Friday, 2017-10-13

openstackgerritSamuel Merritt proposed openstack/swift master: Skip cross-account-copy functest if only one account  https://review.openstack.org/51165500:08
openstackgerritSamuel Merritt proposed openstack/swift master: functests: don't crash if no second account  https://review.openstack.org/51165700:22
*** tovin07_ has joined #openstack-swift00:48
*** m_kazuhiro has joined #openstack-swift00:58
kota_good morning01:19
m_kazuhirokota_: good morning01:36
*** psachin has joined #openstack-swift01:46
kota_m_kazuhiro: hi01:59
kota_Did you get something progressed on symlink?02:00
kota_m_kazuhiro: ^^02:00
m_kazuhirokota_: I and tdasilva talked about recommend content-type of symlink.02:01
kota_m_kazuhiro: ok, how's going on it?02:01
m_kazuhirokota_: Originally, this topic came from symlink middleware's comment of "recommended content-type". And this comment was written by tdasilva. So I asked him the reason why he write about recommended content-type.02:03
m_kazuhirokota_: He said that he didn't remember reason but he thought it was way to distinguish objects.02:05
m_kazuhirokota_: And the talk about recommended content-type finished at the timing because of his schedule.02:07
kota_m_kazuhiro: so that, it's still under going?02:08
m_kazuhirokota_: He also said that he will update symlink patch. So I'm waiting patch update and answer for the recommended content-type.02:08
m_kazuhirokota_: Yes. under going.02:08
kota_m_kazuhiro: what's point tdasilva is now working?02:10
kota_i guess several topics in the symlink improvements in your etherpad...02:10
kota_several topics exist02:10
m_kazuhirokota_: I think his main point is not on etherpad now. He will work for your comments on symlink patch.02:12
kota_m_kazuhiro: to make response comments? or new patch set? I'm not sure I could continue to work on my follow up for now...02:16
m_kazuhirokota_: In my understanding, new patch set. What does "follow up" mean? Your comment? Your following patch?02:19
*** mgagne has quit IRC02:19
kota_https://review.openstack.org/#/c/509149/ m_kazuhiro02:20
patchbotpatch 509149 - swift - Follow up for symlink implementaion02:20
kota_oh, it's already get merge conflict though.02:20
*** zacksh has quit IRC02:20
*** zacksh has joined #openstack-swift02:22
*** mgagne has joined #openstack-swift02:22
*** mgagne is now known as Guest5368002:22
m_kazuhirokota_: I see. Umm, may I update your following patch?02:23
m_kazuhirokota_: (to follow main symlnk patch's update)02:24
kota_m_kazuhiro: i think i can manage my patch for current newest symlink patch set. but not sure on tdasilva work.02:25
m_kazuhirokota_: You say that you can manage your patch for symlink's patch set 60(= current patch set), but you are not sure on symlink's patch set 61 (= next patch set). Correct?02:29
kota_?02:37
m_kazuhirokota_: I'm sorry. I don't understand what you want to say because of my poor english skill.02:38
m_kazuhirokota_: What is your concering point?02:39
kota_m_kazuhiro: just I was thinking of "what/when i can work" on symlink. Hence I'd like to catch up the diff from current patch. However, I couldn't get about the concrete diff so that I'm not sue what/when I can do.02:46
kota_maybe except waiting tdasilva do something?02:47
*** links has joined #openstack-swift02:51
*** ChubYann has quit IRC02:51
m_kazuhirokota_: OK. I see.03:03
*** m_kazuhiro has quit IRC03:15
*** Raymii has quit IRC03:23
*** Raymii has joined #openstack-swift03:23
*** wes_dillingham has quit IRC03:35
*** tovin07_ has quit IRC04:00
*** tovin07_ has joined #openstack-swift04:01
*** mvk has joined #openstack-swift04:20
*** pcaruana has joined #openstack-swift04:27
*** mvk has quit IRC04:31
*** gkadam_ has joined #openstack-swift04:38
*** gkadam_ has quit IRC04:39
*** gkadam has joined #openstack-swift04:40
*** mvk has joined #openstack-swift04:45
*** mvk has quit IRC04:50
*** ChubYann has joined #openstack-swift04:52
*** pcaruana has quit IRC04:56
*** klrmn has quit IRC05:09
*** chlong has joined #openstack-swift05:20
*** ChubYann has quit IRC05:24
*** pcaruana has joined #openstack-swift05:30
*** pcaruana has quit IRC05:39
*** rcernin has joined #openstack-swift05:41
*** spectr has quit IRC05:58
*** spectr has joined #openstack-swift05:58
*** cshastri has joined #openstack-swift06:06
*** hseipp has joined #openstack-swift06:15
*** hseipp has quit IRC06:29
hugokuoIs there known bug of empty container suffix in the partition?06:41
*** okdas has quit IRC07:06
*** pcaruana has joined #openstack-swift07:07
*** cbartz has joined #openstack-swift07:19
*** tesseract has joined #openstack-swift07:22
*** spectr has quit IRC07:26
*** tesseract has quit IRC07:26
*** rcernin has quit IRC07:26
*** pcaruana has quit IRC07:26
*** geaaru has joined #openstack-swift07:27
*** mvk has joined #openstack-swift07:37
*** spectr has joined #openstack-swift07:39
*** pcaruana has joined #openstack-swift07:40
*** rcernin has joined #openstack-swift07:40
*** spectr has quit IRC07:43
*** pcaruana has quit IRC07:44
*** spectr has joined #openstack-swift07:44
*** pcaruana has joined #openstack-swift07:44
*** itlinux has joined #openstack-swift08:33
*** itlinux has quit IRC08:36
acolesgood morning08:45
acolesI just discovered the 'Included In' drop down on gerrit review pages - top right corner - that is very useful08:49
acolessomeone is going to tell me it has been there for years!08:49
*** rcernin has quit IRC08:59
*** spectr has quit IRC08:59
*** pcaruana has quit IRC08:59
acoleskota_: m_kazuhiro I just left a comment re symlink content-type auto-insertion on etherpad https://etherpad.openstack.org/p/swift_symlink_remaining_discussion_points09:12
*** spectr has joined #openstack-swift09:12
*** rcernin has joined #openstack-swift09:12
*** pcaruana has joined #openstack-swift09:13
kota_thx acoles!09:18
*** m_kazuhiro has joined #openstack-swift09:26
m_kazuhiroacoles: Are you here?09:27
acolesm_kazuhiro: hello09:28
m_kazuhiroacoles: Hello. I want to talk about your comments on symlink etherpad page. Do you have time for it?09:28
acolesm_kazuhiro: yes09:29
m_kazuhiroacoles: Thank you.09:29
m_kazuhiroacoles: First, your concerning point of auto insertion is about POST. But I think this topic is only on PUT requests.09:31
m_kazuhiroacoles: So the topic is "if there are recommended content-type, should we insert the content-type automatically when clients PUT symlink without content-type?". But in my current thought, recommended content-type is not needed on symlinks.09:34
acolesm_kazuhiro: my thought is that if a POST can undo any auto-insertion of a specific content-type during PUT, then is there any value in the auto-insertion in the first place?09:34
acolesm_kazuhiro: oh, then maybe I misunderstood the etherpad. I was addressing the question 'Should symlink middleware insert this content-type automatically?'09:34
acoleshmmm, do you mean 'we do not need to recommend any value because we will automatically set it' or ' we do not need to care about the value, just leave it to be whatever the client or default is'?09:36
m_kazuhiroacoles: My thought is second one.09:37
acolesm_kazuhiro: ok. then I do not disagree with you. I do not have a strong opinion about whether there should be a particular content-type, but I do think it is hard to reliably apply any value automatically. So for pragmatic reasons I think it may be better to leave whatever the client sends.09:39
*** chlong has quit IRC09:40
m_kazuhiroacoles: OK. The recommend content-type comment is added by tdasilva. And I asked him about that. His answer was that he doesn't remember but it may be for distinguishing symlinks.09:41
acolesm_kazuhiro: OIC. I just added some more comment on the etherpad - let's see if tdasilva has other thoughts. Seems to me that if we have the symlink target appended to etag then that is sufficient to distinguish a symlink.09:42
m_kazuhiroacoles: I think so. Therefore I think our conclusion will be that recommended content-type comment is removed from symlink patch.09:46
*** itlinux has joined #openstack-swift10:00
*** tovin07_ has quit IRC10:05
acolesm_kazuhiro: ok. have a good weekend!10:20
m_kazuhiroacoles: Thank you!10:21
*** itlinux has quit IRC10:21
openstackgerritAlistair Coles proposed openstack/swift feature/deep: tweak when root is updated with shard range stats  https://review.openstack.org/51180210:29
*** psachin has quit IRC10:33
*** tone_zrt has quit IRC10:35
*** kei_yama has quit IRC10:36
*** m_kazuhiro has quit IRC10:41
*** pcaruana has quit IRC10:48
openstackgerritAlistair Coles proposed openstack/swift feature/deep: save sharding metadata as string  https://review.openstack.org/51180811:09
openstackgerritAlistair Coles proposed openstack/swift feature/deep: delete unused get_metadata_item method  https://review.openstack.org/51180911:11
openstackgerritAlistair Coles proposed openstack/swift feature/deep: tweak when root is updated with shard range stats  https://review.openstack.org/51180211:22
openstackgerritAlistair Coles proposed openstack/swift feature/deep: save sharding metadata as string  https://review.openstack.org/51180811:22
openstackgerritAlistair Coles proposed openstack/swift feature/deep: delete unused get_metadata_item method  https://review.openstack.org/51180911:22
*** hseipp has joined #openstack-swift11:35
tdasilvagood morning11:50
tdasilvakota_: just fyi, i'm working on a new patchset for symlinks and close to submitting it11:51
tdasilvaacoles: thanks for your comments on the etherpad, i think they are very valid11:51
acolestdasilva: NP. good morning!11:52
tdasilvaacoles: I was thinking that leaving just the current recommendation in the docs is sufficient (and harmless) then11:52
acolestdasilva: I don't have a strong opinion but if something is recommended I tend to think that to do otherwise is then a mistake11:52
acoles...unless of course the recommendation comes from a retailer :)11:53
tdasilvaheh11:54
tdasilvaso i think it is good advice...is advice less than a recommendation? ;)11:55
acolestdasilva: it sorts less alphabetically :)11:55
tdasilvalol11:55
acolesso is the advice 'content-type can be anything, but if you're not sure use application/symlink' ?11:57
acolesor is it 'there's a good reason to use application/symlink unless you have a reason not to' ?11:58
*** ^andrea^ has joined #openstack-swift12:08
*** wes_dillingham has joined #openstack-swift12:13
tdasilvaacoles: IMO first one12:26
acolestdasilva: in which case perhaps 'suggested' rather than 'recommended' ... but I don't feel strongly about it12:27
tdasilvaacoles: ack, thanks12:27
* tdasilva goes back to re-read what's in the docs12:27
*** mat128 has joined #openstack-swift12:32
*** links has quit IRC12:45
*** pcaruana has joined #openstack-swift12:48
*** spectr has quit IRC12:57
*** spectr has joined #openstack-swift13:01
*** jistr is now known as jistr|mtg13:08
*** gkadam has quit IRC13:25
*** m_kazuhiro has joined #openstack-swift13:27
*** chlong has joined #openstack-swift13:32
m_kazuhirotdasilva: Are you here?13:38
openstackgerritAlistair Coles proposed openstack/swift feature/deep: use same pending file regardless of broker db path  https://review.openstack.org/51185213:42
*** rcernin has quit IRC13:44
*** d0ugal has quit IRC13:48
tdasilvam_kazuhiro: hi14:01
m_kazuhirotdasilva: Hi!14:02
tdasilvam_kazuhiro: one sec14:02
m_kazuhirotdasilva: ok14:02
*** d0ugal has joined #openstack-swift14:03
tdasilvam_kazuhiro: back14:04
tdasilvam_kazuhiro: what's up? it's really late for you14:05
m_kazuhirotdasilva: I want to talk about some topics.14:05
m_kazuhirotdasilva: about symlink.14:05
m_kazuhirotdasilva: There are 3 topics. First topics is "how about your patch update?"14:06
tdasilvam_kazuhiro: yes, working on it :) hoping to send the patchset today14:08
tdasilvam_kazuhiro: i'm working through kota_ comments14:08
m_kazuhirotdasilva: nice :)   Will you include update for recommended content-type? I read your talk with acoles.14:11
m_kazuhirotdasilva: (recommended content-type is my second topic.)14:12
tdasilvam_kazuhiro: yes, honestly, it will be just a little change in wording on the doc14:12
tdasilvam_kazuhiro: as part of this patchset I took a longer look at the current docs and tried to change things a bit, was just looking for clarifying terms a bit and hopefully make it clearer14:12
m_kazuhirotdasilva: Great. Clear docs are very helpful.14:15
m_kazuhirotdasilva: I think some of kota_'s comments are about container-sync with symlink. So I will make update for the comments after your update.14:16
tdasilvam_kazuhiro: ok, will probably need your help with container-sync14:17
m_kazuhirotdasilva: Then I start 3rd topic... It is today's main topic. I want to share/discuss schedule for progressing symlink review cycle.14:19
m_kazuhirotdasilva: I want to make sure to merge symlink in Queens release.14:20
m_kazuhirotdasilva: So I want to make and share schedule for symlink merge.14:21
*** catintheroof has joined #openstack-swift14:23
m_kazuhirotdasilva: In my understanding, deadline for queens release is end of January.14:25
tdasilvam_kazuhiro: I'd love to see symlinks merge in Queens. In fact, I was just looking at the first patchet yesterday and that was proposed about 2 years ago! But...I've learned that's not something I can control, it's really a community effort to review and commit to maintain the code. All I can do is to try to be as reponsive as possible to reviews14:25
m_kazuhirotdasilva: I believe that we can increase probability of Queens merge , even if we cannot make 100 % proabability.14:31
*** chetna has joined #openstack-swift14:32
m_kazuhirotdasilva: In my thought, if we have schedule, the probability will increase.14:33
*** Dinesh_Bhor has quit IRC14:34
tdasilvam_kazuhiro: sounds good, please do share14:35
m_kazuhirotdasilva: Thanks. I will talk schedule from end. Because I think it will be more easy to explain.14:37
*** cshastri has quit IRC14:38
tdasilvam_kazuhiro: i'm assuming you have dates? do you have something like a table that I can visualize?14:38
m_kazuhirotdasilva: Yes. I have schedule of dates. Umm.. I don't have a table now... Can I write my schedule image on etherpad page?14:40
tdasilvam_kazuhiro: yeah, perfect14:41
m_kazuhirotdasilva: OK please watch tail of https://etherpad.openstack.org/p/swift_symlink_remaining_discussion_points14:41
*** rcernin has joined #openstack-swift14:45
*** spectr has quit IRC14:46
timburkegood morning14:48
m_kazuhirotimburke: good morning14:48
acolestimburke: o/14:48
m_kazuhirotdasilva: I finished to write my current schedule image.14:49
timburkem_kazuhiro: still awake? go to bed! :P14:49
tdasilvam_kazuhiro: ack14:50
m_kazuhirotimburke: OK. I'll go to bed after talk with tdasilva!14:51
timburke:-)14:51
*** chetna has quit IRC14:54
*** chetna has joined #openstack-swift14:54
m_kazuhirotdasilva: I want to do this schedule on etherpad page. And I intend to use my time for this schedule.14:54
tdasilvam_kazuhiro: ok, I think you have a good goal in mind.14:55
*** jistr|mtg is now known as jistr14:58
m_kazuhirotdasilva: Thank you. And I want to be main patch updater for symlink after your next patch update for my schedule... Is it OK?14:59
*** klrmn has joined #openstack-swift14:59
tdasilvam_kazuhiro: absolutely...sorry I jumped the gun on this patchset...how about this, I'll send a new patchset today in whatever form I can get and then I'll mark whatever comments I was not able to address15:01
tdasilvaand you can take it from there and become maintainer15:01
tdasilvasounds good?15:01
tdasilvaI might add some more topics to the etherpad on what is remaining and cross-off whatever is done15:02
*** rcernin has quit IRC15:03
m_kazuhiroOK. That sounds good.15:03
m_kazuhirotdasilva: Thank you very much and I'm sorry for being selfish.15:13
*** acoles has left #openstack-swift15:19
*** acoles has joined #openstack-swift15:19
*** ChanServ sets mode: +v acoles15:19
tdasilvam_kazuhiro: welcome, and you are not being selfish at all. no worries15:19
tdasilvayou are doing the right thing15:19
acolesm_kazuhiro: it is *late* !15:20
tdasilvayes, go to sleep and have a nice weekend15:20
m_kazuhirotdasilva: acoles: Thank you! Have a nice weekend.15:21
*** m_kazuhiro has quit IRC15:23
*** gyee has joined #openstack-swift15:32
*** chlong has quit IRC15:53
acolestimburke: I had similar thought about only ever init'ing the broker with the hash.db path. I'd certainly be ok with that being the norm. we might want to support the other (hash_shard.db) just in case someone wants to grab a path and open a broker, perhaps by always normalising the path that is passed to __init__?15:57
acolestimburke: anything to make it easier to grok all the slight of hand with db names and existence checks :)15:58
timburkewhich is more or less what that last patch does, right? certainly seems like15:58
timburkei agree that it feels a little weird that we're faking out the super().__init__() -- but it might actually be not-so-unreasonable?15:59
acoleshmmm, maybe. it felt dirty when I did it :P15:59
acolesbut maybe it's not so bad16:00
acolesI guess I fear the super __init__ changing in some way that breaks the trickery16:00
timburkeyeah, it's a fair concern...16:01
acolestimburke: mattoliverau I hesitate to claim it - but now we replicate deleted items and flush pending files, I seem to have a shrinking probe test (of sorts) passing16:07
acoleswhat with it being 5pm on a Friday maybe I should quit while I'm ahead16:07
acolesFriday 13th even !16:08
timburkeyay! meanwhile, i just realized that i should be running the existing probe tests more -- acoles, have you noticed failures around https://github.com/openstack/swift/blob/feature/deep/test/probe/test_sharder.py#L303-L30716:08
timburkegood point. have a good weekend acoles!16:08
acolestimburke: hmm, no I haven't run that async_pending test for a while16:09
timburke:-( but when i back it out so *everyone* looks for 60, i get at least one looking for 80... time for me to dig deeper16:10
acolesfun fun fun16:10
acolestimburke: I'll push what I have for shrinking - I added a few more helper methods in the probe test including a direct get to all container nodes16:11
timburkesounds good. i'll aim to get a merge from master up today, just as soon as i've got probe tests in a good state16:11
acolesthanks!16:12
*** saint_ has joined #openstack-swift16:13
*** saint_ has quit IRC16:13
timburkeoh, no! *different* assertion now. i think actually, 3/5 across the board may be right...16:14
*** lnxnut_ has quit IRC16:16
*** klrmn has quit IRC16:19
*** cbartz has quit IRC16:19
*** oshritf has joined #openstack-swift16:22
*** pcaruana has quit IRC16:23
*** hseipp has quit IRC16:24
openstackgerritAlistair Coles proposed openstack/swift feature/deep: WIP Alternative shrinking  https://review.openstack.org/51097816:33
acolestimburke: mattoliverau ^^ I am sure there is plenty to iron out in that still but progress I think. The probe test's shrunken shard is not deleted - but it is in the sharded state and has a delete timestamp16:35
acolesI'm leaving it there for this week16:35
*** wes_dillingham has quit IRC16:42
*** oshritf has quit IRC16:45
openstackgerritMerged openstack/swift feature/deep: tweak when root is updated with shard range stats  https://review.openstack.org/51180216:48
openstackgerritMerged openstack/swift feature/deep: save sharding metadata as string  https://review.openstack.org/51180816:49
openstackgerritMerged openstack/swift feature/deep: delete unused get_metadata_item method  https://review.openstack.org/51180916:50
notmynamegood morning16:51
*** wes_dillingham has joined #openstack-swift16:53
*** hseipp has joined #openstack-swift17:08
*** hseipp has quit IRC17:09
*** klrmn has joined #openstack-swift17:10
*** mvk has quit IRC17:14
*** zaitcev has joined #openstack-swift17:21
*** ChanServ sets mode: +v zaitcev17:21
*** catintheroof has quit IRC17:25
*** catintheroof has joined #openstack-swift17:25
*** mvk has joined #openstack-swift17:27
*** catintheroof has quit IRC17:29
openstackgerritTim Burke proposed openstack/swift feature/deep: Fix up async pending probe test  https://review.openstack.org/51193017:32
*** chetna has quit IRC17:35
*** chetna has joined #openstack-swift17:39
*** dosaboy has quit IRC17:49
openstackgerritMerged openstack/swift feature/deep: use same pending file regardless of broker db path  https://review.openstack.org/51185218:04
*** Guest53680 is now known as mgagne18:14
*** mgagne has quit IRC18:14
*** mgagne has joined #openstack-swift18:14
zaitcevOkay, I'm alive. Don't know how or why, but I saw that Clay reviewed PUT+POST. BTW, here's the actual route, with overnight stops marked black: http://www.zaitcev.us/things/n41475_2017_ferry.jpg18:19
zaitcevEr, I meant to say that I saw that review.18:20
openstackgerritTim Burke proposed openstack/swift feature/deep: Merge branch 'master' into feature/deep  https://review.openstack.org/51194118:22
*** catintheroof has joined #openstack-swift18:33
timburkezaitcev: that's a good sized trip! Austin to just past Rochester? hope you had a good time :-)18:36
zaitcevtimburke: I went the other direction, ferried Carlson home from repairs at the original builder in upstate NY.18:37
timburkeah, gotcha18:37
*** dosaboy has joined #openstack-swift18:48
*** dosaboy has quit IRC18:48
*** dosaboy has joined #openstack-swift18:50
*** geaaru has quit IRC18:50
openstackgerritTim Burke proposed openstack/swift master: Send correct SLO ETag for container updates  https://review.openstack.org/33796018:51
openstackgerritTim Burke proposed openstack/swift master: Autovivify X-Versions-Location container  https://review.openstack.org/26501518:59
*** ChubYann has joined #openstack-swift19:10
openstackgerritTim Burke proposed openstack/swift master: Increase name-length limits for internal accounts  https://review.openstack.org/46577619:25
openstackgerritTim Burke proposed openstack/swift master: Respond 400 Bad Request when Accept headers fail to parse  https://review.openstack.org/50284519:35
openstackgerritTim Burke proposed openstack/swift master: Send correct SLO ETag for container updates  https://review.openstack.org/33796019:45
openstackgerritTim Burke proposed openstack/swift master: Increase name-length limits for internal accounts  https://review.openstack.org/46577619:46
* timburke grumbles about line-length limits...19:47
*** wes_dillingham has quit IRC19:53
*** gyee has quit IRC20:03
*** shaner has quit IRC20:18
*** shaner has joined #openstack-swift20:21
openstackgerritThiago da Silva proposed openstack/swift master: Symlink implementation.  https://review.openstack.org/23216220:42
*** chetna has quit IRC21:09
*** catintheroof has quit IRC21:10
*** wes_dillingham has joined #openstack-swift21:13
*** SkyRocknRoll has joined #openstack-swift21:14
openstackgerritMerged openstack/swift feature/deep: Fix up async pending probe test  https://review.openstack.org/51193021:18
openstackgerritMerged openstack/swift feature/deep: Merge branch 'master' into feature/deep  https://review.openstack.org/51194121:19
*** wes_dillingham has quit IRC21:32
*** mat128 has quit IRC21:32
*** wes_dillingham has joined #openstack-swift21:44
openstackgerritSamuel Merritt proposed openstack/swift master: Add checksum to object extended attributes  https://review.openstack.org/33632321:53
timburketorgomatic: it lives!22:00
torgomatictimburke: seriously22:00
torgomaticI don't remember that patch being so large; last time I looked, it was like 100 lines :)22:01
timburkelooks like an awful lot of it is just dropping skip_if_no_xattrs in a bunch of tests22:03
torgomaticthis is true; it looks a lot bigger when there's 8 months' worth of test rearrangement on master. The conflict hunks are ENORMOUS :)22:03
*** wes_dillingham has quit IRC22:09
*** SkyRocknRoll_ has joined #openstack-swift22:13
*** wes_dillingham has joined #openstack-swift23:11
*** wes_dillingham has quit IRC23:13
*** SkyRocknRoll has quit IRC23:17
*** SkyRocknRoll_ has quit IRC23:18
*** ujjain has quit IRC23:25
openstackgerritTim Burke proposed openstack/swift master: Remove swift-temp-url script  https://review.openstack.org/51198123:29
*** ujjain has joined #openstack-swift23:30
*** ujjain has joined #openstack-swift23:30
openstackgerritSamuel Merritt proposed openstack/swift master: Add checksum to object extended attributes  https://review.openstack.org/33632323:55

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