Friday, 2016-04-22

*** jamielennox is now known as jamielennox|away00:07
notmynameanyone have a rhel/centos box handy?00:11
notmynamewhat is the name of the rpm package that installs rsync?00:11
clarkbnotmyname: just 'rsync'00:12
notmynameclarkb: thanks. (same as dpkg, that's nice)00:12
clarkbrsync.x86_64 : A program for synchronizing files over a network is from yum search rsync00:12
* notmyname is working on other-requirements.txt00:12
clarkbcool, keystone was working on it earlier today as well00:13
clarkbwill be good to see projcets use that to better communicate deps this cycle00:13
notmynameclarkb: I noticed that python isn't in http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/data/bindep-fallback.txt. will that be needed with 16.04? I'm told that's not installed by default now00:15
notmynameie py2700:15
notmynameclarkb: also, do you know if other-reqs is synced by the global reqs bot?00:15
clarkbI would list it, we do preinstlal it as it is required for zuul-cloner and other things. But this way you will work on xenial without our images and it will just noop on our images00:15
clarkbthere is no sync for global reqs and other reqs00:16
openstackgerritJohn Dickinson proposed openstack/swift: created other-requirements.txt  https://review.openstack.org/30921900:16
notmynameok00:16
clarkbat least I don't think there is, gets complicated becuase the distros control versions and stuff00:16
notmynameclarkb: so now what does a test passing on that patch mean? that everythign is good?00:17
clarkbyes, if the unittests pass on that patch it means you successfully installed enough of the dependencies to run your unittests00:17
notmynameyay00:17
notmynameor at least enough were installed00:17
notmynameie something like "python" that's required but might not be specified and it stillworks because it was included by default or something00:18
clarkbya00:18
notmynameclarkb: can you please check the format on that patch, please? I started with the default one, but I'm not really sure that i got it all right00:18
notmynamealso, I'm not 100% sure when the -dev packages are needed. but they were in the default, so I kept them00:19
clarkbyou will need -dev when your python packages link against them00:20
clarkblibxml2-dev and libxslt-dev are required if you depend on lxml for example00:20
notmynameright. I know that one. not sure about the sqlite or libEC dev packages though00:20
notmynamei mean, I think they're probably required (which is why I kept them there). I just don't know exactly where they're used, off the top of my head00:21
clarkbthe ec dev is required since they split it out into distro packages instead of bundling the C headres00:21
clarkbthe python ec package links against the C stuff00:21
clarkband python-dev is needed when python links to C, that list looks fine to me00:22
*** tqtran has quit IRC00:23
notmynameclarkb: thanks00:23
notmynameI compared this against the dependencies swiftstack has in our product for our swift packages00:24
clarkbyou probably don't have the dev stuff there because the packages are prebuilt, but otherwise should line up00:25
notmynameclarkb: so when we get to a point of readding a pypy gate job, we'd have to add pypy to this file, right?00:26
*** hosanai has joined #openstack-swift00:26
*** ChanServ sets mode: +v hosanai00:26
clarkbcorrect00:26
notmynamegot it! thanks :-)00:26
*** asettle has quit IRC00:27
*** asettle has joined #openstack-swift00:27
*** arch-nemesis has quit IRC00:28
*** garthb has quit IRC00:29
*** lyrrad has quit IRC00:34
*** NM has quit IRC00:35
*** NM has joined #openstack-swift00:36
*** itlinux has quit IRC00:43
*** mingdang1 has joined #openstack-swift00:45
*** itlinux has joined #openstack-swift00:54
*** haomaiwang has joined #openstack-swift01:04
*** can8dnSix has quit IRC01:06
*** haomaiwang has quit IRC01:09
*** lcurtis has quit IRC01:09
*** itlinux has quit IRC01:13
*** NM has quit IRC01:13
*** ozialien has quit IRC01:23
*** itlinux has joined #openstack-swift01:25
*** haomaiwang has joined #openstack-swift01:25
kota_hello01:26
kota_timburke, mattoliverau: thanks for reviewing the versioned_writes patch, it seemed to land successfully.01:27
*** mingdang1 has quit IRC01:28
*** itlinux has quit IRC01:28
mattoliveraukota_: morning!01:29
kota_mattolivearu: morning!01:30
*** haomaiwang has quit IRC01:30
*** asettle has quit IRC01:39
*** asettle has joined #openstack-swift01:39
*** jamielennox|away is now known as jamielennox01:44
timburkekota_: yeah, i'm happy with it. definitely making the rebase for https://review.openstack.org/#/c/214922/ painful though :(01:45
patchbottimburke: patch 214922 - swift - Add "history" mode to versioned_writes middleware01:45
*** mwheckmann has joined #openstack-swift01:48
kota_timburke: yeah, I have a expositive patch for versioned_writes (https://review.openstack.org/#/c/209315/) which definitely needs to be rebased :/01:48
patchbotkota_: patch 209315 - swift - Refactor versioned_writes middleware01:48
kota_but it's great we are getting progress the versioned_writes.01:49
timburkeagreed01:49
kota_timburke: ah, and thanks for so many comments on swif3 sigv4 pathc.01:51
kota_patch.01:51
timburkeyeah, i think no matter what, it's going to be a big one01:51
kota_I noticed, but didn't have time to look at yet.01:51
timburkeno worries, i understand. i'll see if i can propose a follow-up (or two?) to address most of the issues before austin01:52
kota_sounds great, it (they?) will be definitely helpful for me :)01:53
timburkeas much as possible, i want to minimize how much you actually *have* to do01:53
kota_and hopefully i wish I have the time to see it too.01:53
kota_exactly,01:53
*** Suyi has quit IRC01:54
kota_in the patch, I was working as a few changes from original authors one but actually, we need to make it to good shape as a little different from the master.01:55
kota_thanks man.01:55
*** ozialien has joined #openstack-swift02:04
openstackgerritTim Burke proposed openstack/swift: Add "history" mode to versioned_writes middleware  https://review.openstack.org/21492202:05
*** _JZ_ has quit IRC02:07
*** lakshmiS has quit IRC02:11
*** EkulTails has quit IRC02:13
*** _JZ_ has joined #openstack-swift02:17
*** ozialien has quit IRC02:18
*** itlinux has joined #openstack-swift02:20
*** dmorita has quit IRC02:29
*** jamielennox is now known as jamielennox|away02:32
*** dmorita has joined #openstack-swift02:33
*** dmorita has quit IRC02:34
*** dmorita has joined #openstack-swift02:35
*** dmorita has quit IRC02:38
*** dmorita has joined #openstack-swift02:38
*** m_kazuhiro has joined #openstack-swift02:41
*** gyee has quit IRC02:48
*** mwheckmann has quit IRC02:48
*** Guest14041 has quit IRC03:07
*** kei_yama has quit IRC03:08
*** _JZ_ has quit IRC03:13
*** rickyrem has quit IRC03:20
janonymousbkeller:yes, please go ahead03:26
jrichlibkeller` ^^03:27
bkeller`alright, thanks :D03:27
*** natarej has quit IRC03:33
openstackgerritJohn Dickinson proposed openstack/swift: created other-requirements.txt  https://review.openstack.org/30921903:33
*** asettle has quit IRC03:40
mahaticgood morning03:42
*** asettle has joined #openstack-swift03:42
mahaticacoles_: thanks for the updates03:42
mattoliveraumahatic: morning03:47
*** garthb has joined #openstack-swift03:47
*** m_kazuhiro has quit IRC03:56
*** mingdang1 has joined #openstack-swift04:02
*** dmorita has quit IRC04:04
*** dmorita has joined #openstack-swift04:05
*** ozialiendoze has quit IRC04:05
*** ozialien has joined #openstack-swift04:06
*** dmorita has quit IRC04:07
*** dmorita has joined #openstack-swift04:07
*** ozialien_ has joined #openstack-swift04:08
*** ozialien has quit IRC04:10
*** ozialien_ is now known as ozialien04:10
*** ozialien10 has joined #openstack-swift04:11
openstackgerritKota Tsuyuzaki proposed openstack/swift: CompositeRing Support  https://review.openstack.org/27192004:11
openstackgerritKota Tsuyuzaki proposed openstack/swift: Fix EC ring validation at ring reload  https://review.openstack.org/26899304:11
mahaticmattoliverau: hello04:14
*** _JZ_ has joined #openstack-swift04:21
*** klrmn has quit IRC04:22
*** esker has joined #openstack-swift04:25
*** haomaiwang has joined #openstack-swift04:26
*** ozialien10 has quit IRC04:28
*** ozialien has quit IRC04:29
*** haomaiwang has quit IRC04:31
*** asettle has quit IRC04:41
*** haomaiwang has joined #openstack-swift04:45
*** haomaiwang has quit IRC04:46
*** haomaiwang has joined #openstack-swift04:46
*** asettle has joined #openstack-swift04:51
*** dmorita has quit IRC04:52
*** asettle has quit IRC04:55
*** nadeem has joined #openstack-swift04:58
*** rcernin has joined #openstack-swift05:03
*** garthb has quit IRC05:04
*** esker has quit IRC05:06
*** haomaiwa_ has joined #openstack-swift05:13
*** haomaiwang has quit IRC05:14
claygredbo: what version of golang are you guys using?  1.4 seems to old 1.6 might be too new?05:16
redboI'm using 1.605:17
clayghrmmmm..05:17
claygimports runtime: C source files not allowed when not using cgo or SWIG: atomic_amd64x.c defs.c float.c heapdump.c lfstack.c malloc.c mcache.c mcentral.c mem_linux.c mfixalloc.c mgc0.c mheap.c msize.c os_linux.c panic.c parfor.c proc.c runtime.c signal.c signal_amd64x.c signal_unix.c stack.c string.c sys_x86.c vdso_linux_amd64.c05:17
clayg1.6.205:17
claygpackage github.com/openstack/swift/go/bench05:17
*** mingdang1 has quit IRC05:18
redboHmm.. Do you have gcc installed?05:21
*** SkyRocknRoll has joined #openstack-swift05:23
clayginternet seemd to think i didn't know how to upgrade go `rm -fr /usr/local/go` and reinstall seemed to help get me to a different error anyway05:23
*** esker has joined #openstack-swift05:26
*** amit213 has quit IRC05:27
*** amit213 has joined #openstack-swift05:27
*** esker has quit IRC05:31
*** bluetefn has joined #openstack-swift05:32
*** haomaiwa_ has quit IRC05:34
claygwhy is $GOPATH/bin/cmd a thing and $GOPATH/src/github.com/openstack/swift/go/bin/hummingbird also a thing?05:37
claygi guess the Makefile is doing some fancyness with -o and the go tools are bitter about src/github.com/openstack/swift not being a go project (you have to go into swift/go and there's not even a main.go there)05:38
redboI hadn't noticed that.  Seems like cmd gets built when "go get" is ran.05:41
claygdifferent things annoy different people i 'spose05:41
bluetefnI'd like to run more than 2 proxy server in a swift cluster  by using haproxy. is it possible?05:43
redboThe cmd subpackage was one of Dave Cheney's suggestions, but it's possible I'm not doing it right.05:43
redbobluetefn: yes, that is pretty normal.  We use haproxy to terminate ssl and load balance to a bunch of proxies.05:44
claygnotmyname: what was the ssl terminator that we were bitching about in bristol - the one that was terrible that used before stud?05:46
notmynamepound05:46
claygah pound :)05:46
claygredbo: if `hummingbird start object` doesn't fork when it spins up the four servers how am I supposed to kill one to test handoffs and stuff in dev?05:49
claygoh wait, hold the phone - it forks05:50
claygjust always the same parent proc showing up in netstat (i guess it listens's before fork and priv drop)05:50
redboI don't think it can do that.  Yet anyway.05:51
redboAt least not easily05:51
claygoh no that's my bad - no fork - just threads05:51
claygi guess - idk how linux works - point is `ps -eLf` shows four guys - but no matter what I kill they all die05:52
portanteresiliency05:52
redboYeah, threads05:52
clayghow many sockets do you have the hummingbird process listening on when you deploy for realzy?05:53
redboTwo, I think?  There's no reason to do it that way, but we already had replication on a separate port in our rings.05:53
claygdo you *have* a 1.conf 2.conf for replication or anything?  of that entire feature of spinning up multiple servers listening on mutiple procs for ever conf/dir in /etc/swift/object-server/ is just a feature for dev05:53
redboYeah, we have two configs05:54
clayg... but you cant manage them seperately with the hummingbird "daemon command" thingy05:54
redboNope05:54
claygk05:54
*** ppai has joined #openstack-swift06:01
*** esker has joined #openstack-swift06:07
redboNot to say it couldn't be added.  It's sort of close to being able to launch multiple processes.  It just doesn't know how to write different named pid files for them.06:08
*** esker has quit IRC06:12
claygredbo: honestly i'd be happiest just plumbing `hummingbird object -d` into swift-init instead of swift-object-server - but i'm sure you have different opinions on that06:12
claygi just hate having to type `swift-init restart proxy account container; hummingbird start object` when `swift-init restart main` is all I really want to say06:13
clayglike I give a %^&* what language it is as an operator writing init scripts?06:13
*** david-lyle has quit IRC06:16
ahaleus operators rarely give a %^&* about anything at all !06:17
redboYou do what I tell you!06:17
redboOr I'll make the hummingbird binary longer and harder to type06:18
ahalehehe06:18
ahaleyou have no idea how much of my day at the moment is spent tying swif^D^D^D^Dhumm06:18
openstackgerritKota Tsuyuzaki proposed openstack/swift: Fix EC ring validation at ring reload  https://review.openstack.org/26899306:20
clayglike i said, different things annoy differnt people06:22
claygi can't imagine living like this - but i'm sort of known for lacking imagination06:22
claygalso fuck sudo and fuck secure_path06:22
ahaleidk - i could put up with almost anything for the replication features hbird has06:24
claygahale: funny thing!  I'm hearing this buzz about "replication features" in hbird and I must have missed a commit somewhere because the last psudeo-interesting thing I saw was the replicator growing a http server and getting a rescueparts command...06:27
claygahale: but I've been thinking a lot about how crappy replication/rebalance is recently and was wondering how many of these conclusions about what needs to be fixed first I had to independently discover06:27
clayg... since ya'll weren't in bristol for us to catch up06:28
clayg... and you don't ever write anything down06:28
ahalei've been using moveparts lots and adjusting the disk_limit's, concurrency and ms_per_part06:28
claygso - here I am trying to dust of my hummingbird deployment!06:28
ahalebut basically, priority replication06:28
claygdiscovering and prioritizing replication work is hard - let ops do it!  :D06:28
ahaleoh no, i mean the hbird "priority replication" feature06:29
claygyeah sorry unless that's dfg's rescue parts command I have no idea what that feature is or how it works?06:29
clayg... but you know ... i'm trying06:30
ahaleits same kinda thing as rescureparts06:30
ahalemoveparts06:30
ahalei mean, move and rescue both use taht new listening stuff and replicate differently, and way better in some situations06:31
redboThe http server just lets you shove replication jobs to the front of the queue.  moveparts is if you rebalance and partitions move from one drive to another, it sends up jobs to move those partitions from that one place to that one new place.06:36
redborestoredevice is for when a drive is replaced, and it prioritizes jobs for the other primaries of partitions that should be on that drive to replicate to it.06:37
*** mingdang1 has joined #openstack-swift06:38
redborescueparts I think just says everyone stop what you're doing and pay attention to this partition because it's under-replicated06:38
redboIn my utopian future, there's a guy who watches for drive swaps, ring changes, and bad dispersion reports and kicks those guys off.06:41
*** itlinux has quit IRC06:42
*** tesseract has joined #openstack-swift06:43
*** tesseract is now known as Guest8301606:43
ahaleyeah for now I just do as you tell me! until you replace me with a small go script06:43
clayginteresting...06:44
claygi mean not so much the part about why there's a need for replicators to have more optomized behaviors depending on the situation - update and updated_deleted just don't cut it06:46
clayg... but that it would be orchestrated by an out-of-band process trying to co-ordinate the whole ordeal instead of just the replicators jobs to think a little harder and ask better questions before the do stupid and useless shit06:46
claygworth some pondering06:47
claygredbo: THANKS FOR WRITING SOMETHING DOWN ;)06:47
*** Guest83016 has quit IRC06:48
*** esker has joined #openstack-swift06:49
redboYeah, a big part of what those guys do is make sure drives don't have more than two things being priority replicated to them at a time.06:49
*** dmorita has joined #openstack-swift06:52
*** haomaiwang has joined #openstack-swift06:53
*** esker has quit IRC06:53
*** dmorita has quit IRC06:57
*** haomaiwang has quit IRC06:57
claygseems like a job for semephores instead of humans - but different things annoy different people i 'spose07:00
claygthe one thing I like about this plan is that it makes running swift require even more institutional devops knowledge!07:00
*** dmorita has joined #openstack-swift07:01
ahalei think it could be interesting if rebalance was able to take hints too, like dispersion reports and recon -d info07:01
*** tesseract- has joined #openstack-swift07:02
claygahale: is recon -d disk fullness?  or part lists?07:02
ahalefullness07:02
redboWell there's a semaphore, but if you sent a bazillion replication jobs and a bazillion - 2 get rejected for concurrency, it's not very useful07:03
*** dmorita has quit IRC07:05
claygi'm not sure we have the same connotation for the cardinality of a bazillion - replication is still part based right?  it's the same number of requests as REPLICATE suffix checks right?07:06
*** Guest57780 is now known as sc07:07
claygjust some of them would say HTTP 420?07:07
claygdon't get me wrong - i'm all for cutting out suffix rehasing on new-primaries that are filling still filling from the old primary - do the new primaries know not to tell the new guys to rehash?07:09
redboThey.. don't know that, really.  They still invalidate their hash any time you sync a file to them.07:10
claygk07:10
claygwell it sounds like next week is going to be tons of fun07:11
ahale:(07:11
claygi still have a bunch to catch up tho - and I can't even seem to manage to get a decent hummingbird devenv up :'(07:11
ahalei think there are two or three different ops efforts going on to get a hbird vagrant all-in-one/many going, probably start appearing on github sooner or later07:13
clayggreat!  i'm trying to shake the rust of vagrant-swift-all-in-one/hummingbird branch07:14
*** dmorita has joined #openstack-swift07:15
claygi'm excited to revert my "fix" to not use conf.d - wtg hbird!07:16
redboThat's still in review :(07:17
redboI'm going to beat david up tomorrow until he approves all my patches07:17
claygoh no!  :'(07:19
claygi thought when unit tests start working again it must have ment that my proxies conf.d was no longer a problem!07:19
*** dmorita has quit IRC07:19
claygredbo: is there really no prescription on line length in golang?07:20
*** mingdang1 has quit IRC07:21
claygi like to side-by-side in my editor and have been *so* spoiled by pep8 telling everyone to wrap at 80 so that I don't have!07:21
redboNot really.  If it feels good, do it.  There are probably some lines in there that are too long.07:21
claygi'm not sure what i'm going to do with lines that 100+ chars long07:21
claygmy entire world might fall apart07:21
claygi don't think i can grok code if I can't side-by-side (you have to remember - i'm rather dull)07:21
redbogholt has a script that sets his terminal to exactly 80 columns wide, and makes his browser exactly fill the rest of the screen.07:22
claygdoes he still hang out on the internet?  maybe in some fuse or nfs channel somewhere?07:23
redboNot that I know of.  He's been twittering a little.07:23
redboAnd he's been talking about making a linedin account for weeks.07:24
claygso explain this one -> http://play.golang.org/p/t8DbarKqpS07:26
claygmaybe I don't understand this def -> func (errorSlice) Error() string07:28
onovygood news everyone: swift/2.5.0 in our production :)07:28
*** ChanServ sets mode: +v cschwede07:29
redboputting an "Error() string" method on an object basically lets you cast it to an error07:30
redboBecause error is an interface with just that method defined07:31
claygonovy: 2.7 is out - get you some concrrent reads - thanks mattoliverau !07:31
-openstackstatus- NOTICE: Logs are failing to be uploaded causing jobs to be marked as UNSTABLE. We are working on repairing the log filesystem and will update when ready. Please do not recheck before then.07:31
*** ChanServ changes topic to "Logs are failing to be uploaded causing jobs to be marked as UNSTABLE. We are working on repairing the log filesystem and will update when ready. Please do not recheck before then."07:31
onovyclayg: yep i know, on my backlog too07:31
claygonovy: :)07:31
redboThen he's mad that an instance of errorSlice that's nil, when cast to an error, is no longer nil.07:31
ahalei really wanna get concurrent reads in prod!07:31
claygonovy: you were one of the folks involved in ionice stuff yeah!?  came up in our office today - i really want to see that in07:32
*** furlongm_ has joined #openstack-swift07:32
onovyclayg: yep, peter and i was woking on it07:32
onovybut still no review :)07:33
*** furlongm has quit IRC07:33
claygit needs some reviews!07:33
onovyalmost ~1 month wihout comment07:34
*** mmcardle has joined #openstack-swift07:35
*** rledisez has joined #openstack-swift07:39
*** hosanai has quit IRC07:41
*** dmorita has joined #openstack-swift07:41
*** bluetefn has quit IRC07:45
*** dmorita has quit IRC07:46
*** m_kazuhiro has joined #openstack-swift07:51
*** nadeem has quit IRC07:52
*** nadeem has joined #openstack-swift07:54
*** nadeem has quit IRC07:57
peterlisakonovy, clayg, everyone: great, I'd appreciate new feedback on that ionice :)07:57
*** nadeem has joined #openstack-swift07:58
*** acoles_ is now known as acoles08:04
claygacoles: morning!08:11
acolesclayg: hey, what you doing around at this time?08:12
acolesclayg: when do you get into Austin?08:13
acolesI' ther emidday Sunday08:13
*** nadeem has quit IRC08:13
*** jmccarthy has quit IRC08:19
*** jmccarthy has joined #openstack-swift08:20
*** mingdang1 has joined #openstack-swift08:21
*** jordanP has joined #openstack-swift08:26
*** mingdang1 has quit IRC08:26
*** haomaiwang has joined #openstack-swift08:26
*** jistr has joined #openstack-swift08:28
*** haomaiwang has quit IRC08:30
openstackgerritKota Tsuyuzaki proposed openstack/swift: CompositeRing Support  https://review.openstack.org/27192008:33
*** Vinsh has quit IRC08:43
openstackgerritMerged openstack/swift: crypto - byok support in keymaster  https://review.openstack.org/30781108:52
*** haomaiwang has joined #openstack-swift08:55
*** haomaiwa_ has joined #openstack-swift08:57
*** haomaiwang has quit IRC09:00
*** haomaiwang has joined #openstack-swift09:00
*** haomaiwa_ has quit IRC09:02
*** ppai has quit IRC09:03
*** stantonnet has quit IRC09:03
claygBy convention, tag strings are a concatenation of optionally space-separated key:"value" pairs. Each key is a non-empty string consisting of non-control characters other than space (U+0020 ' '), quote (U+0022 '"'), and colon (U+003A ':'). Each value is quoted using U+0022 '"' characters and Go string literal syntax.09:04
claygPartition  uint64                `json:"partition"`09:04
claygso weird09:04
claygacoles: yup midday sunday09:04
*** stantonnet has joined #openstack-swift09:05
*** haomaiwang has quit IRC09:07
*** haomaiwang has joined #openstack-swift09:08
*** esker has joined #openstack-swift09:09
*** haomaiwang has quit IRC09:13
*** esker has quit IRC09:14
*** ppai has joined #openstack-swift09:17
*** openstackstatus has quit IRC09:24
*** openstackstatus has joined #openstack-swift09:26
*** ChanServ sets mode: +v openstackstatus09:26
*** _JZ_ has quit IRC09:27
*** david-lyle has joined #openstack-swift09:28
*** furlongm_ has quit IRC09:28
*** furlongm has joined #openstack-swift09:29
*** esker has joined #openstack-swift09:30
*** esker has quit IRC09:35
*** haomaiwang has joined #openstack-swift09:38
*** haomaiwang has quit IRC09:43
*** haomaiwang has joined #openstack-swift09:53
*** haomaiwang has quit IRC09:58
*** ppai has quit IRC10:01
*** haomaiwang has joined #openstack-swift10:07
*** haomaiwang has quit IRC10:11
*** esker has joined #openstack-swift10:11
openstackgerritClay Gerrard proposed openstack/swift: Add User-Agent to REPCONN requests  https://review.openstack.org/30933910:14
*** ppai has joined #openstack-swift10:15
*** esker has quit IRC10:16
*** haomaiwang has joined #openstack-swift10:24
*** natarej has joined #openstack-swift10:24
*** haomaiwang has quit IRC10:28
*** esker has joined #openstack-swift10:32
*** esker has quit IRC10:37
mattoliverauclayg: thanks to you too!10:38
*** haomaiwang has joined #openstack-swift10:41
*** haomaiwa_ has joined #openstack-swift10:45
*** haomaiwang has quit IRC10:45
*** haomaiwa_ has quit IRC10:49
*** cdelatte has quit IRC10:53
-openstackstatus- NOTICE: Log server has been repaired and jobs are stable again. If necessary please recheck changes that have 'UNSTABLE' results.10:54
*** esker has joined #openstack-swift11:14
*** haomaiwang has joined #openstack-swift11:14
*** cdelatte has joined #openstack-swift11:15
*** haomaiwang has quit IRC11:18
*** esker has quit IRC11:19
*** m_kazuhiro has quit IRC11:22
*** Jeffrey4l has joined #openstack-swift11:31
*** ppai has quit IRC11:38
*** esker has joined #openstack-swift11:55
*** haomaiwang has joined #openstack-swift11:59
*** haomaiwang has quit IRC11:59
*** haomaiwang has joined #openstack-swift12:00
*** esker has quit IRC12:00
*** mingdang1 has joined #openstack-swift12:09
*** asettle has joined #openstack-swift12:12
*** asettle has quit IRC12:13
*** esker has joined #openstack-swift12:16
*** esker has quit IRC12:21
*** SkyRocknRoll has quit IRC12:38
*** dmorita has joined #openstack-swift12:50
*** dmorita has quit IRC12:55
*** esker has joined #openstack-swift12:58
*** haomaiwang has quit IRC12:59
*** mingdang1 has quit IRC13:02
*** esker has quit IRC13:03
*** pdardeau has joined #openstack-swift13:09
*** esker has joined #openstack-swift13:10
*** mingdang1 has joined #openstack-swift13:13
*** esker has quit IRC13:18
*** klamath has joined #openstack-swift13:23
*** diogogmt has quit IRC13:24
*** aimeeU has joined #openstack-swift13:34
*** diogogmt has joined #openstack-swift13:35
*** esker has joined #openstack-swift13:38
*** esker has quit IRC13:43
*** adu has joined #openstack-swift13:50
*** zul has quit IRC14:01
*** chuck__ has joined #openstack-swift14:01
*** shakamunyi has joined #openstack-swift14:03
*** sgundur- has joined #openstack-swift14:08
*** dmorita has joined #openstack-swift14:09
*** esker has joined #openstack-swift14:13
*** dmorita has quit IRC14:14
*** shakamunyi has quit IRC14:15
*** arch-nemesis has joined #openstack-swift14:16
*** arch-nemesis has joined #openstack-swift14:16
*** shakamunyi has joined #openstack-swift14:29
*** arch-nemesis has quit IRC14:29
*** lcurtis has joined #openstack-swift14:38
*** arch-nemesis has joined #openstack-swift14:41
*** jmccarthy1 has joined #openstack-swift14:42
*** mingdang1 has quit IRC14:43
*** itlinux has joined #openstack-swift14:57
*** itlinux has quit IRC14:59
*** klrmn has joined #openstack-swift15:00
*** haomaiwang has joined #openstack-swift15:01
-openstackstatus- NOTICE: Log server was repaired as of 10:50 UTC and jobs have been stable since. If necessary, please recheck changes that have 'UNSTABLE' results.15:02
*** tesseract- has quit IRC15:09
openstackgerritAlistair Coles proposed openstack/swift: crypto - Refactor server side copy as middleware  https://review.openstack.org/30946615:14
*** garthb has joined #openstack-swift15:15
acolesjrichli: mahatic tdasilva ^^ all tests pass, put this up as WIP to share but still waiting for it to land on master first15:16
acolesnotmyname: ^^ that's probably as good as we might have hoped to have feature/crypto by today, but I don't propose we merge it15:18
acolesthere's some other patches pending but they are cleanup, tests etc15:19
*** mwheckmann has joined #openstack-swift15:19
tdasilvaacoles: ok15:21
mahaticacoles: okay, thanks15:22
*** jordanP has quit IRC15:25
*** openstack has joined #openstack-swift15:27
*** mmcardle has quit IRC15:29
*** Jeffrey4l has quit IRC15:33
*** haomaiwang has quit IRC15:34
*** ChubYann has quit IRC15:36
*** barra204 has joined #openstack-swift15:37
*** ninja82 has joined #openstack-swift15:42
*** ninja82 has left #openstack-swift15:42
*** ChubYann has joined #openstack-swift15:51
*** ozialien has joined #openstack-swift15:59
notmynamegood morning16:01
*** lakshmiS has joined #openstack-swift16:01
*** jistr has quit IRC16:04
*** nadeem has joined #openstack-swift16:16
*** ozialien has quit IRC16:18
*** rcernin has quit IRC16:19
*** nadeem has quit IRC16:22
*** cschwede has quit IRC16:23
*** tmoreira has quit IRC16:24
*** lyrrad has joined #openstack-swift16:24
*** klrmn has quit IRC16:25
*** dmorita has joined #openstack-swift16:26
*** dmorita has quit IRC16:27
*** nadeem has joined #openstack-swift16:28
*** dmorita has joined #openstack-swift16:29
*** rledisez has quit IRC16:29
*** garthb has quit IRC16:34
*** diogogmt has quit IRC16:34
*** diogogmt has joined #openstack-swift16:36
*** esker has quit IRC16:39
*** gyee has joined #openstack-swift16:39
*** esker has joined #openstack-swift16:39
*** tmoreira has joined #openstack-swift16:42
*** _JZ_ has joined #openstack-swift16:42
*** cschwede has joined #openstack-swift16:43
*** haomaiwang has joined #openstack-swift16:53
*** haomaiwa_ has joined #openstack-swift16:56
openstackgerritJohn Dickinson proposed openstack/swift: JUST A TEST WHERE liberasurecode STUFF ISN'T USED  https://review.openstack.org/30951016:57
*** haomaiwang has quit IRC16:58
*** haomaiwa_ has quit IRC17:01
*** dmorita has quit IRC17:03
*** Suyi has joined #openstack-swift17:06
*** dmorita has joined #openstack-swift17:06
*** mattoliverau has quit IRC17:13
*** matt6434 has joined #openstack-swift17:14
openstackgerritBryan Keller proposed openstack/swift: SwiftLogFormatter will log transaction IDs on INFO level  https://review.openstack.org/30952417:15
*** suyash has joined #openstack-swift17:17
*** Suyi has quit IRC17:19
*** sambhare has joined #openstack-swift17:20
*** suyash has quit IRC17:24
*** nadeem has quit IRC17:25
*** dmorita has quit IRC17:25
openstackgerritBryan Keller proposed openstack/swift: SwiftLogFormatter will log transaction IDs on INFO level  https://review.openstack.org/30952417:27
*** nadeem has joined #openstack-swift17:28
*** nadeem has quit IRC17:29
openstackgerritMerged openstack/swift: swift-[account|container]-info when disk is full  https://review.openstack.org/26872317:31
*** dmorita has joined #openstack-swift17:37
*** chuck__ is now known as zul17:37
*** dmorita has quit IRC17:43
*** tqtran has joined #openstack-swift17:43
*** diazjf has joined #openstack-swift17:43
*** haomaiwang has joined #openstack-swift17:45
notmynameonovy: I'd appreciate your view on patch 30921917:46
patchbotnotmyname: https://review.openstack.org/#/c/309219/ - swift - created other-requirements.txt17:46
*** haomaiwang has quit IRC17:49
*** aimeeU has quit IRC17:53
*** haomaiwang has joined #openstack-swift17:59
*** dmorita has joined #openstack-swift18:02
*** diazjf has quit IRC18:03
*** haomaiwang has quit IRC18:04
*** shakamunyi has quit IRC18:05
*** barra204 has quit IRC18:05
*** shakamunyi has joined #openstack-swift18:06
*** dmorita has quit IRC18:06
*** barra204 has joined #openstack-swift18:07
*** klrmn has joined #openstack-swift18:09
*** acoles is now known as acoles_18:09
*** dmorita has joined #openstack-swift18:13
*** diazjf has joined #openstack-swift18:15
*** dmorita has quit IRC18:17
*** dmorita has joined #openstack-swift18:27
*** haomaiwang has joined #openstack-swift18:35
*** haomaiwang has quit IRC18:41
*** barra204 has quit IRC18:41
*** shakamunyi has quit IRC18:41
*** shakamunyi has joined #openstack-swift18:42
*** barra204 has joined #openstack-swift18:42
*** zaitcev has joined #openstack-swift18:53
*** ChanServ sets mode: +v zaitcev18:53
*** kozhukalov has quit IRC18:54
*** diazjf has quit IRC18:54
*** dmorita_ has joined #openstack-swift18:54
*** diazjf has joined #openstack-swift18:55
*** dmorita has quit IRC18:55
*** arch-nemesis has quit IRC18:56
*** sambhare has quit IRC18:57
*** kozhukalov has joined #openstack-swift19:01
*** Suyi has joined #openstack-swift19:13
*** haomaiwa_ has joined #openstack-swift19:13
*** arch-nemesis has joined #openstack-swift19:18
*** arch-nemesis has quit IRC19:18
*** arch-nemesis has joined #openstack-swift19:18
*** haomaiwa_ has quit IRC19:22
*** arch-nemesis has quit IRC19:24
*** arch-nemesis has joined #openstack-swift19:37
*** mwheckmann has quit IRC19:52
*** jmccarthy1 has quit IRC20:03
claygclam before the storm20:15
clayg... kind of a weird typo20:15
notmynamegoogle says https://handsomefox.com/droveria/files/Bizarro-Clam-Before-the-Storm.jpg20:17
*** diazjf has quit IRC20:17
tdasilvanotmyname: do you still have a signal group going?20:18
notmynameit's just a group chat. there's nothing to set up or anything20:18
notmyname(however, I think I left the group)20:18
tdasilvaok..was just wondering if people were planning to use at the summit...20:19
notmynameprobably everyone will have a corporate slack channel. and personal text messages. and maybe a signal group. or whatsapp. and the sane ones will leave their phone at home and just find people in the hallways ;-20:20
* notmyname is not a sane one20:20
* notmyname goes back to slide prep20:20
claygdfg_: you around?  I'm trying to respond to your ask on the replicator logging - but I think we're crossed-wires20:21
*** diazjf has joined #openstack-swift20:28
dfg_clayg: ya i'm here.20:29
claygdfg_: well i'm almost done typing now!20:29
dfg_ya- i got confused. i thought i added the job_type crap in my current work. but it was in the last patch20:29
claygyeah i don't really know what you were talking about in the first couple of posts - my gist womm20:30
claygfwiw i'm not sure you groked that in my test I was acctually making a job that told a primary part holder to push a part onto a hand-off just to do it and see what happens20:30
dfg_you used to have to pass "job_type" in the json. but nm20:30
dfg_oohh. ya i didn't get that20:31
claygdfg_: probably not relevant then - i mean... is the rpc documented in some test code somewhere?  maybe some probetests or something like that?20:31
claygdfg_: yeah so anyway - i got my repconn logs - i could see the three primaries getting repconns from *somewhere* but it wasn't obvious from whom20:32
claygin python we stick pids in the user agents - but that doesn't work in all this cray single process land20:32
claygso i just grabbed the first unique identifier i could find20:32
claygif the object-server is going to log user-agent we might as well put something useful in it?  not sure if what I cooked up was useful?20:32
claygalso i suppose if you look at the long lines in a multinode system the ip of the source (the replicator node) will be in the log message somewhere?20:33
claygprobably the first field after the colon - it'd have the remote ip?20:33
dfg_ya20:33
dfg_thats the thing- it would only help in a SAIO20:34
claygyeah but... that's useful?  sort of ... required even if ... well idk20:34
dfg_but- i don't really mind all that much. its not a big change or something20:35
dfg_but people do complain about logging or some kind of insight into the system. so i'd like to add something. but i'm not sure if this is that something20:35
clayghow do *you* debug this kind of stuff when you hacking in dev?20:36
claygare you guys just all if it compiles stick it on the lab/staging?20:36
*** cdelatte has quit IRC20:36
dfg_ya- if `make` works i call it done20:37
claygright on20:37
dfg_idk- i do a lot of getting rid of conf files and messing with the ring.20:38
claygso - more seriously - I don't know how to make this one process starts all the replicators make sense in my workflow?20:38
claygahhhh...20:38
claygyeah that seems annoying tho20:38
clayg:)20:38
dfg_ya- it is annoying :)20:39
claygin probetests we do a lot of stop object-server.120:39
*** gyee has quit IRC20:39
claygor cycle object-replicator.220:39
dfg_you can also run a single object server at a time.20:39
claygby moving configs around?  or is there a command syntax to have it shut one of them down?20:39
claygoh... you just mean in the forground20:39
dfg_ya20:40
claygmaybe?20:40
dfg_in forground20:40
dfg_made same typo20:40
claygheh20:41
claygit's because golang only has for loops!20:41
dfg_i don't think you can shut the off one by one like that20:41
dfg_sometimes i miss python...20:42
claygwell - i really want that - breaking things and watching them fix themselves is a big part of how I reason about changes in the consistency engine20:42
claygnah - don't say that20:42
dfg_:)20:42
dfg_there is the lock_device thing20:42
claygI really think I can make swift-init take over the job of orchestrating *all* of swifts backend daemons regardless of the implementation20:42
redboThis almost works: "hummingbird start object -c /etc/swift/object-server/1.conf".  The main problem is it writes object.pid instead of object.1.pid or whatever, so you can't do multiples of those.20:42
claygredbo: why can't i just let swift-init fork into `hummingbird object-server -d -c /blah/1.conf` ?20:43
claygit says it closes stdout?20:43
redboYou can20:43
claygit would mean so much to me to be able to `swift-init start main` I can't even describe it20:44
glangehah20:44
claygi'd really like to be able to tell people "here's how you upgrade to hbird, in [object-server] section of your config write fucking_fast = true" and it just works20:44
redboExcept go can't setuid itself, so swift-init will need to do that.20:45
claygi could preparse the config in swift-init before hitting the subexec and just kick over the right command20:45
claygwat?20:45
claygyou mean drop_priv?20:45
openstackgerritBryan Keller proposed openstack/swift: SwiftLogFormatter will log transaction IDs on INFO level  https://review.openstack.org/30952420:46
redboYeah, go can't drop its own privileges, the "start" command does that when it executes the background guy.20:47
redboI think swift does that after it's started20:47
claygwell *yeah*20:47
claygi think i can't plumb that into the subprocess guy20:47
clayghe has flags for all that crap20:47
claygbut but... what if I need to bind to a privledged port?20:48
claygwhy can go not do this thing?  I'm sure Pike has some great story about how assholes like me aren't smart enough to do it right anyway so they don't need to have that feature20:48
redbosetcap  'cap_net_bind_service=+ep' :)20:48
claygredbo: k, fair enough20:49
claygwell i don't put my object servers on priv'd ports anyway - so doing the priv drop in the subexec would probably work great20:49
claygand you'd be down with all this (patch review pending of course)?20:49
redboThere's a long-running bug for it in go.  Basically the setuid system call only changes the current thread, and libc's setuid uses some complex system of signals and handlers to change all the uid of all the threads owned by the process, but go doesn't have all that yet.20:50
redboSure20:50
claygyawn20:50
claygok great!20:50
redboOR swift-init or whomever could bind the port and pass it as a command-line argument.  But that would take a little work.20:51
*** gyee has joined #openstack-swift20:52
claygjust be all like yo yo yoo this integer here is toally a socket fd I setup for you bro - have at it!20:52
*** ozialien has joined #openstack-swift20:56
*** haomaiwang has joined #openstack-swift21:00
*** haomaiwang has quit IRC21:04
clayggod there's so many outstanding patches to review :'(21:10
*** lakshmiS has quit IRC21:10
*** dmorita_ has quit IRC21:24
*** ozialien has quit IRC21:25
*** haomaiwang has joined #openstack-swift21:29
*** haomaiwang has quit IRC21:34
*** dmorita has joined #openstack-swift21:34
*** diazjf has quit IRC21:38
*** nadeem has joined #openstack-swift21:52
openstackgerritSamuel Merritt proposed openstack/swift: Clean up fallocate tests a little  https://review.openstack.org/30960821:59
*** nadeem has quit IRC22:00
*** gyee has quit IRC22:01
*** dmorita has quit IRC22:05
*** nadeem has joined #openstack-swift22:06
openstackgerritJoel Wright proposed openstack/python-swiftclient: Add new doc structure and contents for swiftclient  https://review.openstack.org/28856622:16
*** lcurtis has quit IRC22:17
*** dmorita has joined #openstack-swift22:22
*** dmorita has quit IRC22:24
*** dmorita has joined #openstack-swift22:31
*** dmorita has quit IRC22:50
*** Suyi has quit IRC22:50
*** Suyi has joined #openstack-swift22:51
*** Suyi has quit IRC22:52
*** Suyi has joined #openstack-swift22:52
*** dmorita has joined #openstack-swift22:54
*** dmorita has quit IRC22:54
openstackgerritMerged openstack/swift: fix fallocate_reserve traceback  https://review.openstack.org/30540523:03
*** nadeem has quit IRC23:17
*** matt6434 is now known as mattoliverau23:19
*** ChanServ sets mode: +v mattoliverau23:20
*** lakshmiS has joined #openstack-swift23:22
*** dmorita has joined #openstack-swift23:32
*** Suyi has quit IRC23:35

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