | Ganesh Maharaj Mahalingam proposed openstack/swift: Unit tests for account/backend.py https://review.openstack.org/240613 | 15:55 |
| *** garthb has joined #openstack-swift | 15:57 | |
| openstackgerrit | Bill Huber proposed openstack/swift: ObjectControllers return application errors as 499 on bad read https://review.openstack.org/236007 | 15:57 |
| wbhuber | clayg: ^^ this should be it | 15:58 |
| *** nadeem has joined #openstack-swift | 15:58 | |
| *** nadeem has quit IRC | 15:58 | |
| *** nadeem has joined #openstack-swift | 15:59 | |
| notmyname | good morning | 16:00 |
| mahatic | good morning | 16:02 |
| notmyname | just checking in before I get on the bus | 16:02 |
| *** trifon has joined #openstack-swift | 16:13 | |
| *** petertr7 is now known as petertr7_away | 16:14 | |
| *** klrmn1 has joined #openstack-swift | 16:16 | |
| *** klrmn1 has quit IRC | 16:17 | |
| *** venkat_p has quit IRC | 16:18 | |
| *** SkyRocknRoll has quit IRC | 16:25 | |
| serverascode | clayg: sorry to bother you again, but yesterday you mentioned "dispersion -v", is that swift-dispersion-report? | 16:28 |
| *** nadeem has quit IRC | 16:28 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Update container on fast-POST https://review.openstack.org/135380 | 16:29 |
| jrichli | wbhuber: thx. I do see there are factors such as the policy used in the test, and the number can be changed. Specifically, I am not getting the right number of "nodes" in the node iterator passed to _make_request in controllers/base.py | 16:29 |
| *** rjaiswal has joined #openstack-swift | 16:33 | |
| *** petertr7_away is now known as petertr7 | 16:36 | |
| *** bobby2 has quit IRC | 16:37 | |
| *** HenryG has quit IRC | 16:37 | |
| *** bobby2 has joined #openstack-swift | 16:38 | |
| *** HenryG has joined #openstack-swift | 16:39 | |
| *** sileht has joined #openstack-swift | 16:41 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Update container on fast-POST https://review.openstack.org/135380 | 16:44 |
| *** pdardeau has quit IRC | 16:45 | |
| haypo | hey. is someone available to review my dnspython3 patch? https://review.openstack.org/#/c/217423/ acoles already put his +2 ;) | 16:47 |
| *** NM has joined #openstack-swift | 16:50 | |
| *** gyee has joined #openstack-swift | 16:52 | |
| *** petertr7 is now known as petertr7_away | 16:53 | |
| *** pdardeau has joined #openstack-swift | 16:53 | |
| *** jmccarthy has quit IRC | 16:53 | |
| clayg | serverascode: swift-ring-builder <something>.builder dispersion -v | 16:56 |
| serverascode | clayg: ah ok thanks again | 16:56 |
| clayg | swift-dispersion-report is great too - but it's more of a cluster health thing | 16:57 |
| *** diazjf has quit IRC | 16:59 | |
| tdasilva | wbhuber: hello | 17:03 |
| acoles | clayg: i fixed what we discussed earlier but I think you have shone a light on a wider issue - see my comments on patch 135380 :/ | 17:03 |
| patchbot | acoles: https://review.openstack.org/#/c/135380/ - Update container on fast-POST | 17:03 |
| *** trifon has quit IRC | 17:03 | |
| acoles | jrichli: ping | 17:06 |
| haypo | acoles: oh cool, cschwede approved the pbr change ;) | 17:06 |
| acoles | haypo: :) | 17:06 |
| haypo | next step: https://review.openstack.org/#/c/199034/ (fix tox -e py34) which may have to wait for pyeclib 1.1 | 17:06 |
| haypo | any idea of the status of the pyeclib 1.1 bump? | 17:06 |
| haypo | (i have to go, i will check messages later) | 17:07 |
| notmyname | haypo: yeah, I'm the blocker there | 17:07 |
| haypo | notmyname: what's the cheat code to unblock you? lol | 17:07 |
| openstackgerrit | Ganesh Maharaj Mahalingam proposed openstack/swift: Add unit tests for direct_client https://review.openstack.org/241291 | 17:07 |
| haypo | notmyname: no serisouly, what's the status? is 1.1 already in global requirements? | 17:07 |
| notmyname | haypo: AFAIK, no. there was an upgrade issue. getting it straightened out is in the top few things on my TODO list | 17:08 |
| notmyname | haypo: my goal is to have it resolved middle of next week | 17:09 |
| clayg | I *like* this Jordan Pittier guy - i don't know who he is - but he gets it -> patch 238799 | 17:10 |
| patchbot | clayg: https://review.openstack.org/#/c/238799/ - Change schedule priority of daemon/server in config | 17:10 |
| *** garthb has quit IRC | 17:10 | |
| clayg | acoles: *no* wider issues - only merging | 17:10 |
| *** mfalatic has quit IRC | 17:11 | |
| *** ChanServ sets mode: +v timburke | 17:11 | |
| notmyname | I'd like to welcome timburke to swiftclient core! | 17:12 |
| clayg | whoot! welcome! | 17:12 |
| timburke | thanks notmyname! | 17:13 |
| gmmaha | congrats timburke | 17:13 |
| hrou | congrats timburke ! | 17:13 |
| tdasilva | timburke: congrats! | 17:13 |
| acoles | timburke: welcome! | 17:14 |
| jrichli | acoles: ack | 17:14 |
| acoles | jrichli: about the content-type meta encoding - can you track down the version where it *was* being encoded? | 17:15 |
| acoles | jrichli: i do remember seeing the whole header being base64 encoded, but I don't think that is a solution. | 17:15 |
| hrou | ^ Yea I thought all content type was base64 encoded at some point ? | 17:15 |
| jrichli | acoles: i will take a look | 17:15 |
| clayg | acoles: I also thought about making the suffix hashing use timestamps instead of filenames | 17:15 |
| acoles | clayg: right. that's another way. | 17:16 |
| acoles | clayg: i'm kicking myself so hard i didn't spot this before :/ | 17:16 |
| clayg | meh | 17:16 |
| clayg | this isn't a problem in the container listsings because they always merge timestamps down yeah? | 17:17 |
| clayg | acoles: I need to imagine how rsync replciation can get itself into this state - because ssync replciation can probably work around it | 17:18 |
| acoles | clayg: no problem there. its just inconsistent filenames. | 17:18 |
| jrichli | acoles hrou: you both are probably right that the whole value was encoded. | 17:18 |
| hrou | jrichli, you were trying to find the patch-set # where that was the case ? | 17:19 |
| clayg | acoles: yeah I just mean that when containers replicate they always merge - there's no chance to have to same info represented two ways - if you merge things down to a single timestamp (with deltas) there's always a canonical form | 17:19 |
| *** blmartin has joined #openstack-swift | 17:20 | |
| acoles | clayg: when i first looked at the ssync_sender test i thought they were unrealistic starting conditions, but concurrent POST requests could get you two valid meta files. | 17:20 |
| jrichli | hrou: havent done that yet - actually in a meeting now. but i just assume if you both remember that, its probalby right :-) | 17:20 |
| acoles | jrichli: well leave the meeting!!! :P :P | 17:20 |
| hrou | acoles, +1, clearly jrichli, that's the only sensible thing to do ; ) | 17:21 |
| jrichli | acoles: its prob bad enough i am dividing my attention ;-) | 17:21 |
| acoles | clayg: plus i got thinking about clusters running a mix of rsync and ssync, or transitioning from one to the other. | 17:22 |
| *** joeljwright has quit IRC | 17:22 | |
| acoles | jrichli: sorry, we'll leave you in peace | 17:22 |
| jrichli | acoles: np, really! | 17:22 |
| *** rjaiswal has quit IRC | 17:22 | |
| *** rjaiswal has joined #openstack-swift | 17:23 | |
| *** nadeem has joined #openstack-swift | 17:26 | |
| *** rjaiswal has quit IRC | 17:27 | |
| *** rjaiswal has joined #openstack-swift | 17:29 | |
| *** peterlisak13 has quit IRC | 17:29 | |
| *** subscope has joined #openstack-swift | 17:30 | |
| openstackgerrit | Merged openstack/swift: Guess content-type on PUT https://review.openstack.org/232572 | 17:32 |
| acoles | woohoo mahatic ^^ | 17:33 |
| *** diazjf has joined #openstack-swift | 17:37 | |
| *** jordanP has quit IRC | 17:37 | |
| jrichli | yay! | 17:38 |
| acoles | notmyname: i'll be around less next week, jury service. | 17:40 |
| *** rledisez has quit IRC | 17:41 | |
| notmyname | acoles: right. thanks for the reminder | 17:41 |
| notmyname | I hope you don't get picked :-) | 17:41 |
| acoles | clayg: ^^ if you have any thoughts please leave on gerrit, i may miss stuff in scrollback | 17:41 |
| acoles | notmyname: i read through the docs but there was no section on being 'de-selected'. will have to wait and see how it works. | 17:42 |
| notmyname | I'm goign to be interested to hear about your experience and compare it to my own | 17:42 |
| hrou | acoles, at least in Canada both lawyers (on both sides) along with the judge have the right to refuse a juror for whatever reason, and if that happens you've done your service. | 17:43 |
| hrou | You just need to be undesirable to one of the sides, which can mean many different things depending on the context : ) | 17:43 |
| notmyname | same here in the US | 17:43 |
| acoles | hrou: i've always tried to make "being undesirable" one of my strengths ;D | 17:44 |
| notmyname | eg like the lady who said she couldn't give an impartial judgement based on the facts presented because she "had certain views on race" | 17:44 |
| acoles | notmyname: yeah, i noted that line. | 17:44 |
| *** mfalatic has joined #openstack-swift | 17:44 | |
| notmyname | or the other lady who was getting "psychic impressions" telling her to go one way or another | 17:45 |
| notmyname | (also, san francisco is weird ;-) | 17:45 |
| hrou | acoles, so just be your self ; - ) jk - its not being "unlikeable" though, usually its some sort of perceived bias | 17:45 |
| hrou | acoles, wait I'm not implying you're "unlikeable" : ) | 17:45 |
| acoles | hrou: stop digging man.. :) | 17:45 |
| hrou | notmyname, haha its funny in Canada they're good at catching people who just say something crazy to get out of it, you need to be somewhat genuine : ) | 17:46 |
| notmyname | hrou: I live in the city where "is that a costume or do you just wear that normally" is a totally valid question | 17:46 |
| *** arnox has quit IRC | 17:47 | |
| *** arnox has joined #openstack-swift | 17:48 | |
| hrou | notmyname, I really really want to visit one day, its on my list; Mind you there were parts of Tokyo I saw that I imagine people may ask the same question : ) | 17:48 |
| blmartin | what like batman costumes? | 17:48 |
| notmyname | blmartin: with all the vigilante-justice batmans around, crime is low | 17:49 |
| *** arnox has quit IRC | 17:49 | |
| blmartin | but are they the heros you deserver? | 17:49 |
| notmyname | (seriously, though, crime isn't low, so there maybe should be more batmans around) | 17:49 |
| notmyname | lol | 17:49 |
| blmartin | deserve* | 17:49 |
| *** garthb has joined #openstack-swift | 17:50 | |
| *** zhill has joined #openstack-swift | 17:51 | |
| *** petertr7_away is now known as petertr7 | 17:53 | |
| *** geaaru has quit IRC | 17:53 | |
| *** jrichli has quit IRC | 17:54 | |
| *** jistr has quit IRC | 17:54 | |
| *** peterlisak13 has joined #openstack-swift | 17:56 | |
| notmyname | summit recap email sent to the dev mailing list | 17:57 |
| *** peterlisak13 has quit IRC | 18:03 | |
| *** klrmn1 has joined #openstack-swift | 18:03 | |
| *** acoles is now known as acoles_ | 18:05 | |
| *** wbhuber_ has joined #openstack-swift | 18:07 | |
| *** wbhuber_ has quit IRC | 18:07 | |
| *** wbhuber_ has joined #openstack-swift | 18:07 | |
| pdardeau | notmyname: i'll be around very little next 3 weeks (training) | 18:08 |
| *** diazjf has quit IRC | 18:09 | |
| *** wbhuber has quit IRC | 18:10 | |
| haypo | timburke: "I'd like to welcome timburke to swiftclient core!" oh congrats! | 18:11 |
| timburke | haypo: thanks! also, thanks gmmaha, hrou, tdasilva, and acoles_! | 18:13 |
| *** pdardeau has quit IRC | 18:21 | |
| *** pdardeau has joined #openstack-swift | 18:22 | |
| notmyname | pdardeau: you're with the intel/rax thing right? | 18:26 |
| openstackgerrit | Tim Burke proposed openstack/python-swiftclient: Add --decode-content option https://review.openstack.org/184957 | 18:31 |
| *** mplex has left #openstack-swift | 18:37 | |
| *** proteusguy_ has quit IRC | 18:37 | |
| pdardeau | notmyname: yes, that's correct | 18:47 |
| *** proteusguy_ has joined #openstack-swift | 18:49 | |
| *** janonymous has joined #openstack-swift | 18:52 | |
| tdasilva | there's something about Fridays and #openstack-swift channel | 18:52 |
| tdasilva | pretty funny stuff | 18:52 |
| *** diazjf has joined #openstack-swift | 18:55 | |
| *** jrichli has joined #openstack-swift | 18:56 | |
| gmmaha | notmyname: are there plans to get openstackclient equipped to be the primary away frm the swift client? | 18:57 |
| *** wbhuber has joined #openstack-swift | 18:58 | |
| *** diazjf has quit IRC | 19:00 | |
| *** wbhuber_ has quit IRC | 19:00 | |
| * notmyname is back from meeting | 19:04 | |
| notmyname | pdardeau: are they making you go to rookie-o? | 19:05 |
| notmyname | gmmaha: I'm having some trouble parsing that question | 19:05 |
| clayg | acoles_: so do you have a failing test in place for the thing you're imaging for inconsistent dir listings? | 19:06 |
| pdardeau | notmyname: yeah, that's the one | 19:08 |
| * notmyname wonders what RAX rookie-o is like without sugar bear | 19:09 | |
| *** aix has quit IRC | 19:15 | |
| mwheckmann | notmyname: clayg: the workaround for my broken rings worked beautifully. Thanks again for your help on that! | 19:21 |
| notmyname | IIRC I ran away to a meeting and clayg actually helped ;-) | 19:22 |
| mwheckmann | sure, but you still stepped up to help while you could. | 19:23 |
| *** mahatic has quit IRC | 19:25 | |
| clayg | acoles_: oh, it looks like maybe https://gist.github.com/alistairncoles/7fa16e567cbacd12a7a6 is the failing test I was looking for | 19:26 |
| clayg | notmyname: the starred list is huge already - and so many are -1'd :\ | 19:27 |
| notmyname | clayg: what's huge is the list of "stuff that just needs one more +2" | 19:29 |
| notmyname | yeah, the issue with the starred list is if *I* want to track something, you end up seeing it too ;-) | 19:31 |
| notmyname | eg the stable/kilo one, the testr one, and the pyeclib one | 19:31 |
| *** openstackstatus has joined #openstack-swift | 19:36 | |
| *** ChanServ sets mode: +v openstackstatus | 19:36 | |
| -openstackstatus- NOTICE: Gerrit will be offline at 20:00-20:15 UTC today (starting 20 minutes from now) for scheduled project rename maintenance | 19:41 | |
| gmmaha | notmyname: i meant to ask whether we are shifting focus to make the openstackclient the primary client for swift or we support both for long | 19:41 |
| notmyname | gmmaha: ah ok | 19:41 |
| notmyname | gmmaha: we tried to shift focus. and we just ended up not focusing on anything | 19:41 |
| gmmaha | :D | 19:41 |
| notmyname | gmmaha: so the current plan is that we will still improve python-swiftclient | 19:42 |
| openstackgerrit | Merged openstack/swift: On py3, use dnspython3 dependency, not dnspython https://review.openstack.org/217423 | 19:42 |
| notmyname | gmmaha: and other work will happen in openstackclient (separately from work on swiftclient) | 19:42 |
| notmyname | gmmaha: with the expectation that both will likely be a thing used for quite some time | 19:42 |
| gmmaha | notmyname: aah cool.. | 19:42 |
| gmmaha | yeah the fade over being long is what i thought | 19:43 |
| gmmaha | thanks.. i was going to start writing some test coverage on the swiftclient and wanted to check if at all i should be doing that | 19:43 |
| clayg | oh nice - i forgot about needs final approval - it's been so long since I made it past priority reviews I forgot what else was down there! | 19:43 |
| notmyname | gmmaha: awesome! yes, you should definitely do that! :-) | 19:44 |
| gmmaha | notmyname: :) for sure.. thanks | 19:44 |
| clayg | notmyname: bah - most of them are py3 | 19:44 |
| notmyname | briancline: if you could take patch 241571 and run your rings through it, that would be great! | 19:44 |
| patchbot | notmyname: https://review.openstack.org/#/c/241571/ - Put part-replicas where they go | 19:44 |
| notmyname | hurricanerix: redbo: ^ same for the RAX rings | 19:45 |
| notmyname | kota_: ^ and NTT | 19:45 |
| notmyname | donagh: ^ and HP | 19:45 |
| clayg | notmyname: can will single core +A py3 changes? I figure if cschwede or acoles_ or torgomatic_ or kota_ are going to take the time to look them over we may as well just get 'em in? | 19:45 |
| clayg | ... maybe that's insane | 19:45 |
| notmyname | clayg: the rule's always been "use your judgement" :-) | 19:45 |
| hurricanerix | notmyname: lol, too bad there isn't a @everyone (or maybe there is, i only know enough irc to get by) =) | 19:46 |
| notmyname | IMO small py3 stuff is probably just fine for a single core to +A | 19:46 |
| notmyname | hurricanerix: yeah, what do you think this is? slack?! | 19:46 |
| hurricanerix | hehe | 19:46 |
| clayg | notmyname: yeah but everyone's judgement normally says that the risk of merging stomething stupid isn't worth the cost of just waiting for another core to see it | 19:46 |
| notmyname | clayg: which is why that's a safe rule ;-) | 19:46 |
| * notmyname wonders how many people in here end up using slack too | 19:47 | |
| clayg | ... but with stuff like doc typo fixes we've had to socialize the idea that the cost is higher than you might think when the risk is so low | 19:47 |
| *** daemontool has quit IRC | 19:47 | |
| clayg | notmyname: I aggree we should single core +A "simple" py3 fixups - I'd like to bring that up next week | 19:48 |
| nottrobin | hurricanerix: everyone may not be official, but my IRC is setup to ping me when someone uses it =) | 19:49 |
| notmyname | clayg: yeah, but it's ok if you start doing that now and then ask for "permission" later | 19:49 |
| clayg | notmyname: my avoidance of py3 changes is purposeful - hence wanting people to merge them faster to get them out of my way | 19:50 |
| *** peterlisak13 has joined #openstack-swift | 19:52 | |
| *** peterlisak13 has quit IRC | 19:57 | |
| pdardeau | notmyname: on the topic of py3, is there a recommended approach for individual developers to easily shift across py2/py3? (without causing too many headaches) | 19:59 |
| mwheckmann | I think I found another small bug in the ring builder: For some reason, I can't use "--id" as a search key for set_info when the id value I want to use is '0'. I works for id's larger than 0. | 20:00 |
| mwheckmann | I had to use another search key as a workaround. | 20:00 |
| notmyname | pdardeau: not that I know | 20:00 |
| mwheckmann | This is with 2.5.0. | 20:00 |
| notmyname | mwheckmann: that sounds familiar. as if I've seen a patch recently for that | 20:02 |
| mwheckmann | ok. Anyway, it's easily worked around. | 20:02 |
| mwheckmann | I also noticed that the use of '--id' is not documented in the help output for the ring builder. | 20:03 |
| -openstackstatus- NOTICE: Gerrit is offline until 20:15 UTC today for scheduled project rename maintenance | 20:04 | |
| *** ChanServ changes topic to "Gerrit is offline until 20:15 UTC today for scheduled project rename maintenance" | 20:04 | |
| notmyname | whelp. guess it's time for lunch then | 20:04 |
| *** diazjf has joined #openstack-swift | 20:06 | |
| clayg | mwheckmann: you get "Invalid <search-value>: '--id'" | 20:07 |
| mwheckmann | yup. | 20:07 |
| *** jroll is now known as tjroll | 20:07 | |
| clayg | mwheckmann: could I trick you into filing a bug that has something like that in the title on https://bugs.launchpad.net/swift/+filebug | 20:08 |
| mwheckmann | sure. | 20:09 |
| clayg | the exact syntax of the commands you ran with the observed output and the expected output would be super ideal | 20:09 |
| clayg | YES! | 20:09 |
| clayg | thanks! | 20:09 |
| mwheckmann | NP | 20:09 |
| mwheckmann | clayg: https://bugs.launchpad.net/swift/+bug/1513955 | 20:19 |
| openstack | Launchpad bug 1513955 in OpenStack Object Storage (swift) ""swift-ring-builder set_info" command doesn't recognize 0 as a search value when "--id" is used as a key" [Undecided,New] | 20:19 |
| *** janonymous has quit IRC | 20:19 | |
| jlhinson | Has anyone else been having trouble with the unit tests? | 20:25 |
| jlhinson | http://paste.openstack.org/show/478214/ | 20:25 |
| jlhinson | upgrading six to 1.10.0 solved my problem | 20:25 |
| gmmaha | jlhinson: think a colleague of mine saw that error as well.. He had to do a upgrade install ( think six package) to get past it | 20:27 |
| jlhinson | I wasn't sure if we needed to update requirements.txt, or if it was just my local copy of six that was bad | 20:27 |
| *** tongli has quit IRC | 20:33 | |
| *** ChanServ changes topic to "Review Dashboard: https://goo.gl/iLt6Ld | Summit etherpads: https://wiki.openstack.org/wiki/Design_Summit/Mitaka/Etherpads#Swift | Summit schedule: http://mitakadesignsummit.sched.org/type/swift | Logs: http://eavesdrop.openstack.org/irclogs/%23openstack-swift/" | 20:37 | |
| *** daemontool has joined #openstack-swift | 20:52 | |
| jlhinson | I'm also unable to parse requirements.txt now... | 20:53 |
| *** lpabon has quit IRC | 20:55 | |
| clayg | mwheckmann: that's awesome! | 20:57 |
| notmyname | jlhinson: haven't seen that yet. I've been messing with pip install -r requirements to do the pyeclib updates. seems to be working for me | 20:58 |
| notmyname | jlhinson: this is on master? | 20:58 |
| jlhinson | yep | 20:59 |
| jlhinson | jrichli ran into the same problem | 20:59 |
| notmyname | finishing a test run now. if it goes well, I'll look at that in few minutes | 21:00 |
| jlhinson | sounds good | 21:00 |
| *** mzhou_ has joined #openstack-swift | 21:01 | |
| *** openstackgerrit has quit IRC | 21:01 | |
| *** openstackgerrit has joined #openstack-swift | 21:01 | |
| briancline | notmyname: having a look at the placement patch. do you just want me to try a rebalance? | 21:02 |
| clayg | briancline: that'd be something! | 21:03 |
| *** mzhou has quit IRC | 21:03 | |
| *** mzhou_ is now known as mzhou | 21:03 | |
| clayg | briancline: if you think the results are "better" or "no worse" or "worse" you should put that on there too | 21:03 |
| clayg | briancline: if you got *really* crazy you could try moving some devices/weights around and see if the results make sense - or even compare the results of the same operations on master. | 21:04 |
| clayg | briancline: experimental observations + analysis (from operataional perspective/experience) would be at least as helpful as going over the code changes line by line and offering nits. | 21:05 |
| briancline | clayg: ok, yeah, we're about to make ring changes and have recently been generating stats about what the movement looks like each time. can run the same scenario for the next one using this version to see how much those stats change | 21:06 |
| *** NM has quit IRC | 21:07 | |
| clayg | @everyone: *but* - for those who feel more comfortable going over the code/test chnages and offering nits - those are welcome too! | 21:07 |
| notmyname | briancline: thanks :-) | 21:08 |
| * clayg considering reverse listings, concurrent gets, and more fast-POST for this afternoon | 21:09 | |
| *** petertr7 is now known as petertr7_away | 21:14 | |
| *** petertr7_away is now known as petertr7 | 21:15 | |
| *** mzhou has quit IRC | 21:19 | |
| *** mzhou has joined #openstack-swift | 21:20 | |
| notmyname | jlhinson: I'm not seeing anything | 21:36 |
| notmyname | but I may have inadvertently gotten past it with an upgrade at some point | 21:36 |
| *** subscope has quit IRC | 21:39 | |
| jlhinson | I was wrong about the version of six causing a problem | 21:39 |
| jlhinson | but I am still having an issue parsing requirements.txt | 21:39 |
| clayg | sudo pip install --upgrade pip | 21:40 |
| clayg | the new bugs are always better than the old bugs | 21:40 |
| notmyname | anyone else getting "WARNING: default min_part_hours may not match the value in the lost builder." printed on unit test runs? | 21:41 |
| clayg | notmyname: i see that sometimes! | 21:41 |
| jlhinson | notmyname: I've seen that before | 21:41 |
| pdardeau | notmyname: yes | 21:42 |
| jlhinson | clay: lol, trying that now | 21:42 |
| clayg | notmyname: you gunna track it down and square it for us? | 21:42 |
| pdardeau | notmyname: that's because of calls to stderr.write | 21:42 |
| clayg | pdardeau: better mock that shit | 21:42 |
| notmyname | clayg: yeah, maybe after I figure out this other test failure I'm seeing ;-) | 21:42 |
| pdardeau | i wrote up a bug on it last week | 21:42 |
| notmyname | see? pdardeau is already all over it :-) | 21:42 |
| clayg | *my* bad | 21:43 |
| notmyname | feels nice being back in a terminal and code editor interface instead of presentation software and emails ;-) | 21:43 |
| clayg | lol @ PTL! | 21:43 |
| notmyname | "This is code! I *know* this" | 21:44 |
| notmyname | https://www.youtube.com/watch?v=dFUlAQZB9Ng | 21:45 |
| clayg | error in setup command: 'install_requires' must be a string or list of strings | 21:45 |
| clayg | ^ this is the failing? | 21:45 |
| *** jerrygb has quit IRC | 21:46 | |
| *** jerrygb has joined #openstack-swift | 21:47 | |
| haypo | [5~[5~[6~ | 21:49 |
| haypo | 2. | 21:49 |
| haypo | oops, sorry | 21:49 |
| clayg | haypo: happens to me all the time | 21:49 |
| timburke | notmyname: i think there may be a reason you run unopposed... | 21:49 |
| notmyname | lol | 21:49 |
| clayg | timburke: it's because he's so funny and charming | 21:49 |
| notmyname | clayg: it's because I make you nice tshirts | 21:50 |
| timburke | clayg: well that's *also* true. but *you're* funny and charming, too, so watch out! | 21:50 |
| haypo | clayg: i'm using a PLC adapter for my LAN and it became very common that I loose the internet connection :-( | 21:50 |
| haypo | clayg: i was trying to kill the SSH connection when IRC just came back /o\ fail | 21:51 |
| * clayg thinks back to one of the working sessions when notmyname (rightfuly frustrated by the devolution of some topic to bike sheading) went on a tear about us needed to come back to the problem at hand and make progress ( and in the middle he slipped in something about be tired cranky and hungry or something) | 21:51 | |
| haypo | "A PLC connection has many advantages to a wireless connection, however the quality of the connection will still depend on the quality of the domestic electrical system. Improper wiring and circuit breakers in between the connected cables can negatively affect the performance, and can cause connection interruptions." | 21:51 |
| notmyname | that was fun. good times | 21:51 |
| *** jerrygb has quit IRC | 21:51 | |
| clayg | good times | 21:52 |
| clayg | this unable to parse requirements.txt - NOT good times - wtf is going on? | 21:52 |
| clayg | JESUS! I'M ON PIP *SEVEN* | 21:52 |
| clayg | what are you guys doing?! | 21:52 |
| clayg | you know what - backwards compatibility is hard - let's just *ALWAYS* increment the major version number | 21:53 |
| clayg | *washes hands* | 21:53 |
| *** peterlisak13 has joined #openstack-swift | 21:53 | |
| haypo | clayg: i really like the transparent binary wheel cache introduced in pip 6, tox is much faster with that | 21:53 |
| clayg | haypo: sounds fancy | 21:53 |
| clayg | i probably don't like it | 21:53 |
| notmyname | timburke: clayg is being funny and charming again ;-) | 21:54 |
| haypo | clayg: my pbr requirement was merged into swift, but i don't see how it's related to pip in your case (i guess that you are talking about something different?) | 21:54 |
| clayg | funny is probably debable - but it's definately not charming - I like to put it all out there so others can vent through me vicariously and only *one* of us has to come across as an a#$%^&ole | 21:55 |
| clayg | Expected ',' or end-of-list in dnspython>=1.12.0;python_version<'3.0' at ;python_version<'3.0' | 21:55 |
| clayg | ^ doubt it's *unrelated* | 21:56 |
| *** alejandrito has quit IRC | 21:56 | |
| timburke | clayg: yep, need to upgrade one or more of pip, setuptools, and pbr | 21:56 |
| clayg | looks to be my favorite :) | 21:56 |
| notmyname | haypo: what pbr patch did you have? | 21:56 |
| openstackgerrit | Merged openstack/swift: bug/1270114. Missing documentation for Bulk Middleware. https://review.openstack.org/232688 | 21:57 |
| *** mwheckmann has quit IRC | 21:57 | |
| clayg | I'm going to add `pip install --upgrade setuptools pbr pip` to my .bashrc | 21:57 |
| haypo | notmyname: i'm trying to always get the latest versions of pbr, tox and pip :) for pbr, it's simply not installed on my system | 21:57 |
| clayg | oh i'm sorry - were trying to get some work done!? | 21:57 |
| *** peterlisak13 has quit IRC | 21:58 | |
| haypo | notmyname: it's only installed inside tox venvs | 21:58 |
| *** amoturi has quit IRC | 21:58 | |
| haypo | clayg: i can try to help you, but i don't understand your issue. what are you doing, what is the error message? | 21:58 |
| haypo | clayg: acoles_ had issues because he had an old version of setuptools and/or pbr, sorry i don't recall and i was unable to reproduce his issue :-( | 21:59 |
| haypo | right now on my fedora 23, i have setuptools (18.2), tox 1.8, pip 7.1.2, (pbr is not installed) | 21:59 |
| clayg | haypo: `pip install --upgrade pbr` was the fix for me | 22:00 |
| notmyname | I'm listening to someone give an intro to swift talk (yes, while I'm looking at this test failure). first, it's probably a bad idea to talk about the CAP theorem in an intro swift talk. second, instead of "eventually consistent" I'm going to start saying "highly available" | 22:00 |
| haypo | clayg: i guess that you are not using tox, but setup.py directly? | 22:00 |
| haypo | clayg: because tox should now take care of the pbr version, i added "pbr>=1.6" in requirements.txt | 22:00 |
| clayg | what does tox have to do with anything? | 22:01 |
| haypo | clayg: usually, i use tox to run tests, i never install anything on my system | 22:01 |
| clayg | THEN HOW DO YOU KNOW IT WORKS | 22:02 |
| clayg | lol | 22:02 |
| notmyname | clayg: we've tried nothing, and it works every time | 22:03 |
| clayg | look - i'm up and running again - in a few minutes someone will try to test out branch that's been rebased on master and thye'll have the same problem I did - we'll all say - "oh yeah someone made something better, everyone is broken until everything is upgrade" and they'll go back to work - it's fine | 22:03 |
| haypo | clayg: i expected that setup.py and tox have the same behaviour, but it looks like setup.py is unable to upgrade itself pbr :-/ | 22:04 |
| clayg | then later we'll all have to *package* this mess and curse again for a little while and once we get it working we'll forget that pain too and move on | 22:04 |
| clayg | it's great | 22:04 |
| notmyname | ah. that's why I'm not seeing these problems. I didn't have the very latest code! | 22:04 |
| clayg | notmyname: "oh yeah someone made something better, everyone is broken until everything is upgrade" | 22:04 |
| *** petertr7 is now known as petertr7_away | 22:05 | |
| notmyname | so wait. why did we add pbr to requirements.txt?!? we spent a *lot* of effort on making sure it was a build-only dependency | 22:05 |
| clayg | notmyname: who cares | 22:05 |
| clayg | notmyname: it's over | 22:05 |
| clayg | notmyname: we lost | 22:05 |
| clayg | notmyname: all of python lost | 22:06 |
| clayg | notmyname: just package pbr every other week, and pip install -e . | 22:06 |
| haypo | notmyname: "why did we add pbr to requirements.txt?!?" i'm not sure that pbr is required at runtime, but i added pbr>=1.6 to ensure that we have at least pbr 1.0 to support environment markers | 22:08 |
| notmyname | haypo: wasn't that handled in the setup.py stanza? | 22:08 |
| notmyname | the cap on the version there was removed a long time ago so it would get the latest | 22:09 |
| *** jerrygb has joined #openstack-swift | 22:09 | |
| haypo | notmyname: if you run "tox -e update ../swift", setup.py is updated to add "pbr>=1.8" | 22:10 |
| haypo | notmyname: but acoles_ got issues with this change, so i tried to remove this requirements to fix his usecase | 22:10 |
| clayg | notmyname: you can try to dig into it if you want, but eventually you'll find out no one pushing all this pain out onto the entire python ecosystem *cares* - you didn't blow away your entire deployment and install everything fresh from the very latest version?! wow. | 22:10 |
| *** jerrygb_ has joined #openstack-swift | 22:10 | |
| notmyname | and it looks like probe tests in the community cluster failed too | 22:11 |
| clayg | LOL | 22:11 |
| notmyname | haypo: what sort of meta-deployment magic is this? we're supposed to run deployment code to generate/edit deployment files? | 22:11 |
| notmyname | what happened to explicit is better than implicit? | 22:11 |
| haypo | if i recall correctly, putting a requirements in setup.py is not enough. but sorry, i don't recall all minor details of python packaging. for me, it's just a big mess | 22:12 |
| * clayg remembers when all we hade was setuptools0.6c11 and we *made due* | 22:13 | |
| *** jerrygb has quit IRC | 22:14 | |
| haypo | notmyname: sorry, i don't understand your point. my goal is to install dnspython3 on python 3 and dnspython on python 2. the best way to do that (for many reasons) is to use environment markers | 22:14 |
| haypo | notmyname: so somehow, we need to ensure that everything used to package and install swift support environment markers | 22:14 |
| haypo | notmyname: maybe pbr>=1.6 in requirements.txt is not the best way to do that, but i recall vaguely some corner cases with pbr not explicitly required in requirements.txt | 22:15 |
| haypo | my expert for pbr is lifeless who understand these things much better than me :) | 22:15 |
| clayg | :) | 22:16 |
| haypo | i hate pbr, but i don't know anything better, so... | 22:16 |
| clayg | once openstack took over pip and setuptools didn't the *need* for pbr to work around upstream issues decrease dramatically? why is it still a thing? | 22:18 |
| haypo | clayg: python packaging evolves very slowy, i mean it takes years to fix issues | 22:20 |
| haypo | clayg: but IMHO pip made huge progress last months, it's really impressive | 22:20 |
| notmyname | new version of something complains about /tmp being writable | 22:20 |
| *** vinsh has quit IRC | 22:21 | |
| *** vinsh has joined #openstack-swift | 22:22 | |
| blmartin | ring balancing question: if I add many devices to a ring and rebalance, is it expected and natural that my partitions may not be equally distributed across all devices? | 22:23 |
| haypo | clayg, notmyname : so do you think that we need to change something for pbr? or is it ok? | 22:23 |
| blmartin | or is there a doc I need to read | 22:23 |
| notmyname | haypo: dealing with another patch, I ended up upgrading my saio before I got this patch, so it's hard for me to explicitly test the upgrade path right now | 22:26 |
| *** vinsh has quit IRC | 22:26 | |
| *** August1914 has joined #openstack-swift | 22:26 | |
| timburke | fwiw on the env marker stuff, there was a bug for something similar in swiftclient... https://bugs.launchpad.net/python-swiftclient/+bug/1476585 | 22:27 |
| openstack | Launchpad bug 1476585 in python-swiftclient "Error when installing python-swiftclient release version 2.5.0 : error in setup command: 'tests_require' must be a string or list of strings containing valid project/version requirement specifiers" [Medium,Fix released] - Assigned to Christian Schwede (cschwede) | 22:27 |
| timburke | (not that we can use the same solution - remove the need for env markers - here, though) | 22:27 |
| *** rjaiswal has quit IRC | 22:27 | |
| haypo | notmyname: ok. don't hesitate to come back to me if you have issues. according to acoles_ & clayg, upgrading pbr fixes the issue | 22:27 |
| notmyname | haypo: I think my frustration is that there is an issue. not that it's fixable | 22:28 |
| notmyname | timburke: yeah, i thought we had environ markers somewhere else | 22:29 |
| openstackgerrit | Bill Huber proposed openstack/swift: ObjectControllers return application errors as 499 on bad read https://review.openstack.org/236007 | 22:29 |
| wbhuber | tdasilva: your comments tended to ^^ | 22:29 |
| haypo | timburke: it looks like the fix was to... remove environment markers :-) | 22:30 |
| *** dustins has quit IRC | 22:33 | |
| torgomatic_ | upgrading pbr is never fast for me | 22:34 |
| torgomatic_ | it always drags in newer dependencies, and then stuff breaks and I have to figure out why, and I sometimes just end up rebuilding my SAIO VM | 22:35 |
| clayg | torgomatic_: you didn't blow away your entire deployment and install everything fresh from the very latest version?! wow. | 22:36 |
| torgomatic_ | clayg: and I'm not even using docker. I think that makes me a dinosaur. | 22:37 |
| clayg | RAWARRRR | 22:37 |
| * clayg has choosen reverse listings | 22:39 | |
| clayg | step #0 create a container with 25K objects | 22:40 |
| *** blmartin has quit IRC | 22:50 | |
| *** jlhinson has quit IRC | 22:55 | |
| *** nadeem has quit IRC | 23:02 | |
| notmyname | torgomatic_: we've got simplejson in requirements.txt. can that be removed now? | 23:04 |
| *** haypo has left #openstack-swift | 23:04 | |
| torgomatic_ | notmyname: nope; I still get a bunch of hits from `git grep simplejson` | 23:04 |
| notmyname | torgomatic_: oh you haven't fixed all those yet? ;-) | 23:04 |
| notmyname | ok yeah. I see them too. should have looked there first :-) | 23:05 |
| torgomatic_ | it's subtle; each change requires some careful thought | 23:05 |
| notmyname | just teasing :-) | 23:05 |
| timburke | notmyname: torgomatic_: it's that what zackmdavis_'s patch 240596 is for? | 23:06 |
| patchbot | timburke: https://review.openstack.org/#/c/240596/ - remove remaining simplejson uses | 23:06 |
| *** diazjf has quit IRC | 23:09 | |
| *** hrou has quit IRC | 23:12 | |
| *** wbhuber has quit IRC | 23:18 | |
| *** pdardeau has quit IRC | 23:33 | |
| openstackgerrit | Merged openstack/swift: Re-organise ssync tests https://review.openstack.org/220198 | 23:43 |
| *** jerrygb_ has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!