Tuesday, 2016-11-15

*** Trixboxer has joined #openstack-swift00:00
kota_acoles: thanks for reviewing my patch 283946, i will be circle back to the review remembering my draft idea at that time. but unfortunately, I have to be in my company seminor programs for a couple of days so probably my response will be slow, sorry.00:14
patchbothttps://review.openstack.org/#/c/283946/ - swift - Fix stats calculation in object-reconstructor00:14
mattoliveraukota_: morning00:22
kota_mattoliverau: morning00:22
*** catintheroof has quit IRC00:47
*** catintheroof has joined #openstack-swift00:48
*** catintheroof has quit IRC00:52
*** klamath_ has quit IRC00:52
claygtimburke: swift info --json | jq .swift.policies[].name <- first try!02:13
timburkew00t!02:13
*** klrmn has quit IRC02:14
timburkenow to add `alias swift_policies='swift info --json | jq .swift.policies[].name'` to your .profile so you never need remember it again02:15
*** two_tired has joined #openstack-swift02:22
*** duobei has joined #openstack-swift02:35
*** tqtran has quit IRC02:48
*** links has joined #openstack-swift02:49
*** clu_ has quit IRC02:50
*** bkopilov_ has quit IRC02:57
*** jamielennox is now known as jamielennox|away03:07
*** klrmn has joined #openstack-swift03:18
*** jamielennox|away is now known as jamielennox03:21
openstackgerritJason Wells proposed openstack/python-swiftclient: Allow _retry to continue if there's an SSLError in --insecure mode.  https://review.openstack.org/39750903:23
*** dmorita has quit IRC03:23
*** duobei has quit IRC03:46
*** duobei has joined #openstack-swift03:58
*** psachin has joined #openstack-swift04:09
*** jamielennox is now known as jamielennox|away04:15
*** jamielennox|away is now known as jamielennox04:17
*** bkopilov_ has joined #openstack-swift04:17
*** SkyRocknRoll has joined #openstack-swift04:19
*** david_c has quit IRC04:26
*** oshritf has joined #openstack-swift04:27
*** oshritf has quit IRC04:27
*** dmorita has joined #openstack-swift04:34
*** dmorita has quit IRC04:38
*** jamielennox is now known as jamielennox|away04:40
*** jamielennox|away is now known as jamielennox05:03
*** duobei has quit IRC05:14
*** ppai has joined #openstack-swift05:19
*** duobei has joined #openstack-swift05:29
*** ppai has quit IRC05:35
*** dmorita has joined #openstack-swift05:35
*** ppai has joined #openstack-swift05:39
*** dmorita has quit IRC05:40
*** ppai has quit IRC05:40
*** tqtran has joined #openstack-swift05:48
*** tqtran has quit IRC05:52
*** duobei has quit IRC06:00
*** Jeffrey4l has quit IRC06:03
*** two_tired has quit IRC06:23
*** sams-gleb has joined #openstack-swift06:41
*** Jeffrey4l has joined #openstack-swift06:41
*** ppai has joined #openstack-swift06:42
*** dmorita has joined #openstack-swift06:45
*** dmorita has quit IRC06:47
*** ChubYann has quit IRC07:12
*** siva_krish has joined #openstack-swift07:14
*** klrmn has quit IRC07:22
*** rcernin has quit IRC07:23
*** dmorita has joined #openstack-swift07:23
*** dmorita has quit IRC07:28
*** rvasilets___ has joined #openstack-swift07:36
*** hseipp has joined #openstack-swift07:45
*** oshritf has joined #openstack-swift07:47
*** duobei has joined #openstack-swift07:50
*** sams-gleb has quit IRC07:53
*** sams-gleb has joined #openstack-swift07:53
*** sams-gleb has quit IRC07:58
*** sree_ has joined #openstack-swift08:01
*** sree_ is now known as Guest5915808:01
Guest59158hii all, By default how many requests that proxy-server will accept?08:02
*** sams-gleb has joined #openstack-swift08:10
*** cbartz has joined #openstack-swift08:12
*** duobei has quit IRC08:19
*** rledisez has joined #openstack-swift08:21
*** duobei has joined #openstack-swift08:21
*** pcaruana has joined #openstack-swift08:22
*** amoralej|off is now known as amoralej08:22
*** zul has quit IRC08:33
*** duobei has quit IRC08:34
*** dmorita has joined #openstack-swift08:36
*** geaaru has joined #openstack-swift08:39
*** dmorita has quit IRC08:40
*** zul has joined #openstack-swift08:44
*** kei_yama_ has joined #openstack-swift09:05
*** kei_yama has quit IRC09:05
*** joeljwright has joined #openstack-swift09:08
*** ChanServ sets mode: +v joeljwright09:08
winggundamthMy coworker ask me about Swift project. Why it named Swift? anybody knows history about this?09:12
*** mvk has quit IRC09:13
*** mvk has joined #openstack-swift09:15
*** openstackgerrit has quit IRC09:18
winggundamthSo our micro-service that interacts with Swift named Taylor now :P09:18
*** openstackgerrit has joined #openstack-swift09:19
*** clsacramento has joined #openstack-swift09:21
*** acoles_ is now known as acoles09:26
acoleswinggundamth: people who might know the answer to "why swift?" will be asleep now ;) The project originated from Rackspace cloud files team in Texas.09:28
acoleskota_: no problem09:29
winggundamthacoles: I watch this right now https://www.youtube.com/watch?v=Dd7wmJCDh4w ;)09:29
cbartzHello, we currently observe an asymmetry on our cluster: 3 (of 6) nodes are running under high load (20+), and the 3 others with very small load (3). Its always the same set of  nodes (so it is not the case that sometimes the load on the one drops and then on another increases). How can this be? Are there any tools to dig further in this? Originally I wanted to post this on ask.openstack.org, but I am not able to post this, m09:37
*** rvasilets___ has quit IRC09:38
*** duobei has joined #openstack-swift09:42
*** Trixboxer has quit IRC09:47
*** dmorita has joined #openstack-swift09:48
rlediseznotmyname: i was just reading your update on the wiki page about small files09:49
rlediseznotmyname: especially "Why not talk to the block device itself? At what point do we need to invent a whole filesystem ourselves, and at that point, what benefit do we have over using what's available to us already?"09:50
rlediseznotmyname: i was writing yesterday a summary of our experimentations for the wiki09:50
rlediseznotmyname: i just posted it: https://wiki.openstack.org/wiki/Swift/ideas/small_files/experimentations09:50
rlediseznotmyname: what might interest you is the last part (using dmu of zfs). from all our tries, it's for now clearly the best option09:51
rlediseznotmyname: i'm interested of your thoughts on this kind of solution. eg: is there a chance this kind of solution could be upstreamable? because we don't want to work on something we wouldn't be able to upstream, it would be a lost of time for everybody09:52
*** dmorita has quit IRC09:53
*** Trixboxer has joined #openstack-swift09:59
Guest59158acoles: are you around10:07
*** duobei has quit IRC10:10
*** siva_krish has quit IRC10:13
*** asettle has joined #openstack-swift10:22
Guest59158hii all, By default how many requests that proxy-server will accept?10:28
openstackgerritAlistair Coles proposed openstack/swift: Use per policy diskfile manager in object replicator  https://review.openstack.org/39725910:31
*** dcourtoi has joined #openstack-swift10:35
*** deep_1 has joined #openstack-swift10:52
deep_1Is there any way to use credentials from ldap for swift and s3 ? I want to avoid openstack credential create for every user from ldap ?10:54
*** mvk has quit IRC11:16
*** deep_1 has quit IRC11:39
*** links has quit IRC11:43
*** mvk has joined #openstack-swift11:47
openstackgerritOndřej Nový proposed openstack/swift: object-replicator cleanup  https://review.openstack.org/39161712:22
openstackgerritOndřej Nový proposed openstack/swift: Use direct_get_suffix_hashes in the reconstructor  https://review.openstack.org/39455112:23
*** catintheroof has joined #openstack-swift12:28
*** Guest59158 has quit IRC12:29
*** bkopilov_ has quit IRC12:31
*** deep_1 has joined #openstack-swift12:44
*** SkyRocknRoll has quit IRC12:46
*** kei_yama_ has quit IRC12:56
*** admin6 has quit IRC13:05
*** admin6_ has joined #openstack-swift13:05
*** sams-gle_ has joined #openstack-swift13:07
*** sams-gleb has quit IRC13:07
*** duobei has joined #openstack-swift13:08
*** duobei has quit IRC13:12
*** asettle is now known as her-royalness13:32
*** acoles is now known as acoles_13:37
*** ndk_ has quit IRC13:44
*** ppai has quit IRC13:53
*** ndk_ has joined #openstack-swift13:54
tdasilvagood morning13:55
*** her-royalness is now known as asettle13:57
*** acoles_ is now known as acoles14:05
acolestdasilva: o/14:05
*** klamath_ has joined #openstack-swift14:05
*** klamath_ has quit IRC14:05
tdasilvaacoles: \o14:06
*** klamath_ has joined #openstack-swift14:06
*** StraubTW has joined #openstack-swift14:09
*** pcaruana has quit IRC14:13
*** psachin has quit IRC14:18
*** vinsh has quit IRC14:25
*** bkopilov_ has joined #openstack-swift14:29
*** _JZ_ has joined #openstack-swift14:33
*** sams-gle_ has quit IRC14:36
*** sams-gleb has joined #openstack-swift14:37
*** sams-gleb has quit IRC14:41
*** tuan_luong has joined #openstack-swift14:42
*** amoralej is now known as amoralej|lunch14:43
*** tqtran has joined #openstack-swift14:53
*** sams-gleb has joined #openstack-swift14:53
*** tqtran has quit IRC14:57
*** vinsh has joined #openstack-swift15:04
*** nikivi has joined #openstack-swift15:10
*** ouchkernel has joined #openstack-swift15:12
*** siva_krish has joined #openstack-swift15:13
*** vinsh has quit IRC15:14
*** vinsh has joined #openstack-swift15:15
*** david_c has joined #openstack-swift15:15
*** vinsh has quit IRC15:16
*** vinsh has joined #openstack-swift15:17
*** vinsh has quit IRC15:18
*** vinsh has joined #openstack-swift15:21
*** klamath_ is now known as klamath15:37
*** deep_1 has quit IRC15:42
*** openstackgerrit has quit IRC15:48
*** amoralej|lunch is now known as amoralej15:48
*** openstackgerrit has joined #openstack-swift15:48
acolesbkeller`: IIRC you were asking about maybe cleaning up/wsgifying some middleware recently - here is a good candidate - bug 164198015:52
openstackbug 1641980 in OpenStack Object Storage (swift) "Invalid copy destination header can cause 500 response" [High,New] https://launchpad.net/bugs/164198015:52
bkeller`thanks, i'll take a look15:54
*** pcaruana has joined #openstack-swift16:02
notmynamegood morning16:35
*** sgundur_ has joined #openstack-swift16:35
notmynamerledisez: oh, nice. there's a lot to read on that page :-)16:35
rlediseznotmyname: good morning16:38
*** tuan_luong has quit IRC16:40
notmynamerledisez: ok, quick read done. thanks for writing it up and telling me about it16:44
*** abalfour has quit IRC16:45
*** abalfour has joined #openstack-swift16:45
*** sgundur_ has quit IRC16:47
*** siva_krish has quit IRC16:48
*** nikivi has quit IRC16:52
*** tqtran has joined #openstack-swift16:52
openstackgerritMahati Chamarthy proposed openstack/swift: remove empty db hash and suffix directories  https://review.openstack.org/33160116:53
*** sgundur_ has joined #openstack-swift17:01
*** nikivi has joined #openstack-swift17:02
*** joeljwright has quit IRC17:05
*** klrmn has joined #openstack-swift17:13
*** hseipp has quit IRC17:17
*** ChubYann has joined #openstack-swift17:25
*** calebb has quit IRC17:27
*** dmorita has joined #openstack-swift17:29
*** Jeffrey4l has quit IRC17:34
*** rledisez has quit IRC17:36
*** siva_krish has joined #openstack-swift17:36
*** mvk has quit IRC17:38
*** sgundur_ has quit IRC17:40
*** calebb has joined #openstack-swift17:46
timburkegood morning17:48
*** sgundur_ has joined #openstack-swift17:58
*** nikivi has quit IRC18:02
*** asettle has quit IRC18:10
*** klamath has quit IRC18:13
*** portante has quit IRC18:19
*** portante has joined #openstack-swift18:20
*** portante has quit IRC18:26
*** portante has joined #openstack-swift18:28
*** amoralej is now known as amoralej|off18:40
*** acoles is now known as acoles_18:46
*** geaaru has quit IRC18:48
*** hseipp has joined #openstack-swift18:50
*** joeljwright has joined #openstack-swift18:53
*** ChanServ sets mode: +v joeljwright18:53
*** joeljwright has quit IRC18:54
*** jordanP has joined #openstack-swift18:56
*** sgundur_ has quit IRC18:59
*** rcernin has joined #openstack-swift19:02
*** nikivi has joined #openstack-swift19:05
openstackgerritJanie Richling proposed openstack/swift: Document X-Openstack-Request-Id in all responses  https://review.openstack.org/39789619:22
notmynamecool. backblaze just published their Q3 hard drive stats https://www.backblaze.com/blog/hard-drive-failure-rates-q3-2016/?utm_source=Social&utm_medium=Twitter&utm_term=blog%2Bpost&utm_content=tweet19:29
*** hseipp has quit IRC19:36
openstackgerritMerged openstack/swift: Better error messages in go/probe  https://review.openstack.org/39090319:42
openstackgerritThiago da Silva proposed openstack/swift: Symlink implementation.  https://review.openstack.org/23216219:53
*** david_c_ has joined #openstack-swift19:55
*** david_c has quit IRC19:56
*** david_c_ is now known as david_c19:56
*** sgundur_ has joined #openstack-swift20:05
*** sgundur_ has quit IRC20:09
*** klamath has joined #openstack-swift20:17
*** sgundur_ has joined #openstack-swift20:17
*** klamath has quit IRC20:23
*** nikivi has quit IRC20:27
*** klrmn has quit IRC20:31
*** david_c has quit IRC20:37
*** sgundur_ has quit IRC20:37
*** sams-gleb has quit IRC21:11
*** sams-gleb has joined #openstack-swift21:12
*** sams-gleb has quit IRC21:16
mattoliveraumorning21:25
openstackgerritChristian Hugo proposed openstack/swift: Use direct_get_suffix_hashes in the reconstructor  https://review.openstack.org/39455121:27
tdasilvamattoliverau: o/21:27
*** pcaruana has quit IRC21:32
*** sgundur_ has joined #openstack-swift21:49
notmynamelifeless: any chance you're around? I have a question about testrepository22:08
notmynamewell, two, actually22:10
notmyname1 -- is there an equivalent to `nosetests -x` ie stop after the first failure?22:10
notmyname2 -- is there any differences in behavior with raising SkipTest (ie different from what nose would do)22:10
*** mvk has joined #openstack-swift22:14
*** catinthe_ has joined #openstack-swift22:18
*** catintheroof has quit IRC22:21
*** sgundur_ has quit IRC22:32
*** catintheroof has joined #openstack-swift22:35
openstackgerritTim Burke proposed openstack/swift: Fix up X-Trans-Id-Extra docs  https://review.openstack.org/39797922:36
openstackgerritTim Burke proposed openstack/swift: Add X-Openstack-Request-Id to Access-Control-Expose-Headers  https://review.openstack.org/39798022:36
*** catinthe_ has quit IRC22:36
*** joeljwright has joined #openstack-swift22:36
*** ChanServ sets mode: +v joeljwright22:36
notmynamentata: didn't you work on the change that made functional tests use testr? it was a while back22:42
*** rcernin has quit IRC22:42
mattoliveraunotmyname: I think it's -f to stop after first failure22:46
notmynamemattoliverau: ah? cool22:47
mattoliverauhttps://wiki.openstack.org/wiki/Testr#How_can_I_exit_a_test_run_after_the_first_failure.3F22:47
notmynamentata: I don't remember what that conversion was like, but IIRC it wasn't too bad22:50
notmynameok, so want to hear a funny story?22:52
*** StraubTW has quit IRC22:52
notmynamethe intro to this hilarious story is https://review.openstack.org/#/c/336323/22:53
patchbotpatch 336323 - swift - Add checksum to object extended attributes22:53
notmynamewherein we remove the xattr mock22:53
notmynameand you'd think, "cool! less mocking, more real testing!". and you'd be right22:53
notmynameexcept that unittests that need xattrs in the tmpdir fail because now the success is dependent on the underlying FS to support xattrs22:54
notmynameand ext4 breaks with more than 4k of xattrs (hello encryption!) and tmpfs just flat out doesn't support xattrs22:54
notmynameno problem, let's just mount /tmp as xfs22:55
notmynameturns out, that can't be done, since that is a change to the common testing guidelines and the common gate image from infra22:55
notmynamebut there's a plan!22:55
notmynamewe'll just set TMPDIR to something that is xfs and everythign will work22:55
notmynamenow, all we gotta do is change any hard-coded reference of /tmp in our code to tempfile.gettempdir()22:56
notmynameeasy22:56
notmynamewell, except for the extra code22:56
notmynameand the new gate jobs22:56
notmynamebut that's ok22:56
notmynameexcept there's also the in-process functional tests22:56
notmynamewhich, of course, we're importing test.unit and getting the xattr mock. which doesn't exist now22:57
notmynamewhich means that in process func tests now break if they needed xattrs (in the same cases I mentioned above)22:57
notmynameand if you think that's funny, we haven't even gotten to the punchline yet! hillarious!22:57
notmynameso no worries, i'll just add some SkipTest() to the in process functional tests22:58
notmynamegreat, looks like it's working. we skip abotu 415 of 456 fuctional tests on my SAIO and tests "pass"! winning!22:58
notmynameexcept22:58
notmynameexcept that I was using nosetests to run the tests. .functests (and the gate) use ostestr22:59
notmynameit turns out that there exists nose.SkipTest, unittest.SkipTest, and unittest2.SkipTest22:59
notmynameand if you run under nose, then nose.SkipTest and unittest.SkipTest are just fine22:59
notmynameand if you run under ostestr, unittest2.SkipTest is fine23:00
*** vinsh has quit IRC23:00
*** klrmn has joined #openstack-swift23:00
notmynameso, and here's the *really* funny part... remember back when we switched *just* the functests to use ostestr instead of nose? and remember how functional tests import stuff from unittests?23:00
notmynamethat means that if we fix unittests to SkipTest when there's no xattr support, we break the `raise SkipTest` under the functest run. and if we fix the functest run to raise the right SkipTest, then unittests fail because it's the wrong SkipTest23:02
notmynamehillarious!23:02
timburkedefine yet-another SkipTest that inherits from both of them?23:04
notmynameooohh! I like door #323:04
notmynameI was thinking of 1) change unittests to use ostestr 2) internally duplicate my code to be slightly different for unittests and functests23:05
*** jordanP has quit IRC23:10
ntatanotmyname, sorry. I had very limited knowledge on testr.. I only used it to list tests and run23:13
ntata:(23:13
notmynamentata: ah, no worries. I'm likely misremembering23:14
notmynamefor those following at home, timburke's idea works! https://gist.github.com/notmyname/ecbafa4ed7185f45a59aa7d5ed78458b (but not in the reverse order of inheritance)23:14
*** vinsh has joined #openstack-swift23:30
*** vinsh has quit IRC23:30
*** vinsh has joined #openstack-swift23:31
*** kei_yama has joined #openstack-swift23:33
jrichlithanks for the quick review, timburke23:34
timburkejrichli: it only spawned two more :P23:37
*** siva_krish has quit IRC23:37
jrichlitimburke: i was just looking at those :-)23:37
timburkethanks!23:37
jrichliso, the one in controllers/base for expose_headers : I am not familiar with Access-Control-Expose-Headers.  I guess I will have to educate myself on it23:39
timburkeyeah, CORS is kinda weird. as i recall, it defines which headers a piece of javascript running in the browser can access when making cross-domain requests23:41
jrichlioh, ok.  thx23:41
timburkeif you go to the trouble of getting that loaded into your head, https://review.openstack.org/#/c/274303/ might be worth checking out at the same time23:42
patchbotpatch 274303 - swift - Add Vary: headers for CORS responses23:42
timburke...although i never responded on your previous comment. oops23:42
jrichli:-) so would that functest be a good idea?23:44
*** joeljwright has quit IRC23:44
timburkecertainly couldn't hurt. i'll go look at what CORS-related functests we've got currently23:44
jrichlisounds good.  ok, i am gonna take off now - for awhile at least.  have a good night!23:45
zaitcevI'm trying to figure out how Fragment Archives are organized by reading swift/proxy/controllers/obj.py and I do not understand A THING.23:49
zaitcevHundreds little yielding iterators all alike23:49
zaitcevwell, maybe dozens23:49
timurdoes anyone use the vagrant-swift-all-in-one here? I'm running into an issue where it's prompting me for a password... which seems like the wrong thing to happen23:49
zaitcevClay does23:49
timburketimur: on trusty?23:49
timurtimburke: on xenial, it appears23:50
timursince that's the default now?23:50
timburkehrm. idk then; i think i ran into a similar issue, fell back to trusty, then had to adjust my localrc per https://github.com/swiftstack/vagrant-swift-all-in-one/blob/master/localrc-template#L223:51
timurtimburke: I noticed that using the username "ubuntu" works, whereas otherwise it attempts to use my username and then fails23:52
timurI guess that's a temporary workaround for me23:52
*** catintheroof has quit IRC23:52
*** catintheroof has joined #openstack-swift23:54
*** catintheroof has quit IRC23:58
notmynameI just learned how to redirect all of stdout and stderr for a script to somewhere else. I think this will make probetest error output a lot cleaner https://gist.github.com/notmyname/a7e55a7e46eeb6b5332b1f1a800e741023:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!