Thursday, 2017-08-24

« Wednesday, 2017-08-23 Index Friday, 2017-08-25 »
*** rcernin has quit IRC00:12
openstackgerritTim Burke proposed openstack/swift master: proxy-server: errors talking to backend servers don't need tracebacks  https://review.openstack.org/49696500:12
*** cshastri has joined #openstack-swift00:55
*** itlinux has joined #openstack-swift00:57
*** eckesicle has quit IRC01:01
*** two_tired has joined #openstack-swift01:03
*** klrmn has quit IRC01:07
*** itlinux has quit IRC01:09
*** cshastri has quit IRC01:12
*** kiennt has joined #openstack-swift01:13
*** cshastri has joined #openstack-swift01:13
*** shuyingya has joined #openstack-swift01:19
*** vint_bra has joined #openstack-swift01:37
*** vint_bra has quit IRC01:39
*** klrmn has joined #openstack-swift02:23
*** silor has joined #openstack-swift03:10
*** links has joined #openstack-swift03:18
*** gkadam has joined #openstack-swift03:24
*** itlinux has joined #openstack-swift03:42
*** rmcall has quit IRC03:55
*** rmcall has joined #openstack-swift03:57
*** itlinux has quit IRC04:00
*** itlinux has joined #openstack-swift04:02
*** two_tired has quit IRC04:09
*** mkrcmari__ has joined #openstack-swift04:22
mahaticnotmyname: thanks for the info on meeting04:24
* notmyname wonders about that more people respond about *not* meeting than actually attend the meeting04:25
notmyname;-)04:25
mahaticnotmyname: lol04:27
mahaticwell we don't get a special ping for our nicks when there's a meeting ;)04:27
notmynameheh04:28
mahaticwe just show up04:28
notmynameI should add that to patchbot or something04:28
mahaticI think we're used to the timings by now. I have a reminder on my phone for the meeting though :)04:29
mahatickota_: happy birthday!04:30
*** gyee has quit IRC04:31
*** mvk_ has joined #openstack-swift04:45
*** mkrcmari__ has quit IRC04:49
*** itlinux has quit IRC04:54
zaitcevI hate Oslo so much. Stupid crapware.04:54
zaitcevGnocchi uses it, unfortunately. As a result, one cannot rebuild gnocchi on a system that happens to have gnocchi installed already, because Oslo's config generator invokes stevedore, and that one ignores PYTHONPATH, loads installed gnocci instead of the one being built, blows up.04:55
zaitcevWhy did I agree to this04:56
mattoliverauzaitcev: :(05:10
zaitcevalso, never IRC when you're about done and need to sleep05:10
*** rcernin has joined #openstack-swift05:11
*** rcernin has quit IRC05:27
*** silor has quit IRC05:28
*** silor has joined #openstack-swift05:33
*** kiennt has quit IRC05:34
*** kiennt has joined #openstack-swift05:55
*** kiennt has quit IRC06:09
*** rcernin has joined #openstack-swift06:12
*** mvk has joined #openstack-swift06:12
*** mvk_ has quit IRC06:16
*** ChubYann has quit IRC06:16
*** pcaruana has joined #openstack-swift06:21
*** skudlik has joined #openstack-swift06:36
*** klrmn has quit IRC06:38
*** joeljwright has quit IRC06:45
*** hseipp has joined #openstack-swift06:46
*** oshritf has joined #openstack-swift07:04
*** hseipp has quit IRC07:04
*** cschwede_ has joined #openstack-swift07:11
*** tesseract has joined #openstack-swift07:20
*** shuyingy_ has joined #openstack-swift07:39
*** shuyingya has quit IRC07:39
*** cbartz has joined #openstack-swift07:39
*** gkadam_ has joined #openstack-swift07:40
*** geaaru has joined #openstack-swift07:42
*** gkadam has quit IRC07:42
openstackgerritOpenStack Proposal Bot proposed openstack/swift master: Imported Translations from Zanata  https://review.openstack.org/49613707:42
*** gkadam_ has quit IRC07:45
*** tesseract-RH has joined #openstack-swift07:49
*** tesseract-RH has quit IRC07:50
*** tesseract-RH has joined #openstack-swift07:50
*** tesseract has quit IRC07:52
*** silor has quit IRC07:53
*** tovin07_ has joined #openstack-swift07:59
*** tesseract-RH has quit IRC08:01
*** kiennt has joined #openstack-swift08:02
*** tesseract-RH has joined #openstack-swift08:02
*** hseipp has joined #openstack-swift08:02
acolesgood morning08:03
acoleskota_: happy birthday !08:03
*** bkopilov has joined #openstack-swift08:22
*** rmcall has quit IRC08:34
*** psachin has joined #openstack-swift08:36
*** joeljwright has joined #openstack-swift08:50
*** ChanServ sets mode: +v joeljwright08:50
*** joeljwright has quit IRC08:52
*** joeljwright has joined #openstack-swift08:52
*** ChanServ sets mode: +v joeljwright08:52
*** mvk has quit IRC09:03
*** cshastri has quit IRC09:06
*** mvk has joined #openstack-swift09:07
*** tesseract-RH has quit IRC09:18
*** mvk has quit IRC09:20
*** tesseract has joined #openstack-swift09:23
*** psachin has quit IRC09:31
*** kei_yama has quit IRC09:33
*** tovin07_ has quit IRC10:10
*** kiennt has quit IRC10:37
*** eckesicle has joined #openstack-swift10:51
*** mat128 has joined #openstack-swift11:39
*** silor has joined #openstack-swift12:22
*** shuyingy_ has quit IRC12:35
*** mat128 has quit IRC12:55
*** catintheroof has joined #openstack-swift13:02
*** eckesicle has quit IRC13:12
*** lucasxu has joined #openstack-swift13:15
*** baojg has joined #openstack-swift13:20
*** baojg has quit IRC13:40
*** baojg has joined #openstack-swift13:42
*** baojg has quit IRC13:43
*** links has quit IRC13:46
*** baojg has joined #openstack-swift13:48
*** shuyingya has joined #openstack-swift13:53
*** skudlik has quit IRC14:00
*** shuyingya has left #openstack-swift14:02
*** mat128 has joined #openstack-swift14:03
*** mat128 has quit IRC14:05
*** baojg has quit IRC14:14
*** gyee has joined #openstack-swift14:17
*** baojg has joined #openstack-swift14:21
*** mat128 has joined #openstack-swift14:21
*** baojg has quit IRC14:23
*** baojg has joined #openstack-swift14:26
openstackgerritPeter Lisák proposed openstack/swift master: Recon's get_diskusage and get_unmounted return all devs.  https://review.openstack.org/34880314:28
*** srbgupta has joined #openstack-swift14:36
*** mat128 has quit IRC14:36
*** jistr|trng is now known as jistr15:09
*** cbartz has quit IRC15:17
*** baojg has quit IRC15:19
*** baojg has joined #openstack-swift15:24
*** baojg has quit IRC15:36
*** SkyRocknRoll has joined #openstack-swift15:36
*** baojg has joined #openstack-swift15:39
*** oshritf has quit IRC15:43
*** mat128 has joined #openstack-swift15:43
*** mat128 has quit IRC15:45
*** vint_bra has joined #openstack-swift15:48
*** mat128 has joined #openstack-swift15:48
*** vint_bra has quit IRC15:52
*** itlinux has joined #openstack-swift15:52
*** rcernin has quit IRC15:54
*** pcaruana has quit IRC15:57
*** ianychoi has joined #openstack-swift16:00
*** tesseract has quit IRC16:03
*** itlinux has quit IRC16:03
*** openstackgerrit has quit IRC16:04
*** links has joined #openstack-swift16:06
*** joeljwright has quit IRC16:10
*** itlinux has joined #openstack-swift16:13
*** oshritf has joined #openstack-swift16:17
*** vint_bra has joined #openstack-swift16:20
*** silor has quit IRC16:22
*** hseipp has quit IRC16:26
timburkegood morning16:29
*** oshritf has quit IRC16:30
*** mvk has joined #openstack-swift16:34
*** links has quit IRC16:38
notmynamegood morning16:38
*** oshritf has joined #openstack-swift16:38
*** chsc has joined #openstack-swift16:40
*** chsc has joined #openstack-swift16:40
acolestimburke: notmyname o/16:40
*** geaaru has quit IRC16:46
*** psachin has joined #openstack-swift16:51
*** vint_bra has quit IRC16:55
*** oshritf has quit IRC16:56
tdasilvatimburke: hi! just started playing with swift3 and I'm running into some "The request signature we calculated does not match the signature you provided" error, got any ideas?17:09
tdasilvatimburke: this is a swift3+keystone env17:09
timburketdasilva: v2 or v4 signatures? do you know? also, out of curiosity, what version of keystone?17:10
tdasilvai do not know about v2 or v4, but keystone is pike17:10
tdasilvaI can see the error is returned by swift3 middleware and not s3token17:11
timburkeyay! it has https://github.com/openstack/keystone/commit/3ec1aa4 then. ocata is sad :-(17:11
timburketdasilva: got some logs for me to look at?17:12
timburkemy first guess ('cause it comes up a decent bit) is that we're working with v4 signatures and the region doesn't match17:13
tdasilvatimburke: http://paste.openstack.org/show/619341/17:13
timburkei stand corrected -- that string to sign is definitely v217:16
timburkewhat's the keystone log look like? also, you might want to switch to using the version of s3token shipped with swift317:19
*** klrmn has joined #openstack-swift17:23
*** rcernin has joined #openstack-swift17:27
tdasilvatimburke: one sec17:28
timburkealso, what's the pipeline? did you *need* to disable auth_pipeline_check? i always felt like that was mostly for non-keystone, non-tempauth auth systems...17:29
tdasilvapipeline = catch_errors healthcheck proxy-logging cache ratelimit bulk tempurl formpost authtoken keystone swift3 s3token staticweb copy container_quotas account_quotas slo dlo versioned_writes proxy-logging proxy-server17:31
tdasilvain swift3 it's hardcoded to be something like keystoneauth17:32
tdasilvaso i just disabled auth_pipeline_check17:32
timburkegot it. i'm pretty sure we need something like ... authtoken swift3 s3token keystone ... though -- s3token just populates the env similar to authtoken with delay_auth_decision, and relies on keystoneauth to make final decisions (with the extra swift context of ACLs and the like)17:35
tdasilvacrap, i knew that, copy and paste error17:38
tdasilvaok, so i made that change and will also change to use the s3token shipped with swift317:38
*** catinthe_ has joined #openstack-swift17:44
*** catintheroof has quit IRC17:45
tdasilvatimburke: http://paste.openstack.org/show/619345/17:52
tdasilvalol i think i delete line 160 by mistake: auth_pipeline_check=False is still there17:54
tdasilvait's there on the config file, deleted on the paste by mistake17:54
notmynametdasilva: he wants you to know he's getting coffee17:55
tdasilvanotmyname: heh, thank you...17:56
timburke:-( that log seems so anaemic...17:59
timburkei'm so used to seeing things like "Calling S3Token middleware." "Connecting to Keystone sending this JSON: ..." "Keystone Reply: Status: 200, Output: ..."18:00
timburkemaybe i'm just too used to debug level from gates like http://logs.openstack.org/85/482485/3/check/gate-swift3-tox-keystone-ubuntu-xenial/d714017/tox/proxy.log.txt.gz ?18:01
timburkeoh! *there's* the problem! of course...18:01
timburke> s3curl: StringToSign='GET\n\n\nThu, 24 Aug 2017 14:53:16 +0000\n/192.168.24.11/bucket'18:02
timburke> Aug 24 14:53:14 localhost proxy-server: {... 'string_to_sign': 'GET\n\n\nThu, 24 Aug 2017 14:53:16 +0000\n/bucket', ...18:02
tdasilvatimburke: but is it possible that the request is not even getting to s3token? it seems like i'm getthing this: https://github.com/openstack/swift3/blob/master/swift3/request.py#L29118:02
timburkegotta convince s3curl that the ip address isn't part of the bucket name18:02
timburkedon't think so... again, that string to sign looks like you're using v2 signatures, and that error only comes up for v418:04
*** itlinux has quit IRC18:08
claygtimburke: not sure on patch 496535, I put it on the meeting agenda18:17
patchbothttps://review.openstack.org/#/c/496535/ - swift - Simplify testing for logging at error vs exception18:17
timburkei did it through the log lines interface because we almost never use the raw log dict interface. as a result, we'd often end up with tracebacks for things where a traceback just isn't useful, as evidenced by the string of patches leading up to that one18:21
timburkefwiw18:21
timburkeand of the places where we *do* test with the log_dict, i don't think *any* of them test the negative -- they only ever check that we *do* log a traceback18:24
timburkethat patch is *also* why i hesitated in adding tests with https://review.openstack.org/#/c/496965/ -- it seems like they'd surely conflict, and i didn't want to make that patch chain any longer than it already was18:26
patchbotpatch 496965 - swift - proxy-server: errors talking to backend servers do...18:26
timburkethough i guess since i tweaked the message format, maybe it isn't *so* bad...18:28
claygtimburke: ^ considering trying to capture some of that reasoning/justification in gerrit for that patch - it'll probably be useful context later.18:37
*** cschwede_ has quit IRC18:47
tdasilvatimburke: so the way it chooses v1 or v4 is just based on the headers sent with the request, right?18:52
tdasilvasorry..v2 or v418:53
timburketdasilva: yup. v2 requests look like `Authorization: AWS <user>:<sig>` while v4 looks like `Authorization: AWS4-HMAC-SHA256 Credential=<user>/<date>/<region>/s3/aws4_request, SignedHeaders=<headers>, Signature=<sig>`18:57
*** rmcall has joined #openstack-swift18:58
*** ChubYann has joined #openstack-swift19:06
*** openstackgerrit has joined #openstack-swift19:09
openstackgerritTim Burke proposed openstack/swift master: proxy-server: errors talking to backend servers don't need tracebacks  https://review.openstack.org/49696519:09
tdasilvatimburke: what's interesting is that I tried testing with s3cmd, and just running `s3cmd ls` creates a v2 request, but `s3cmd ls s3://bucket` creates a v4 request19:13
timburkecurious...19:14
tdasilvai've tried enabling debug logging in keystone, but i'm not really sure how to do it19:14
timburkelooks like it's coming out of this logic: https://github.com/s3tools/s3cmd/blob/master/S3/S3.py#L155-L16319:16
tdasilvayep19:18
*** SkyRocknRoll has quit IRC19:30
timburkeboth should work though -- are you running into trouble with one or the other?19:34
tdasilvaboth19:37
tdasilvabut different errors19:37
tdasilvawith v4 i get InvalidAccessKeyId19:37
tdasilvawhich is weird, because I can't see that even being used in the swift3 code source at all19:38
tdasilvaso i'm still trying to get my head around the code19:38
timburkeyeah, that InvalidAccessKeyId makes me wonder if it's hitting actual s3...19:43
timburkein swift3 we really don't have a good way to check whether a user exists or not19:44
tdasilvait was not, it was my config mistake, it is corrected now, but back to same previous error of SignatureDoesNotMatch19:52
tdasilvaso at least s3curl and s3cmd are matching??? fwiw19:52
timburkewhat's the config look like? are the clients still claiming that the HTTP path in the string to sign should be "/192.168.24.11/bucket"?19:54
timburkemight work better to have an entry in /etc/hosts so you can use a hostname?19:55
*** psachin has quit IRC19:57
*** psachin has joined #openstack-swift19:59
tdasilvatimburke: so now with s3cmd I'm sure that the request is a v4 request...getting some logs for you...20:00
tdasilvatimburke: you might notice some extra debug statements in the log:  http://paste.openstack.org/show/619360/20:02
timburkeclayg: i can't decide whether to love or hate https://github.com/openstack/swift/blob/master/swift/common/utils.py#L173520:03
timburkemaybe you're right and i should carve out an exception for BadStatusLine there -- it seems like we should *never* use logger.error()? or we should just say error = exception in the class def20:03
timburkeand make sure that *all* of our normal testing infra includes an adapter?20:04
timburkesuck a giant yak!20:04
notmynametimburke: once you shave the yak, you'll have plenty of hair to make the paintbrushes for recoloring the bikeshed20:08
timburketdasilva: why's the canonical request say "location=" for the query string when the actual request that's later sent is for "/bucket/?location" (no equals-sign)? i know i've gotten other reports about this... but it seems like a client bug20:08
timburkeheh. such* a giant yack. freudian slip20:09
*** mat128 has quit IRC20:11
tdasilvatimburke: idk?? i'm pretty certain it's some config error on my part otherwise things would not be working on the gate, just not sure how to track it down yet: http://paste.openstack.org/show/619361/20:12
timburkeclayg: like, i understand the convenience of having some centralized logging infra like that for whether or not to write a traceback, but why does it break from the stdlib like that? https://docs.python.org/2/library/logging.html#logging.Logger.exception reads *just* like Logger.error except for the "Exception info is always added to the logging message."20:12
claygtimburke: since it's been there longer than you or I have been alive we don't have to have a love/hate feeling toward it - we can just accept it as "being"20:13
timburketdasilva: gate? which gate? maybe there's some difference in s3cmd versions?20:14
tdasilvatimburke: i meant to say i'm keystone+swift3 func test gate. so i don't think it's a problem with the software itself..more like how i'm configuring it...but yeah, could be s3cmd itself...20:16
tdasilvatimburke: i just grabbed the latest from s3cmd github...20:16
timburkeclayg: if we're going to accept it as "being" (rather than rip it out and follow stdlib conventions), then i'm thinking we should just alias error to exception20:17
timburkeand throw in a elif exc is not None before setting call = self._exception20:19
claygtimburke: that doesn't sound so unreasonable to me... basically just further merge the two... main difference being you could call logger.error w/o having anything useful in sys.exc_info and maybe something weird... idk - you could try it - sounds like more yak20:29
*** itlinux has joined #openstack-swift20:30
*** lucasxu has quit IRC20:31
timburkethe yak is there regardless, and i can't unsee it. at least the tracebacks from https://review.openstack.org/#/c/496150/ got cleaned up. which only happened because we were using the root logger!20:31
patchbotpatch 496150 - swift - Socket errors don't warrant tracebacks when talkin... (MERGED)20:31
timburkehttps://github.com/openstack/swift/blob/master/swift/common/utils.py#L1735 doesn't do us *any good* if we're just using logging.exception20:32
tdasilvatimburke: need to step out, thanks for the help, will continue later and let you know what i found20:33
*** psachin has quit IRC20:36
*** zaitcev has quit IRC20:40
*** zaitcev has joined #openstack-swift20:41
*** ChanServ sets mode: +v zaitcev20:41
openstackgerritMerged openstack/swift master: Use more specific asserts in test/unit/account tests  https://review.openstack.org/34276620:44
*** catinthe_ has quit IRC20:53
*** catintheroof has joined #openstack-swift20:54
*** catintheroof has quit IRC20:58
*** itlinux has quit IRC21:03
*** itlinux has joined #openstack-swift21:05
*** rmcall has quit IRC21:07
*** rcernin has quit IRC21:19
*** itlinux has quit IRC21:24
*** itlinux has joined #openstack-swift21:25
*** srbgupta has quit IRC21:57
*** catintheroof has joined #openstack-swift21:59
*** itlinux has quit IRC23:00
mattoliveraumorning23:04
*** chsc has quit IRC23:12
*** kei_yama has joined #openstack-swift23:39
*** mbirru has quit IRC23:39
*** mbirru has joined #openstack-swift23:40
*** catintheroof has quit IRC23:46
« Wednesday, 2017-08-23 Index Friday, 2017-08-25 »

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