Thursday, 2017-07-20

*** tonanhngo has joined #openstack-swift00:06
*** tonanhngo has quit IRC00:10
*** tonanhngo has joined #openstack-swift00:15
openstackgerritTim Burke proposed openstack/swift master: functests: Only convert headers that should be ints to be ints  https://review.openstack.org/48538100:18
*** tonanhngo has quit IRC00:20
*** lucasxu has joined #openstack-swift00:24
*** frank_young has joined #openstack-swift00:26
*** frank_young has quit IRC00:31
*** lucasxu has quit IRC00:39
*** frank_young has joined #openstack-swift00:40
*** frank_young has quit IRC00:40
*** lucasxu has joined #openstack-swift00:43
*** lucasxu has quit IRC00:53
*** lucasxu has joined #openstack-swift00:54
*** klrmn has quit IRC01:03
*** frank_young has joined #openstack-swift01:15
*** kiennt has joined #openstack-swift01:25
*** JimCheung has quit IRC01:30
*** lucasxu has quit IRC01:31
*** JimCheung has joined #openstack-swift01:35
*** vint_bra has joined #openstack-swift01:52
*** JimCheun_ has joined #openstack-swift02:05
*** JimCheung has quit IRC02:05
*** vint_bra has quit IRC02:12
*** bkopilov has quit IRC02:16
*** klrmn has joined #openstack-swift02:25
*** caiobrentano_ has joined #openstack-swift02:32
*** caiobrentano has quit IRC02:32
*** JimCheun_ has quit IRC02:52
*** ukaynar has joined #openstack-swift02:58
*** Sukhdev_ has joined #openstack-swift03:00
*** Sukhdev has quit IRC03:01
*** JimCheung has joined #openstack-swift03:15
*** gkadam has joined #openstack-swift03:36
*** JimCheung has quit IRC03:43
*** SkyRocknRoll has joined #openstack-swift03:47
*** caiobrentano_ has quit IRC03:58
*** JimCheung has joined #openstack-swift03:59
*** JimCheung has quit IRC04:03
*** klrmn has quit IRC04:06
*** JimCheung has joined #openstack-swift04:15
*** JimCheung has quit IRC04:19
*** chlong_ has joined #openstack-swift04:20
*** SkyRocknRoll has quit IRC04:20
*** frank_young has quit IRC04:21
*** gyee has quit IRC04:21
*** frank_young has joined #openstack-swift04:31
*** rcernin has joined #openstack-swift04:34
*** frank_young has quit IRC04:36
*** JimCheung has joined #openstack-swift04:54
*** JimCheung has quit IRC04:58
*** JimCheung has joined #openstack-swift04:58
*** JimCheung has quit IRC05:05
*** JimCheung has joined #openstack-swift05:06
*** JimCheung has quit IRC05:10
*** ukaynar has quit IRC05:23
*** chlong_ has quit IRC05:26
*** frank_young has joined #openstack-swift05:32
*** rcernin has quit IRC05:36
*** frank_young has quit IRC05:37
*** kiennt has quit IRC05:44
*** cshastri has joined #openstack-swift05:51
*** frank_young has joined #openstack-swift05:53
*** Sukhdev_ has quit IRC05:56
*** psachin has joined #openstack-swift05:57
*** frank_young has quit IRC05:58
*** kiennt has joined #openstack-swift05:58
*** rcernin has joined #openstack-swift05:59
*** frank_young has joined #openstack-swift06:14
*** winggundamth has joined #openstack-swift06:49
openstackgerritKota Tsuyuzaki proposed openstack/swift master: Ring rebalance respects co-builders' last_part_moves  https://review.openstack.org/47700006:56
openstackgerritKota Tsuyuzaki proposed openstack/swift master: WIP: Consideration for avoiding RingBuilder change as possible  https://review.openstack.org/48546906:56
kota_whooa?06:57
kota_i updated patch 477000 too?06:57
patchbothttps://review.openstack.org/#/c/477000/ - swift - Ring rebalance respects co-builders' last_part_moves06:57
kota_ugh! It seems i rebased it to the master unintentionally07:00
kota_sorry acoles07:02
*** itlinux has quit IRC07:04
xrbacoles: thx!07:12
openstackgerritjunbo.li proposed openstack/swift master: Update and optimize documentation links  https://review.openstack.org/48547407:14
*** tesseract has joined #openstack-swift07:19
acolesgood morning07:26
*** ChubYann has quit IRC07:29
*** cschwede_ has joined #openstack-swift07:33
kota_acoles: good morning07:33
acoleskota_: o/07:33
kota_acoles: i just pushed my comments07:34
acoleskota_: ok, looking...07:34
kota_assuming I will not able to work too much on that today due to that I'm joining OpenStackDaysTokyo (summit like conference for OpenStack in Japan)07:35
kota_acoles: basically your work looks awesome to me07:35
kota_acoles: one point, I still want to keep current code on RingBuilder is just I'm messing around there in another patch. Though it needs more work to catch up current rebalance algorithms07:37
acoleskota_: I just looked at https://review.openstack.org/#/c/485469 - it works, but it would mean that there was a different behaviour for _can_part_move in the subclass w.r.t. the superclass i.e. one can be called before rebalance, the other cannot. Do you have a concern that the change int he subclass may break something? or just prefer to not have a change in RingBuilder? I think it just makes the method safe to call07:38
acoles when on master it may blow up if called before rebalance.07:38
patchbotpatch 485469 - swift - WIP: Consideration for avoiding RingBuilder change...07:38
acoleskota_: thanks for working on this I hope you openstack day is good07:38
kota_ah, i look at the behavior change found in your test case briefly07:40
kota_basically, your test looks to hook the parent class call but my change at patch 485496 skips the parent call depends on the rebalance stat (i.e. non-rebalanced ring won't be called) so...07:41
patchbotNo data found for patch 48549607:41
kota_patch 48546907:42
patchbothttps://review.openstack.org/#/c/485469/ - swift - WIP: Consideration for avoiding RingBuilder change...07:42
kota_it seems safe???07:42
kota_ah07:42
kota_wait, I probably miss read your comments at irc07:43
kota_ah... super class CANNOT be called but child CAN be called.07:44
kota_hmmm07:44
*** oshritf has joined #openstack-swift07:45
acoleskota_: yes, it is subtle but with your change 485469 it is ok to call the subclass method at any time but the superclass method can only be called after rebalance() inits the array07:46
kota_acoles: fair point07:46
kota_hmmm, i don't have clear answer, sorry.07:47
acolesobviously the subclass behaviour is also different because it calls cobuilders. Is there any time we would *want* the superclass method to blow up if called too early? I should take a look at you other patch 47922307:49
patchbothttps://review.openstack.org/#/c/479223/ - swift - Fix converging unbalance state at ring builder reb...07:49
kota_n.b. not only directly related to your patch but... it tries to change rebalance behaivor07:50
kota_s/only//07:51
kota_perhaps, it may be orthogornal07:53
acoles479223 doesn't seem to touch _can_part_move07:53
*** geaaru has joined #openstack-swift07:56
*** caiobrentano has joined #openstack-swift07:59
acoleskota_: I don't think my patch changes the result of _can_part_move other than avoiding a TypeError if called before rebalance - are you concerned that it *does* change the result?08:01
kota_acoles: honestly not so much concerned but a little when starting to review your patch.08:02
charzkota_: just one question for swift3, when swift3 issue an account/container/SLO?multipart-manifest?delete request with internal client, does SLO able to handle that?08:02
*** caiobrentano has quit IRC08:03
kota_charz: when using multi-upload?08:04
*** clarkb has quit IRC08:04
kota_woops? swift3 with internal client?08:05
*** clarkb has joined #openstack-swift08:05
charzkota_: https://github.com/openstack/swift3/blob/master/swift3/controllers/multi_delete.py#L104-L10808:10
kota_charz: basically swift3 works just issuing the multipart-delete expecting slo middleware to delete all segments08:11
charzWhen client try to use POST (multiple object delete api) to delete a SLO object, it seems like I don't see the mulitpart-delete request in logs08:12
kota_charz: not sure it works correctly in *INTERNAL CLIENT* pipeline. I don't think we had the test in swift3 repo08:12
charzkota_: okay, it works for normal object.08:13
charzkota_: I saw the multipart-delete request sent by swift3, but I don't see any proxy/object logs for that.08:15
charzkota_: let me create a bug and put some logs, hope that can help.08:16
kota_er -08:18
kota_charz: let me know if the delete request occurred but not loggedl right?08:19
kota_s/loggdl/logged/08:19
charzkota_: correct, I saw proxy into https://github.com/openstack/swift3/blob/master/swift3/controllers/multi_delete.py#L106 , but I don't see the request likes `DELETE account/conatiner/object?multipart-manifest?delete` to delete the SLO.08:22
charzkota_: And I can see the SLO(manifest/segment) still in there.08:23
kota_oh, so it sounds the request not occurred?08:23
charzWhen I removed remove the `query=query` in https://github.com/openstack/swift3/blob/master/swift3/controllers/multi_delete.py#L106,08:24
charzThe manifest just deleted, but still have segments, it makes sense.08:24
charzBut I can't understand why add query (`multipart-manifest?delete`) won't work.08:25
kota_so it sounds, it gets wrong query for slo?08:25
* kota_ have to look at recent slo changes08:25
charzkota_: maybe, but when I check the response code in the delete request, I got `200 ok` response form internal client.08:27
charz:q!08:28
kota_charz: hmm... it seems the test for delete completed multi upload object was missing :/08:35
kota_https://github.com/openstack/swift3/blob/master/swift3/test/unit/test_multi_upload.py <- looking at this08:36
kota_perhaps, adding the test there is a good starting point to know what happans.08:36
charzkota_: agreed! just created a bug https://bugs.launchpad.net/swift3/+bug/170540908:41
openstackLaunchpad bug 1705409 in Swift3 "Delete Multiple Objects API won't able to delete multipart objects" [Undecided,New]08:41
*** mathiasb has quit IRC08:41
kota_charz: thx!08:41
kota_woops? my accidental rebasing broke something on patch 477000?08:42
patchbothttps://review.openstack.org/#/c/477000/ - swift - Ring rebalance respects co-builders' last_part_moves08:42
kota_the failure test looks not related to anything... :/08:42
*** mathiasb has joined #openstack-swift08:42
acoleskota_: I'm looking08:53
kota_acoles: thx08:53
openstackgerritKota Tsuyuzaki proposed openstack/swift master: WIP: Consideration for avoiding RingBuilder change as possible  https://review.openstack.org/48546908:54
kota_the following my patch doesn't get failed at the py27 tox (i just had a fault on pep8, though) so it looks periodic failure or something?08:55
acoleskota_: I can't reproduce so recheck09:01
* kota_ is being offline for openstack days tokyo booth crawling and an after event. will up to look at patch 477000 again with a fresh head tommorow morning09:02
patchbothttps://review.openstack.org/#/c/477000/ - swift - Ring rebalance respects co-builders' last_part_moves09:02
kota_acoles: i had issued the recheck already :|09:02
kota_acoles: but thanks for checking on your env too09:02
acoleskota_: have a good night09:05
*** hoonetorg has joined #openstack-swift09:10
*** oshritf has quit IRC09:11
*** winggundamth has quit IRC09:15
*** oshritf has joined #openstack-swift09:25
*** cshastri has quit IRC09:34
*** oshritf has quit IRC09:34
*** oshritf has joined #openstack-swift09:41
*** mvk has quit IRC09:43
*** cshastri has joined #openstack-swift09:47
openstackgerritbhavani proposed openstack/swift master: Stop using deprecated 'message' attribute in Exception  https://review.openstack.org/48330010:04
*** mvk has joined #openstack-swift10:10
*** kiennt has quit IRC10:30
*** frank_young has quit IRC10:32
*** cshastri has quit IRC10:33
*** frank_young has joined #openstack-swift10:42
*** frank_young has quit IRC10:48
*** hoonetorg has quit IRC10:52
*** oshritf has quit IRC11:02
*** frank_young has joined #openstack-swift11:03
*** oshritf has joined #openstack-swift11:05
*** oshritf has quit IRC11:12
*** oshritf has joined #openstack-swift11:26
*** hoonetorg has joined #openstack-swift11:29
*** oshritf has quit IRC11:29
*** caiobrentano has joined #openstack-swift11:33
openstackgerritbhavani proposed openstack/swift master: Stop using deprecated 'message' attribute in Exception  https://review.openstack.org/48330011:33
*** oshritf has joined #openstack-swift11:35
openstackgerritAlistair Coles proposed openstack/swift master: Add reconstructor section to deployment guide  https://review.openstack.org/48556911:38
openstackgerritAlistair Coles proposed openstack/swift master: Add more structure to the deployment guide  https://review.openstack.org/48557011:38
*** oshritf has quit IRC11:43
*** caiobrentano has quit IRC11:45
*** oshritf has joined #openstack-swift11:49
*** hoonetorg has quit IRC12:03
*** MVenesio has joined #openstack-swift12:04
*** chlong_ has joined #openstack-swift12:16
*** gkadam has quit IRC12:28
*** caiobrentano has joined #openstack-swift12:32
*** catintheroof has joined #openstack-swift12:36
*** oshritf has quit IRC12:42
*** oshritf has joined #openstack-swift12:47
*** lucasxu has joined #openstack-swift13:19
*** frank_young has quit IRC13:35
*** chlong_ has quit IRC13:40
*** frank_young has joined #openstack-swift14:07
*** frank_young has quit IRC14:12
*** vint_bra has joined #openstack-swift14:21
*** oshritf has quit IRC14:32
*** lucasxu has quit IRC14:39
*** hoonetorg has joined #openstack-swift14:56
*** hoonetorg has quit IRC14:57
*** chlong_ has joined #openstack-swift14:57
*** hoonetorg has joined #openstack-swift14:57
*** caiobrentano has quit IRC15:01
*** caiobrentano has joined #openstack-swift15:02
*** caiobrentano has quit IRC15:04
*** caiobrentano has joined #openstack-swift15:04
*** frank_young has joined #openstack-swift15:05
*** caiobrentano has quit IRC15:08
*** frank_young has quit IRC15:09
*** lucasxu has joined #openstack-swift15:22
*** rcernin has quit IRC15:25
*** caiobrentano has joined #openstack-swift15:26
*** hoonetorg has quit IRC15:27
timburkegood morning15:29
*** caiobrentano_ has joined #openstack-swift15:30
*** caiobrentano has quit IRC15:30
*** hoonetorg has joined #openstack-swift15:35
tdasilvatimburke: good morning15:38
tdasilvatimur: around?15:39
timburkewhooo! isa_l_rs_cauchy's verified clean!15:40
*** psachin has quit IRC15:41
tdasilva\o/15:43
*** Renich has joined #openstack-swift15:44
acolestimburke: good to hear :)15:49
acolestimburke: I guess we'll both reviewing patch 478416 - I already have one diff that I think is necessary, maybe I'll leave it 24 hours and then if we have others I will apply them all and push a new version of the patch (in clayg's absence)?15:59
patchbothttps://review.openstack.org/#/c/478416/ - swift - Add multiple worker processes strategy to reconstr...15:59
acolesnotmyname: cc ^^16:00
openstackgerritAlistair Coles proposed openstack/swift master: Add more structure to the deployment guide  https://review.openstack.org/48557016:02
*** itlinux has joined #openstack-swift16:19
*** chlong_ has quit IRC16:19
*** mvk has quit IRC16:21
*** itlinux has quit IRC16:21
*** itlinux has joined #openstack-swift16:22
*** saint_ has joined #openstack-swift16:23
notmynamegood morning16:24
notmynamethat's great news on the cauchy matrix!16:25
notmynametdasilva: timur isn't around yet, but I'll guess he'll be at his desk momentarily16:25
tdasilvanotmyname: no worries, i started looking at this patch and just left him a quick question there16:26
notmynameack16:26
*** frank_young has joined #openstack-swift16:27
*** frank_young has quit IRC16:31
*** SkyRocknRoll has joined #openstack-swift16:32
*** tesseract has quit IRC16:32
*** rcernin has joined #openstack-swift16:33
*** kallenp has joined #openstack-swift16:38
*** kallenp has left #openstack-swift16:38
*** SkyRocknRoll has quit IRC16:44
*** lucasxu has quit IRC16:51
*** caiobrentano_ has quit IRC16:56
*** tonanhngo has joined #openstack-swift16:57
*** lucasxu has joined #openstack-swift16:58
*** klrmn has joined #openstack-swift17:07
*** SkyRocknRoll has joined #openstack-swift17:13
*** tesseract has joined #openstack-swift17:18
*** SkyRocknRoll has quit IRC17:22
*** chlong_ has joined #openstack-swift17:32
*** SkyRocknRoll has joined #openstack-swift17:32
tdasilvamathiasb, jrichli, timburke: just fyi..the python-barbicanclient patch has merged...18:04
jrichliyay!  thanks tdasilva18:05
timburkeinteresting... i wonder what all that plumbing will look like if there's ever a v2... i suppose the osc_plugin will probably just grow a `version='v' + version` kwarg18:08
notmynametdasilva: nice18:10
tdasilvatimburke: yeah, that's what I assumed...sort of just punted the problem for later18:13
timurtdasilva: hey! I left a pointer to one of your questions. The functional tests I'll try to get to today or tomorrow18:14
timurI have to finish fighting with an internal bug first :(18:14
tdasilvatimur: yeah, saw that, I did some testing and will leave some comments18:14
tdasilvai think we are missing just tests and docs18:14
*** frank_young has joined #openstack-swift18:15
tdasilvado we also want to get slo patch in for this release?18:15
timurtdasilva: yes, I think both need to go in18:15
timurbecause with the first patch the failure mode when you submit a stream > maximum object size is not great18:15
tdasilvaoh cool, yeah, i did not test that big of a file ;)18:16
tdasilvaprobably should have18:16
timurI was debating on squashing the two changes -- let me know if that would be easier to review or not18:16
timurtdasilva: well, you can probably change your cluster limits ;)18:16
tdasilvaI probably should18:17
tdasilvatimur: thinking very quickly (without looking at the slo patch) i'd imagine it makes sense to squash just so that for example documentation is all together, otherwise you would have to break them up18:18
tdasilvaprobably makes sense to write it all together...idk...or just keep the docs in the second patch18:18
*** frank_young has quit IRC18:19
timurtdasilva: I can see that argument. Or I can add documentation to each one or write a separate patch. I don't know what the "swift" way would be. My tendency is to try to make small incremental, self-contained changes, which is how this started18:20
tdasilvatimur: either way is fine, small enough patch, i'm just we can make it work (and yes, i do like the idea of keeping patches small and self-contained)18:21
*** geaaru has quit IRC18:25
*** Sukhdev has joined #openstack-swift18:28
*** hoonetorg has quit IRC18:49
*** SkyRocknRoll has quit IRC18:50
*** ChubYann has joined #openstack-swift18:53
*** frank_young has joined #openstack-swift18:54
*** frank_young has quit IRC18:58
*** itlinux has quit IRC19:38
*** openstack has joined #openstack-swift19:59
timurtdasilva: about the --no-slo flag (or something like it), I've thought about that too. I think it'd be nice to add it, because otherwise there is no way to say "I know what I want and that is not an SLO. I promise my stream is < max object size"20:01
timurThat's a new shell option, though, that's oddly specific to stdin uploads :\20:02
tdasilvatimur: yeah, that's honestly i'm a little worried about automating the --use-slo for *just* from stdin20:02
tdasilvaI'm trying to weight in if it would make sense to keep it consistent with the current behavior that user must pass --use-slo20:03
tdasilvaor maybe change behavior for both20:03
tdasilvado you know what S3 clients (e.g., s3cmd) does?20:04
timurwhat happens if a user tries to upload a file larger than max size today?20:04
tdasilvatimur: I'd expect it would throw an error, but i haven't tried tbh - timburke ???20:04
tdasilvai guess i really need to adjust my ansible-saio to build bigger drives :)20:05
timburkeas i recall, we just error out on the 41320:06
timburketdasilva: you could also just crank down the max_object_size20:07
tdasilvatimburke: duh20:07
timburkeer, max_file_size, rather20:09
*** openstackgerrit has quit IRC20:17
*** frank_young has joined #openstack-swift20:23
*** frank_young has quit IRC20:27
*** Sukhdev has quit IRC20:53
*** lucasxu has quit IRC21:02
*** catintheroof has quit IRC21:10
*** chlong_ has quit IRC21:16
*** itlinux has joined #openstack-swift21:19
*** mvk has joined #openstack-swift21:20
tdasilvaso...when using keystone middleware, if a user has the admin role, does that mean that has R/W permission for other projects rather than his own?21:27
tdasilvaI thought only ResellerAdmin would have that21:27
notmynameyeah, that's the default (ResellerAdmin)21:28
tdasilvawhile running func test again master, i'm seeing errors here: https://github.com/openstack/swift/blob/master/test/functional/test_versioned_writes.py#L54221:29
tdasilvai'm assuming this is an error in my configuration since we are not seeing failures on the gate21:29
tdasilvabut can't figure out where21:29
*** frank_young has joined #openstack-swift21:42
*** frank_young has quit IRC21:46
*** tonanhngo has quit IRC22:19
*** tonanhngo has joined #openstack-swift22:19
*** tonanhngo has joined #openstack-swift22:26
*** vint_bra has quit IRC22:26
*** tonanhngo has quit IRC22:30
*** tesseract has quit IRC22:37
*** openstackgerrit has joined #openstack-swift22:48
openstackgerritJohn Dickinson proposed openstack/swift master: 2.15.0 authors and changelog update  https://review.openstack.org/48583922:48
*** frank_young has joined #openstack-swift23:00
*** frank_young has quit IRC23:05
*** Sukhdev has joined #openstack-swift23:15
*** Sukhdev has quit IRC23:15
*** Sukhdev has joined #openstack-swift23:16
*** Renich has quit IRC23:20
*** Sukhdev has quit IRC23:48
*** frank_young has joined #openstack-swift23:50
openstackgerritJohn Dickinson proposed openstack/python-swiftclient master: 3.4.0 authors/changelog update  https://review.openstack.org/48585923:50
*** frank_young has quit IRC23:54

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!