kota_ | good morning | 00:12 |
---|---|---|
*** esker has joined #openstack-swift | 00:15 | |
*** esker has joined #openstack-swift | 00:17 | |
*** chlong has joined #openstack-swift | 00:19 | |
kota_ | mattoliverau: morning | 00:26 |
mattoliverau | kota_: morning | 00:26 |
kota_ | thanks mattolivearu to kick my small patch :) | 00:36 |
*** takashi has joined #openstack-swift | 00:55 | |
mattoliverau | kota_: no worries, it was a very easy review ;) | 00:56 |
openstackgerrit | Merged openstack/swift: Import BrainSpliter directly in expirer probe https://review.openstack.org/315836 | 00:58 |
*** vint_bra has quit IRC | 00:59 | |
*** itlinux has quit IRC | 01:08 | |
*** itlinux has joined #openstack-swift | 01:18 | |
*** itlinux has quit IRC | 01:23 | |
*** itlinux has joined #openstack-swift | 01:25 | |
*** jamielennox is now known as jamielennox|away | 01:26 | |
*** itlinux has quit IRC | 01:26 | |
*** jamielennox|away is now known as jamielennox | 01:28 | |
*** itlinux has joined #openstack-swift | 01:35 | |
*** ouchkernel has joined #openstack-swift | 01:35 | |
*** chlong has quit IRC | 01:56 | |
*** sgundur_ has joined #openstack-swift | 01:58 | |
*** tanee is now known as tanee_away | 02:03 | |
*** tanee_away is now known as tanee | 02:03 | |
jrichli | mattoliverau: did you find out why the saio is not ok with the 1% default? If not, is there anything I can help test for that - like what sort of test.conf changes should I try? | 02:04 |
*** itlinux has quit IRC | 02:04 | |
*** itlinux has joined #openstack-swift | 02:05 | |
mattoliverau | jrichli: not yet, started looking at it. but was the weekend so got distracted. I was playing with changing the max_file_size constraint in the test.conf. But I'll continue looking today, once meetings are over :) | 02:06 |
jrichli | mattoliverau: ok. let me know if you want me to run something. | 02:08 |
jrichli | hope you had a good weekend :-) | 02:08 |
*** itlinux has quit IRC | 02:10 | |
mattoliverau | jrichli: will do, and yup i did. Planning on the move to the beach house, so busy but exciting :) | 02:10 |
mattoliverau | jrichli: Enjoying your weekend? | 02:10 |
jrichli | mattoliverau: ah yes - beach house! That sounds really great. My weekend was good, I suppose. Pretty typical :-) | 02:11 |
*** sgundur_ has quit IRC | 02:14 | |
*** itlinux has joined #openstack-swift | 03:03 | |
*** asettle has joined #openstack-swift | 03:05 | |
*** jamielennox is now known as jamielennox|away | 03:06 | |
*** asettle has quit IRC | 03:10 | |
*** psachin has joined #openstack-swift | 03:22 | |
*** sgundur_ has joined #openstack-swift | 03:25 | |
notmyname | hellow world | 03:29 |
*** sgundur_ has quit IRC | 03:29 | |
mahatic | good morning | 03:31 |
mahatic | notmyname: hello | 03:31 |
*** jamielennox|away is now known as jamielennox | 03:36 | |
*** furlongm has quit IRC | 03:40 | |
*** janonymous has joined #openstack-swift | 03:40 | |
*** sams-gleb has joined #openstack-swift | 03:42 | |
*** sams-gleb has quit IRC | 03:46 | |
*** psachin has quit IRC | 03:48 | |
*** links has joined #openstack-swift | 03:49 | |
mattoliverau | notmyname, mahatic: o/ | 03:58 |
*** psachin has joined #openstack-swift | 04:07 | |
openstackgerrit | Mohit Motiani proposed openstack/swift: Updating help of swift-init: https://review.openstack.org/316448 | 04:13 |
openstackgerrit | Mohit Motiani proposed openstack/swift: Updating --help of swift-init: https://review.openstack.org/316448 | 04:15 |
*** eranrom has quit IRC | 04:32 | |
*** jrichli has quit IRC | 05:03 | |
openstackgerrit | Merged openstack/swift: Clean up fallocate tests a little https://review.openstack.org/309608 | 05:10 |
*** rcernin has joined #openstack-swift | 05:16 | |
*** rcernin has quit IRC | 05:21 | |
*** suyash has joined #openstack-swift | 05:28 | |
*** suyash has quit IRC | 05:28 | |
*** rcernin has joined #openstack-swift | 05:38 | |
*** sams-gleb has joined #openstack-swift | 05:43 | |
*** sams-gleb has quit IRC | 05:48 | |
*** eranrom has joined #openstack-swift | 05:57 | |
*** klrmn1 has quit IRC | 06:07 | |
*** sheel has joined #openstack-swift | 06:12 | |
openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Fix probe failure and small things https://review.openstack.org/315918 | 06:19 |
*** janonymous has quit IRC | 06:58 | |
*** asettle has joined #openstack-swift | 07:35 | |
*** sams-gleb has joined #openstack-swift | 07:37 | |
*** jmccarthy has quit IRC | 07:39 | |
*** jmccarthy has joined #openstack-swift | 07:40 | |
*** rvasilets has quit IRC | 07:43 | |
*** rvasilets has joined #openstack-swift | 07:44 | |
*** rvasilets has quit IRC | 07:44 | |
*** rvasilets has joined #openstack-swift | 07:45 | |
*** asettle has quit IRC | 08:03 | |
*** McMurlock has joined #openstack-swift | 08:06 | |
*** asettle has joined #openstack-swift | 08:17 | |
*** dosaboy has quit IRC | 08:26 | |
*** McMurlock has quit IRC | 08:27 | |
*** Shashikant86 has joined #openstack-swift | 08:28 | |
*** McMurlock has joined #openstack-swift | 08:29 | |
*** Shashikant86 has quit IRC | 08:30 | |
*** jistr has joined #openstack-swift | 08:30 | |
*** dmk0202 has joined #openstack-swift | 08:32 | |
*** dosaboy has joined #openstack-swift | 08:36 | |
*** acoles_ is now known as acoles | 08:49 | |
*** McMurlock has quit IRC | 08:57 | |
*** takashi has quit IRC | 08:59 | |
*** ouchkernel has quit IRC | 09:04 | |
*** ouchkernel has joined #openstack-swift | 09:05 | |
openstackgerrit | Andy McCrae proposed openstack/swift: [Docs] Document prevention of disk full scenarios https://review.openstack.org/315610 | 09:14 |
*** asettle has quit IRC | 09:17 | |
*** janonymous has joined #openstack-swift | 09:19 | |
*** sams-gleb has quit IRC | 09:19 | |
*** sams-gleb has joined #openstack-swift | 09:22 | |
rsFF | a | 09:24 |
rvasilets | hi I can't figure out what I'm doing wrong I have installed dependences from http://docs.openstack.org/developer/swift/development_saio.html Additionally install tox. Install requirements and test-requirements. But my tox -r pep8 not working. The error is http://paste.openstack.org/show/497187/ | 09:25 |
ahale | looks like you need to install python-dev | 09:25 |
ahale | and probably some kind of pip / setuptools stuff | 09:26 |
*** mvk has quit IRC | 09:28 | |
rvasilets | python-dev already installed | 09:28 |
rvasilets | ahale, ^ | 09:29 |
rvasilets | setup tools too) | 09:30 |
ahale | hm not sure then :( | 09:30 |
rvasilets | this is the same behaviour for me at working machine and at home) | 09:31 |
ahale | this line looked kinda suspicious- L27 c/_cffi_backend.c:2:20: fatal error: Python.h: No such file or directory | 09:31 |
*** natarej has joined #openstack-swift | 09:32 | |
rvasilets | Yes) | 09:33 |
rvasilets | its strange | 09:33 |
rvasilets | But I have python) | 09:33 |
rvasilets | and googled this error | 09:33 |
*** natarej_ has quit IRC | 09:34 | |
rvasilets | I found that people solved by installing python-dev or libffi-dev | 09:34 |
rvasilets | But I have them already( | 09:34 |
*** sams-gleb has quit IRC | 09:43 | |
openstackgerrit | Mahati Chamarthy proposed openstack/swift: crypto - Add "disable_encryption" config option https://review.openstack.org/315588 | 09:55 |
*** mvk has joined #openstack-swift | 09:57 | |
*** sams-gleb has joined #openstack-swift | 09:59 | |
*** daemontool has joined #openstack-swift | 10:01 | |
*** natarej_ has joined #openstack-swift | 10:02 | |
*** natarej has quit IRC | 10:04 | |
openstackgerrit | Andy McCrae proposed openstack/swift: [Docs] Document prevention of disk full scenarios https://review.openstack.org/315610 | 10:09 |
*** natarej_ has quit IRC | 10:10 | |
*** natarej has joined #openstack-swift | 10:14 | |
*** natarej has quit IRC | 10:26 | |
*** sams-gleb has quit IRC | 11:01 | |
*** sams-gleb has joined #openstack-swift | 11:01 | |
*** _fortis has quit IRC | 11:13 | |
openstackgerrit | Alistair Coles proposed openstack/swift: crypto - Don't encrypt zero size object etags https://review.openstack.org/316732 | 11:14 |
*** kei_yama has quit IRC | 11:15 | |
*** janonymous has quit IRC | 11:46 | |
openstackgerrit | venkatamahesh proposed openstack/python-swiftclient: Migrated contributing file from .md to .rst https://review.openstack.org/316758 | 11:53 |
openstackgerrit | venkatamahesh proposed openstack/python-swiftclient: Migrated contributing file from .md to .rst https://review.openstack.org/316758 | 11:55 |
*** raildo-afk is now known as raildo | 12:06 | |
*** diogogmt has quit IRC | 12:15 | |
*** psachin has quit IRC | 12:19 | |
*** psachin has joined #openstack-swift | 12:19 | |
*** diogogmt has joined #openstack-swift | 12:20 | |
*** asettle has joined #openstack-swift | 12:22 | |
*** vint_bra has joined #openstack-swift | 12:27 | |
*** vint_bra has quit IRC | 12:28 | |
*** diogogmt has quit IRC | 12:29 | |
*** diogogmt has joined #openstack-swift | 12:34 | |
*** NM has joined #openstack-swift | 12:40 | |
*** sams-gleb has quit IRC | 12:40 | |
*** sams-gleb has joined #openstack-swift | 12:41 | |
*** psachin has quit IRC | 12:46 | |
*** sams-gleb has quit IRC | 12:50 | |
*** links has quit IRC | 12:55 | |
openstackgerrit | Anne Gentle proposed openstack/swift: Adds migrated API reference files https://review.openstack.org/312315 | 12:58 |
*** pauloewerton has joined #openstack-swift | 13:01 | |
*** psachin has joined #openstack-swift | 13:11 | |
*** sams-gleb has joined #openstack-swift | 13:11 | |
*** aj701 has quit IRC | 13:24 | |
*** aj701 has joined #openstack-swift | 13:26 | |
*** vint_bra has joined #openstack-swift | 13:29 | |
*** bill_az has joined #openstack-swift | 13:31 | |
*** esker has quit IRC | 13:34 | |
*** _JZ_ has joined #openstack-swift | 13:40 | |
*** klamath has joined #openstack-swift | 13:42 | |
*** klamath has quit IRC | 13:42 | |
*** klamath has joined #openstack-swift | 13:42 | |
*** t4nk983 has joined #openstack-swift | 13:43 | |
*** sheel has quit IRC | 13:45 | |
*** CaioBrentano has joined #openstack-swift | 13:50 | |
*** jrichli_ has joined #openstack-swift | 13:51 | |
*** jrichli_ has quit IRC | 13:53 | |
CaioBrentano | hi all! | 13:53 |
CaioBrentano | mlocate.db on my object server is getting huge. Is there any downside to swift if I add objects path to prunepath in /etc/updatedb.conf ? | 13:54 |
*** jrichli has joined #openstack-swift | 13:56 | |
ahale | no downside to swift at all - probably make it all a bit better than when having mlocate indexing disks too. we just dont run it at all | 13:56 |
CaioBrentano | thanks ahale! | 13:56 |
*** ametts has joined #openstack-swift | 13:57 | |
*** diogogmt has quit IRC | 14:02 | |
*** sgundur_ has joined #openstack-swift | 14:06 | |
openstackgerrit | Alistair Coles proposed openstack/swift: crypto - make_crypto meta a property of CryptoContext https://review.openstack.org/316863 | 14:07 |
*** sgundur__ has joined #openstack-swift | 14:07 | |
*** sgundur_ has quit IRC | 14:10 | |
*** sgundur__ has quit IRC | 14:12 | |
*** links has joined #openstack-swift | 14:15 | |
*** esker has joined #openstack-swift | 14:20 | |
*** esker has quit IRC | 14:20 | |
*** esker has joined #openstack-swift | 14:21 | |
sgundur1 | rvasilets: hi, were u able to resolve the tox errors ? | 14:24 |
*** wadeholler has joined #openstack-swift | 14:26 | |
sgundur1 | rvasilets: pls check the owner for the swift.egg-info ; | 14:26 |
wadeholler | hi folks; trying to learn something and help out openstack-ansible; should I be concerned with the following error in object-error.log: "object-replicator: Can't find itself 172.30.244.2 with port 6000 in ring file, not replicating" | 14:27 |
*** Jeffrey4l has joined #openstack-swift | 14:28 | |
*** thumpba has joined #openstack-swift | 14:30 | |
*** rnowling has joined #openstack-swift | 14:31 | |
glange | wadeholler: I don't know why, but I can't find that error message anywhere in the current swift codebase | 14:35 |
wadeholler | glange: I find it here https://github.com/openstack/swift/blob/master/swift/common/db_replicator.py - line 650 I think. | 14:36 |
glange | ah, I had stale code | 14:39 |
ahale | oO that code looks weird :) does that mean no replication on a device thats had all its paritions removed in a rebalance and is trying to drain data | 14:41 |
glange | it might, it's a little hard to follow | 14:42 |
glange | well, it just prints the error message and keeps going, right? | 14:44 |
ahale | oh yeah, nm :) | 14:45 |
glange | but it's not telling the truth, I think? because it does actually replicate? | 14:47 |
wadeholler | glange: I have verified that an object that I uploaded has been replicated to 3 different nodes per the policy. So yes I think. I don't know what db replication is and obviously very new at swift | 14:50 |
rvasilets | sgundur1, Looks like I need to do extra step to install swift but http://paste.openstack.org/show/497227/ I was hope that I could run units without doing the whole manual http://docs.openstack.org/developer/swift/development_saio.html | 14:51 |
glange | accounts and containers have dbs - the container db has the listing for its objects - the account db has a list of all it containers | 14:51 |
glange | this replciation is a replcation of one of those db types | 14:51 |
glange | wadeholler: I can't find where logger.error() is - if that doesn't throw an exception, it looks like to me your db will be replicated despite the error message (and that is how it should work) | 14:52 |
*** psachin has quit IRC | 14:52 | |
glange | wadeholler: probably we want it to work that way | 14:53 |
wadeholler | glange: ok. so safely ignore this. understood and thank you very much | 14:53 |
glange | wadeholler: basically what it is saying is your server/device isn't in the ring | 14:53 |
redbo | If find_local is false, then it won't be replicating anything. | 14:53 |
redbo | There's no way for find_local to be false and dirs to be populated | 14:54 |
glange | redbo: because error() throws an exception or because something that happens furhter down in the code? | 14:54 |
redbo | because the "dirs" array that is what it's replicating gets filled in after find_local is set to true | 14:54 |
*** dmorita has joined #openstack-swift | 14:54 | |
sgundur1 | rvasilets: the swift.egg-info was owned by root in my case, after changing to swift user it worked for me and also to install I generally use the comd - $ sudo python setup.py develop | 14:55 |
glange | ah | 14:55 |
glange | redbo: so he needs to add his device/server to the ring? | 14:55 |
redbo | Looks like it | 14:55 |
glange | wadeholler: ^^ | 14:55 |
wadeholler | redbo: glange: thank you; will investigate how to do that. | 14:57 |
glange | wadeholler: and this is either the account or container ring | 14:57 |
glange | because only those are subject to db replicaiton | 14:57 |
rvasilets | sgundur1, The same is for develop but I have deleted ;python_versions.. | 14:58 |
rvasilets | from requirements | 14:58 |
rvasilets | and sudo python setup.py develop work | 14:58 |
ahale | glange: yeah but the error log line did say object-repli | 14:58 |
rvasilets | sgundur1, After egg was generated I can find there owner field | 14:59 |
*** arch-nemesis has joined #openstack-swift | 15:00 | |
*** diogogmt has joined #openstack-swift | 15:00 | |
glange | ahale: yeah, that's confusing | 15:00 |
glange | swift/common/db_replicator.py: self.logger.error("Can't find itself %s with port %s in ring " | 15:00 |
*** esker has quit IRC | 15:00 | |
glange | that's the only place I find that error message in the current code base | 15:00 |
sgundur1 | rvasilets: ok, so r u trying to run unittests now ? | 15:01 |
*** eranrom has quit IRC | 15:01 | |
*** itlinux has quit IRC | 15:02 | |
rvasilets | yes | 15:03 |
rvasilets | sgundur1, result is the same | 15:03 |
*** pdardeau has quit IRC | 15:04 | |
rvasilets | I want to run at least units) Deploy will be later | 15:04 |
*** ntata- has quit IRC | 15:04 | |
*** mmotiani_ has quit IRC | 15:05 | |
*** yarkot has quit IRC | 15:05 | |
*** joeljwright has joined #openstack-swift | 15:05 | |
*** ChanServ sets mode: +v joeljwright | 15:05 | |
*** sgundur1 has quit IRC | 15:05 | |
*** pdardeau has joined #openstack-swift | 15:06 | |
*** yarkot has joined #openstack-swift | 15:07 | |
*** sgundur1 has joined #openstack-swift | 15:09 | |
*** ntata- has joined #openstack-swift | 15:10 | |
*** mmotiani_ has joined #openstack-swift | 15:10 | |
*** sgundur- has joined #openstack-swift | 15:12 | |
*** klrmn has joined #openstack-swift | 15:15 | |
*** Jeffrey4l has quit IRC | 15:17 | |
*** itlinux has joined #openstack-swift | 15:18 | |
*** zaitcev has joined #openstack-swift | 15:24 | |
*** ChanServ sets mode: +v zaitcev | 15:24 | |
*** McMurlock has joined #openstack-swift | 15:25 | |
*** nadeem has joined #openstack-swift | 15:29 | |
*** esker has joined #openstack-swift | 15:31 | |
*** esker has quit IRC | 15:31 | |
*** esker has joined #openstack-swift | 15:32 | |
*** lcurtis has joined #openstack-swift | 15:36 | |
*** dmorita has quit IRC | 15:40 | |
*** Jeffrey4l has joined #openstack-swift | 15:41 | |
*** esker has quit IRC | 15:44 | |
*** dmk0202 has quit IRC | 15:44 | |
*** McMurlock has quit IRC | 15:44 | |
*** sams-gleb has quit IRC | 15:45 | |
*** sams-gleb has joined #openstack-swift | 15:45 | |
notmyname | FYI i've updated patch 312267 | 15:46 |
patchbot | notmyname: https://review.openstack.org/#/c/312267/ - governance - add golang as an approved language | 15:46 |
acoles | mahatic: jrichli: I managed to workaround some of the failing probe tests (patch 316924) without needing any client generated override flag. | 15:47 |
patchbot | acoles: https://review.openstack.org/#/c/316924/ - swift (feature/crypto) - crypto - make some probe tests compatible with crypto | 15:47 |
acoles | container sync tests will still fail of course, as expected. | 15:48 |
acoles | mahatic: jrichli: if this approach ^^ works then imho it is better that having an override header that *any* client could use to disable crypto - I'm not sure we want that. | 15:48 |
jrichli | the container sync tests succeed if you add crypto to the internal-client pipeline. | 15:48 |
jrichli | or they did before we added byok | 15:49 |
acoles | jrichli: yeah, that makes sense. | 15:49 |
jrichli | ok, good. wanted to be sure we didn | 15:49 |
jrichli | t regress | 15:49 |
*** lakshmiS has joined #openstack-swift | 15:49 | |
jrichli | thanks, will take a look. sounds good! | 15:49 |
acoles | jrichli: but then with our current keymaster you'd need the root secret in a file on your container server which is where the db's are with etags at rest :/ | 15:50 |
acoles | jrichli: I didn't try the sync tests with internal client modified - can you do that easily with patch 316924 - I didn't touch those tests so should be ok | 15:50 |
patchbot | acoles: https://review.openstack.org/#/c/316924/ - swift (feature/crypto) - crypto - make some probe tests compatible with crypto | 15:50 |
jrichli | acoles: yes, we talked about that in tokyo. IIRC, it was said that the proxy config on container servers would be on a separate drive - we need to spell out the need for that i guess | 15:51 |
jrichli | i mean, the need for the separation | 15:51 |
jrichli | acoles: yes, i will try that | 15:52 |
acoles | jrichli: yes re separation. | 15:52 |
acoles | thanks | 15:52 |
jrichli | acoles: i will also make a note on docs review about separation and container sync | 15:52 |
*** wadeholler has left #openstack-swift | 15:52 | |
*** sams-gleb has quit IRC | 15:54 | |
acoles | jrichli: ok thanks, IIRC I put something in doc about separation, but not for sync | 15:55 |
*** sheel has joined #openstack-swift | 15:56 | |
*** asettle has quit IRC | 15:57 | |
jrichli | acoles: oh, really? i just put some comments there. mentioned on adding container-sync and internal-client pipeline, and the need for separation. | 15:57 |
*** lyrrad has joined #openstack-swift | 16:01 | |
*** pdardeau_ has joined #openstack-swift | 16:08 | |
*** gyee has joined #openstack-swift | 16:09 | |
openstackgerrit | Mohit Motiani proposed openstack/swift: Add description of server options (all, main, rest) for swift-init help https://review.openstack.org/316448 | 16:09 |
acoles | jrichli: https://review.openstack.org/#/c/312096/3/doc/source/overview_encryption.rst@55 - like I said, "something" ;) | 16:10 |
patchbot | acoles: patch 312096 - swift (feature/crypto) - crypto - add overview doc (WIP) | 16:10 |
acoles | cuold be beefed up | 16:10 |
*** wadeholler has joined #openstack-swift | 16:10 | |
jrichli | ah, i see. thx! | 16:11 |
ober37 | If someone has time for a 1-liner review, it would be appreciated. https://review.openstack.org/#/c/316303/ | 16:17 |
patchbot | ober37: patch 316303 - swift - Bypass Account Mgmt Functional Cases When Disabled | 16:17 |
*** klrmn has quit IRC | 16:25 | |
acoles | jrichli: the patch to allow MAXHEADERS to be varied says the 100 limit is for python 2.7.9 on, patch 195940, which may explain why its not an issue for me (I am on 2.7.6) | 16:40 |
patchbot | acoles: https://review.openstack.org/#/c/195940/ - swift - Increase httplib._MAXHEADERS (MERGED) | 16:41 |
*** t4nk983 has quit IRC | 16:41 | |
acoles | jrichli: but where the tox-func job fails I see python2.7.6 in the job logs?? | 16:41 |
jrichli | acoles: oh, ok thx. as for the tox-func job, I had changed to override it in conf - and it still failed on top of that. | 16:42 |
acoles | jrichli: yeah, i don't understand that failure | 16:43 |
*** Jeffrey4l has quit IRC | 16:43 | |
jrichli | acoles: I am seeing 3 probetest errors with patch 316924 | 16:43 |
patchbot | jrichli: https://review.openstack.org/#/c/316924/ - swift (feature/crypto) - crypto - make some probe tests compatible with crypto | 16:43 |
jrichli | just now looking into that | 16:43 |
acoles | jrichli: unless something about the in process setup faisl to set the max headers? | 16:44 |
jrichli | with this patch, i am getting the "too many open files" error from test_*_reload (test.probe.test_wsgi_servers.TestWSGIServerProcessHandling) | 16:45 |
jrichli | I think notmyname ran into something similar with probetests recently | 16:46 |
acoles | well then it's not my fault ;P | 16:46 |
*** links has quit IRC | 16:47 | |
jrichli | acoles: if i am right :-) his too many open files could have been something different | 16:47 |
jrichli | acoles: if the limit is now 100, should our max test not change to be testing that limit instead of 90? | 16:48 |
timburke | good morning | 16:48 |
jrichli | acoles: my understanding is that we will always be over the limit in the test against the max - by doubling the headers with crypto-meta. the extra header count default is still 0. | 16:50 |
notmyname | good morning | 16:50 |
*** sams-gleb has joined #openstack-swift | 16:51 | |
acoles | jrichli: you confused me there, your patch increases the limit | 16:51 |
jrichli | right, i dont know why the functest fails on the patch. but just going back to not understanding why we no longer need to set the extra header count in configs - technically, that number is added ot the MAX that comes from bufferedhttp | 16:54 |
*** lakshmiS has quit IRC | 16:54 | |
jrichli | acoles: actually, the 90 for MAX_META_COUNT is set in constraints.py | 16:55 |
acoles | jrichli: why do you say "we no longer need to set the extra header count in configs"? Doesn't your patch show that we do? at least for gate-dsvm-swift-functional? | 16:57 |
jrichli | when you say that the max is 100 for a certain version of python, how is that related to the "MAX_META_COUNT + 32 + max(EXTRA_HEADER_COUNT, 0)" calculation? | 16:57 |
*** nadeem has quit IRC | 16:57 | |
jrichli | acoles: locally, neither you or I need to set that value anymore. sitll donno what's up with the gate test | 16:57 |
*** nadeem has joined #openstack-swift | 16:57 | |
acoles | "anymore" - I never have, but like I said I am on 2.7.6 and the patch for extra headers etc said its an issue with >=2.7.9 | 16:58 |
*** sams-gleb has quit IRC | 16:58 | |
acoles | jrichli: httplib in python 2.7.6 imposes a limit of 100 headers but swift is able to increase that limit (and does since 100 would not be enough for max user metadata plus other headers) | 16:59 |
acoles | jrichli: https://hg.python.org/cpython/file/v2.7.9/Lib/httplib.py#l278 | 17:00 |
jrichli | acoles: I have to assume that we have no tests right now that test using the maximum which ends up being used. because crypto doubles the usage, and it that should result in a failure for a test that goes to the upper bound. | 17:01 |
*** jistr has quit IRC | 17:02 | |
mmotiani | jrichli: My func tests are still failing after changing fallocate_reserve to 0 | 17:04 |
acoles | jrichli: ?? this failed http://logs.openstack.org/19/316419/1/check/gate-swift-dsvm-functional/af13736/console.html#_2016-05-15_04_47_40_391 and then you increased the extra header count and it passed | 17:04 |
mmotiani | timburke: df ouput looks like this: http://paste.openstack.org/show/497241/ | 17:04 |
acoles | jrichli: the mystery is why tox -e func failed on your 3rd patchset where you do have extra header count | 17:05 |
jrichli | acoles: but you have never had to change the extra header count. and i have tried locally and no longer need to do that. the functest gates have been a bit weird wrt config lately, IMO, so I dont trust yet that I know what is happening there. | 17:06 |
acoles | jrichli: what python version are you using? | 17:06 |
jrichli | 2.7.6 | 17:06 |
acoles | so that explains why the test does not fail on your machine | 17:07 |
*** dmorita has joined #openstack-swift | 17:07 | |
*** ajiang has quit IRC | 17:07 | |
jrichli | mmotiani: when i tested, i changed the default used in the code: wsgi.py and daemon.py | 17:07 |
* jrichli is leaving for lunch - brb | 17:07 | |
*** bhakta has quit IRC | 17:09 | |
mmotiani | jrichli: Yes, I changed them only. | 17:09 |
*** pdardeau_ has quit IRC | 17:11 | |
*** dmorita has quit IRC | 17:11 | |
*** david-lyle_ has quit IRC | 17:12 | |
*** david-lyle has joined #openstack-swift | 17:13 | |
*** esker has joined #openstack-swift | 17:14 | |
*** esker has quit IRC | 17:14 | |
*** esker has joined #openstack-swift | 17:14 | |
*** klrmn has joined #openstack-swift | 17:17 | |
*** lakshmiS has joined #openstack-swift | 17:22 | |
*** dmorita has joined #openstack-swift | 17:24 | |
*** dmorita has quit IRC | 17:26 | |
*** dmorita has joined #openstack-swift | 17:26 | |
*** suyash has joined #openstack-swift | 17:31 | |
*** esker has quit IRC | 17:33 | |
*** mvk has quit IRC | 17:33 | |
*** rcernin has quit IRC | 17:47 | |
openstackgerrit | Alistair Coles proposed openstack/swift: Add Encryption to the default pipeline https://review.openstack.org/316419 | 17:54 |
*** eranrom has joined #openstack-swift | 17:55 | |
acoles | jrichli: ^^ that might sort out tox -e func on jenkins but since I cannot reproduce easily locally we'll have to wait to see. The httplib._MAXHEADERS had been set with 122 before the in process setup applied the extra_header_count in swift.conf-sample to the test swift.conf. | 17:56 |
*** sams-gleb has joined #openstack-swift | 17:56 | |
acoles | jrichli: away now | 17:56 |
jrichli | acoles: thx! looking now | 17:56 |
*** acoles is now known as acoles_ | 17:57 | |
*** pdardeau_ has joined #openstack-swift | 17:57 | |
*** pdardeau_ has left #openstack-swift | 18:00 | |
*** sams-gleb has quit IRC | 18:03 | |
*** delattec has joined #openstack-swift | 18:09 | |
*** cdelatte has quit IRC | 18:12 | |
*** delatte has joined #openstack-swift | 18:14 | |
*** delattec has quit IRC | 18:17 | |
*** StraubTW has joined #openstack-swift | 18:18 | |
*** dmorita has quit IRC | 18:18 | |
*** dmorita has joined #openstack-swift | 18:20 | |
*** dmorita has quit IRC | 18:20 | |
*** dmorita has joined #openstack-swift | 18:20 | |
*** sgundur1 has left #openstack-swift | 18:25 | |
pdardeau | jrichli: mmotiani: i started with fresh vsaio and had 22 functional tests fail. changed out the 2 occurrences of fallocate_reserve from 1% to 0 and now all functional tests pass | 18:29 |
mmotiani | pdardeau: jrichli: Yup, works fine for me too. thanks! | 18:45 |
mmotiani | Is anyone working on a patch for removing error 503 from func test? | 18:46 |
openstackgerrit | Tim Burke proposed openstack/swift: Use the same key for memcache and env['swift.infocache'] https://review.openstack.org/310566 | 18:51 |
*** ouchkernel has quit IRC | 18:59 | |
*** sams-gleb has joined #openstack-swift | 19:00 | |
*** tqtran has joined #openstack-swift | 19:04 | |
*** ouchkernel has joined #openstack-swift | 19:05 | |
pdardeau | mmotiani: maybe propose one if you'd like | 19:05 |
*** sams-gleb has quit IRC | 19:09 | |
mmotiani | pdardeau, yes, I will do that now | 19:14 |
*** arch-nemesis has quit IRC | 19:17 | |
openstackgerrit | Mohit Motiani proposed openstack/swift: Revert fallocate_reserve to fix the functional tests https://review.openstack.org/317063 | 19:29 |
*** ouchkernel has quit IRC | 19:38 | |
*** rcernin has joined #openstack-swift | 19:42 | |
*** ouchkernel has joined #openstack-swift | 19:44 | |
*** sheel has quit IRC | 19:45 | |
jrichli | mmotiani: mattoliverau was looking into possible changes to test.conf file for this. You can get some information on that from scrollback. | 19:49 |
*** bsdkurt has quit IRC | 19:54 | |
*** NM has quit IRC | 19:59 | |
*** NM has joined #openstack-swift | 19:59 | |
*** sams-gleb has joined #openstack-swift | 20:06 | |
*** sams-gleb has quit IRC | 20:13 | |
*** dmk0202 has joined #openstack-swift | 20:14 | |
*** pauloewerton has quit IRC | 20:30 | |
*** raildo is now known as raildo-afk | 20:33 | |
*** raildo-afk is now known as raildo | 20:33 | |
*** NM1 has joined #openstack-swift | 20:33 | |
*** pauloewerton has joined #openstack-swift | 20:34 | |
openstackgerrit | Thiago da Silva proposed openstack/swift: remove double checks on account/container info https://review.openstack.org/317080 | 20:35 |
*** NM has quit IRC | 20:37 | |
*** NM1 has quit IRC | 20:39 | |
*** NM has joined #openstack-swift | 20:39 | |
*** mminesh has joined #openstack-swift | 20:39 | |
*** pauloewerton has quit IRC | 20:41 | |
*** pauloewerton has joined #openstack-swift | 20:42 | |
notmyname | jrichli: what do you think of this? https://gist.github.com/notmyname/129bafae14381108e135f98f24be8c64 | 20:44 |
jrichli | notmyname: thanks, good call! | 20:45 |
jrichli | will add | 20:45 |
notmyname | I kinda wish we could add that for the hash prefix and suffix in swift.conf too, but it's probably too late | 20:45 |
timburke | notmyname: might still be worth printing a giant warning, even if we can't raise ValueError | 20:46 |
jrichli | notmyname: but .. how will things run in automated tests? | 20:46 |
notmyname | timburke: "You made a terrible mistake and there's nothing you can do about it now" <-- doesn't seem like a good warning | 20:47 |
jrichli | for automated tests, if it always gets a modified config, you can do this. but if it uses the sample ... we cant error | 20:48 |
timburke | notmyname: if you're still in POC, you can still fix it | 20:48 |
notmyname | jrichli: hmmm...cause that's what's used for comparing the crypto output and key derivation, right? | 20:48 |
notmyname | timburke: how about a friendly "RTFM n00b!" | 20:49 |
jrichli | well, this says error if the default config value (the one in our sample) is not overriden, right? | 20:49 |
notmyname | jrichli: right | 20:49 |
notmyname | jrichli: so if it's set, it better not be set to the example given in the sample config file | 20:49 |
jrichli | but sometimes our tests are setup to use the values from the sample. | 20:50 |
jrichli | then those types of tests will always fail with crypto | 20:50 |
notmyname | jrichli: what about a warning? | 20:50 |
jrichli | i'll go with that! | 20:50 |
*** CIA has quit IRC | 20:51 | |
openstackgerrit | Merged openstack/swift: Make info caching work across subrequests https://review.openstack.org/280977 | 20:58 |
*** NM has quit IRC | 21:06 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 21:09 |
*** sams-gleb has joined #openstack-swift | 21:10 | |
openstackgerrit | Tim Burke proposed openstack/swift: Make get_info requests useful with recheck_*_existence == 0 https://review.openstack.org/310216 | 21:13 |
*** sams-gleb has quit IRC | 21:19 | |
notmyname | redbo: what would it take to move your golang tests from your jenkins server into the -infra servers? or reworded: I'd like to help get those moved over and if you can give me the script you run on your server, I can start working on that | 21:21 |
*** raildo is now known as raildo-afk | 21:22 | |
*** nadeem has quit IRC | 21:29 | |
*** rnowling has quit IRC | 21:30 | |
*** lakshmiS has quit IRC | 21:34 | |
*** joeljwright has quit IRC | 21:47 | |
*** pauloewerton has quit IRC | 21:54 | |
redbo | notmyname: I don't know how the infra servers work. If we already know how to get swift running, and we can get go on there swift, it shouldn't be too hard. | 22:09 |
notmyname | I know a little about how the infra servers work. I just don't know what to run on them :-) | 22:09 |
notmyname | together, we will overcome! | 22:09 |
redbo | notmyname: these are my dockerfile and the test script I use. https://gist.github.com/redbo/d7133a069341a2f17119 | 22:09 |
*** dmk0202 has quit IRC | 22:10 | |
notmyname | redbo: perfect, thanks. | 22:11 |
*** jasongalyon has joined #openstack-swift | 22:13 | |
*** sams-gleb has joined #openstack-swift | 22:16 | |
*** _JZ_ has quit IRC | 22:22 | |
*** rickyrem has joined #openstack-swift | 22:22 | |
*** sams-gleb has quit IRC | 22:23 | |
*** dmorita has quit IRC | 22:25 | |
*** NM has joined #openstack-swift | 22:26 | |
*** jasongalyon has quit IRC | 22:27 | |
*** thumpba has quit IRC | 22:30 | |
*** dmorita has joined #openstack-swift | 22:34 | |
mattoliverau | jrichli: I know what is causing it. The max_file_size test makes a few requests with content_length of max_file_size +/- some size for bounds testing, but then doesn't send anything and timesout. It works great, execpt when you have fallocate_reserve on, in which case the fallocate syscall is made based on the length (content_length) in the request. This means the bytes are temp reserved while the request timesout. So | 22:34 |
mattoliverau | you either need to change the max_file_size of your SAIO swift.conf to something smaller (And that depends on your xfs loopback or partitions (say to a 3rd). Or turn off fallocate_reserve. I'll change the SAIO's object-servers to have it turned off and add a note about this to the saio documentation. | 22:34 |
mattoliverau | notmyname: ^ | 22:34 |
mattoliverau | oh and good morning | 22:35 |
openstackgerrit | Brian Ober proposed openstack/swift: Storage Account Object Count Quota Support https://review.openstack.org/312603 | 22:36 |
*** thumpba has joined #openstack-swift | 22:38 | |
*** wadeholler has left #openstack-swift | 22:41 | |
*** ametts has quit IRC | 22:41 | |
*** StraubTW has quit IRC | 22:42 | |
tdasilva | mattoliverau: o/ | 22:44 |
openstackgerrit | Tim Burke proposed openstack/swift: Make get_info requests useful with recheck_*_existence == 0 https://review.openstack.org/310216 | 22:47 |
notmyname | mattoliverau: I only vaguely remember the details of what you were looking in to | 22:47 |
mattoliverau | notmyname: lol, nps, Let me put together a SAIO doc update then you'll understand :) | 22:49 |
notmyname | mattoliverau: I mean, I kinda understand what you just said (small drive + large max_obj_size + some buffering = 507 error) | 22:50 |
notmyname | drive size < 2*max_obj_size < (drive size + fallocate_reserve size) | 22:51 |
notmyname | right? | 22:51 |
notmyname | actually... | 22:51 |
notmyname | (2*max_obj_size) < (drive size - fallocate_reserve size) | 22:51 |
mattoliverau | Yeah, something like that :) | 22:55 |
openstackgerrit | David Goetz proposed openstack/swift: go: have replicateDevice go routine loop per device https://review.openstack.org/317125 | 22:55 |
mattoliverau | we already have disable_fallocate in the saio configs. So I'll confirm it's actually working and then make a doc update to warn peeps. | 22:55 |
mattoliverau | its like: if free < fallocate_reserve: 507 else fallocate(). Where free == drive_free_space - content-length. Cause we are making sure of the state post the object is put down. In the test one of the replicas causes the 507, cause the first one has fallocated an amount albeit only temporary. | 22:58 |
*** klamath has quit IRC | 23:02 | |
*** rnowling has joined #openstack-swift | 23:05 | |
*** rcernin has quit IRC | 23:08 | |
*** rnowling has quit IRC | 23:13 | |
*** CaioBrentano has quit IRC | 23:15 | |
*** sams-gleb has joined #openstack-swift | 23:20 | |
*** lakshmiS has joined #openstack-swift | 23:23 | |
*** lcurtis has quit IRC | 23:25 | |
*** tqtran has quit IRC | 23:26 | |
*** sams-gleb has quit IRC | 23:29 | |
*** diogogmt has quit IRC | 23:29 | |
*** wadeholler has joined #openstack-swift | 23:30 | |
*** kei_yama has joined #openstack-swift | 23:32 | |
*** dmorita has quit IRC | 23:42 | |
*** dmorita has joined #openstack-swift | 23:43 | |
*** dmorita has quit IRC | 23:43 | |
*** dmorita has joined #openstack-swift | 23:43 | |
*** vint_bra has quit IRC | 23:48 | |
*** jamielennox is now known as jamielennox|away | 23:51 | |
*** hosanai has joined #openstack-swift | 23:51 | |
*** ChanServ sets mode: +v hosanai | 23:51 | |
*** dmorita has quit IRC | 23:51 | |
*** dmorita has joined #openstack-swift | 23:52 | |
*** dmorita has quit IRC | 23:53 | |
*** dmorita has joined #openstack-swift | 23:53 | |
*** NM has quit IRC | 23:54 | |
openstackgerrit | Merged openstack/swift: Fix up get_account_info and get_container_info https://review.openstack.org/280978 | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!