Tuesday, 2017-04-04

*** chsc has quit IRC00:05
*** catintheroof has joined #openstack-swift00:25
claygtdasilva: i'm not sure how exactly but this is 100% related to us -> http://lists.openstack.org/pipermail/openstack-dev/2017-April/114918.html00:52
claygIf can figure out a sane way by which openstack infra can provide rpm and deb packages of the code openstack needs to deploy the infra they need to test on we've like solved the whole digum00:52
claygbonus points if we optionally test that old versions of openstack continue to work on the old versions of our dependencies packaged for the distributions available at the time those old versions of openstack was released00:53
claygand super awesome when a release of openstack includes a list of *all* of it's dependencies versions not just python and people who want to deploy openstack can solve those dependencies or figure out someway to steal upstream packages on the distros upstream tested with00:55
kota_morning00:56
notmynamehello kota_00:56
kota_hello notmyname00:57
kota_notmyname: how was going on the tc meeting in this morning?00:57
kota_this?00:57
notmynameit's tomorrow00:57
kota_eh - ok00:57
notmynamewell, "tuesday" and it's already tuesday for you ;-)00:57
kota_notmyname: kk00:57
notmynamewhatever, it's about 20 hours from now00:57
notmynameso far the proposal has positive feedback00:58
kota_yeah, i saw the gerrit patch you pasted in the last night briefly00:58
notmynamehttps://review.openstack.org/#/c/451524/00:59
patchbotpatch 451524 - governance - Use case analysis for Golang addition to Openstack00:59
kota_notmyname: that one00:59
kota_it looks like no negative vote so far01:00
notmynamebased on previous experience, I won't count on it passing until it actually does ;-)01:01
kota_notmyname: exactly01:01
*** catintheroof has quit IRC01:02
kota_clayg: thanks for the comment on affinity per policy (patch 382766), I am able to address the acoles's comments in this week01:08
patchbothttps://review.openstack.org/#/c/382766/ - swift - Support Read/Write Affinity Per Policy01:08
claygkota_: sounds great!01:09
claygkota_: thanks01:09
claygi've been looking at composite rings all day01:09
claygwell off and on01:10
claygi keep getting pulled into stuff that's been going on while I was out last week01:10
kota_clayg: thanks! I am just about to say the composite ring one to ask you to review again :D01:10
*** klrmn has quit IRC01:18
*** JimCheung has quit IRC01:55
*** bkopilov has quit IRC02:09
*** vint_bra has joined #openstack-swift02:12
*** klrmn has joined #openstack-swift02:27
*** chsc has joined #openstack-swift02:47
*** chsc has joined #openstack-swift02:47
*** chsc has quit IRC02:53
*** vint_bra has quit IRC02:59
*** links has joined #openstack-swift03:08
claygkota_: I'll sleep on composite rings - see if either your or acoles have time to read over and assuage my fears03:33
claygi need to bolt for now03:34
kota_clayg: alright, will look at, thanks!03:36
*** bkopilov has joined #openstack-swift03:39
*** vint_bra has joined #openstack-swift03:41
*** SkyRocknRoll has quit IRC03:50
*** gkadam has joined #openstack-swift03:55
*** m_kazuhiro has joined #openstack-swift04:00
m_kazuhirogood morning04:03
*** vint_bra has quit IRC04:06
*** m_kazuhiro has quit IRC04:13
*** vint_bra has joined #openstack-swift04:15
*** m_kazuhiro has joined #openstack-swift04:17
mahaticgood morning04:18
mahaticm_kazuhiro: o/04:18
m_kazuhiromahatic: o/04:18
*** abhitechie has joined #openstack-swift04:22
*** klrmn has quit IRC04:22
*** psachin has joined #openstack-swift04:24
*** vint_bra has quit IRC04:29
*** SkyRocknRoll has joined #openstack-swift04:45
*** rcernin has joined #openstack-swift04:46
*** gkadam has quit IRC04:51
*** oshritf has joined #openstack-swift05:05
*** oshritf has quit IRC05:06
*** gkadam has joined #openstack-swift05:07
*** John341 has joined #openstack-swift05:09
*** rcernin has quit IRC05:09
*** mathiasb_ has joined #openstack-swift05:10
*** szaher_ has joined #openstack-swift05:10
*** rcernin has joined #openstack-swift05:11
*** sgundur1 has joined #openstack-swift05:11
*** supamatt_ has joined #openstack-swift05:12
*** mlanner_ has joined #openstack-swift05:12
*** balajir_ has joined #openstack-swift05:13
*** szaher has quit IRC05:14
*** Dinesh_Bhor has quit IRC05:14
*** mathiasb has quit IRC05:14
*** supamatt has quit IRC05:14
*** mlanner has quit IRC05:14
*** sgundur has quit IRC05:14
*** John341_ has quit IRC05:14
*** balajir has quit IRC05:14
*** supamatt_ is now known as supamatt05:14
*** mlanner_ is now known as mlanner05:14
*** jamielennox is now known as jamielennox|away05:15
*** Dinesh_Bhor has joined #openstack-swift05:21
*** ChubYann has quit IRC05:23
*** chsc has joined #openstack-swift05:24
*** chsc has joined #openstack-swift05:24
*** chsc has quit IRC05:29
mattoliveraum_kazuhiro, m_kazuhiro morning05:31
*** jamielennox|away is now known as jamielennox05:37
*** madgoat has joined #openstack-swift05:43
*** madgoat has left #openstack-swift05:44
*** mtreinish has quit IRC05:49
*** mtreinish has joined #openstack-swift05:50
*** mtreinish has quit IRC05:52
*** mtreinish has joined #openstack-swift05:56
*** jaosorior_away is now known as jaosorior06:01
*** abhitechie has quit IRC06:13
*** cshastri has joined #openstack-swift06:18
m_kazuhiromattoliverau: morning!06:23
*** oshritf has joined #openstack-swift06:25
*** tesseract has joined #openstack-swift06:26
*** oshritf has quit IRC06:33
*** oshritf has joined #openstack-swift06:34
*** abhitechie has joined #openstack-swift06:37
m_kazuhiroeveryone: If you are interested in swift automated tiering, please visit http://forumtopics.openstack.org/cfp/details/37 and leave comments or questions. I want to get many opinions about it before boston summit.06:38
*** oshritf has quit IRC06:40
*** winggundamth has joined #openstack-swift06:42
*** oshritf has joined #openstack-swift06:43
*** hseipp has joined #openstack-swift06:45
*** oshritf has quit IRC06:49
*** sputnik13 has quit IRC06:50
*** oshritf has joined #openstack-swift06:50
*** sputnik13 has joined #openstack-swift06:54
*** sputnik13 has quit IRC06:54
*** oshritf has quit IRC06:55
*** sputnik13 has joined #openstack-swift06:56
*** oshritf has joined #openstack-swift06:58
openstackgerritMatthew Oliver proposed openstack/swift master: Sharding: use container backends _record_to_dict to simplify _generate_object_list  https://review.openstack.org/45303207:02
mattoliverauacoles: ^ that is one way of simplifiing the sharders _generate_object_list to use the already existing _record_to_dict that the container broker uses. Which can be passed in as a row_transform_func in your patch.. so obviously I'm adding to your chain here :)07:04
*** pcaruana has joined #openstack-swift07:20
*** m_kazuhiro has quit IRC07:20
*** dja has quit IRC07:22
*** m_kazuhiro has joined #openstack-swift07:28
*** oshritf has quit IRC07:31
*** adriant has quit IRC07:32
*** oshritf has joined #openstack-swift07:34
*** geaaru has joined #openstack-swift07:37
*** jamielennox is now known as jamielennox|away07:45
*** amoralej|off is now known as amoralej07:53
acolesgood morning07:58
kota_acoles: good morning07:59
*** cbartz has joined #openstack-swift08:08
mattoliverauacoles: morning08:17
mahaticacoles: good morning08:18
acoleskota_: mattoliverau mahatic o/08:23
kota_clayg: I dumped some thought on your comments at patch 441921 but I need more time to think deeply to answer to your question at patchset 11 so I'll continue to review it tommorow08:54
patchbothttps://review.openstack.org/#/c/441921/ - swift - Add Composite Ring Functionality08:54
kota_acoles: and I think you may have a opinion too.08:55
* kota_ has to leave the desk now today08:55
acoleskota_: thnks yes I'll look at it too08:55
*** abhitechie has quit IRC08:58
*** abhitechie has joined #openstack-swift08:59
*** abhitechie has quit IRC09:02
*** abhitechie has joined #openstack-swift09:02
*** szaher_ has quit IRC09:19
*** szaher has joined #openstack-swift09:19
*** dja has joined #openstack-swift09:29
*** mvk has quit IRC09:31
*** oshritf has quit IRC09:48
*** mvk has joined #openstack-swift10:00
*** winggundamth has quit IRC10:04
*** m_kazuhiro has quit IRC10:32
*** SkyRocknRoll has quit IRC10:34
*** jamielennox|away is now known as jamielennox11:03
*** bkopilov has quit IRC11:12
*** mathiasb_ is now known as mathiasb11:44
*** torgomatic has quit IRC12:01
*** catintheroof has joined #openstack-swift12:05
*** amoralej is now known as amoralej|lunch12:08
*** chlong has joined #openstack-swift12:10
*** oshritf has joined #openstack-swift12:25
*** oshritf has quit IRC12:28
*** oshritf has joined #openstack-swift12:29
*** oshritf has quit IRC12:43
*** oshritf has joined #openstack-swift12:50
*** oshritf has quit IRC12:52
*** oshritf has joined #openstack-swift12:52
*** oshritf has quit IRC12:54
*** oshritf has joined #openstack-swift13:00
*** stradling has joined #openstack-swift13:00
*** oshritf has quit IRC13:03
*** abhitechie has quit IRC13:03
*** abhitechie has joined #openstack-swift13:04
*** links has quit IRC13:04
*** kei_yama has quit IRC13:04
*** oshritf has joined #openstack-swift13:08
*** oshritf has quit IRC13:09
*** bkopilov has joined #openstack-swift13:09
*** abhitechie has quit IRC13:09
*** zul has quit IRC13:12
*** _JZ_ has joined #openstack-swift13:13
*** amoralej|lunch is now known as amoralej13:14
*** oshritf has joined #openstack-swift13:22
*** oshritf has quit IRC13:24
*** SkyRocknRoll has joined #openstack-swift13:26
*** oshritf has joined #openstack-swift13:31
*** zul has joined #openstack-swift13:32
*** oshritf has quit IRC13:33
*** jaosorior is now known as jaosorior_away13:36
*** oshritf has joined #openstack-swift13:37
*** SkyRocknRoll_ has joined #openstack-swift13:42
*** SkyRocknRoll has quit IRC13:45
*** oshritf has quit IRC13:48
*** hseipp has quit IRC13:58
*** oshritf has joined #openstack-swift13:58
*** oshritf has quit IRC14:01
*** oshritf has joined #openstack-swift14:03
*** vint_bra has joined #openstack-swift14:04
*** oshritf has quit IRC14:11
*** oshritf has joined #openstack-swift14:15
*** gkadam has quit IRC14:19
*** dewanee_ has joined #openstack-swift14:39
*** geaaru has quit IRC14:41
dewanee_question about partitions and rebalancing14:42
dewanee_anyone up for it?14:42
dewanee_I'll write it out:14:42
*** cshastri has quit IRC14:43
dewanee_I have two ring of the same cluster at two different moment in time. If I try to rebalance them to end up in the same final state (i.e. all device with the same weight) in one case I have the same number of partitions per device while in the other end up with a quite steep unbalance (about 1:4) between devices.14:45
dewanee_object rings14:45
dewanee_any idea why?14:45
dewanee_is there any fixed limit on number of partitions it can rebalance?14:47
*** xinli has joined #openstack-swift14:59
*** oshritf has quit IRC15:00
xinliI am a new one in this community, and we will support swift service in our cloud enviorment, I want to build a devSteck dev and unit test environmnet. Any document link to build the environment?15:02
*** klrmn has joined #openstack-swift15:03
openstackgerritAlistair Coles proposed openstack/swift master: Add id to RingBuilder to differentiate rings in composite  https://review.openstack.org/44931015:04
acolesxinli: many devs use a 'swift-all-in-one' for testing https://docs.openstack.org/developer/swift/development_saio.html15:04
acolesxinli: however, that does not use devstack, just installs swift15:05
xinli+acoles: if I want to review the code or patticipate fixing bug, which platform should I use?15:06
acolesxinli: there is also this resource, which I have not yet tried myself, https://github.com/swiftstack/vagrant-swift-devstack15:07
xinli+acoles: I will try both of them. If you want to submit new patch, how do you create your local repository?15:08
acolesxinli: you should read this https://github.com/openstack/swift/blob/master/CONTRIBUTING.rst15:08
xinli+acoles: cool, I got three links, thanks15:09
acolesand this https://docs.openstack.org/developer/swift/first_contribution_swift.html15:09
acolesfour links :)15:09
xinli:)15:09
tdasilvaxinli: welcome15:09
acolesxinli: also, ask questions here (like you already have)15:09
xinli+tdasilva thanks a lot.15:10
acolesxinli: note that at some point you will need to go through the general openstack signup process dcoumented here https://docs.openstack.org/infra/manual/developers.html so it is worth working through those steps.15:11
xinli+acoles: I have worked in cinder for one year and delivered several patches, it seems the same process for configre the Gerrit15:29
acolesxinli: yes it will be, just different project repo15:31
xinli+acoles: I have the global git and git review setup in my VM for submit the cinder patch, is there any extra steps I need to follow?15:31
xinli+acoles: got it, thanks15:32
acolesxinli: not that I recall, global git config should be ok for all projects15:32
xinli+acoles: great.15:32
openstackgerritAlistair Coles proposed openstack/swift master: Add id to RingBuilder to differentiate rings in composite  https://review.openstack.org/44931015:33
acoleskota_: clayg: ^^ I revived this, left as a follow on to make easier to review but if we think this works then we can squash it in (it replaces the source attr)15:34
*** klamath has joined #openstack-swift15:36
*** klamath has quit IRC15:36
*** klamath has joined #openstack-swift15:36
*** klamath has quit IRC15:41
*** vinsh has quit IRC15:44
*** tonanhngo has joined #openstack-swift15:54
notmynamegood morning15:54
notmynamexinli: welcome15:54
*** jlvillal_pto is now known as jlvillal15:54
xinli+notmyname: thank you15:55
*** tonanhngo has joined #openstack-swift15:56
notmynameclayg: tdasilva: I didn't realize the ubuntu cloud archive would solve all our problems. that's really interesting15:56
*** SkyRocknRoll_ has quit IRC15:59
*** pcaruana has quit IRC16:01
tdasilvanotmyname, tdasilva: just looked up UCA, it looks similar to what RDO is doing on centos side?? where they make a repo avaiable for each openstack release???16:05
tdasilvalol16:05
tdasilvanotmyname, clayg ^^^16:05
tdasilvai'm not crazy and typically don't talk to myself16:06
notmyname"typically"16:06
notmynameI'm not really that familiar with UCA (or RDO) but yeah if it means that we can be more in control of what's packaged and also what's tested, that seems good16:07
tdasilvaclayg: so I don't know how *we* can have more control, but from what I know in the RDO case, for swift for example, they package whatever we (swift) mark as the version for that release.16:11
tdasilvaback to clayg's question about being able to test older releases, the different RDO repos would allow for that I believe. So you could install the "mitaka" repo and that would contain the swift, libec, pyeclib of that release, if I understand correctly16:13
tdasilvaplus fixes we have made in stable releases...16:14
*** SkyRocknRoll has joined #openstack-swift16:19
notmynamehmmm... comment from zaneb on https://review.openstack.org/#/c/367915/16:26
notmynameI've been talking with the docs team this morning (hi asettle)16:30
notmynamewe've agreed that in The Future (tm) the swift team will maintain our config option docs at https://docs.openstack.org/developer/swift/deployment_guide.html#common-configuration and sample configs, and the docs team will link to them instead of keeping their docs manuall updated16:30
* asettle waves16:31
asettleThanks notmyname :)16:31
notmynameasettle: did I capture that corrently?16:31
clarkbtdasilva: yes that basically it, other things to consider are that once a new LTS comes out they stop making UCA repos for the old LTS. And they won't make a UCA repo for the current LTS until the next openstack release after the LTS release16:32
asettleGot it in one notmyname :)16:32
asettleANy questions, fire at me (or like, don't, I'm sensitive)16:32
*** psachin has quit IRC16:32
asettleI gotta roll - thanks for the update notmyname :D16:32
notmynamenoooo! patch 44810416:32
patchbothttps://review.openstack.org/#/c/448104/ - requirements - Revert "Update eventlet uc to 0.20.1"16:32
notmynameasettle: good night16:32
tdasilvaclarkb: ack, interesting. I wonder how that affects the gate testing. Basically we cant add a new LTS node flavor until there's an openstack release for it???16:34
clarkbtdasilva: so far for our normal gate testing we've just entirely ignored UCA because historically its been unreliable, but apparently its what people use in the real world and old libvirt has problems so I've proposed we start using it in devstack at least (thread on mailing list)16:35
clarkbtdasilva: if you look there is currently a pike repo but all of its contents are the ocata contents from what I can tell so its a place holder until the next release happens (I think, canonical/ubuntu would have to confirm that)16:36
tdasilvaclarkb: looking up the thread16:37
*** chsc has joined #openstack-swift16:43
*** chsc has joined #openstack-swift16:43
notmynameis the TC meeting at 1900 or 2000 UTC?16:48
*** vinsh has joined #openstack-swift16:57
*** vinsh_ has joined #openstack-swift16:58
*** vinsh has quit IRC16:58
*** cbartz has quit IRC16:59
*** stradling has quit IRC17:00
*** stradling has joined #openstack-swift17:01
*** tesseract has quit IRC17:04
*** ChubYann has joined #openstack-swift17:06
*** mvk has quit IRC17:15
*** amoralej is now known as amoralej|off17:20
*** tonanhngo has quit IRC17:29
xinli+notmtname: in the link of "All in one" I am using a vm, there is no /dve/sd?, should I use loop device or I could use the virtual disk /dev/vd?17:31
xinlihttps://docs.openstack.org/developer/swift/development_saio.html17:31
acolesxinli: I use a VM with a second virtual block device. BTW you mistyped his nick - notmyname17:33
xinli+acoles: thanks17:34
notmynameI prefer to have VirtualBox expose "real" drives to the SAIO17:34
notmynamebut using a loop device works too17:34
clayggood morning17:44
*** d0ugal has quit IRC17:48
claygnotmyname: it's more like clarkb and all of openstack really has exactly the same problems with non-python dependencies.  I'm sure to some extent OpenStack has been involved in fixing libvert and openvswitch bugs (maybe driving features) - but openstack hasn't really had a way to specificy a dependency on a new non-python release and test it in the gate.  clarkb is looking at UCA, tdasilva is looking at RDO, it's all the same17:48
*** d0ugal has joined #openstack-swift17:49
*** mvk has joined #openstack-swift18:01
clarkbpart of the problem too is we don't want to go in our corner and make our own distro that requires openstack users to fabricobble some sort of working set of deps on their own18:02
clarkbwhich means relying on distros as they provide the software the vast majority of our users are going to use18:02
tdasilvanotmyname: 2000 UTC18:04
claygclarkb: I think I grok that - and understanding when we say we "require" version xyz of thing - it's still a valuable point of fact to understand "do we still *work* on LTS old distro xyz's packaged version of thing" - then you make the call to carry the cruft or abandon that distro to the version they currently ship18:13
*** silor has joined #openstack-swift18:14
claygbut what we did with python moved forward regardless of the distro's and didn't look back18:14
*** stradling has quit IRC18:20
*** stradling has joined #openstack-swift18:21
acoleskota_: clayg : I update https://etherpad.openstack.org/p/composite_rings with a section title "Proposals" to try to clarify next steps18:25
acoleshmm, dropped all my 'd's in that sentence18:26
*** chosafine has joined #openstack-swift18:30
notmynamefor those that like graphs with dubious practical utility, I've created one that is the count of log files by date for the QA cluster job logs18:31
*** hseipp has joined #openstack-swift18:31
notmynamehttp://d.not.mn/logs_by_day.png18:32
notmynameafter lunch today, I'll be deleting a bunch of these. currently, i'm guessing i'll delete anything before 2017. and i'll probably add some expiry to the logs to only keep them for a month (or two? idk)18:35
notmynameplease speak up if you have some reason to keep older logs around18:35
*** oshritf has joined #openstack-swift18:38
*** oshritf has quit IRC18:44
*** hseipp has quit IRC18:48
*** oifrag has joined #openstack-swift19:04
*** oifrag has left #openstack-swift19:08
*** xinli has quit IRC19:18
*** silor has quit IRC19:30
*** chosafine has quit IRC19:38
*** SkyRocknRoll has quit IRC19:50
*** ediardo has quit IRC20:00
*** ediardo has joined #openstack-swift20:01
notmynametdasilva: I just got back from lunch. lurking in the tc meeting20:10
tdasilvanotmyname: me too20:10
tdasilvanotmyname: lurking, lunch was a while back ;)20:10
claygacoles: the write up on the etherpad was grea!  excellent work!  I'm sorry I let the implementation get so far along w/o realizing the implications of some of the choices20:29
claygFWIW I feel like the module is shaping up to be really robust - much better than I imagined - I'm feeling better and better about moving forward with swift-ring-composer20:29
claygkota_: acoles: great work and KUDOS!20:29
claygI'll look at patch 449310 next20:30
patchbothttps://review.openstack.org/#/c/449310/ - swift - Add id to RingBuilder to differentiate rings in co...20:30
*** stradling has quit IRC20:30
*** _JZ_ has quit IRC20:36
*** rcernin has quit IRC20:55
*** dja has quit IRC21:26
*** chosafine has joined #openstack-swift21:38
*** chosafine has quit IRC21:38
*** dja has joined #openstack-swift21:52
*** dja has quit IRC21:58
*** dja has joined #openstack-swift22:00
*** daxtens has joined #openstack-swift22:05
*** dja has quit IRC22:05
*** dja has joined #openstack-swift22:10
*** daxtens has quit IRC22:10
*** vint_bra has quit IRC22:19
*** dja has quit IRC22:23
mattoliverauMorning22:26
*** dja has joined #openstack-swift22:42
*** hoonetorg has quit IRC22:43
*** hoonetorg has joined #openstack-swift22:44
*** hoonetorg has quit IRC22:47
*** hoonetorg has joined #openstack-swift22:49
*** hoonetorg has quit IRC22:54
*** catintheroof has quit IRC22:57
*** hoonetorg has joined #openstack-swift22:58
*** JimCheung has joined #openstack-swift23:05
*** kei_yama has joined #openstack-swift23:24
*** chsc has quit IRC23:38
*** JimCheung has quit IRC23:50
*** JimCheung has joined #openstack-swift23:50
*** JimCheung has quit IRC23:54

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